(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Annual report : National Institutes of Health. Division of Computer Research and Technology"



vu' 



-T" a-t 





















g^rv^-TB5 















5«- -> ^ 



mm 








NATIONAL INSTITUTES OF HEALTH 
NIH LIBRARY 



AUG 5?00 



BLDG 10, 10 CENTER DR 
BETHESDA,MD 20892-1150 



DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



FISCAL 

YEAR 

1981 



ANNUAL 
REPORT 



rVDT T 

DCDCRTK" 



DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



FISCAL ANNUAL VOLUME 1 

YEAR REPORT 

1981 



_ rv r\ 
rVHT T 

CDCKTKT 

1> !AJ 



Foreword 

The work of the Division of Computer Research and 
Technology covers a large spectrum of activities. It 
ranges from doing research in biology, statistics, 
mathematics, and computer science to providing 
computer facilities and services for the NIH. 

The several DCRT laboratories and branches 
embody and integrate this variety of talents. Each 
has a major functional focus. But the success of the 
Division's work arises from the interaction of 
members of each group with others across 
organizational and disciplinary lines. Many projects in 
the Division require the expertise of people from 
several segments of the spectrum. 

DCRT's collaborative projects link its staff to 
professionals both inside and outside NIH. The result 
is a balance in emphasis to provide the work done 
by DCRT at NIH with the benefits of collaborations 
outside of NIH. 

While DCRT does not have money for grants, it does 
provide occasional support for meetings on scientific 
topics related to its work. 

This year's annual report is presented in two 
volumes: 

Volume 1 gives a summary overview of the 

work of each group and highlights its 

accomplishments. 

Volume 2 includes detailed projects and 

activities of each group. 
If you have comments on the report or suggestions 
for improving future annual reports, please send 
them to: 

DCRT Information Office 

Building 12A, Room 3027 

Division of Computer Research 

and Technology 
National Institutes of Health 

Bethesda, Maryland 20205 




Dr. Arnold Pratt, DCRT Director, is internationally known for 
his research in computational linguistics. 



From the Director 



The Division of Computer Research and Technology 
(DCRT) was established in 1964 to make computers 
useful at NIH. Over the years, the work of the 
Division has become an integral part of the conduct 
and management of NIH research programs. 

It is a pleasure to report once again the 
accomplishments of the people in an exceptionally 
competent and productive group of laboratories, 
branches, and offices. All parts of the DCRT 
program are important for that balance of science 
and technology that has made the Division and 
computing throughout NIH so successful. 

Three brief glimpses into the stream of fiscal year 
1981 history, help to show that computing thrives and 
supports progress in all the NIH programs. 
• NIH biomedical scientists got wide recognition 
for using computers. A single June issue of the 
NIH Record carried three examples. 
One article told of a young NIH staff physician 
becoming professor of radiology at an academic 
medical center. He first came to NIH in 1973 as 
a medical student for our course on computers 
in medicine; after graduating, he returned to 
work in DCRT and in the NIH Nuclear fvledicine 
Department, creating and using new computing 
techniques. 

Another reported an award by the Endocrine 
Society to an NIH medical scientist for 
outstanding leadership and service. The award 
was for his development of a series of computer 
programs that are used in laboratories 
throughout the world as well as widely at NIH. 
The third story told of a prestigious European 
prize for distinction in advancement of 
knowledge about diabetes mellitus. The picture 
in the Record showed the recipient working at 
a computer terminal, using one of the several 
advanced DCRT systems he exploited 
successfully while he was a Visiting Scientist at 
NIH. 



• NIH can pay its suppliers promptly. A computer 
system can now process vouchers for goods 
and services as soon as they come to the NIH 
Accounts Payable Section. In the past, manual 
processing labored under week- or month-long 
backlogs. The administrative data base system 
created by DCRT staff cuts out most of the 
paper work in voucher payment, and it also 
automatically records obligations and accruals in 
the central accounting record. 

• The NIH Computer Center obtained contracts 
for new equipment. This feat came after 
seemingly endless years of preparation and 
negotiation. Over the next five or more years, 
the 8,000 users of the central NIH systems will 
benefit in a timely way from proven advances in 
computing technology. 

DCRT looks forward with enthusiasm to the coming 
year as an opportunity to work with NIH scientists 
and administrators in creating still more powerful and 
useful systems. The challenge lies in building strong 
intellectual links from computers, mathematics, and 
engineering to the substance of science and the art 
of administration. After 17 years the frontiers of 
computing remain open for new advances in all NIH 
programs. Research and development programs 
within DCRT will continue to help lead the way. 



InA*^' 



M 




Arnold W. Pratt, M.D. 

Director 

Division of Computer Research 
and Technology 



Contents 



Division of Computer Research 
and Technology 



Laboratory of Statistical and Mathematical 
Methodology 8 

Provides statistical and mathematical help in the 
computer analysis of biomedical data; offers 
statistical and mathematical packages for users; 
develops methodology in multivariate analysis, curve 
fitting, biological shape and pattern theory. 

Computer Systems Laboratory 14 

Provides consultation and collaboration in the design 
and implementation of specialized computer systems 
for laboratory and clinical applications. 

Laboratory of Applied Studies 20 

Relates mathematics, statistics, and computer 
science to such biomedical problems as ECG 
analysis, evaluation of physiological systems in 
health and disease, modeling of the microcirculation, 
and estimation problems in laboratory medicine. 

Physical Sciences Laboratory 24 

Conducts research in mathematical theory and 
practical instrumentation to explain biological 
phenomena in terms of chemistry and physics at the 
subcellular molecular levels. 

Data Management Branch 28 

Serves as a central resource of systems analysis, 
design, and programming for data processing 
projects relating to scientific, technical, management, 
and administrative data. 

Computer Center Branch 32 

Designs, implements, and operates the NIH 
Computer Center; provides assistance, training and 
technical communications to the more than 8,000 
users of the Central Utility. 



Office of ADP Policy Coordination 39 

Coordinates the complex Federal policies and 
procedures that govern getting and using computers 
at NIH. 

Office of Administrative Management 41 

Provides general administrative management support 
for the Division's w^ork. 

Office of Scientific and Technical 42 
Communications 

Serves as a central source of information about 
DCRT activities and about computer-related 
disciplines. 




LSM combines research in mathematical statistics, 
mathematics, and computer and information science. 

Research projects in LSM vary from studies of natural 
language processing for medical information systems to 
studies in statistical methodologies for biomedical 
applications. 




Laboratory of Statistical and 
i\/lathematical l\/lethodology 



James E. Mosimann, Chief 



Function 

The Laboratory of Statistical and Mathematical 
Methodology (LSM) combines research in 
mathematical statistics, mathematics, computer, and 
information science with collaboration and service in 
these areas to NIH researchers and administrators. 

In addition to the position of chief, the laboratory has 
fifteen full-time professional positions distributed 
among four sections: 

• The Statistical Software Section (SSS) 
provides consultation to and collaboration with 
NIH researchers and administrators in all 
computational aspects of biomedical data 
analysis, including selection and support of large 
systems/packages. Three specialists in scientific 
programming are led by a computer systems 
analyst whose specialty is statistics. 

• The Biomatiiematics and Computer Science 
Section (BCS), directed by a mathematician, 
performs independent research and provides 
consultation and collaboration in the specialties 
of its five computer and mathematical scientists. 

• The Statistical Methodology Section (SMS) 
works closely with the Statistical Software 
Section. Two professionals in mathematical 
statistics provide biostatistical consultation and 
do independent research. 

• The Medical Information Science Section 
(MIS) investigates and develops methods for 
application of information and computer science 
to medical language data processing. Two 
computer specialists work under the direction of 
a computer systems analyst who is an expert in 
computational linguistics. 

Scope of Work 

LSM staff interact with all NIH institutes, with other 
Federal agencies outside HHS, and with biomedical 
researchers worldwide. FY81 was LSM's seventh 
year as a separate entity within DCRT. The volume 
of its computational and consultative services 
continued to expand; its research activities 
maintained about the same level as the preceding 
year. 

A major part of LSM activity is the offering of 
statistical and mathematical systems/ packages to 
the NIH user community. LSM accepts responsibility 



for evaluation of new systems/packages and their 
suitability for NIH. When it offers a system/package 
to the NIH community, LSM makes three basic 
commitments: 

1. Maintenance of the package, with adequate 
documentation, through NIH computer system 
changes, system/package updates, and corrections. 

2. Rapid response to queries concerning user 
access to a system/package program, including job 
control language and program parameters. 

3. Assistance in interpretation of results. 
During this year, as in the past year, the Statistical 
Software Section of LSM maintained the following 
systems/packages and programs on the IBM 370 
system of the DCRT Computer Center: 

BMD (BMDP, Biomedical Computer Programs, 
UCLA) 

SPSS (Statistical Package for the Social Sciences, 
SPSS, Inc.) 

SAS (Statistical Analysis System, SAS Institute, 
Inc.) 

P-STAT (Statistical Package, P-STAT, Inc.) 

IMSL (International Mathematical and Statistical 
Libraries, IMSL, Inc.) 

MSTAT1 (Collection of Mathematical and 
Statistical Programs, DCRT) 
There is a major commitment to maintain these 
systems/packages and respond to queries about 
their use. In FY81 alone the SSS staff responded to 
over 4,500 calls. Also during this year, every system/ 
package went through at least one major update. 

The Biomathematics and Computer Science Section 
maintains several systems/packages and specialized 
systems on the DECsystem-10 of the Computer 
Center. Foremost in use is the interpretive system 
MLAB, designed (by LSM scientists) for 
biomathematical modeling. BCS supports the 
PROLOG system (for non-numerical data 
processing) and also a DECsystem-10 version of 
IMSL whose S/370 version is supported by SSS. C- 
LAB, a system on the DEC-10 for pattern recognition 
and clustering (written by an LSM scientist) is 
supported by SMS. The Unified Generator Package, 
written and maintained by a BCS staff member, is on 
DCRT's IBM System 370. 

As a result of LSM's policy of not only supporting the 
use of these systems/packages but also aiding in 



9 



42000 -T 

37800 - 

33600 - 

29400 - 

25200 - 

21000 - 

16800 - 

12600 - 

8400 - 

4200 - 

1000 - 



Uses per month of 

Statistical paclcages supported by LSIVi* 




JUN'75 JUN'76 JUN 77 JUN '78 JUN '79 JUN '80 JUN '81 

'Packages supported by the Statistical Software Section only. Does not include 
packages supported by the Biomathematics and Computer Sciences Section. 



the interpretation of their output, the statisticians of 
the Statistical Methodology Section provide 
consultation over a wide range of scientific fields. 
Some very brief consultations are very successful 
because there is a known answer to the question at 
hand. Other consultations involve extensive time and 
statistical/mathematical/computer science research 
as well. 

Research projects in LSM vary widely from studies 
of natural language processing for medical 
information systems and studies of efficient 
algorithms for information retrieval to studies in 
mathematics and statistical methodologies for 
biomedical applications. 



10 



Highlights of the Year's Activities 

Computation 

In FY81 LSM continued to expand teaching and 
documentation for supported systems/packages. 
LSM taught four introductory courses for SAS, two 
for BMDP, two for SPSS, and one for IMSL. In 
addition, two introductory courses and one advanced 
course for MLAB, and one introductory course for C- 
LAB were taught. Five articles on MLAB techniques 
and one article on mathematical modeling of 
chemical kinetics systems appeared in INTERFACE. 
A new edition of the MLAB Applications Manual was 
released, with additional sections on bivariate density 
analysis (including 3-dimensional graphics and 
contour mapping methods in MLAB) and on delta 
modulation encoding of signals. 

Two new systems/packages were offered on an 
experimental basis to the NIH computer user 
community. SCSS, an interactive version of SPSS, 
and SAS/GRAPH, a graphics package from the SAS 
Institute, were installed and tested at DCRT. 

A DECsystem-10 utility program for interactive 
computer generation of an index to a user's 
document was completed. It is being production- 
tested on new editions of the MLAB Reference 
Manual and the MLAB Beginner's Guide, now in 
preparation. 

MLAB was enhanced by technical improvements to 
increase speed and reduce memory requirements 
and by the addition of Fourier transform and inverse 
transform operators. 

Consultation, Collaboration, and Research 

As in FY80, LSM consultation and research in FY81 
was closely tied to the use of the computer. Most 
consultations (55 percent) involved statistical advice 
combined with considerable computer use. Others 
(40 percent) involved computer use alone and a 
small fraction (5 percent) involved mathematical or 
statistical advice with only limited computer use. 

In FY81, LSM research, collaborative, and 
consultative efforts merged more closely and were 
less distinguishable among themselves. In a number 
of studies, statistical methodologies were developed 
for, or modified to suit, specific biomedical problems. 
For example, in a collaboration with A. DeBlas, 
NHLBI, discrete distribution models were used to 



analyze data from monoclonal hybridoma 
experiments. In a collaboration with A. Grimes, CC, 
an unordered paired-data method was modified to 
study monaural versus binaural amplification in 
hearing-impaired children. In a third example, a 
collaboration of several years duration (with Drs. G. 
Hirschman, R. Wineman, and M. Wolfson, NIADDK) 
on complications of dialysis for patients with end 
stage kidney disease was completed, using a special 
method for adjusting hospitalization rate by the 
patient's length of time at risk. Results from all three 
collaborations have been published or accepted for 
publication. 

Other areas in which LSM consultation led to 
productive collaborative efforts include studies on 
schistosomiasis with Dr. A. Cheever, NIAID, with 
three papers published, as well as studies on the 
pain syndrome Causalgia with Dr. A. Tahmoush, 
NINCDS. In the latter case a paper, using linear 
model methods, for unbalanced data, has been 
prepared on measurements of sympathetic nervous 
system activity in patients versus controls. 

Statistical research on simultaneous confidence 
intervals for ratios continued in FY81. The results, of 
wide applicability, are accepted for publication. Other 
research included a study of the connection between 
statistical and algebraic independence. These results 
are applicable to the sample covariance matrix of 
multivariate data. This latter is important in statistical 
discriminant analyses, which is a subject of LSM 
research in collaboration with Drs. J. Darroch, 
Flinders University, South Australia, and H. Hoffman, 
DRS. Discriminant methods adapted for size and 
shape variables are being used to study genetic 
variation in laboratory mice. A separate study of 
independence of size and shape variables before 
and after scale change is in press, along with other 
LSM studies of statistical distributions. 

In computer science, research on computer 
generation of scientific manuscripts led to 
development of an interactive software system. This 
system has the capability of generating complex 
displays of mathematical formulas and MLAB 
graphics on a high-resolution printer-plotter. Recent 
research on procedures for managing extendible 
array files was completed and published. Research 
also continued on the 'symmetric axis' method for 



11 




Staff members teach courses on the use of various program 
packages that LSM supports. 

The Statistical Software Section (SSS) provides consultation 
to and collaboration with NIH researchers and 
administrators. 





describing biological shapes. This included 
augmentation of software, publication of a paper or 
the geometry of the three-dimensional case, and 
collaboration with Dr. R. Webber, NIDR, in the 
analysis of human mandible images. Mathematical 
studies of equivalence of theories of modules over 
ring, and also of convex homogeneous cones in 
finite-dimensional vector spaces over the real field 
were completed. The latter studies have potential 
application in a variety of areas, including size and 
shape variables. In linguistic analysis an extensive 
set of IBM 370 assembly language programs was 
developed for the automatic identification of prefixe 
and suffixes in French medical terminology. In 
collaboration with Drs. J. Costa and D. Henson, NC 
studies of the automatic processing of natural 
language surgical pathology reports continued. 



12 



I 



Future Plans 

No major shift in laboratory service or research is 
anticipated in the coming year. Current levels of 
statistical and mathematical systems/packages 
support, consultation, and user assistance will be 
maintained. Research projects will be continuations 
of those already initiated and reported here. 



Publications 



Blum. H : 3-D Symmetric Axis Coordinates: An Overview and Prospectus In 
Badler. N . Bajcsy. R and Otto. G. (Eds.): Three Dimensional Ob/ecl 
Representation New YorK. London, and Heidelberg. Spnnger-Veriag, 
1981 

Carlson. R , and Malley, J D.: Job Satisfaction of Staff RN'S in Primary and 
Team Nursing Delivery Systems Research in Nursing and Health, 

Cheever, A. W.. Duvall, R. H.. Minker, R. G,: Extrahepatic Pathology in 
Rabbits Infected with Japanese and Philippine Strains of Schistosoma 
Japonicum and the Relation of Intestinal Lesions to Passage of Eggs in 
the Feces The American Journal of Tropical Medicine and Hygiene 29: 
1316-1326. 1980 

Cheever, A W. Duvall. R H , Minker. R. G . and Nash T E.: Hepatic 
Fibrosis in Rabbits Infected with Japanese and Philippine Strains of 
Schistosoma Japonicum The American Journal of Tropical Medicine 
and Hygiene 29: 1327-1339. 1980. 

Cheever. A, W , Duvall R H,. and Minker. R. G.: Quantitative Parasitologic 
Findings in Rabbits Infected with Japanese and Philippine Strains of 
Schistosoma Japonicum The American Journal of Tropical Medicine 
and Hygiene 29: 1307-1315. 1980. 

DeBlas. A, L,. Ratnaparkhi. M. V.. Mosimann, J. E.: Estimation of the 
number of monoclonal hybridomas in a cell fusion expenment. Journal 
of Immunological Methods (in press) 

Garcia-Hidalgo. I , and Dunham, G.: An experiment in English-Spanish auto- 
mated translation of medical language data Methods of Information in 
Medicine 2^: 38-46. 1981, 

Gnmes. A. M.. Mueller, H G. and Malley. J, D.: Examination of binaural 
amplification in children. Ear and Hearing (in press). 

Hirschman. G. H,. Wolfson, M., Mosimann, J. E,. Clark, C B.. Dante. M. L., 
and Wineman. R. J.: Complications of dialysis. Clinical Nephrology 
15:66, 1981 

Hutchinson. G : A complete logic for n-permutable congruence lattices. 
Algebra Universalis (in press). 

Knott. G. D.: Procedures for managing extendible array files Software 
Practice and Experience 11: 63-84, 1981. 

Malley. J. D.: Simultaneous confidence intervals for ratios of normal means. 
Journal of The American Statistical Association (in press), 

Minker. J., and Minker. R. G.: Optimization of Boolean Expressions-Histori- 
cal Developments, Annals of the History of Computing 2: 227-238. 
1980 

Mosimann. J. E.. and Malley, J. D,: The Independence of Size and Shape 
Before and After Scale Change, In Taillie. C. Patil. G P . and Baldes- 
sari. B. (Eds): Statistical Distributions in Scientific Work, Vol 4. Models, 
Structures and Characterizations Dordrecht, Holland. D Reidel Pub- 
lishing Co (in press). 

Norton. L M : A note about Laplace transform tables for computer use. 
SIGSAM Bulletin 14: 30-31. 1980. 

O'Connor. M A : Invariant metrics on cones. Proc of the Conference on 
Invariant Metrics and Holomorphic Maps. Rome. Italy. Istituto di Alta 
Matematica F. Seven di CNR, (in press) 

Ratnaparkhi. M V : Some bivanate distributions of (X.Y) where the condi- 
tional distnbution of Y. given X is either beta or unit-gamma In Taillie, 
C. Patil. G. P . and Baldessari. B (Eds ). Statistical Distributions in 
Scientific Work. Vol 4. Models. Structures and Characterizations. Dor- 
drecht, Holland. D Reidel Publishing Co (in press). 

Roux. J. J J . and Ratnaparkhi. M V On matnxvanate beta type I distribu- 
tion and related charactenzation of Wishart distribution. In Tail'ie. C, 
Patil. G, P., and Baldessari. B. (Eds ) Statistical Distributions in Scientif- 
ic Work. Vol. 4. Models. Structures and Charactenzations. Dordrecht, 
Holland. D, Reidel Publishing Co. (in press). 



13 




Harold Ostrow, a CSL engineer, contributed to tine 
development of the heart probe, which allows non- invasive 
study of cardiac function. 

Dr. Benes Trus (CSL) and Dr. Alasdair C. Steven (NIADDK) 
collaborated to study the molecular organization of beet 
necrotic yellow vein virus (Virology, in press). Left: original 
electron micrograph; middle: computer reconstruction; right: 
computer model. 




14 



Computer Systems Laboratory 



Alan M. Demmerle, Chief 



Function and Scope of Work 

The Computer Systems Laboratory, a group of about 
30 professionals representing the disciplines of 
engineering, computer science, medicine, and 
chemistry, is the major source of expertise at NIH for 
minicomputer and microcomputer technology. CSL 
engineers and scientists, in collaboration with NIH 
laboratory and clinical investigators, apply this 
technology in the areas of laboratory automation and 
patient care. Most work supports intramural research 
programs, although some projects have been 
undertaken with extramural programs and, more 
rarely, with other Federal agencies. In addition to 
supporting ongoing research, CSL also investigates 
new applications of computers to biomedical 
research and identifies appropriate technology for 
use in these applications. 

Small computers that can be used in the laboratory 
or at the patient's bedside are increasingly important 
to biomedical and clinical research because of more 
complex research methods and instrumentation. 
Laboratory automation, data acquisition from 
analytical instruments, and patient monitoring are 
examples of functions to which CSL has applied 
small computers at NIH. Computers may be used 
only in an adjunctive manner, for example, as a more 
convenient means to acquire laboratory data; or they 
may be integral parts of an elaborate instrumentation 
system such as the computer-controlled mass 
spectrometer. 

Recent technological advances are contributing to 
changes in the nature of CSL's work. Foremost 
among these advances are developments in large 
scale circuit integration (LSI)--the microprocessor 
revolution--that have brought about the 
miniaturization of computer components and a 
dramatic decline in their prices and power 
requirements. One result has been a greater 
availability of computers for biomedical research. 
Thus, CSL engineers are now able to use 
microprocessors to solve problems that once were 
avoided because of cost, size, or manpower 
constraints. 

The Laboratory brings together professionals with 
diverse backgrounds to apply computer technology 
effectively to NIH programs and to offer research 
and development capabilitites responsive to NIH 
needs. Engineers, computer scientists, and 



mathematicians evaluate and apply new electronics 
and computer technology to solve biomedical 
problems. Personnel with backgrounds in medicine, 
biology, and chemistry communicate effectively with 
biomedical investigators and clinicians and identify 
potential computer applications. This multidisciplinary 
approach aids the recognition of problem areas that 
will benefit from automation and the interpretation of 
research needs in terms of computer methods. 

This year, CSL engineers and scientists worked on 
approximately 28 projects, representing collaboration 
with almost all of the Institutes. Many of these 
projects were continued from previous years; 9 
projects were begun this year and there were 
significant changes in scope in four others. Projects 
range in size from consulting activities of a few days 
or weeks duration to large scale efforts taking a 
number of man-years. Because much work involves 
the development of new methods or technology, or 
is influenced strongly by the changing needs of 
research, it is often difficult to predict the long term 
scope of a project. 

Fiscal Year 1981 Highlights 

During the past year the scope of CSL computing 
has broadened considerably. Modest computing 
requirements posed by individual scientists, or local 
groups of investigators, are being met increasingly 
through the use of microcomputers, while at the 
other end of the computational spectrum CSL is 
involved in the design, development, and 
management of large minicomputer systems, 
including, for example, those directed toward 
providing general image processing services to a 
significant segment of the NIH community. 

Two projects that typify the exploitation of 
microcomputer technology in clinical and laboratory 
settings are, respectively, the Automated Pulmonary 
Physiology Testing Project and the Molecular 
Interaction Laboratory Data System. The Automated 
Pulmonary Physiology Testing Project utilizes a 
Digital Equipment Corporation DECLAB MINC 11/03 
microcomputer system in support of pulmonary 
dysfunction diagnostic testing. Analog signals are 
rapidly acquired, digitized, and analyzed, providing 
the physician with intermediate results that can be 



15 



used to determine the course of further testing. 
Reports suitable for the medical record are produced 
locally and, together with any raw data desired, 
archived in machine readable form for retrospective 
analysis. These data will eventually be incorporated 
into the Pulmonary Branch data base being 
developed by DCRT's Data Management Branch on 
the NIH IBM 370 computers. Telephone data 
transmission to either the IBM 370 or the 
DECsystem-10 computers of the NIH Central Facility 
will then afford the Pulmonary Laboratory offline 
access to a full array of scientific, mathematical, and 
database manipulation functions. Specific procedures 
already implemented include static compliance and 
inspiratory muscle strength. Treadmill stress testing, 
dynamic compliance, and work of breathing are 
currently being added to the system repertoire. 
Future plans anticipate the establishment of breath- 
by-breath studies and closed-loop exercise 
procedures where such parameters as patient heart 
rate may be held constant by varying treadmill speed 
and grade. Use of this system has resulted in 
decreased time and increased accuracy and quality 
of the procedures performed, thus benefitting patient 
and physician alike 




CSL collaborative efforts put minicomputers and 
microcomputers to work for a variety of scientific needs 
throughout NIH. 

The recently completed Molecular Interaction 
Laboratory Data System features a Digital Equipment 
Corporation PDP-1 1 /03 microcomputer system 
interfaced to a Beckman Model E analytical 
ultracentrifuge and a Gary Model 61 circular dichroic 
spectropolarimeter owned by the Molecular Disease 
Branch, NHLBI. Ultracentrifuge data are conditioned, 
converted, formatted, and graphically displayed on a 



video display terminal. A system operator selects 
data for further processing and creates a file from 
these data, and from data required for analysis 
functions that have been previously entered into a 
log file. This file of preprocessed data is then 
transferred to the NIH DECsystem-10 where 
preestablished command sequences, invoked under 
the modeling laboratory program MLAB, are used to 
compute molecular weights and to assess molecular 
interactions. Circular dichroism spectra generated by 
the Cary 61 spectropolarimeter are digitized and high 
frequency noise is removed by digital filtering 
techniques. Spectra may be added, subtracted or 
averaged prior to transfer to the NIH DECsystem-10 
for further analysis. Typically the samples analyzed 
by this technique are proteins that the investigator 
wishes to characterize with respect to their three- 
dimensional conformations under various conditions. 
Using recently published data on the relationship 
between circular dichroism spectra and conformation 
in terms of the contributions to its structure by the 
four basic types of tertiary structure: helix, beta 
sheet, beta turn, and random coil. 

A significant increase in computational complexity 
over that represented by microcomputer projects is 
found in long-term CSL support for the flow 
microfluorimetry/cell sorter users at NIH. This work 
started as a minicomputer-based project to acquire, 
display, and analyze data from a Becton-Dickinson 
flow microfluorimeter (FMF) and a Los Alamos 
Scientific Laboratory FMF. Successful system 
development and operation has led to reproduction 
of the system for a number of new users and to 
incorporation of microcomputer technology to meet 
expanded requirements. The basic minicomputer 
system is implemented on a Digital Equipment 
Corporation (DEC) PDP-1 1/34 computer and 
features CSL instrument interfaces and an extensive 
software package that runs under DEC's single user 
RT-1 1 operating system. During the past year a 
system was installed for NHLBI and two were 
installed for NCI, thus bringing to five the number of 
systems supported by CSL. Copies of the CSL 
design are also being used by several organizations 
outside of NIH. 

To facilitate the simultaneous acquisition and 
processing of FMF/cell sorter data in a high 
workload environment, development is underway to 
replace the DEC RT-1 1 operating system with the 



16 



DEC RSX-11M multiuser operating system. An LSI- 
11 microcomputer is being introduced into the 
system to remove the data acquisition task from the 
PDP-11/34 computer. Besides providing independent 
operator interaction during parameter entry and 
acquisition phases, the microcomputer will produce a 
permanent hardcopy of the 'laboratory notebook.' It 
is planned that only completed data files will be 
transferred across the direct memory access link 
between the LSI-11 microcomputer and PDP-11 /34 
computer. 

An increasing number of scientific projects at NIH 
produce visually-based experimental data such as 
electron micrographs, stained tissue sections, gel 
electrophoresis, autoradiographs, etc., which require 
objective, accurate, quantitative analysis. Utilizing 
DCRT's powerful DEC PDP-11/70 -driven Evans and 
Sutherland Image Processing System as the key 
resource, CSL has, for the past three years, 
collaborated with scientists of several Institutes in a 
variety of image processing projects. Although these 
projects were directed toward specific goals, an 
important supplementary aim was the creation of 
general purpose software easily adaptable to a wide 
range of image processing needs. This latter role 
has expanded significantly during FY81 as CSL was 
assigned a major role in managing the Evans and 
Sutherland Image Processing System. 

Current CSL involvement with the Evans and 
Sutherland Image Processing System is focused 
primarily on the development of biomedically- 
oriented image processing software, complemented 
by the support necessary to make such packages 
usable to NIH scientists. Typical programs included 
PIC, CINT, and (VIONKEY. The PIC software package 
facilitates examination of collagen structure from 
electron micrographs, performs quantitative analysis 
of one-dimensional gels, and elucidates virus 
structures. CINT automatically locates and integrates 
up to 1200 spots on a two-dimensional gel picture 
and provides limited analysis of autoradiographs. 
MONKEY is a general image enhancement package 
that can be used to evaluate all or part of an image 
and includes operations such as smoothing, 
sharpening, statistics, and linear combinations. The 
Evans and Sutherland System supports both 
molecular graphics and image processing users. 
Demand upon the system has escalated to such a 
degree that severe scheduling problems exist not 



only during prime time but throughout most of the 
evening period as well. Many image processing 
problems, moreover, are not amenable to solution on 
the powerful but relatively dated Evans and 
Sutherland System. CSL is therefore in the process 
of designing and procuring a new image processing 
system. This system will consist of a powerful 32-bit 
computer with a mixture of medium and high 
resolution video displays. The microdensitometer 
associated with the Evans and Sutherland System 
will not be required by molecular graphics users and 
will be used to provide precise digitization of x-rays, 
micrographs and other images. The computer and its 
peripherals have been purchased and delivery is 
anticipated in FY82. Procurement of the display 
subsystem is being held up pending the availability of 
funds. 

Continuing in its more traditional role CSL has 
pursued a number of collaborative image processing 
projects. A new project, undertaken with scientists 
from the NEI and Harvard Medical School, is 
track cataract disease history. Images produced by 
split-lens lamp photography are digitized by means 
of microdensitometry and entered into the computer 




CSL engineers and scientists evaluate and apply electronics 
and computer technology to solve biomedical problems. 



concerned with measurement of opacity of the 
human eye lens along the visual axis in order to 
for analysis. Present work is oriented to determine 
the effects of variables-such as camera type, 
photographer, photographic processing, and 
microdensitometry factors-on expenmental 
accuracy. 



17 



Another project new this year focuses the combined 
power of Computer Assisted Tomography (CAT) and 
Positron Emission Tomography (PET) upon a study 
of brain cell metabolism and its relationship to 
disease associated with aging. Conducted in 
conjunction with NIA scientists, the project's initial 
goal is to delineate brain substructures represented 
in sequential CAT scan images and to determine 
metabolic activity in these substructures from 
corresponding sequential PET scan images. 

Ongoing image analysis work with the NHLBI 
includes topographic analysis of arterial 
atherosclerosis formation and graphical 
representation of myocardial blood flow. Further 
refinement of the block model used by NINCDS 
scientists to provide three-dimensional graphic 
representation of the neuronal structure of the cat 
brain stem now make it useful for arbitrary plane 
viewing of other spatially-sequenced images. 

Good R&D programs require good tools. For many 
years CSL has had an electronics laboratory. More 
recently, a computer laboratory has been added. The 
computer laboratory facilities include 'development 
systems'-minicomputer and microcomputer 
systems--that are used to investigate and implement 
new software and hardware for eventual use in 
target computer systems, i.e., systems intended for 
use in specific projects. These development systems 
permit work to begin on a project long before the 
target system ordered for that project is delivered, 
thus improving productivity and reducing the time 
until project completion. Development systems are 
indispensable to microprocessor projects, because 
software development work is often not possible on 
the microprocessor hardware itself. The computer 
laboratory facilities were expanded again this year so 
that more people can use the development systems 
at the same time. Virtually all minicomputer and 
microcomputer work can now be done using 
development systems. 

CSL is called upon for consultative assistance in its 
areas of expertise by both the intramural and 
extramural programs. Usually such assistance in 
limited to providing expert advice--the conventional 
definition of consulting. Occasionally however, a 
consulting role may lead to engineering or software 
work, or even to an extensive project. During the 



past year consultation activities with the 
Rehabilitation Medicine Department of the Clinical 
Center led to a survey of the state-of-the-art in gait 
laboratory automation. Subsequent to this survey, 
requirements, possible design approaches, and costs 
for automating the NIH Gait Laboratory were studied. 
A collaborative project culminating in the 
computerization of NIH Gait Laboratory activities 
appears imminent. 

The Computer Systems Laboratory continued its 
support of the annual international conference, 
Computers in Cardiology. This conference provides a 
forum for direct interaction and exchange among 
physicians, computer scientists, and engineers who 
use computers to assist research or clinical care in 
the field of cardiology. 

Future Plans/Trends 

In common with many organizations, CSL can expect 
FY82 and beyond to be clouded by many of the 
uncertainties that were encountered this year 
concerning resources. Technological developments 
in large scale circuit integration promise an increase 
in the number of biomedical applications for which 
microcomputers and minicomputers can be useful 
and affordable; however, anticipated budget 
reductions combined with an unpredictable rate of 
inflation may well offset these advances. The 
proposed new Image Processing System poses a 
striking example of the potential conflict. Hailed just 
one year ago as a project that owed its feasibility to 
technological progress and declining costs, it has 
remained virtually dormant during the past year, a 
hostage to the unavailability of funds. 

An expected continuation of stringent personnel 
restrictions serves to further obscure a precise 
forecast of future directions. Limitations in staffing 
tend to encourage an escalation in demands for 
automation; however, all new NIH computer 
initiatives will have to be reconciled with uncertain 
funding and uncertain CSL support. Moreover, CSL 
has for many years required Institutes to provide 
some limited computer expertise as a condition for 
assistance. At a time when CSL is fully extended, 
the commitment of Institute personnel assumes 
greater than usual significance; participation by the 
Institutes will prove difficult as so many computer 



18 



projects permit increasingly complex research 
activities rather than conserve manpower. An 
excellent illustration of staffing problems can be 
seen in the NIH Library Project. Automation of the 
Library has become imperative if an exploding 
volume of literature is to be managed by a fixed or 
shrinking Library staff. Nevertheless, after 
languishing for one year in a CSL project queue 
awaiting the availability of appropriate personnel, the 
project has suffered another equally lengthy delay 
due to the existence of a similar staffing plight in the 
NIH Contracts Department. 

The substantial uncertainties and delays that have 
plagued many recent projects make effective use of 
a limited CSL staff an imposing challenge. A number 
of actions adopted to confront the resource 
problems appear to hold promise and may become 
trends for the future. 

A contract to provide software support has been 
established and already two projects are expected to 
benefit from this type of support. Provided that 
adequate funding can be maintained, the contract 
mechanism offers an attractive supplement to CSL 
programming capability, particularly for well defined 
projects. 

As was the case during the past year, continued 
heavy involvement in providing consultation services 
can be expected for the future. During the past year, 
for example, substantial specialized advice was 
given to the National Toxicology Program, NHLBI 
Framingham Longitudinal Study, Clinical Center 
Department of Critical Care Medicine, Clinical Center 
Nuclear Medicine Department, NHLBI Surgery 
Branch, NHLBI Cardiology Branch, Clinical Center 
Department of Rehabilitation Medicine, and NCI 
Surgery Branch. Although some consultations have 
been, and undoubtedly will continue to be, 
substitutes for actual work on indefinitely delayed 
projects, consultation services can offer, in 
appropriate circumstances, substantial rewards 
relative to I'mited personnel investment. 

Finally CSL will continue to deploy staff on those 
projects promising maximum impact to the NIH 
Community. Current examples of this policy are the 
Flow Microfluorimetry Project with its duplicated 
systems and the Image Processing Project with its 
general purpose developments. 



Publications and Presentations 

Allen, S.. Songco, D.. Plexico. P.. and Mortord. R : A Voice Output Module 
Developed tor a Blind Programmer. Journal ol Visual Impairment and 
Blindness. New York, American Foundation lor the Blind, April 1981. 
pp. 157-161 

Bacharach, S , Green, M , Bonow. R . Findley, S , Ostrow, H., and Johnston. 
G Measurement ol Ventricular Function by ECG Gating During Atrial 
Fibrillation. J. Nuclear Med 22226-23 1,1 981, 

Bacharach, S,, Green. M , Ostrow. H . and Johnston. G.: Developments in 
Nuclear Medicine's Computer System-Applications to Cardiology IEEE 
Transactions in Nuclear Science 22:1095-1 101.1980, 

Barrett, W.: An Interactive Algorithm lor Multiple Threshold Detection. IEEE 
Proceedings. Pattern Recognition and Image Processing (in press) 

Barrett, W,, DeLeo, J.. Cornhill, J , and Fry, D.: A System lor Automated 
Analysis ol Plaque Formation in Expenmental Atherosclerosis 53rd 
Scientilic Sessions ol the American Heart Association, November 19, 
1980 

Computer Systems Laboratory of the National Institutes of Health. NIH 
Publication No, 81-1926, June 1981, 12pp. 

Green, M,, Ostrow, H.. Bacharach, S.. Allen, S., Bonow, R., and Johnston. 
G.: Real-time Scintillation Probe Measurement ol Lett Ventncular Func- 
tion. Proceedings ol the Meeting in Freiburg (in press). 

Hauser, S , and Almeida, A.: A Control and Data Processing instrument lor 
Kidney Tubule Research Biomedical Sciences Instrumentation 1713- 
19, 1981. 

Marlino. R., Kempner, K., McClellan. J., and McLees. B.: Automation ol a 
Medical Intensive Care Environment with a Flexible Conliguration ol 
Computer Systems. In O'Neill, J,T {E6.):Proceedings ol the Fourth 
Annual Symposium on Computer Applications in Medical Care. New 
York, IEEE, 1980, pp. 1562-1568. 

Nadel, L.: Automated Pulmonary Analysis by an On-line Microcomputer In 
Nair, S (Ed): Proceedings of the Conference on Computers in Cntical 
Care and Pulmonary Medicine (in press). 

Nikodem, V., Trus. B.. and Rail. J.: Two-dimensional Gel Analysis ol Rat 
Liver Nuclear Proteins alter Thyroidectomy and Thyroid Hormone 
Treatment Proc. Natl. Acad Sci USA (in press), 

Padikal, R., Lichter, A., Tepper, J,. Gladstein. E.. Schwade. Jr.. Fredrickson, 
H.. Risso, W., Roberson, L, Van de Geijn, J , and Kinsella, T.: Expen- 
ence with a CT Based Treatment Planning System. In O'Neill. J.T. 
(Ed.): Proceedings ol the Fourth Annual Symposium on Computer Ap- 
plications in Medical Care. 1980, pp. 83-88, 

Plexico, P.: Microcomputer Application in Biomedical Research, Association 
lor the Advancement of Medical Instrumentation 14:307-310.1980, 

Powell, J,, Fico. R-, Jennings. W., O'Bryan, E , and Schultz. Jr.. A.: A Local 
Network lor Distributed Laboratory Microcomputer Proceedings ol the 
Twenty-first IEEE Computer Society International Conference. 1980, pp. 
185-190. 

Steven, A., Trus. B.. Putz. C, and Wurtz, M.: The Molecular Organization ol 
Beet Necrotic Yellow Vein Virus, Virology (in press). 

Trus, B., and Steven, A,: Digital Image Processing ol Electron Micrographs- 
The PIC System. Journal of Ultramicroscopy (in press). 



19 





In collaboration with the NIH Nuclear Medicine Department 
LAS researcher Margaret Douglas is producing diagnostic ' 
images (see insert) that allow scientists to detect motion , 
defects in the walls of the heart. i 

Special Achievement Awards were earned by Martha 
Norton and Margaret Douglas for their contributions toward 
developing a text- and command-entry technique for 
computerized typesetting printing systems. The awards 
were presented by Dr. Eugene Harris, Chief, LAS 



20 



Laboratory of Applied Studies 



Eugene K. Harris, Chief 



Functions 

The Laboratory of Applied Studies (LAS) has three 
main purposes: 

1. in collaboration with biomedical scientists, to 
apply mathematical theory and computing science to 
the construction, testing, and improvement of 
mathematical models of physiological processes-- 
particularly reaction-diffusion kinetics, transport of 
substrate to tissues, and the control of metabolism 
within cells and tissues; 

2. In collaboration with clinicians, to develop and 
apply mathematical or statistical theory and special- 
purpose computing procedures (analog or digital as 
required) to facilitate research projects aimed at 
improving the diagnosis of disease and assessment 
of treatment; 

3. to engage in independent research in applied 
mathematics, statistics, and computer systems 
necessary to provide a sound theoretical basis for 
collaborative studies, and to insure that state-of-the- 
art mathematical and computational methods are 
available as research tools at NIH. 

Two sections carry out these primary LAS functions: 

Applied Mathematics Section-AMS-CJohn E. 
Fletcher, Ph.D., Chief). This staff of five includes 
specialists in applied mathematics, computer 
science, biomathematics, and medicine. 

Medical Applications Section -MAS-(James J. 
Bailey, M.D., Chief). This five-member staff includes 
physician-scientists, electronic engineers, and 
computer systems analysts. 
The Chief, LAS, is a biostatistician with training in 
public health and the basic medical sciences. 

Scope of Work 

The Laboratory of Applied Studies works on projects 
in basic and clinical biomedical science. Largely, 
these involve collaboration with other groups at NIH, 
elsewhere in the U.S.A., and abroad. The 
collaborating investigators this year included: 

• biochemists and pharmacologists at NIH, at the 
Medical College of Virginia, and at other 
universities in the U.S.A. and in France working 
on models for receptors of drugs or other 
ligands, on the kinetics of enzymes in 
membranes and on other problems in tissue 
metabolism 



• physiologists and chemical engineers in the 
U.S. A and Europe studying the transport of 
substrate within the microcirculation and the 
regulation of tissue perfusion 

• clinicians in the cardiology, pulmonary, and 
hematology branches of the NHLBI; in the 
arthritis and rheumatism branch of the NIADDK; 
and in the medical intensive care unit and the 
departments of diagnostic radiology and 
diagnostic imaging of the Clinical Center 

• clinical chemists and pathologists at NIH 
(Clinical Pathology Department, Clinical Center) 
and elsewhere in the U.S.A., in Europe, and in 
Japan engaged in the collection and study of 
reference values in laboratory medicine 

• electrocardiologists and biomedical engineers in 
the U.S.A., Canada, and Europe concerned with 
improved algorithms for computer-based 
interpretation of ECG's and evaluation of ECG 
interpretative programs. 

During FY81 LAS staff members participated in 
various teaching and consulting, or advisory, 
activities. 

J. Fletcher continues to serve as Chairman of the 
fvlathematics and Computer Science Departments, 
Foundation for Advanced Education in the Sciences. 

J. Bailey continues as a member of an NHLBI site- 
visiting team concerned with computer analysis of 
exercise ECG's. He also serves as consultant on 
common standards for quantitative 
electrocardiography for an EEC-sponsored program 
in medicine and public health. 
E. Harris continues to be a consultant in applied 
statistics to the Food and Drug Administration's 
Division of Medical Devices and Diagnostic Products. 
Dr. Harris also serves as consultant statistician to 
the College of American Pathologists and to the 
International Federation of Clinical Chemistry (Expert 
Panel on the Theory of Reference Values), and is a 
member of the Board of Editors of Clinical 
Chemistry. 

During this reporting year, Dr. Adelin Albert, a 
mathematical statistician from the University of 
Liege, Belgium, joined the laboratory as a Fogarty 
International Research Fellow under Dr. Harris' 



21 



ALKALINE PHOSPHATASE 



Charts like this one, which illustrates blood chemistry 
differences in healthy individuals, aid Dr. Harris in his 
collaborative studies. 



preceptorship. Dr. Albert is developing and testing 
appropriate statistical techniques to utilize 
multivariate, serial data for optimal prediction of 
patient outcomes. 

Highlights of Year's Activities 

This year has seen substantial progress in all active 
LAS collaborative research projects, based in many 
cases on the technical advances reported last year 
in developing and implementing various computer 
systems. Some major accomplishments this year 
include: 

• A joint project with the Pulmonary Branch and 
the Clinical Hematology Branch, NHLBI on the 
computerized analysis of pulmonary gas 



exchange in normal volunteers and selected 
patient groups has been accelerated through the 
efforts of R. Burgess and E. Pottala of the 
Applied Mathematics and Medical Applications 
Sections. Complete redesign of analog circuitry 
has produced a reliable gas analysis system. 
The computer and computer-controlled exercise 
testing equipment has been specified, ordered, 
and, in large part, received. Dr. Burgess has 
also substantially increased the precision and 
sample size of the renal scintigraphy study on 
dogs, initiated during the prior reporting year. 

• A major advance in the mathematical analysis of 
blood and substrate supply to tissue has been 
initiated by J. Fletcher (Chief, AMS). Incorrect 
solutions found in the existing literature for 
mathematical models of perfused tissue 
experiments have been corrected. The correct 
solutions to these and more general models are 
now being explored and will be used to predict 
distributed substrate levels for comparison with 
experimentally obtained microelectrode 
measurements. 

• A comparative study of two ECG computer 
programs, based on ECG-independent clinical 
documentation of the presence or absence of 
various forms of heart disease in 284 patients 
has been completed and published by J. Bailey 
(Chief, MAS), M. Horton (MAS), and 
cardiologists at the Royal Infirmary in Glasgow. 
This study should provide a model for future 
evaluations of widely-used ECG programs 
because of the careful clinical documentation 
and the methods used to categorize ECG 
diagnostic output statements, allowing 
recognition of semantic equivalents despite 
differences in the terms used in different 
programs. The LAS investigators have extended 
their studies to the evaluation of serial 
computer-based ECG interpretations in a joint 
project initiated this year with Dr. D. Savage of 
the Framingham Heart Study, NHLBI. 

• The PICTUR image processing computer 
package, developed and implemented last 
reporting year by M. Douglas (MAS) has been 
extensively applied, in collaboration with Dr. J. 
Costa (NIMH), to determination of the spatial 
distribution of fluorine within the dense bodies of 



22 



blood platelets. If continuing work proves 
successful, this fluorine-dense body system 
could become a reliable model for intracellular 
monitoring of therapeutic drugs in patients. 
• M. Douglas and M. Horton (MAS) shared a 
DHHS special service award for their work with 
N. Crawford and V. A. Parsegian (Physical 
Sciences Laboratory) on computerized 
typesetting of scientific papers. 

Future Plans 

Installation of the computer-based gas analysis/ 
exercise system will be completed and testing of 
normal volunteers will begin in cooperation with the 
pulmonary and clinical hematology branches of 
NHLBI. The scintigraphic studies of renal 
hypertension in dogs will be completed and the 
results will be prepared for publication. Expected 
progress in other active projects will include: 

--application of generalized mathematical models 
to experimental data on organ perfusion with the 
goal of identifying the ranges of critical parameters 
controlling organ response to physiologic challenge, 

-expanded collaboration with NIH laboratory 
scientists on application of network simulation 
programs to improving the understanding of 
molecular transport across membranes, 

"development of appropriate physical and 
statistical theory to support the continued study of 
intracellular distributions of physiologically important 
molecules through electron beam analysis, and 

-completion of comparative statistical analysis of 
multivariate and univariate time series applied to 
clinical laboratory data, using both real and simulated 
results of patient monitoring. 



Publications and Presentations List 

Bailey, J J : The future ol gold standards and computenzed electrocardio- 
graphy In Tolan, GD, and Pryor. T A (Eds): Computerized Interpreta- 
tion of the ECG. V. New York, Engineering Foundation, 1980, pp. 229- 
233. 

Bailey J J.. Berson, AS., Jackson, L.K., Milliken, J.A., Stevens, J.M., Tolan, 
GO. and Wolf, H.K.: Evaluation Methodologies for ECG diagnostic 
systems In Bonner. RE, and Pryor, T A (Eds ): Computerized Inter- 
pretation of the ECG. VI (in press). 

Bailey, J J , and Harris, E,K ; Evaluation of ECG interprelalionTrulh versus 
beauty. In Tolan, GD., and Pryor, T,A. (Eds): Computerized Interpreta- 
tion ol the ECG. V New York, Engineering Foundation, 1980, pp. 179- 
182 

Bailey, J J , and Norton, MR: Type A electrocardiogram data bases: Pur- 
pose and development In Wolf. H K , and Macfarlane, P.W. (Eds): 
Optimization ol Computer-ECG Processing. New York, Norlh-Holland 
Publishing Company, 1980, pp. 189-195. 

Berk, P.D., Blaschke, T.F., Shrager, R.I,, and Waggoner, J.G.: Phenobarbitol 
does not increase hepalichema turnover in man. Gastroenterology 
79:1004, 1980. 

Bunow, B : Cellular Enzymology: Effect of compartmentation on steady 
stale kinetics J. Theor Biol 84: 611-628, 1980 

Bunow, B.: Turing and the physico-chemical basis of biological patterns. In 
Prewitt, J (Ed): IEEE Turing Memorial (in press). 

Bunow, B., Kernevez. J.P , Duban, MC, Jolly, G,, and Thomas, D,: Pattern 
formation by reaction-diffusion instabilities: Application to morphogene- 
sis in Drosophila, J. Theor Biol 84 629-649, 1980, 

Bunow, B,, and Mikulecky, DC : On the feasibility of using flux meas- 
urements to distinguish among active transport models. Polish Winter 
School of fi/lembrane Transport (in press). 

Douglas, M.A., Hui, S,W,, Costa, J,L,, and Bailey. J, J,: Computerized proc- 
essing and subtraction of energy-filtered electron images as an aid to 
elemental analysis. In Bailey, G,W (Ed): Proceedings of the Thirty- 
eighth Meeting EMSA. Baton Rouge, Claitor's Publishing Division. 
1980, pp 128-129, 

Fletcher, J,E , and Jolly, M,: The Compulation of Substrate Levels in Per- 
fused Tissues, Proceedings of the Annual SIAM Meeting (in press), 

Fletcher, J,E,. and Schubert, R.W,: The Theoretical Prediction ol Substrate 
Levels and Their Histograms in Cell Free Perlused Tissues. Proceed- 
ings of the International Meeting of the Society of Oxygen Transport to 
Tissue (ISOTT) (in press), 

Harris, E,K,: Further applications of lime series analysis to short series of 
biochemical measurements. Proceedings of Workshop on Reference 
Values in Clinical Pathology (m press), 

Harris, E,K,: Regression, least squares and correlation. In Seligson, D., M.D. 
(Ed): Handbook ol Clinical Chemistry (in press) 

Harris. E.K.: Statistical aspects ol relerence values in clinical pathology. In 
Stefanini, M.. and Benson, E. (Eds.): Progress in Clinical Pathology. 
New York, Grune & SIralton. Inc., 1981, Vol VIII, pp. 45-66. 

Harns. E.K,: Use ol statistical models to detect subject-specific changes. 
Proceedings of International Conference on Automated Multiphasic 
Health Testing & Services (in press). 

Kernevez. J.P,, Jolly, G,, Thomas, D,, and Bunow, B,: Pattern formation and 
wave propagation in the S-A system In Bardos, C. Lasry, J,M . and 
Schatzman, M. (Eds.): Lecture Notes in Mathematics 782: 201-221, 
Spnnger-Verlag, 1980. 

Macfarlane, P.W., Chen, C.Y., and Bailey, J. J.: A comparison of point 
scoring techniques for the diagnosis of LVH. In Macfarlane, P.W. (Ed.): 
New Frontiers in Eleclrocardiology (in press). 

Macfarlane. P.W., Melville, D,l . Horton. MR., and Bailey, J. J.: Comparative 
evaluation of the IBM (12 lead) and Glasgow Royal Infirmary (3 ortho- 
gonal lead) ECG computer programs. Circulation 63 354-359. 1981. 

Sharan, M.. Kernevez. J.P,, and Bunow, B,: On numerical exploration of 
bifurcating branches of solutions in reaction diffusion equations model- 
ing enzymatically active artificial membranes Research report of the 
Department of Applied Mathematics. Compiegne, France, University of 
Technology of Compiegne, 50 pp. 



23 



%'n.-. t 




Nancy Crawford shared an award with LAS members for 
her work with Dr. Adrian Parsegian on the computerized 
typesetting of scientific papers. 

PSL carries out research to understand biological 
phenomena in terms of physics and chemistry. 



LS 


lI^I 


B^l 


Pi 


Mm^ 




t- 


■kflHV^vKa^lHH 


^ 1 



24 



Physical Sciences laboratory 



George H. Weiss, Chief 



Function 

The Physical Sciences Laboratory has three principal 
functions: 

• to carry out research in the physical sciences in 
order to understand biological phenomena in 
terms of physics and chemistry 

• to develop theory and practical instrumentation 
for biomedical experiments, and in particular to 
relate these to the capabilities of modern 
computer technology 

• to provide consulting services to other scientists 
at NIH in physics, theoretical chemistry, and 
several fields in applied mathematics. 

The staff of the Physical Sciences Laboratory 
consists of seven professionals who work in the 
areas of general biophysics, nuclear magnetic 
resonance, applications of light scattering techniques 
in biomedical experiments, the physical chemistry of 
polyelectrolytes, and problems in applied 
mathematics. 

Scope of Work 

The Physical Sciences Laboratory has a combined 
program of research projects internal to the 
laboratory and collaborative projects with scientists 
at NIH and at other institutions. The collaborative 
projects, done jointly with approximately twenty-five 
other investigators, include two major projects with 
data being generated by off-campus scientists. 

Highlights of the Year's Activities 

Although a large amount of time was devoted to 
getting bugs out of the 360 MHz spectrometer, 
useful experimental information is now becoming 
available. The advantages of being able to do pulsed 
and two-dimensional Fourier transform spectrometry 
are being exploited by Dr. James Ferretti in several 
ongoing investigations. He is investigating cross 
relaxation pathways in rigid organic molecules. Dr. 
Ferretti is also developing both theoretical and 
experimental aspects of Fourier transform 
spectroscopy allowing the deduction of internuclear 
distances and rotational correlation times by the use 
of cross relaxation. He has also concluded an 
investigation of the relation between errors in 
chemical shifts due to random and digitization errors. 
Specifically Dr. Ferretti has contrasted several 



strategies for filtering and smoothing resulting data, 
finding optimal strategies for many cases of practical 
interest. 

Dr. Adrian Parsegian and his collaborators have 
made several significant advances in the 
understanding of intermolecular forces in proteins 
and nucleic acids. Methods similar to those 
developed earlier by Dr. Parsegian to measure 
membrane-membrane forces are now being applied 
to the study of the aggregation, gelation, and 
crystallization of proteins, and to the packing of 
parallel double helical strands of DNA. This project 
will continue with the development of theory and 
experiments to systematically collect thermodynamic 
data on gelation and crystallization. 

Dr. Ralph Nossal and his collaborators have 
performed experiments on experimental analogues 
of the flow of blood cells confirming an earlier 
theoretical development that allows one to interpret 
light scattering experiments in vivo. The combination 
of theory and experimental verification sets the stage 
for clinical applications of light scattering and in 
particular for the measurement of blood flow in 
capillaries, which has heretofore been impossible. 

A theory has been developed by Dr. Nahum 
Gershon of the PSL and Dr. B. Aizenbud of MIT for 
diffusion on curved surfaces, with particular 
reference to fluorescent photobleaching recovery 
experiments. The major result of this investigation is 
that curvature does not affect estimated diffusion 
constants to any significant degree. 

Dr. Parsegian was appointed Editor of the 
Discussions of the Biophysical Society and Dr. Weiss 
was appointed Mathematics Editor of the Journal of 
the Washington Academy of Sciences. 



25 




PSL develops theory and practical instrumentation for 
biomedical experiments. 



Future Plans 

With few exceptions, plans for the future involve 
building on scientific insights that have been gained 
in the course of PSL current research. Now that 
nnost of the bugs have been eliminated from the 360 
MHz spectrometer, the next hope is both to develop 
methodology for two dimensional Fourier transform 
spectroscopy and to apply this technique to specific 
structure and configuration problems for biologically 
interesting molecules. 

Studies of forces in membranes will continue with 
further analyses of the properties of sickle cell 
hemoglobin. Another project coming to fruition in the 
coming year will be a study of protein and nucleic 
acid contact using the molecular graphics crystal 
structures. These will enable investigators to vary 
configuration parameters to elucidate the contacts 
stabilizing protein dimers and tetramers. 



26 



Publications 



Aizenbud, B. M., and Gershon. N.D.: Hydrodynamic equations and VH light 
scattenng from viscoelaslic (solid-like and tluid-like) systems. Pheno- 
menological approach. Physica A (in press). 

Bonner, R., and Nossal, R.: A model (or laser Doppler measurements of 
blood flow in tissues. Applied Optics 20:2097-2108. 1981 

Brenner. S. L.. and Korn, E D.: Stimulation of actin ATPase activity by 
cytochalasins provides evidence for a new species of monomeric aclin. 
J. Biol. Chem. (in press). 

Ciarkowski, J. E., Ferretti. J. A., and Marshall, G. R.; Comparative conforma- 
tional studies of angiotensin II and two stencally constrained analogs 
by 600 MHz proton spectroscopy. J. Am. Chem. Soc. (in press). 

Egan. W.. Ferretti. J. A., and Marshall, G. R.: Relaxation parameters and 
motional properties in biological macromolecules. Bull Magnet. Reso- 
nance 2.\bA7 . 1981. 

Gershon, N.D., Smith, R. M., and Jarett, L.: Computer assisted analysis of 
ferritin-insulin receptor sites on adipocytes and the effect of cytochala- 
sin B on groups of insulin receptor sites. J. Membr. Biol. 58:155-160, 
1981 

Gupta, R. K., Ferretti, J, A., Becker, E. D., and Weiss, G. H.: A modified fast 
inversion recovery technique for spin-lattice relaxation measurements. 
J. Magnet. Resonance 38:447-452, 1981. 

Kiefer, J. E., and Weiss, G. H.: A comparison of tvro methods for accelerat- 
ing the convergence of Fourier series. Computers and Mathematics (in 
press). 

Lindenberg, K., Seshadri, V. E.. Shuler, K. E.. and Weiss. G. H.: Lattice 
random walks for sets of random walkers. J. Statistical Physics 23:11- 
25, 1980 

Us, L. J., Parsegian, V. A., and Rand, R. P.: Binding of divalent cations to 
dipalmitoylphosphatidylcholine bilayers and its effect on bilayer interac- 
tion. Biochemistry 20.M&\-\770. 1981. 

Lis, L. J., Lis, W. T., Parsegian, V. A., and Rand, R. P.: Adsorption of 
divalent cations to a variety of phosphatidylcholine bilayers. Biochemis- 
try 20:MT\-M77. 1981. 

Lis, L. J., McAlister, M., Fuller. N,. Rand, R. P., and Parsegian, V. A.: 
Interactions between neutral phospholipid bilayer membranes. Biophys. 
J. (in press). 

Lis, L. J., McAlister, M.. Fuller, N., Rand, R. P., and Parsegian, V.A.: 
Measurement of the lateral compressibility of several phospholipid bi- 
layers. Biophys. J. (in press). 

Mohr, J. P., Weiss, G. H., Caveness, W. F., Dillon, J, D., Meirowsky, A, M., 
and Rish, B. L.: Language and motor deficits following penetrating head 
iniunes in Vietnam. Neurology 30:1273-1279, 1980. 

Meirowsky, A. M., Caveness, W. F., Rish, 8. L., Dillon, J. D., Mohr, J. P., 
Kistler, J. P., and Weiss, G. H.: Cerebrospinal fluid complicating missile 
wounds of the brain. J. Neurosurgery Si:ii-i7 , 1981. 

Nossal, R.: Mathematical theories of topotaxis. In Jager, W., Post, H,, and 
Tautu, P. (Eds): Biological Growth and Spread. Mathematical Theories 
and Applications. Heidelberg, Spnnger-Verlag, 1980, pp. 410-440. 

Nossal, R.: Quasielastic light scattenng from polymer gels. In Chen, S. H.. 
Chu, 8., and Nossal, R. (Eds): Scattering Techniques Applied to Supra- 
molecular and Nonequilibrium Systems. Plenum Publ. Corp., New York 
(in press). 

Nossal, R., and Jolly, M.: Shear waves in cylindrical gels. J. Appl. Phys. (in 
press). 

Oppenheim, I.. Shuler, K. E.. and Weiss, G. H.: Stochastic processes. In 
Lerner, R. G., and Tngg, G. L. (Eds): Encyclopedia o1 Physics. New 
York, Addison-Wesley, 1980, pp. 964-967. 

Parsegian, V. A.: Forces between membranes approaching contact. Scandi- 
navian J. of Clinical Investigation 41:156, 39-47, 1981. 

Parsegian, V. A., Rand, R. P., and Slamatoff, J.: Perturbation of membrane 
structure by uranyl acetate labeling. Biophys. J. 33:475-478. 1981. 

Parsegian. V. A., and Weiss, G. H.: Spectroscopic parameters for computa- 
tion of van der Waals forces. J. Colloid Interface Sci. 81:285-289. 1981. 



Rish, B. L.. Dillon. J. D.. Caveness. W. F . Mohr. J. P.. Kistler. J. P.. and 
Weiss, G. H.: The evolution of craniotomy as a debridement technique 
for penetrating craniocerebral injunes. J. Neurosurgery 53:772-775. 
1980. 

Weiss. G. H.: Asymptotic form for random walk survival probabilities on 3-D 
lattices with traps Proc. Natl Acad Sci USA 77:1273-1274. 1980. 

Weiss. G. H,: Can one measure rate constants using chromatographic 
methods? Separation Sci 16 75-80. 1981 

Weiss, G- H.: First passage time problems for one dimensional random 
walks. J. Statistical Physics 24581-589, 1981. 

Weiss, G. H.: Note on lattice random walks with an excluded point. J. 
Mathematical Physics 22:562-563, 1981. 

Weiss, G. H.. and Darvey. I. G.: A note on the choice of substrate concen- 
tration in enzyme kinetics. J. Theoretical Biology 90:437-439. 1981 

Weiss. G. H.. Ferretti. J A., and Kiefer. J E : A study of precision in the 
measurement of chemical shifts. J. Magnet Resonance (in press). 

Weiss. G. H., and Shiesinger. M F.: On the expected number of distinct 
points in a subset visited by an N-step random walk. J. Statistical 
Physics (in press). 



27 





Data base management technology provides both online 
and batch processing to meet various operational 
accounting and management control needs of NIH. 

The Branch creates and maintains user-oriented tools, like 
RMAG, to speed building of programs and to improve 
operation of applications systems. 

DMB supports the NIH data base management system, 
which, when complete, will handle nearly all NIH materiel 
and financial transactions. 




y 




28 



Data Management Branch 



J. Emmett Ward, Chief 



Functions 

The Data Management Branch (DMB) provides 
advice and assistance to research investigators, 
progrann officials, and administrators throughout NIH 
In planning for and obtaining computer data 
processing services. In this role the branch is a 
central NIH resource for systems analysis, design, 
and programming. There are currently 47 permanent 
full time employees whose disciplines include 
computer science, mathematics, and statistics. 

Scope of Work 

DMB staff design and create computer-based data 
management systems that provide practical solutions 
to the unique mix of administrative, scientific and 
management data processing problems encountered 
at the NIH. Each new computer system user is 
provided comprehensive training in all system 
facilities and functions. In addition DMB staff teach 
courses about data management and DMB 
programming tools; they provide advice on data 
management techniques to NIH programmers; they 
serve as consultants to the B/I/D's for obtaining and 
monitoring contracting services for computer 
systems development; and they create and maintain 
general purpose, user-oriented programming tools to 
speed building and improve operation of applications 
systems. 

The DMB comprises five sections. The Applied 
Systems Programming Section (ASPS) and the 
Scientific Applications Section (SAS) provide 
general computer systems analysis and programming 
services for all of the B/I/D's. The ASPS supports 
general data management, and the SAS handles 
those projects which require scientific data analysis. 

The Data Base Applications Section and the Data 
Base Enhancement and Control Section develop 
and maintain the central administrative data base for 
NIH materiel and financial management. The Clinical 
Support Section develops and maintains the 
Clinical Information Utility as a data base for 
research and patient care in the Clinical Center. 



FY81 Accomplishments 

In the NIH Administrative Data Base system, during 
the first quarter, the demands for ad-hoc reporting 
and the significant increase in maintenance of the 
backup and general reporting functions for data base 
management required another shift of Branch 
personnel. The Software Support Section was 
abolished and its primary function was transferred to 
the Office of the Chief. Two people were transferred 
to the Data Base Applications Section to assume the 
maintenance and reporting functions. 
As a result of this transfer, major progress is being 
made to streamline the backup and maintenance 
functions and to standardize software for both ad- 
hoc and recurring reports. A commitment source 
data system, which: 

1. controls telephone charge agreements, basic 
purchase agreements, and indefinite delivery 
contracts 

2. monitors records of call and telephone charge 
orders against these agreements and contracts 

3. validates and summarizes fund and 
accounting information 

4. enables automatic generation of vendor 
identification and address on purchase documents, 

has also been installed on the administrative data 
base. This provides managers, purchasing 
personnel, and financial clerks with fully 
synchronized information on the status of all source 
commitment documents. 

A delegated procurement and receiving entry system 
was completed during the year and is now in the 
process of being installed on site in the B/I/D's. 
When it is completely installed, the movement of ail 
paper actions relating to delegated purchasing and 
receiving will end, and the computer data base will 
be the source for all information except the locally 
signed, legally required hard copy documents. 

In the accounts payable area, DMB has provided a 
full invoice inventory procedure which ties the 
invoice process to the purchase order. When orders 
are eligible to be paid, invoices are automatically 
displayed to the voucher examiner, preliminary 
payment schedules are electronically prepared and 
presented for review, and final schedules to Treasury 
are released through the computer system. The 
experience gained in working with staff of the 
Division of Financial Management (DFM) has also 



29 



led to the streamlining of manual and machine 
procedures which optimize the entire payment 
process. 

In the NIH Clinical Information Utility (CIU), 
development this year has concentrated on 
implementing the weekly cumulative laboratory 
summary and on reducing the size of the hard copy 
medical record. It became clear that producing 
complete retrospective summaries each week not 
only required an unacceptable level of elapsed 
computer time to produce, but also literally inundated 
the medical records staff with weekly volumes of old 
and new paper. To eliminate both problems, it was 
agreed that only permanent medical record 
replacements would be provided every six months. A 
sophisticated computer system was developed to 
provide this capability. This new system balances the 
volume of the permanent replacement paper flow 
and collates the permanent and weekly summaries 
in a way that assures ease of handling by medical 
records personnel. 

Further reduction of the medical record is now in the 
requirements analysis stage. Specific consideration 
is being given to: identification of information 
required legally and medically in the record, 
requirements for access and display of high and low 
use medical records data, and methods and modes 
of electronic, hard copy, and other alternative forms 
of storing medical records data. The final design of 
either a system or systems to accomplish a 
reduction in the bulk and optimization of access and 
display of medical records data is hoped to be 
completed by January 1982. 

Another clinically-oriented project, which eventually 
will require an interface with the CIU, is the 
extension and enhancement of the BRIGHT system 
to provide Clinical Center investigators the ability to 
perform online analysis of their own clinical data. A t- 
test module and a plotting module have been added 
recently to BRIGHT. Other modules--to provide 
descriptive statistics, chi-square test, linear 
regression, ANOVA, normality test, non-parametric 
tests, and life table analysis-are planned. Modules 
will be added as new requests are received from 
investigators. 

In addition to this work on central NIH administrative 
and medical data management facilities, the 



significance and breadth of DMB's involvement in 
the NIH mission is evident in many computer 
systems it has developed to support individual 
scientific, administrative, and management projects 
during the past year. Virtually every B/l/D has 
benefitted from services provided by the DMB. Each 
of these systems has served a vital segment of NIH 
and, when viewed together, they illustrate DMB's 
very reason for existence as a central resource for 
all of NIH. 

In the area of general support for NIH activities, 
DMB continued to maintain and teach courses on 
the Inquiry and Reporting System (IRS) and MARK 
IV; to support NIH use of Chemical Biological 
Activities (CBAC) and Biosciences Information 
System (BIOSIS) awareness searches on a biweekly 
and semimonthly basis, respectively; to maintain and 
distribute the NCI Survival System; and to consult 
with and assist NIH programmers and contractors, 
enabling facile use of DCRT computer facilities. 

The Scientific Applications Section (SAS) is developing a 
computer system that will enable Clinical Center 
investigators to analyze their own data. 




30 



Future Plans 



Publications 



The Clinical Support Section will begin, by January 
1982, software development for the system or 
systems to reduce the size and optimize the use of 
medical records data. As a by-product of this effort, 
it is anticipated also that requirements for an 
integrated CIU data base will be defined and the 
software development for this effort can begin 
shortly afterwards. As each new benefit becomes 
available, it will be phased into the day-to-day 
functions of the Clinical Center. 

New functions anticipated to be added to the NIH 
Administrative Data Base system include: stock 
requisitioning, central and self-service stores 
inventory, open market requisitioning, accounts 
receivable, and vendor credits. The requirements 
study for an upgraded financial management system 
should be completed by the end of this calendar 
year; a system should begin shortly after that. 

In addition DMB will continue its primary role as a 
central resource for computer applications 
development to all components of NIH that need this 
service. 



Hams, E.K., Yasaka, T,, Horton. MR., and Shakarji, G.; Comparing Multivar- 
iate and Univariate Subject-specific Reference Regions for Blood Con- 
stituents in Healttiy Persons (in press) 

Hirscfiman, G H . Wolfson, M., Mosimann. J.E.. Clark, C.B., Dante, ML, and 
Wineman, R J : Complications of Dialysis. Clinical Nephrology 15:66, 
1981 

Rodbard, D, Cole. B.. and Munson, P.J.: The Need for innovative Ap- 
proaches to Radioimmunoassay Quality Control. In Wilson. D. W. (Ed.): 
Quality Control ol Radioimmunoassays (in press). 




31 




Bill Jones, Carol Kahl, Jennifer Fajman, and Roger Fajman 
were some of the computer professionals involved in 
designing the new version of WYLBUR. 

The NIH Computer Utility provides services to over 8,000 
users and processes about 21,000 job sessions each day. 




32 



Computer Center Branch 



Joseph D. Naughton, Chief 



Function 

The Computer Center Branch (CCB), the largest 
connponent of DCRT, designs and operates the NIH 
Central Computer Utility and its associated online 
telecommunications facilities, in support of scientific 
and administrative programs throughout NIH. 

Two large multi-computer facilities, the IBM System 
370 and the DECsystem-10, form the nucleus of the 
Computer Utility. These are linked by 
communications facilities and connected by 
telephone lines to hundreds of remote interactive 
terminals located in research laboratories and 
administrative offices throughout NIH. 
Complementing this array of systems hardware is a 
complex set of software, either designed and 
implemented by Center personnel or acquired from 
other sources and adapted to meet the unique 
requirements of the NIH biomedical research 
program. 

Approximately 140 professional, technical, and 
administrative personnel ensure the smooth 
functioning of the NIH Computer Utility 24 hours a 
day. The computer specialists, programmers, and 
systems analysts design, implement and maintain 
the complex computer systems software that 
monitors and controls the flow of work through the 
system. They also design and conduct extensive 
training courses, write and publish technical 
documentation on the use of the Utility, assist users 
in problem diagnosis, and maintain and schedule 
recurring production applications. Experienced 
technicians operate the computer systems and 
auxiliary equipment and provide data entry services. 
The remainder of the staff provides the necessary 
administrative support for this complex work. 

Research and development projects are active in the 
areas of scientific image processing, computer 
networking and communications, text editing, display 
of biomedical objects, and utilization of mass storage 
devices. 

To augment the function of the Central Computer 
Utility, the Computer Center provides systems 
programming support, consultation, documentation, 
and training. 



Scope of Work 

The NIH Computer Utility provides services to over 
8,000 authorized users. These include research 
scientists and program managers from every area of 
NIH. The IBM System 370 facility is used as a 
Federal Data Processing Center for biomedical and 
statistical computation by authorized staff in 24 other 
Federal agencies. All services are provided on a 
cost-recovery, fee-for-service basis through the NIH 
Service and Supply Fund. 

A variety of programming languages-including 
FORTRAN, COBOL, PASCAL, BASIC, SPEAKEASY, 
PL/I, and SAIL-are available, as well as a data 
base/data management system (IMS) and a 
comprehensive library of utility programs. Direct 
interactive computing and batch job services are 
available through WYLBUR, TSO, and through 
similar interactive systems on the DECsystem-10. 
The Center provides several facilities for job output 
on paper and microfiche and has programs for 
creating two- or three-dimensional graphic displays 
for advanced projects such as those involving 
macromolecular structures. 

The work load of the Computer Utility has grown 
steadily since it opened in 1968, and FY81 has been 
no exception. The Computer Utility processed an 
average of 21,000 job-sessions per day during the 
past year. Workload on the IBM System 370 
exceeded half a million job-sessions per month for 
the first time in October 1980 and in March 1981 
reached 544,248. The DECsystem-10 timesharing 
facility, utilizing new equipment acquired late last 
year, continued to expand its services and workload. 



33 



Highlights of the Year's 
Accomplishments 

The past year has been an especially important one, 
with major achievements being made in a number of 
areas. 

--The long-awaited new and greatly enhanced NIH 
version of WYLBUR became operational on January 
8, 1981 for all users of the Utility. An 11 -week 
transition period, during which both old and new 
versions were available, proved to be exceptionally 
smooth; user reaction to the new features of this 
interactive, terminal-oriented facility was enthusiastic. 
The improved document formatting and new 
command procedures facilities proved to be 
especially popular. The document formatting 
capability facilitates the production of all types of 
documents. Command procedures can prompt for 
and validate data and can automate repetitive work 
within WYLBUR, providing improved productivity in 
many types of computing applications. 

-One of the largest procurements ever conducted 
by NIH was brought to a close this year as a ten- 
year, 'total system' contract was formally signed with 
IBM Corporation. IBM agreed to provide hardware, 
software, maintenance, and support services to the 
NIH Computer Center throughout the 1980's, 
enabling the Center to meet NIH's information 
processing needs efficiently without the lengthy 
procurement delays that occurred frequently in the 
past. The contract allows rapid utilization of new 
technologies as they are developed by industry and 
provides the flexibility to adapt to changes in 
workload, thus permitting improved cost 
effectiveness. 

"Advances were made in the renovation of 
Buildings 12 and 12A. The plotter and the bursting 
equipment were moved closer to the computer room 
on the first floor of Building 12A, resulting in 
improved turnaround for these services. Construction 
of modern classrooms and a new graphics systems 
area began in the basement of Building 12A. A new, 
enlarged user terminal area and the new classroom 
facilities are expected to be completed in early 1982. 
"Major work was accomplished with the molecular 
graphics system this year. Activities included 
determination of structures by crystallographic 
methods, display of known structures, and modeling 
of hypothetical structures. The coordinates for two 
virus structures were determined during the year. 



Techniques for extending known protein structures 
to other members of the same family were also 
developed. This enabled modeling of hypothetical 
structures of immunoglobulins, myoglobins, and the 
clotting factor proteins. The most exciting experiment 
involved trying to predict the correct architecture for 
proteins where only the primary structure is known. 
The first result at this level is a model for the 
structure of the human leucocyte interferon. 

--In FY81 as in previous years, the Computer 
Center evaluated available hardware and software 
compoments. It selected and installed those that 
serve best to help the Center meet computing needs 
of NIH. This year additions include: 

• A PASCAL/VS compiler for the IBM 370 to 
make PASCAL, a language used widely by 
computer scientists, available on both systems 
of the Computer Utility 

• SPEAKEASY, a language designed at the 
Argonne National Laboratory for scientific and 
mathematical problem solving 

• DISSPLA and TELL-A-GRAF, commercial 
systems that provide extensive facilities for 
creating graphic output and make these facilities 
available to users having graphic display 
terminals. 



34 



NIH COMPUTER UTILITY 



SYSTEM 370 SERVICES 



o 

^ 

m 

:z; 

o 

I — I 

m 
m 
w 
c/o 

Q 

Iz; 

<; 

in 
pq 
o 

•-3 




1967 1868 1968 1970 1971 1872 1973 197* 1978 1976 1977 1978 1979 1980 1861 
CALENDAR YEAR 



35 



Future Plans 



Completing the installation and integration of the 
new IBM equipment and developing the full 
operating capacity of both the DECsystem-10 and 
the IBM System 370 will remain important goals of 
the coming years. Installation of new IBM equipment 
will continue into the coming year as old equipment 
is gradually phased out and replaced with newer, 
faster, and more reliable technology. For instance, 
the 3380 disk drive, IBM's newest disk drive, will be 
installed toward the end of 1981. The new disk drive, 
which will provide both local disk workspace and 
shared permanent storage, will triple the storage 
capacity of the system and provide greatly enhanced 
operating speeds as well as increased reliability. 

Improving the reliability, availability, and performance 
of the Computer Utility is a continuing goal of the 
Center. However, the Computer Center faces a 
major challenge as it endeavors to continue 
improving services while responding to the current 
need to reduce operating expenses and limit staff 
size. To meet this challenge, the Computer Center 
has established a special support level for non- 
critical software. While essential software will 
continue to receive full support service, less widely- 
used products will be designated to receive limited 
support. This lower level of support will exclude 
consulting and assistance with problem diagnosis 
and will provide only limited maintenance service. 
The new policy will allow the Computer Center to 
continue expanding software offerings without 
jeopardizing quality of service and responsiveness. 

The coming year will also see completion of the 
building program, installation of security facilities to 
improve data security, and development of simpler 
capabilities for output processing on the JES2 
facilities. Several new training courses will help users 
take maximum advantage of new features of 
VVYLBUR; these include an introductory WYLBUR 
course for programmers, courses on document 
formatting and command procedures, and a series of 
seminars on advanced topics in WYLBUR. 




I I I I I I I 



36 



Two-dimensional computer graphics like this one, which 
shows the nucleic acid sequence for an RNA fragment help 
scientists in their investigative work throughout NIH 



Publications 



Feldmann, R. J., and BIng, D H.: Teaching Aids for Macromolecular Struc- 
ture. New York, The Taylor-Merchant Corp.. 1980, 98 pp. 



37 



The three DCRT Offices complement the work of the 
six Laboratories and Branches by: 
• coordinating the complex Federal policies and 

procedures that govern getting and using 

computers at NIH 



providing general administrative management 

support for the Division's work 

serving as a central source of information about 

DCRT activities and about computer-related 

disciplines. 




The DCRT Library maintains a collection of about 6,000 
books and technical reports on computer science, 
mathematics, statistics, engineering, management, and 
Information science. 

Gloria Crawford, the DCRT Administrative Officer, 
supervises all travel, training, forms management. Privacy 
Act, and business functions of the Division. 

Mike Reed and Julia Neel of the Financial Management 
Section oversee the financial operation of DCRT. 



38 



1 



Office of ADP Policy Coordination 



Functions 

The Office of ADP Policy Coordination, under the 
direction of the Assistant Director of the Division, has 
three closely related functions: 

1. it is a focus for NIH-wide coordination of 
automatic data processing (ADP) policy matters. 

2. It serves as a central NIH point of contact on 
policy and regulatory questions with the Public 
Health Service, the Department of Health and 
Human Services, other HHS Agencies, and the 
General Service Administration. 

3. It provides advice and assistance concerning 
the internal operations of DCRT in matters of ADP 
policies and regarding interagency sharing 
agreements with other Federal agencies. 



Scope 



The role of the office includes: 

• advising the Director of DCRT and through him 
the Director of NIH on ADP policy matters 

• reviewing and evaluating proposals from NIH B/ 
l/D's for procurements and contracts related to 
computing and ADP 

• directing the development of the annual NIH 
ADP Plan 

• assisting the NIH Division of Management Policy 
on questions relating to its responsibility for 
administrative and management computer 
applications 

• representing NIH in PHS and DHHS policy 
formulation efforts 

• working with GSA staff to obtain necessary 
approvals for NIH on procurements and 
contracts 

• coordinating interagency agreements with other 
Federal agencies that use DCRT facilities, and 

• answering inquiries from scientists and 
administrators who are confused by the whole 
process. 

This office has grown over the years in the breadth 
of its activities but not in its size or cost. This fact is 
strikingly analogous to the growth of the ADP 
technology itself over the last decade and a half. 
The functions of the office have evolved to solve 
problems as new problems arose, to adapt to 
change as changes occurred and to fill gaps as they 



Henry J. Juenemann, Chief 



became obvious. This is strikingly similar to the 
parallel changes in computer technology itself, where 
every year has been marked by increased capacity 
and capabilities applicable to a wider range of 
problem areas but still attainable with resource 
expenditures of similar or smaller magnitude as 
those of the older technologies being replaced. 

Highlights of FY81 Activities 

A major highlight of the year was completion of the 
complex reprocurements of both systems comprising 
the NIH central computing facility. After a protracted 
five year process and in spite of a GAO protest, 
award of a contract was made representing a full 
reprocurement of the general purpose IBM 370 
system. The resulting contract provides NIH 
computer users with a long penod of stability while 
also permitting flexibility to respond promptly to 
technological and workload changes. Likewise, 
reprocurement of the DECsystem-10 scientific time 
sharing system was completed to provide a 
combination of a period of stability with the same 
flexibility to adjust to change. Both reprocurements 
were accomplished in a way that had no adverse 
impact on their user communities. 

During the year this office reviewed nearly 600 
proposals for acquisition of ADP equipment and/or 
services. Each was reviewed to ensure that it was 
justified and was in conformance with PHS, HHS, 
GSA, and 0MB guidelines. Suggestions and 
assistance were provided to the NIH Procurement 
Branch and to contracting officers in Research 
Contract Branches as to the most expeditious 
procurement route to follow. In many cases one or 
more of the Laboratories and Branches of DCRT 
assisted by providing expertise to help in the review 
of technical aspects of the proposals. 

The office arranged the transfer of a computer 
system from DCRT to the National Library of 
Medicine, for relocation in a computer room in the 
Lister Hill Center building. This action resulted in a 
significant increase in NLM capacity to support its 
information retrieval activities and made possible 
savings estimated to be nine million dollars over the 
next three years. 



39 



Arrangements were also made for the system thus 
discontinued by NLM to become the primary support 
system for the NIH Clinical Center Medical 
Information System. It replaced a much older and 
more limited contractor-operated system for meeting 
day-to-day patient care information needs of the NIH 
Clinical Center. The transfer included arrangements 
for the Clinical Center to make use of the computer 
room facilities formerly used for the NLM system. 
The transfer satisfied for the first time the vital need 
for backup redundancy in a system which-24 hours 
a day, 7 days a week-serves all of the wards and 
clinical services in a patient care environment. This 
relocation resulted in an anticipated saving of 1.5 
million dollars over a 28-month period. 

During the year a number of cell sorter and image 
processing computer system procurements were 
expedited as well as procurements of a variety of 
other automated equipment including, this year, 
many microcomputers. An RFP was issued for the 
automation of the NIH Library; a contract award is 
expected in the last few months of the year so that 
implementation of the initial phases of the project 
can begin. 

At year end the most time-consuming undertaking of 
this office was the effort to extend DCRT's program 
to supply users with several varieties of 'NIH 
Standard' terminals. Having standard terminals 
maximizes efficiency of the Center and of the users 
who access the NIH central computing facilities. 
Extension of contracts for the three existing types of 
standard terminals was being sought as was 
expansion of the program to cover those needed for 
the new NIH Administrative Data Base System. 

The Annual ADP Plan-which combines projections 
of new ADP initiatives and required ADP 
expenditures for all bureaus, institutes, divisions, and 
offices of NIH-was completed. It details an NIH ADP 
program projected to be 62 million dollars and 783 
work years in FY82 growing to 88 million dollars and 
833 work years by FY87. Although the accuracy of 
the out-year projections must be regarded with 
caution, the trend of ADP and computing 
involvement in the scientific and managerial life of 
NIH is unmistakable. 



Future Plans ^M 

FY82 will be marked by major changes in the 
structure, staffing, and focus of NIH's overall ADP 
Policy Coordination functions. These changes, the 
nature of which are not predictable at this writing, 
will be accomplished during FY82. 



40 



Office of Administrative 
iVIanagement 



L. Lee Manuel, Chief 



Function and Scope of Work 

The Office of Administrative IVlanagement, under the 
direction of the Executive Officer, consists of 15 
people, organized functionally into three sections: 
finance, personnel, and general administration. The 
office serves as liaison between these functions and 
the NIH Office of Administration, Office of Research 
Services and with other NIH, PHS, and DHHS 
offices. It handles a broad range of administrative 
managerial functions for an NIH research division of 
almost 300 people. 

Fiscal Year 1981 Accomplishments 

The Administrative Office processed a vast number 
of administrative actions and acquired approximately 
30 million do^lars in supplies and equipment during 
FY8I. Day-to-day management activities conducted 
by this staff included: procurement purchases and 
contracts; travel; training; the administration of 
property, space, and communications; payroll; and 
mail/messenger services. As a result of new and 
pending delegations, the position descriptions of the 
staff were reviewed and restructured as necessary to 
better take advantage of these authorities. An 
automated system was developed to track and 
report travel plans and obligations. 

During FY81 the Project Control Office conducted a 
major update of its files of information on 8,000 
users under some 2,000 project accounts. The 
Project Control Officer also was appointed Assistant 
Systems Security Coordinator for the Division and 
was charged with operational responsibility for 
meeting departmental guidelines and reporting 
requirements relating to ADP systems security. 

The Budget Office spent most of the year coping 
with a continually decreasing Management Fund 
budget and increasing reporting requirements to NIH 
program officials on various detailed levels, such as 
travel and consultant services. The office studied the 
impact of estimated 1981 operating expenses in the 
fee-for-service areas along with workload/income 
projections. Division cost center managers 
negotiated rates for various services with the 
Division of Financial Management. 

DCRT acquired its own personnel staff as a result of 
the Division of Personnel Management 
decentralization. As a result a DCRT Personnel 



Officer was appointed for the first time and several 
delegated authorities, such as position classification, 
were acquired. The Personnel Office provided DCRT 
with advice and assistance in several areas. 

During FY81, approximately 400 personnel actions 
that included promotions, reassignments, temporary 
appointments, excepted appointments, and transfers 
were processed. In February 1981, hiring outside the 
Department was restricted to hardship appointments 
and clinical case positions only; DCRT was not 
affected. Recruitment efforts have been solely 
departmental for vacancies we have been able to fill. 
The hiring freeze and Reduction In Force within PHS 
increased the supply of applicants for the computer 
specializations, but have made it difficult to retain our 
computer operators, an area in which we have 
experienced a large number of losses. 

DCRT is serving as the NIH-wide focal point in the 
development of Performance Elements and 
Standards for the Computer Specialist, Technician, 
Mathematics, and Statistics occupational groups. 
This effort is being carried out for implementation of 
the Performance Management System (PMS) to be 
instituted October 1981. The Personnel Officer also 
took the leadership role in implementing the Factor 
Evaluation System for the 334 (computer specialist) 
series and for the PMS. 

Future Plans/Trends 

Contracting, procurement, travel, and consultant 
service requirements as mandated by PHS and 
Office of the Secretary, DHHS involve intensive 
monthly and quarterly reporting of financial 
procurement plans vs. the actual commitment and 
numerous other reporting requirements for 
consultant sen/ices, travel, budget reductions in 
equipment, and other items. To meet this need, we 
plan to cross-train existing staff and to decentralize 
the entering of procurement actions into the NIH 
Materiel Management System. 

The Financial Management/ Project Control Section 
will develop budgets at the laboratory level within the 
Division and will implement a system reporting to the 
lab chiefs on expenditures. A revised Project Control 
Office form is being developed to ease the annual 
process of updating files. 



41 



Office of Scientific and 
Technical Communications 



William C. Mohler, M.D., Chief 



^ 



Functions 

The DCRT Office of Scientific and Technical 
Communications (OSTC), under the direction of the 
Associate Director, DCRT, includes: 

• The DCRT Library, which maintains a collection 
in computer science and mathematics, statistics, 
engineering, information science, and 
management 

• The DCRT Information Office, which serves as 
the focus for providing the NIH community and 
the general public with information about 
DCRT's activities and their relationship to 
biomedical research 

• Scientists assigned to this office, working in 
related areas of pattern recognition, 
multidimensional information processing, and 
applications to medical decision making. 

Scope of Activities 

The DCRT Library is a small, independent, special 
library, staffed by the Librarian and a library 
technician. The staff members provide a full range of 
library activities and have access to a wide variety of 
online information services and data bases. The 
collection of monographs, periodicals, and other 
documents covers subjects related to the work of 
DCRT. These include computer science, 
mathematics, statistics, electronic engineering, 
information science, and management. 

The Library supports the work of the DCRT staff and 
serves as a resource for employees in the rest of 
NIH. It is an integral part of the Washington area 
network of special libraries and cooperates with 
libraries outside the area to share resources. It does 
this through organizations such as the Interlibrary 
Users Association of the Washington/Baltimore 
Area, the Metropolitan Washington Library Council, 
FEDLINK (a Federal library consortium), and--at the 
national level-the OCLC (Online Computer Library 
Center) network. 

The DCRT Information Office, too, is small and 
handles the full range of activities of an NIH 
Information Office. The Information Officer, assisted 
by a Public Information Specialist, answers inquiries, 
produces and distributes print and audiovisual 
materials, and arranges briefings for visitors. They 
coordinate special events, work with members of the 



42 



■ 



media and provide advice, assistance, and 
educational resources on communications for the 
DCRT staff. The office is responsible for all Freedom 
of Information requests coming to DCRT. 

A significant part of the Information Office program is 
directed toward improving within NIH an ^H 

understanding of the Division's work and the ^| 
application of computing to biomedical research. But 
the scope of its communications includes Federal 
agencies, schools, libraries, private industry, medical 
organizations, and a wide variety of individual 
scientists and lay persons. The Information Officer is 
active in the NIH information community and in 
Washington area associations of communication 
professionals. 

The other professional activities of OSTC derive from 
the interests of its scientific professionals. They work 
with other professionals at NIH and with medical and 
technical groups, government and private, outside of 
NIH. They have research and development projects 
on image processing and decision analysis. 

Highlights of FY81 

The DCRT Library maintained its excellent services 
to users during FY81 in spite of transient vacancy in 
its library technician position and a period of 
extended sick leave for the Librarian. Indeed the 
number of books circulated and online bibliographic 
searches made surpassed those of the previous 
year. The new Library Technician, Anita Florentino, 
came to DCRT with six years experience from the 
Technical Services Branch of the NIH Library. She 
and the Librarian, Mrs. Chu, cleaned up the 
accumulated backlog in processing library materials 
by closing the Library to users one afternoon a week 
during the early spring. Preliminary design and 
drafting is complete on a new Library user's 
brochure to replace the old (1971) version. 

The computer-based Circulation System and ^M 

Document Indexing Systems installed in previous ^" 
years by the Library with the help of DCRT staff 
continued to function well and to assure timely 
availability of books and documents to DCRT staff 
and to other Library users. The Librarian was a guest 
speaker on the realities of library automation at a 
meeting of the Special Libraries Association in 



I 



Washington, D.C., and served with other DCRT staff 
on the Advisory Task Force on Automation for the 
NIH Library. She completed her year as Chairperson 
of the delegation from FEDLINK (the Federal Library 
Network) to the Users Council of (OCLC Online 
Computer Library Center), a national cooperative 
library network. She was elected a Director and 
member of the Executive Board of the District of 
Columbia Library Association. 

With the assistance of the DCRT Library Advisory 
Committee, the Librarian completed a review of 
journal holdings, made major revisions to the list of 
journals for subscription in FY82, and set new criteria 
for holdings of back year copies as hard copy or 
microform. The Librarian participated with other 
DCRT and NIH staff at meetings to advise on 
collection policies for books in mathematics, 
statistics, engineering, and physics at the NIH 
Library. The consensus held that the NIH Library 
should have an up-to-date, if limited, set of books in 
these areas for all NIH. The DCRT Library assumed 
responsibility for the collection of highly theoretical 
and advanced works. 

The DCRT Information Office had a busy and 
productive year. 

The office produced a 22-minute videotape program 
documenting two DCRT projects dealing with 
computer voice technology. The first of these two 
systems dealt with a computer voice output system 
developed by CSL and now in use by a blind 
computer programmer. The tape also dealt with an 
existing computer voice input system being used by 
a quadriplegic programmer employed by DMB. 
During the International Year of Disabled Persons 
(1981), the videotape was distributed nationwide to 
100 organizations serving the handicapped, to 
several other Federal agencies, and to a local media 
outlet. The tape was made available to 
approximately 180 public television stations 
throughout the United States and its territories. PBS 
member station WETA-TV in Washington, D.C. aired 
the program in March, 1981. 

A short sound/slide show describing DCRT and the 
work of its laboratories and branches was also 
created. 

The Information Office prepared new brochures for 



LSM, DMB, and the DCRT Library, and assisted in 
the production of the CSL brochure. The office 
reprinted Computers at NIH: Tools for the 
Advancement of tvledicine and issued an update of 
Computing Resources, the compendium of who does 
what in DCRT, designed to help NIH staff 
understand how DCRT can aid NIH scientists and 
administrators. 

Coordinating and producing the FY81 DCRT Annual 
Report was another major job. The FY80 version 
received an Award of Achievement from the Society 
for Technical Communication, Washington Chapter. 

A third communications course, 'Effective Listening,' 
was organized. Thirty-five Division staff members 
received this training. This course complements the 
'Effective Speaking' and 'Effective Writing' courses 
given in previous years. 

Both public inquiries and publication distribution 
remained at levels similar to those of last year. The 
Information Office handled approximately 30 special 
inquiries and distributed about 300 publications each 
month. The general public and other B/I/D's made 
up the majority of the DCRT audience, but a 
significant number of students and schools also 
received materials. Several groups received a DCRT 
briefing and tour, including scientists from China and 
Japan. Information Office staff members also visited 
several nearby colleges to discuss DCRT's role at 
NIH and the application of computers to biomedical 
research. 

The Information Officer continued independent 
research work on preparing text in magnetic format 
for direct typesetting by GPO; this may result in a 
system ready for NIH-wide use by the end of the 
calendar year. She was appointed Chair of the NIH 
Printing Committee. In addition, she was also active 
in outside professional organizations, heading a 
panel on evolving technologies in editing for the 
annual conference of the National Association of 
Government Communicators. She also served as 
president of the 350-member Washington chapter of 
Women in Communications, Inc., a national 
organization for public relations, journalism, 
broadcasting, and communications professionals. 

Work of the scientists assigned to the office 
continued along lines it has followed in previous 



43 



years. The areas covered are best seen in the list of 
publications and presentations. Dr. Judith M. S. 
Prewitt, research mathematician in the Office of the 
Director, was elected Fellow of the Institute of 
Electrical and Electronics Engineers (IEEE). She is 
also an officer of several scientific societies and an 
editorial board member of four scholarly journals: 
IEEE Transactions on Pattern Analysis and Machine 
Intelligence, Computer Graphics and Image 
Processing, Medical Decision Making, and Analytical 
and Quantitative Cytology. She is a National Visiting 
Lecturer of the Society for Industrial and Applied 
Mathematics as well as its representative to the 
American Association for the Advancement of 
Science, Mathematics and Statistics Sections. She 
has just been appointed first Chairman of the new 
IEEE Computer Society Technical Committee on 
Computational Medicine. 

Plans 

The Library and the Information Office will support 
the needs of DCRT and NIH within the limits set by 
available staff and funds. The Library will continue to 
work on changing catalog records to conform to the 
new forms of the revised Anglo-American Cataloging 
Rules (AA CR 2). It will explore the applicability of 
computer systems to support this task and to keep 
up with the effects of automation by the Library of 
Congress on its catalog. As a first step, the Librarian 
and DCRT volunteers reviewed older monographs to 
weed the collection. The retrospective conversion on 
the OCLC system will begin during the summer of 
1981. Catalog records for items added to the library 
prior to 1976 will be tagged for inclusion on the 
DCRT holdings tape. The DCRT Library will use 
computer programs written by the Computer 
Systems Laboratory for preliminary processing of the 
OCLC tapes. 

The Information Office will continue to develop and 
improve materials to tell people what DCRT does 
and how computers are used in biomedical research. 
This may include some new materials for use at NIH 
and perhaps for export in response to requests from 
outside NIH. The Information Office will survey both 
DCRT and NIH to bring up to date the list of needs 
for communications about computers at NIH. 



Publications and Presentations 

Computers at NIH: Tools for the Advancement of Ivledlcine. NIH Publication 

No. 81-1039, reprinted 1981, 24 pp. 
Computing Resources of the Division of Computer Research and Technol- 
ogy. NIH Publication No. 81-1698, reprinted April 1981, 28 pp. 
Data Management Branch. NIH Publication No. 81-1927 (in press). 

Dwyer, A., Prewitt, J. M. S., Ecker, J., and Plunkett, J.: The Use of the 
Hazard Rate to Allay the Peril of Inappropriate Followup: An Optimiz- 
ation Approach. Third Annual Meeting of Society for Medical Decision 
Making, Philadelphia, PA, October 19-21, 1981. 

Herron, R., Dwyer, S. J., and Prewitt, J. M. S.: Computer Graphics for 
Medical Imaging. NCGA '81 Tutorial T-11: Summary (in press). 

Laboratory of Statistical and Mathematical Methodology. NIH Publication 
No. 81-1930, July 1981, 12 pp. 

Miller, P. O.: Text-to-tape Copy Preparation. Communication in the 80's: 
Meeting the Challenge (in press). 

Prewitt, J. M. S.: Computerized Cell Classification and Counting: The Auto- 
mation of Obsolescence and Uncertainty? Washington Bioengineering 
Group, American Red Cross, Bethesda, MD, December 11, 1980. 

Prewitt, J. M. S.: Mathematical Methods Applied to Image Processing in 
Medicine. In Cardus, D., and Vallbona, C. (Eds.): First Conference on 
Mathematics at the Service of Man. Berlin, Springer-Verlag, 1981, pp. 
24-97. 

Prewitt, J. M. S.: Operations Guide for the IEEE Computer Society Techni- 
cal Committee on Computational Medicine or Computer Science and 
Engineering in Medicine. New York, IEEE Computer Society Press, 
1981, 18 pp. 

Prewitt, J. M. S.: The Diagnostic Performance of Leukocyte Counters. Ninth 
Northeast Bioengineering Conference. New Brunswick, NJ, March 20, 
1981. 

Prewitt, J. M. S., Lander, B., and Roelofs, L: Computer Graphics for the 
Intelligent Microscope. National Computer Graphics Association, Inc. 
Second Annual Conference and Exposition, Baltimore, MD, June 15-18, 
1981. 

Prewitt, J. M. S., Plantholt, M., Simpson, M., Edberg, T., and Sanfeliu, A.: 
The Graph-Theoretic Characterization of Tissue Textures. Eighth Con- 
ference on Analytical Cytology and Cytometry, Society for Analytical 
Cytology, Portsmouth, NH, May 19-25, 1981. 

Prewitt, J. M. S., Ranade, S., and Kohler, M. S.: Segmentation of Cell 
Images: Art or Science? Eighth Conference on Analytical Cytology and 
Cytometry, Society for Analytical Cytology, Portsmouth, NH, May 19-25, 
1981. 



44 



* U.S. GOVERNMENT PRINTING OFFICE: 1981-720-020/6581 REGION 3-1 



DIVISION OF COMPUTER RESEARCTTTCND TECHNOLOGY 



FISCAL 

YEAR 

1981 



ANNUAL 
REPORT 



VOLUME 2 



DaXRTK 

U !AJ 



f 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



FISCAL ANNUAL VOLUME 2 

YEAR REPORT 

1981 



rv rk 

r rvT% 

rVHT T 

DCDCKTKr 



Foreword 



The work of the Division of Computer Research and 
Technology covers a large spectrum of activities. It 
ranges from doing research in biology, statistics, 
mathematics, and computer science to providing 
computer facilities and services for NIH. 

The several DCRT laboratories and branches 
embody and integrate this variety of talents. Each 
has a major functional focus. But the success of the 
Division's work arises from the interaction of 
members of each group with others across 
organizational and disciplinary lines. Many projects in 
the Division require the expertise of people from 
several segments of the spectrum. 

DCRT's collaborative projects link its staff to 
professionals both inside and outside NIH. The result 
is a balance in emphasis to provide the work done 
by DCRT at NIH with the benefits of collaborations 
outside of NIH. 

While DCRT does not have money for grants, it does 
provide occasional support for meetings on scientific 
topics related to its work. 

This year's annual report is presented in two 
volumes: 

• Volume 1 gives a summary overview of the 
work of each group and highlights its 
accomplishments. 

• Volume 2 includes detailed projects and 
activities of each group. 

If you have comments on the report or suggestions 
for improving future annual reports, please send 
them to: 

DCRT Information Office 

Building 12A, Room 3027 

Division of Computer Research 

and Technology 
National Institutes of Health 
Bethesda, Maryland 20205 



Contents 



Laboratory of Statistical and Mathematical 
Methodology 

Summary of Activities 7 
Computation 7 
Consultation 8 

Research Projects 10 
Automated Data Processing of Medical Language 10 
Cluster Analysis 1 1 

Research Topics in Computer Science 12 
Discrete Mathematics and Applications 13 
Biological and Visual Shape 14 
Multivariate Statistical Analysis 14 
Linear Methods in Statistics 15 
Non-numerical Programming Techniques and 

Applications 16 
Topics in Geometry and Analysis 1 7 



Computer Systems Laboratory 

Summary of Activities 19 

Research Projects 24 
Computer Support for Flow Microfluorimetry 'Cell Sorters 

(FMF) 24 
Distributed Laboratory Data Acquisition and Control 

System 25 
Molecular Interactions Laboratory Data System 27 
Californium-252 Plasma DesorptJon Mass Spectrometer 

Data System 27 
Combined EDS-WDS X-ray Analysis Scanning Electron 

Microscope System 28 
Cardiac Scintillation Probe 28 
Medical Intensive Care Unit Patient Monitoring Computer 

System 29 
Cardiac Intensive Care Unit Patient Monitoring Computer 

System 31 
Computerized Radiation Therapy 32 
Image Processing Facility 33 

Digital Imaging Applications in Cardiovascular Research 34 
Cerebral Metabolic Imaging 35 
Computer Analysis of Gel Electrophoresis 36 
Automated Analysis of Plaque Formation in Experimental 

Atherosclerosis 37 
Computerized Three-Dimensional Model of the Cat's 

Brain Stem 38 
Rehabilitation Medicine Department Computer System 39 
Image Processing of Electron Micrographs 40 
Positron Emission Tomography (PET) Scan Image 

Analysis in Aging Studies 41 
Computer Analysis of Autoradiographic Images of 

Recombinant DNA Colonies 42 
Cataract Grading via Computenzed Slit-lamp Image 

Analysis 43 



Image Analysis in Computerized Tomography (CT) 

Contrast Material Evaluation 44 
Image Analysis in Automated Radiotherapy Treatment 

Planning 45 
Automated Pulmonary Physiology Testing 45 
Measurement of Transepithelial Resistance of Kidney 

Tubule 47 
Electron Microanalysis Facility 48 
Computer Assistance for Blind Computer Users 49 
Metabolic Energy Measurements 50 
Medical Information Technology Project 51 



Laboratory of Applied Studies 

Summary of Activities 53 

Research Projects 55 
Statistical Research in Clinical Pathology 55 
Mathematical Models of Binding Equilibria 57 
Mathematical Modeling of Substrate Transport in 

Physiological Environments 57 
Analysis of Coupled Transport and Biochemical Kinetics 59 
Nonlinear Equations 62 
Numencal Approximation Techniques for the Solution of 

Reaction-Diffusion Systems in Biology 63 
Computer-based Studies in Pulmonary Pathophysiology 

and Respiratory Disease 65 
Investigation of Hybrid Computing for the Construction of 

Simulation Models and for the Analysis of Physiologic 

Signals 66 
Computer Systems for Nuclear Medicine 67 
Computer-Aided Analysis of Electrocardiograms 69 
Computer-based studies in Ultrasonography 70 
Computer Based Analysis and Image Processing in 

Electron Microscopy and X-ray and Electron-Loss 

Spectroscopy 71 



Physical Sciences Laboratory 

Summary of Activities 75 

Research Projects 77 
Consulting Sen/ices 77 

Theory of Biochemical Separation Techniques 77 
Actin in Nonmuscle Cells— Biophysical and Biochemical 

Studies 78 
Theory and Application of Nuclear Magnetic Resonance 

Spectroscopy 79 
Correlation Function Spectroscopy/Laser Light Scattering 81 
Cell IVIotility and Chemotaxis 81 
Theory and Ivleasurement of Intermolecular Forces 82 
Studies in tvlathematics and Statistics 83 
Quantitative Analysis of the Electronmicroscopy of Cells 

and their Plasma Ivlembrane 84 
Diffusion of r\/lolecules on Cell Surfaces and Light 

Scattenng from Fluids 85 
Computerized Typesetting of Scientific Papers 87 

Data Management Branch 

Summary of Activities 89 

Computer Center Branch 

Summary of Activities 95 

Research Projects 97 
Nucleic Acid Structure Synthesis and Display 97 

Office of the Director 

Summary of Activities 98 

Research Projects 99 
Text-to-tape Copy Preparation 99 



1 



Laboratory of Statistical and 
Mathematical Methodology 



James E. Mosimann, Chief 



Summary of Activities 

LSM activities can be divided into three areas: 
computation, consultation, and research. 

Computat ion 
A major part of LSM activity is the offering of 
statistical and mathematical systems/packages to 
the NIH user community. LSM accepts responsibility 
for evaluation of new program packages and their 
suitability for NIH. When LSM does support a 
system/package for the NIH community, it provides 
maintenance, documentation, instruction, and 
assistance for users to interpret the results. 

Statistical Systems/Packages Support. Dunng 
this year, as in the past year, the Statistical Software 
Section of LSM maintained the following program 
packages and programs: 

BMD, BMDP: Biomedical Computer Programs, 
UCLA. 

SPSS: Statistical Package for the Social Sciences, 
SPSS, Inc. 

SAS: Statistical Analysis System, SAS Institute, 
Inc. 

P-STAT: Statistical Package, P-STAT, Inc. 

IMSL: International Mathematical and Statistical 
Libraries, IMSL, Inc. 

MSTAT1: Collection of Mathematical and 
Statistical Programs, DCRT. 

Dunng the year every system/package went through 
at least one major update. The SSS staff answered 
over 4,500 calls for assistance, and taught a total of 
eight courses on these systems/packages; two each 
on the SPSS and BMDP packages and four courses 
on the SAS system. 

The use of program packages continues to increase. 
The average accesses per month of all the statistical 
packages rose from around 23,000 during FY80 to 
over 33,000 in FY81. For the fifth year in a row SAS 
experienced the largest increase of any of the 
packages. SAS averages over 24,000 accesses per 
month, up from 15,000 per month in FY81. The 
average number of accesses per month for SPSS 
increased from 5,500 to 6,000. The average 



combined accesses of the BMDP and BMD 
packages rose from 1,700 accesses per month in 
FY80 to around 2,600 this year. As an example of a 
package used for specialized purposes, PSTAT 
averaged 60 accesses per month, down from 70 
average accesses per month in FY80. The mam 
programs and subroutines in MSTAT1 averaged 650 
accesses per month, in contrast with 450 in FY80. 
Accesses to the IMSL package cannot be counted, 
but it is estimated that usage increased during FY81. 

In addition, two new systems/packages were offered 
on an expenmental basis to the NIH computer user 
community. SCSS, an interactive version of SPSS, 
and SAS/GRAPH, a graphics package from SAS 
Institute, were installed and tested at DCRT. 

MLAB Support. The Biomathematics and Computer 
Science Section maintains the DECsystem-10 
interpretive program MLAB for biomathematical 
modeling at NIH. This package, designed and 
implemented by BCS staff, is used by several 
hundred NIH researchers each year for various 
modeling and graphical display tasks. It has been 
sent to many universities and research centers at 
their request. It is part of the NIH-funded Prophet 
system, SUMEX-AIM at Stanford University, and the 
NIH-EPA Chemical Information System. 

During FY81, a revision (Third Edition) of the MLAB 
Applications Manual was prepared and distributed. A 
series of tutorial articles on MLAB in INTERFACE 
continued this year, and six more articles have 
appeared. Two beginning MLAB courses and one 
advanced course were taught during FY81. MLAB 
was enhanced by improvements making it more 
efficient and by the addition of Fourier transform and 
inverse transform operations. 

Support of C-LAB. SMS has assumed support of 
C-LAB, a computer system/package for pattern 
recognition and cluster analysis. A course on C-LAB 
was taught during the fiscal year. C-LAB techniques 
were modified to follow the latest published 
methods, and compatibility with MLAB was 
maintained. 



Support for the Unified Generator Pacloge. This 
package, developed by a BCS staff member, 
generates IBM S/370 assembly language programs. 
The compatibility of the package with new WYLBUR 
was maintained. As before, assistance was provided 
for users on request. 

Support for REDUCE and PROLOG. BCS has 

continued support for the REDUCE system (obtained 
from the University of Utah) for manipulation of 
algebraic formulas. 

The PROLOG software system (obtained from the 
University of Edinburgh) is used for non-numerical 
data processing. PROLOG has been applied in LSM 
research in linguistic analysis. 

Indexing Program. An interactive DECsystem-10 
program for generating the index of a document file 
was completed. It is undergoing production testing 
on MLAB documentation now in preparation. 

Consultation 
As in previous years there was considerable 
variation in the amount of time required for an LSM 
consultation. Some very brief consultations are 
successful, and are brief precisely because there is a 
known answer to the question posed. Other 
consultations involve extensive time and statistical/ 
mathematical/computer science research as well. 

LSM consultations in FY81 were of the following 
types: 

• Mathematical, statistical and computer science 
advice with limited computer use (5%). 

• Mathematical or statistical advice with 
considerable computer use (55%). 

• Computational advice alone (40%). 

The large computer use in these figures results from 
the continued availability and use of general purpose 
statistical and mathematical packages like SAS and 
MLAB. 

The diverse nature of LSM consulting is indicated by 
the projects and activities listed below. 

Hemodynamic and Plasma Catecholamine 
Responses to Hyperthermic Cancer Therapy in 
Humans. Y. Kim (CC/ANES). Cancer patients 
treated by induction of hyperthermia under thiopental 
and fentanyl anesthesia, respond with attenuated 
hemodynamic changes compared with those 
reported on normal volunteers. Measured plasma 
catecholamine at hyperthermic condition showed 
evidence of sympathetic nerve response to 
hyperthermia. Statistical procedures used include 
multiple regression analysis, anova, and descriptive 
statistics. 

8 



I 



Sleep Analysis. W. Duncan (NIMH/BDP). The 
analysis focuses on examining the relationship 
between mental illness and sleep disorders. 
Discriminant analysis was employed to evaluate 
possible contribution of sleep variables in 
distinguishing between groups of normal subjects 
and depressed (unipolar or bipolar) or insomniac 
patients. 

Non-parametric Tests. W. Schniderwind (CC). This 
study involves use of the non-parametric median and 
Mann-Whitney tests for comparing the effect of staff 
education on medical and patient care encounters. ^, 

Spectral Analysis of Mood Cycles. F. Putnam ^ 
(NIMH). A frequency-domain spectral analysis study 
was done of possible connections between mood 
cycles and physiological measurements in disturbed , 
patients. 

Automatic Processing of Natural Language 
Pathology Reports. D. Henson (NCI/BCPC). A 
data base for the Clinical Center surgical pathology 
data is being created using automatic encoding of 
natural language reports. An initial updating of the 
systematized Nomenclature of Pathology (dubbed 
SNOP-NIH) has been completed. 



I 



Automatic Encoding of Surgical Pathology 
Reports. E. Jaffe (NCI/LP). In connection with the 
automatic encoding of surgical pathology reports, the 
new malignant lymphoma classification currently 
used at NIH was incorporated into SNOP. 

Atlanta Autopsy Data Base . T. O'Leary (NCI/LP). 
The data and lexicographic changes consistent with 
SNOP were incorporated into the current Atlanta 
autopsy data base. Test runs were conducted. 

Metal Ion Protein Binding. C. Chatterji (NIAID/ 
LC). Optical absorbance experiments measured 
metal ion binding to a protein constituent of snake 
venom. LSM designed MLAB procedures to curve-fit 
mathematical models to the data. 

Analysis of Simultaneous Binding Reactions. L. 

Jacobson (NICHD/LCP). Simultaneous binding 
reactions were studied by obtaining NMR scanner 
absorbances at specific frequencies. LSM assisted in 
mathematical modeling. 

Ultracentrifuge Analysis. M. Lewis (DRS/BEI). A 
mathematical model was developed for studying the 
distribution of molecules in an organic solvent during 
ultracentrifugation. LSM assisted in model 
modifications to represent compressibility effects. 

Compartmental Analysis of Drug Action. R. 

Burns (NIMH/LCS). Effects of drugs with radioactive 
tracers were studied in experimental animals. LSM 



i 



assisted in modeling drug action by developing a 
three-compartment differential equation system. 

Interferon Measurements. M. Morin (DRS/VR). 
Groups of monkeys were injected with various 
compounds, and the yield of interferon was 
determined by measuring viral inhibition. Sample 
sizes required to produce significant differences 
between groups of compounds were calculated. 

Potency Analysis. G. Krishna (NHLBI/IR CP). 
Maximum likelihood estimates of relative potencies 
of groups of compounds administered to a strain of 
mice were calculated. Log potency probit analysis 
techniques were applied. 

Mapping Enzyme Cutting Sites on Circular 
Plasmid DNA. M. Huberman (NHLBI/MH). 
Specimen circular plasmid DNA is subjected to 
enzymes that cut the DNA at specific sites. 
Fragment lengths from one- and two-enzyme 
complete digest experiments are measured by 
electrophoresis. LSM designed and tested a pilot 
MLAB procedure to generate DNA maps consistent 
with the fragment data. An assessment of a Stanford 
computer program for constructing DNA maps from 
fragment data is scheduled for this fiscal year. 

Continuous, Constant-Volume Diafiltration 
Methods. K. Roy (NIADDK/LMB). A laboratory 
technique for continuous, constant volume 
diafiltration measurements of binding parameters 
was designed and tested on several nucleic acid 
monomer-polymer systems. LSM assisted in 
modeling, designing, and testing of MLAB 
procedures. These procedures perform such 
functions as automatic elimination of bad data, 
curve-fitting of models to data, and generation of 
graphical displays of raw data, selected good data, 
and curve-fitted models. 

K-means Clustering. J. Wunderlich (NCI). K-means 
clustenng in CLAB was used to study the genetic 
control of immune response in mice, to identify high 
and low responders. 

Depletion of Lymphocytes in Circulation. J E. 

French (FDA/DBBP). Analysis of covariance was 
done for data pertaining to the depletion of 
circulating lymphocytes by leukapheresis in dogs. 

Cell Microfilament Networks. N. Gershon (DCRT/ 
PSL). Three-dimensional graphical displays of cell 
microfilament networks were prepared and 
presented at a scientific conferenceiposter session. 
LSM assisted in using MLAB facilitie^ for computer 
generation of three-dimensional graphical displays. 

Free Run Cycling of Firefly Pacemaker. John 
Buck (NIADDK/LPB). The interflash duration of 



fireflies was used as a measure of endogenous 
pacemaker timing behavior. Since Student's t-test, 
needed in analyzing free run pacemaking, requires 
that interflash durations have a Gaussian distribution, 
chi-square analyses of truncated samples were used 
to ascertain whether this requirement was fulfilled. 

Protein Frequency Profiles. H. Saroff (NIADDK/ 
LBP). Chi-square goodness-of-fit tests of the 
Gaussian, binomial, and Poisson distributions were 
applied to distnbutions of the number of random 
matches for amino acid sequences obtained from 
Monte Carlo experiments. 



kinoaiiu KIEKi IMCMlllOi IICMKI 

MjiCT raau p. tr «. wi. .►»••! 


."iiiSiii!™*! 


r;;:;,-ms„ 








Autcnated Data Processing of Medical Language 






PI: H. G. Pacak Computer Systems Analyst LSH 
A. M. Pratt Director 00 

Other: G, Donhan Conp-jter Progranner LSH 
S. Harper Computer Progranner LSH 
I. Gar«ia-llld«lgo Guest ^Jorker LStl 
^. Graitson Guest Worker LSH 


OCRT 
OCRT 
DCRT 
OCST 
OCRI 
DCRT 


MOhUIIK Will {i> .',) 


Ijihnrjttnry a( <;tAt l^tir^l and l^th^matjcal I^ethodoloov 


Medical Infomation Science Section 


ikitMun «wtoc4ii» 






,.,...™. 




The wJor objective of the project is the developtnen 
automatic processing of natural nedical language. A 
is the progran for informtion storage and retrieval 
for the Laboratory of Pathology, nCI. The entire co 


major appl 


icSl" 



ZOl CT 00008-07 ISM 



October 1. 1980 through September 30. 1931 



and Itathematical Hethodolo! 



, Bcthesda. Haryland 20305 



patumt or retinal i 



i developed and t 
talyiing spaclal poir 



sloped for testing 



Research Projects 

Automated Data Processing of Medical 
Language 

The major objective of the project is the 
development of methods for the automatic 
processing of natural medical language. A major 
application is the program for information storage 
and retrieval of pathology data for the Laboratory of 
Pathology, NCI. The entire corpus of surgical 
pathology will be encoded. 
The continued collaboration with the Laboratory of 
Pathology, NCI, to create a data base for the Clinical 
Center surgical pathology data indexed by automatic 
encoding has reached the following stage. 

With Dr. Donald E. Henson (NCI/BCPC), the initial 
updating of the Systematized Nomenclature of 
Pathology (dubbed SNOP-NIH) was completed. This 
now permits indexing of almost all SNOP 
topographic descriptions found in the surgical 
pathology reports. The SNOP category of 
morphological diagnoses has been similarly updated 
for surgical pathology data. With Dr. Elaine S. Jaffe 
(NCI/LP), the new malignant lymphoma classification 
currently in use at NIH has been incorporated into 
SNOP-NIH. 

The current surgical pathology data base has been 
automatically encoded, and some changes have 
been made in the automatic encoding program to 
reduce redundant and erroneous encoding. 

This year, the Atlanta autopsy data base project was 
revived under the leadership of Dr. T. J. O'Leary, 
(NCI/LP) . On the basis of test runs of these data, 
some lexicographic changes consistent with SNOP 
were incorporated into the current dictionary. 

Work continued on the study of morphosemantic 
structuring of medical Greek-Latin derived forms. 
The goal is development of a generalized system for 
automatic morphosemantic analysis of medical 
compound words. 

Work was elaborated on a project to determine 
automatically the productive morphemes (prefixes, 
infixes, suffixes) used in the formation of the terms 
appearing in the French translation of SNOP. The 
morphemes were isolated by procedures involving 
pairwise and setwise comparisons of terms. Nearly 
17,000 terms were segmented. 

An English-to-Spanish translation procedure and its 
associated dictionaries were developed and 
implemented for 1,426 terms of the morphology 
section of the International Classification of Diseases 



10 



for Oncology. Morphological substitutions and 
respelling rules permit translation of the ICD-0 terms 
derived from Greek and Latin, which are cognate in 
the source and target languages, without 
construction of a large lexicon. This work was 
accomplished in cooperation with a guest worker, 
Isabel Garcia-Hidalgo. 

The stem dictionary for the 'hepatitis' data base for 
NLM was completed. 

A list of semantically productive terminal morphemes 
in medical language was prepared for the MEDLINE 
system at NLM. 

Page copies of computer oriented medical 
dictionaries (SNOP, ICD) were made available to 
three medical institutions. 

Future efforts: 

a. Continuation of research studies in medical 
language at present level (morphology, syntax, 
semantics). 

b. Creation of a lexicographic data base to be 
used for merging of medical dictionaries and 
extraction of microglossaries. 

c. Continuation of collaboration in the encoding of 
surgical pathology data with the Laboratory of 
Pathology, NCI, to refine the medical dictionary and 
study the language of diagnoses. 

Publications: 

Garcia-Hidalgo. I , and Dunham. G.; An experiment in English-Spanish auto- 
mated translation of medical language data. Methods of Intormalion in 
Medicine 20: 38-46, 1981. 



the distribution of cones is non-random, tending 
towards regularity. 

Statistical methods, based on nearest neighbor 
distances between cones, were used to determine 
which type of regular pattern, with known error, is 
matched most closely. Voronoi regions (convex 
polygons surrounding each point) were computed 
and are being used as another approach to 
measuring regularity. 

Proposed Course: Earlier work on measuring the 
effectiveness of cluster tendency algorithms will be 
completed. 

More retinal cone data will be collected, and spatial 
distributions modeled, based on a statistical analysis 
of the data. A wider range of cluster analysis 
algorithms will continue to be developed and applied. 

Publications: None 



Cluster Analysis 

Cluster analysis algorithms based on the latest 
published research and extensions to it were 
developed and tested. 

Algorithms for analyzing spatial point patterns were 
developed for testing patterns of retinal cones for 
regularity. 

Background and Objectives: The main objective is 
the development of computer programs and 
methods for cluster analysis and related problem 
areas for use by NIH researchers. 

Progress in FY81: New algorithms for representing 
multivariate data graphically have been programmed 
and will be added to the C-LAB cluster analysis 
package. These include Andrews plots, biplots, and 
probability plots. Recent work in the analysis of point 
patterns is being applied to the distribution pattern of 
cones from monkey retinas to determine if there is 
an underlying regularity and to measure the degree 
of regularity. Standard statistical tests indicate that 



, 1980 through Scptcnbor 31. 19S1 



> Topks In Corputer Scl( 



Computer Spcctallsl 



Laboratory of Stflt)stic«l and Hatheratlcal Hethodology 



and Conputer Scier 



Various storage i 



n benefit fron iraproved • 



Currently, a study or Che hashing storage and retr 



Research Topics in Computer Science 

Various storage and retrieval algorithms have been 
studied. The development of flexible and efficient 
storage and retrieval algorithms is very useful; such 
algorithms are used in almost all computer programs. 
Thus biomedical computation in particular can 
benefit from improved storage and retrieval methods. 

Currently, a study of the hashing storage and 
retrieval methods is underway. This has resulted in 
the analysis of the performance of the hashing 
method that resolves collisions using direct-chaining 
with coalescing lists. 

Project Description: The object of this project is to 
develop theoretical bases for new computer 
methods which will expand and improve the use of 
computing in biomedical computation. The methods 
used are the application of known algorithms and the 
development of new pertinent theorems involving 
combinatoric and other related mathematics. 
Research work in storage and retrieval algorithms 
and their efficiency has been the primary topic of 
concern. 

Concurrently, an exhaustive survey of storage and 
retrieval methods is underway. This includes the 
recently-introduced k-d tree method. Various 
improvements and refinements in both the 
algorithms and their analysis are being studied. 

Much effort has gone into studying the B-Tree data 
structure for large files and developing a deletion 
algorithm to efficiently remove items from B-Trees. 

Routines to store, retrieve, and delete items in a 
hash table that employs direct-chaining with and 
without coalescing lists have been prepared. The 
analysis of these algorithms is an active area of 
study. 

Publications: 

Knott, G. D.: Procedures for managing extendible array files. Software 
Practice and Experience 1 1 : 63-84, 1 981 . 



12 



siir^riSiS'SJ'-r^Mis: isrtr* 


■I'ii^.'Higii'iii-l 


Z01 CT 00011 07 LSH 


October 1. IWO throuqh Septenbtr JO. 1961 1 


OKcete Kathe^ticv and Ap|>1lc«tioni 


PI; G. ttutchlnson teiearch tlalheiMttctan LSI XRT ! 


OtC*yit«iilO Syitws Tcm. CCS. KM 


taboroior/ of Statijtical and ItatrwwtlMl HUhodoloqy 


BlomathwMttci and Computer Science Section 


DCRT. NIH, Sethetda. Ilaryland 20?0& 


1.3 1 0.8 1 O.S 




Inclutton r«1atlont between vector ipaees and related problcas concerning 
•nodwies over rings were itodled. 

Preparation of utenltflc nanuscriptt by conputer graphics nethods using 
printer-plotters on nJnicowputeri MS Investigated. 



Discrete Mathematics and Applications 

Inclusion relations between vector spaces and 
related problems concerning modules over rings 
were studied. 

Preparation of scientific manuscripts by computer 
graphics methods using printer-plotters on 
minicomputers was investigated. 

Project Description: The project objective is to 
develop mathematical theory and computational 
techniques using discrete mathematics (algebra, 
combinatorics and graph theory), and to apply such 
methods to appropriate problems of biomedical 
research and computer science. 

Methiods Employed and Major Findings: Studies of 
inclusion relations between modules over a ring (a 
mathematical concept generalizing vector spaces 
and commutative groups) continued. A study giving 
five characterizations of the case that two rings have 
the same module inclusion theories (plus three more 
characterizations for finite nngs) was completed and 
submitted for publication. 

In computer science, previously developed 
minicomputer software for generating graphical 
displays of scientific and mathematical text was 
augmented. New facilities include: high-resolution 
hard copy output of scientific and mathematical text 
obtained using a Varian 9211 printer/plotter, direct 
incorporation of figures and graphs generated on the 
DECsystem-10 using MLAB or OMNIGRAPH into 
pages of scientific text, and many new notations and 
character fonts implemented. It is expected that the 
first complete version of this system will be available 
duhng this fiscal year. Parallel with this work, the 
computer program TEX developed at Stanford 
University for computer generation of mathematical 
text was adapted for NIH by the Laboratory Systems 
Unit. Experimental trials have used TEX to generate 
mathematical text on the printer/plotter. 

Proposed Course: Study of the equivalence of 
different approaches to module theory will be 
continued. The most important unsolved problem is 
the classification of nngs that lead to the same 
restricted theory of modules. 

Research on computer generation of scientific 
manuscripts will continue, with most work 
concentrated on improvement of computer input 
methods so that mathematical notations can be 
described without placing heavy burdens on the 
users. Experimental work involving the TEX system 
will continue. 

Publications: 

Hutchinson. G A complete logic (or n-permulable congruence lattices 
Algebra UniversalisXm press). 

13 



sEra.g'is'jrr! 


RISTT' 


«.;*,i-^':sr;.?,.c.» 


MOjfci imnn 






■ liillilf hS »WCT 


10) CT 0012-0? LSH 


rutwcmuu 


1 






•" *' '"•' 


Bioloqlcat 4n<l Vtiud 


1 Shape 


MIS LtMuian «M iRtii 


»n unki»iMi, MC iiuu v nmciru ihviiiuiioiii uo *u oimi* 


MOilUIMU n««.tt (MU 


It) M TNI MOJiel 




Research Cen. Phys. Sefentlsl LSH DCRT 




Chief, Clinical InveiHgations 








Chief. Genetics Unit VR OHS 






ocRT, mm, MS 










:kAt «nd HathButlcal Kethodolooy 


SioMthtiMtlcs and Comiuter Science Section | 








Maryland 20205 




|««,»,«,u. pu.. 


MCI t—ofim KiiUt) 




J {.) «i. M^OI 


□ t«)"*tti-iw<i a(.)N(iihE« 


n {.-»•.«•» a(.') '«( 


IIUI 




The overall objecti 


ve IS to develop a fomal descriptive languaqe natural to 


bioloqicdl itidp«s i 


id apply thi^ langoaoe to i "tmber of problems arising in 




ne and biology. This would allow for the automation of 


niny shape oroces^e 




undentaniUng of it 


oc and shape development for biological and medical 


porpoiei. 






WJ^lMU ' IS?W SIT" 


;:!.Tfl" 


.,.:;'.• ."'Sir;,?.,.!! 


"""' 


"""" 














IIMUUMAL aUtMCH flOJIGT 


















iber ]0, 1981 








■ M 111! 








KjUWartate Sutlit 


cil ta 


,sis 


















"'•'" 


OJltl 






CI: J. I. Hoslwn 




Chier, LSH 


IW 


DCRT 


0th**-: «. V. fiatnawrkht 


'tsitlnq Associote 


IW 




J. 0. miitr 






ISH 


DCST 












A. L. MUl 




SUNY. Stonybrook, NV 
Unlv*r5Uy of Ue^lern 
Uiltralli, Hedlandl. 




1 


■="•""•' "" 1 t 








; 


«»» 











^ Laboratory of Sutlsttcal and Hjthwjti 


a1 Hetfiodology 


Office of Uit Chief 


: V.H1. NIH. eethesda, Maryland 2O20S 


^Li 1 1.2 


r- 






Biological and Visual Shape 

The overall objective is to develop a formal 
descriptive language natural to biological shapes and 
apply this language to a number of problems arising 
in main areas of medicine and biology. This would 
allow for the automation of many shape processes 
now done by humans and permit better modeling 
and understanding of shape and shape development 
for biological and medical purposes. 

Progress in FY81: A general program for extracting 
symmetric axis descriptions is now available at 
DCRT. The mathematics for extending this geometry 
to three-dimensional data, such as will be coming 
from a variety of new scanners, is continuing. In 
addition, the development of mathematics for 
extracting these descriptions from gray scale data 
(for example, tissue sections and x-rays) is 
continuing. 

This geometry is being applied to the study of growth 
and development of the human mandible. Some 
shape invariants have been found and are being 
studied. This work will possibly be continued in two 
directions. The first is a continued examination of 
individual human mandibles. The second is the study 
of genetic and environmental influences of shape 
development in the mandibles of laboratory mice. 



Proposed Course: 
mandibles. 



Finish worl< on human 



Publications: 

Blum, H.: 3-D Symmetric Axis Coordinates: An Overview and Prospectus. In 
Badler, N., Bajcsy, R. and Otto. G. (Eds.): Three Dimensional Object 
Representation. New York, London, and Heidelberg, Springer-Verlag, 
1981. 

Multivariate Statistical Analysis 

The overall objective of this project is the study of 
multivariate statistical methods for the analysis of 
data that take the form of ratios or proportions. 

Study continues on multivariate statistical methods 
(size-shape methods) for analyzing ratios following a 
multivariate lognormal distribution. Studies also 
continue on ratios which follow an Inverted Dirichlet 
distribution. Studies of discriminant analyses for size 
and shape variables (with J. N. Darroch) continue. 
The Principal Investigator presented a review of his 
work along with a tutorial seminar for biologists at 
the Florida State University, Tallahassee. A study on 
bivariate distributions where the conditional 
distribution of Y given X is a Beta distribution (by M. 
V. Ratnaparkhi) is in press. 

A study of complications of dialysis (with G. 
Hirschman, et al.) was published during this year. 
This study used methods for analyzing proportions of 



xss-JiS'K-urrr!:?: js^* 


I'^wav"^^ 


MMCTWM* 

ZOl CT 00039-04 ISM 


October 1. 1960 through Septetrter 30. 1981 


I.ILI tf HMICT (N .».r..t,». » 1m.) 


'<;iiMU. mnamu. uuced • mi fMJUT 

J. 0. N«n«y Research HsthCMtlCtdn ISM OCRT 

tner: Now 


Hone 


Ubor«torv or StdtistiMi «nd HtthiMttc«l HeChodolOQv 


SUtUttc«l Hethoilolooy $«cttofl 


KRT. NIW. Betheidj. HtrvUnd ?0205 


0.6 I 0.6 1 


,{,} -WMMJtCII D (•)•«»• nuMi ::(<)MiMi 


The ov*r*H obJ»cttv» of ihU proJ«t Is thr study of }int*r nelhods in 
•Miytcs durlnq the year. Nhtch (re of Midespread use 4t HUi. 



patients hospitalized in relation to time at risk. 

Statistically similar methods were also applied to the 
design of experiments for the production of 
hybridomas (A. DeBlas). 

Publications: 

DeBlas. A L . Ralnaparkhi. M V. Mosimann. J. E.: Estimation ol the 
number o( monoclonal hybridomas in a cell fusion expenment Journal 
ol Immunological Methods (in press) 

Hirschman. G H . Woltson. M , Mosimann, J E.. Clark, C. B.. Dante. M L , 
and Wineman. R. J : Complications of dialysis. Clinical Nephrology 
1566. 1981. 

Mosimann. J. E.. and Malley. J D The Independence of Size and Shape 
Before and After Scale Change In Taillie. C . Patil, G P . and Baldes- 
sari. B (Eds): Statistical Distributions in Saeniilic Work. Vol 4, Models. 
Structures and Characterizations Dordrecht. Holland. D Reidel Pub- 
lishing Co (in press) 

Ralnaparkhi. M V Some bivariale distnbutions of (X.Y) where the condi- 
tional distribution of Y. given X is either beta or unit-gamma In Taillie. 
C. Patil. G P . and Baldessari. B (Eds ) Statistical Distributions m 
Scienti/ic Work. Vol 4. Models. Structures and Characterizations Dor- 
drecht. Holland. D, Reidel Publishing Co (m press) 

Roux. J J J . and Ralnaparkhi. M V On matnx-variate beta type I distnbu- 
tion and related charactenzation of Wishart distnbution In Taillie. C . 
Patil. G P . and Baldessari. B (Eds); Statistical Distributions m Scientil- 
ic Work. Vol 4. Models. Structures and Characterizations Dordrecht. 
Holland. D Reidel Publishing Co (in press) 

Linear Methods in Statistics 

The overall objective of this project is the study of 
linear methods in analyses during the year, which 
are of widespread use at NIH. 

Progress in FY81: Linear methods in statistics 
continue to be studied. Theoretical results on 
algebraic independence and statistical independence 
were obtained. These are of the sample covariance 
matrix of multivariate measurements. Previous 
research on Simultaneous Confidence limits for 
ratios was refined and consolidated into a single 
report for publication. These latter methods are of 
broad application, for example, with multiple 
regression and discriminant analyses. 

New research into studies of conditional probability 
were undertaken and are in progress. Studies of the 
application of linear model with unbalanced data (a 
common type of data at NIH) continued. Some 
stress was put on repeated measures anaylses using 
the new versions of the statistical systems SAS and 
SPSS. 

Publications: 

Carlson. H . and Malley. J D Job Satisfaction of Staff RNS in Primary and 
Team Nursing Delivery Systems Research in Nursing and Health. 
1981 

Gnmes. A. M . Mueller. H G and Malley. J D : Examination of binaural 
amplification in children Ear and Hearing (in press) 

Malley. J D : Simultaneous confidence intervals for ratios of normal means 
Journal ol The American Statistical Association (in press) 

Mosimann. J E . and Malley. J D The Independence of Size and Shape 
Before and After Scale Change In Taillie. C . Patil. G P . and Baldes- 
sari. B (Eds ) Statistical Distnbutions m Scientitic Work. Vol 4. Models. 
Structures and Characterizations Dordrecht. Holland. D Reidel Pub- 
lishing Co (in press) 

15 



— • 


.rr-T 


IR iE.T' 


llTRUtMU 


•jili'fi'I 


ZOl CT 00047-0 


3.S« 


oitoSV" 


1° 1980 thr 


ouqh September 30. 


1981 






»»-„u»e 


Jlt, (.0 ...,. 




and Applicjt 


o«s 





F Statistical and Mathematical Methodology 



DCfiT. HIH. Bethesda. Maryland 20205 



Non-numerical Programming Techniques and 
Applications 

The special-purpose computer language PROLOG is 
being used to explore a potential research project in 
computational linguistics and artificial intelligence. 
The ultimate goal of this project is to develop a 
methodology for automatically transforming the 
information presented in a textbook into a form 
which may be used in an appropriate manner by a 
computer. 

A textbook on BASIC programming has been chosen 
and a heavily edited version of its first chapter has 
been used to avoid numerous problems of 
computational linguistics not directly relevant to the 
specific task of manipulating knowledge 
representations. To date, a PROLOG program has 
been developed that can analyze an input of three 
paragraphs in English that describe how to write 
expressions in the BASIC language. As a result of 
analyzing these paragraphs, the PROLOG program 
automatically synthesizes a program that could 
determine whether or not strings of characters form 
legal BASIC expressions. The use of PROLOG is 
essential to this project. It is appropriate for tasks in 
computational linguistics and artificial intelligence, 
and for the representation of knowledge (particularly 
procedural knowledge) as well. 

A paper has been prepared which discusses the 
aims, methods, and initial progress of this project. 

Publications: 

Norton, L M.: A note about Laplace transform tables for computer use. 
SIGSAM Bulletin 14: 30-31, 1980. 



I 



16 



SoStei*NSi»s'Si'»ir«. 


k? 


:rf 


inUHMl MMilCH IMJieT 


;oi ci-ooo;9-oi lw ] 


October 1. 1980 Ihr 




n April 


!0. nei 




Topic* tn Geometry 




««jlyi 


H 




«mi, Li»«>*tMn UD HIT 


" 


».o.. 


SUff Fello* LSM 


IJC«T 







1 


».n. 


l»t.orjtory o< Stitfltlcal jnJ "•thepiatlcil WtHoilolosy 


Bi«««th«n«tKs and Computer Science Section 


DCRI. HIH. SlU...u.. ».,,..... ■"« 1 


in.i^«.,iai. p>ii>..H. |0t.I>. 1 


3 (.) loau luijcni □ 1.) a>i> IitUll 


„.,..,.«. 


1 


for the syinetric ails transfom is under develop 




cal fonullui 



Topics in Geometry and Analysis 

Metrics for convex homogeneous cones were 
studied. A mathematical formalism for the symmetric 
axis transform is under development. 

Progress in FY81: Four invanant (under the group 
of linear automorphisms) metrics on convex 
homogeneous cones have been studied. Bounds on 
those have been obtained, and an equivalent 
expression for each has been found, which allows its 
explicit calculation in a simple closed form involving 
generalized 'eigenvalues' of the cone's Vinberg 
algebra representation. A parametrization of an 
interesting family of cones has been obtained, which 
offers the possibility of applications to the 
generalized symmetric spaces, and further study of 
this has been initiated. 

A mathematical formalism for the symmetric axis 
transform is being developed. The techniques 
involved allow generalizations from open sets in two- 
dimensional Euclidean space to open sets in n- 
dimensional Euclidean space and the study of 
convergence properties of symmetric axis 
transforms. By methods analogous to those used in 
cut sets of Riemannian manifolds, topological 
invariants of the manifold have been shown to be 
inherited by its symmetric axis. A study of local 
differential geometric properties of the axis has been 
initiated. 

Publications: 

O'Connor. MA Invariant metrics on cones. Proc. ol the Conference on 
Invanant Melncs and Holomorphic Maps. Rome, Italy, Istitulo di Alta 
Matematica F Seven di CNR (in press) 



17 



n 



Computer Systems 
Laboratory 



Alan M. Demmerle, Chief 



Summary of Activities 

Computer Support for Flow Microfluorimetry/Cell 

Sorters[FMF] (NCI, NHLBI). This project provides 
support for acquisition and processing for four 
Becton-Dickinson FACS II and one Coulter 
instrument. Of these, three were new in FY81. FACS 
ll/PDP-11/34 systems were installed for IR, EA, 
NHLBI and EEB, NCI. A Coulter PDP-11/34 system 
was installed for VA MOB, NCI. All CSL systems are 
currently using DEC'S RT-11 single user operating 
system. CSL is developing an RSX-11M multi-user 
system to replace RT in some high volume 
applications. This system will feature an LSI-1 1 
microcomputer that will independently interact with 
the FMF operator dunng parameter entry and will 
acquire data. Benchmark testing and design of the 
LSI-1 1 buffer was accomplished in FY81. Both the 
RT and RSX systems use DEC'S VT-1 1 graphics 
terminal, which is no longer available for purchase 
and on which guarantee maintenance is scheduled 
to be eliminated. After a thorough evaluation, a 
Tektronix T4025 was selected as a replacement. 
CSL IS pursuing a contract to develop software to 
respond to the existing VT-1 1 graphic calls and drive 
the T4025. 

Cardiac Scintillation Probe(NM, NHLBI). CSL has 
continued the development of its Cardiac Scintillation 
Probe System begun in 1977. This non-imaging 
ECG-Gated scintillation probe, when used in 
conjuction with left ventricular (LV) catheterization, 
permits simultaneous quantification of the variation 
of LV volume and pressure. Derived parameters 
such as LV compliance can also be monitored, in 
addition to such measurements as ejection fraction, 
filling and ejection rates, and temporal relationships. 
This year realtime pressure-volume measurements 
were used to study the effects of nephidipine and 
verapermil on patients with asymmetric septal 
hypertrophy. New hardware and software have been 
developed to allow online calculation of new 
parameters and permit the system to be easily used 
on a routine basis by the Clinical Center personnel. 
Development is continuing to increase the detection 
efficiency of the probe and to quantify the limitations 
of the technique. 



Nuclear Medicine Computer Systems(CC). CSL 

has continued consultation and support for the 
imaging systems located in the Nuclear fvledicine 
Department. This involved working with Nuclear 
IVIedicine to assess their clinical computing 
requirements in regard to their changing needs and 
anticipated growth with the new Ambulatory Care 
Research Facility. This year, due to the increased 
demand for routine clinical cardiac studies, the 
Nuclear Medicine Department has doubled the 
capabilities in this area by purchasing an additional 
camera system and expanding its computer system. 
CSL helped specify a camera system that, in 
addition to performing present cardiac studies, would 
permit the investigation into high efficiency camera 
systems and their potential applications. 

Computerized Radiation Therapy(NCI). CSL has 
developed a computer system, now in clinical 
operation in the Radiation Oncology Branch, NCI, to 
use the detailed contour and density information 
available from computer assisted tomography to 
improve radiation treatment planning. This system for 
external beam treatment planning is based on a 
generalized 3-D dose field model that covers photon, 
electron, and neutron beams. 

The computer program and most of its clinical 
implementation has been completed for the photon 
and electron fields available from the local 6 MV and 
12 MV linear accelerators. The current capabilities 
include interactive simulation of most irradiation 
techniques, including the effect of most beam 
modifying devices. The system enables the display 
of dose distributions computed in several transverse 
contours and overlaid on corresponding CT scans. 

Cardiac Intensive Care Unit Patient Monitoring 
Computer System(NHLBI). The Surgery Branch of 
NHLBI has collaborated with CSL on a continuing 
basis in the development of automated techniques 
for monitoring patients in a post-operative cardiac 
intensive care environment. Previous annual reports 
describe the details of the system's functional goals 
and project milestones. The system was maintained 
in continuous operation until the end of December 
when the computer was removed from sen^ice, in 

19 



anticipation of the Surgery Branch's pending 
relocation to new facilities in the Clinical Center. The 
new Intensive Care Unit, constructed this year, had 
been planned with the design features necessary to 
support computerization. The decision as to which of 
several approaches to online patient monitoring is to 
be implemented will be deferred until next year. 

Medical Intensive Care Unit Patient Monitoring 
Computer System(CC). Automation of the Medical 
Intensive Care Unit required the development of a 
multiple-computer system configuration to provide for 
the measurement, analysis, control, and 
recordkeeping functions that are dictated by the role 
of the unit. The departure of the unit's senior staff at 
the end of last year resulted in the temporary 
cessation of hardware/software development, 
although the system was maintained in continuous 
operation. With the arrival of a new Department 
Chief, a reconsideration of system goals was 
undertaken. Particular emphasis was placed on 
upgrading the system's cardiac catheterization 
capabilities. Data collection and retrieval functions of 
the primary patient data management system are 
being reconfigured to support anticipated research 
protocols. 

The Biomedical Image Analysis Project(NHLBI, 
NEI, NCI, NIDR, NIADDK). This project is oriented 
toward the development of general-purpose 
algorithms and techniques for image digitization, 
contrast enhancement, edge detection, contour 
extraction, contour following, contour coordinate data 
compression, and three-dimensional representation. 
The resultant general-purpose capability is being 
accomplished through work with a number of NIH 
researchers who encounter relatively similar classes 
of problems in unique individual settings. 

Automated Electrocardiogram Processing 

System(CC). A computer system for the online 
collection, analysis, storage, and retrieval of 
diagnostic electrocardiograms was procured this year 
to automate the Clinical Center's Heart Station. The 
system was installed in refurbished facilities at the 
site of the dismantled cardiac intensive care unit 
patient monitoring computer system, and acceptance 
testing was rapidly completed. The delivery and 
installation of special NIH-specified software 
paralleled the training of the operational staff and 
the education of all clinical center physicians 
regarding the switchover to the automated 
electrocardiogram system. 

Molecular Graphics(NIDR). Three projects use the 
Evans and Sutherland Picture System for modeling 
of protein structures. 

20 



1. From the amino acid sequence of collagen, 
detailed structures can be analyzed and compared 
with x-ray diffraction and electron microscopy data. 

2. As the amino acid sequence of the helical | 
portion of myosin is obtained from cyanogen i 
bromide fragments, these pieces can be modeled as 
double stranded alpha helices. When the complete 
sequence is known, more sophisticated models 
including groups of molecules can be designed. 

3. A study of the structure and possible models of 
streptococcal M proteins is also underway in 
collaboration with the Rockefeller University. These 
models are based on a double stranded alpha helix. 

Rehabilitation Medicine Department Computer 
System(CC). This project involves the development 
of computer techniques in rehabilitation medicine in 
collaboration with the Rehabilitation Medicine 
Department of the NIH Clinical Center. CSL has 
recommended computer techniques that can be 
used to automatically acquire anatomical and 
physiological information from patients, perform 
calculations on the data obtained, and display the 
necessary results to the medical staff. The 
automated techniques include the measurement of 
body forces (hand and ground reaction forces), 
muscle activity (monitoring the electromyogram of 
muscles), and body kinematics (the position and 
angles of the limbs and joints in space and time). 
The system will also allow the medical staff to 
access a data base with computer generated forms 
displayed on a terminal screen, and to perform 
inquiries and generate reports using the accumulated 
data. In FY82 CSL will continue the work begun in 
FY81 including the specification of the computer 
system, the evaluation of methods to perform the 
desired measurements, the selection of the 
necessary transducers and instrumentation, and the 
specification of the required software. 

Automated Pulmonary Physiology 

Testing(NHLBI). Lung compliance and inspiratory 
muscle strength procedures performed in the 
pulmonary physiology/exercise laboratory of the 
Pulmonary Branch have been automated using a 
MINC 11/03 computer system. Data is acquired and 
analyzed in realtime, with graphical and textual 
reports being produced at the completion of each 
procedure. An exercise testing procedure has been 
partially automated. Data from the test is entered 
into the computer manually. Analysis and report 
generation are then completed automatically by the 
computer. Work is in progress to enable automatic 
realtime acquisition of exercise data. A scheme has 
been developed to locally store patient results from 
the above tests on disk for retrospective reference. 



Eventually, this data will be transmitted to the central 
PB data base scheduled for development by the 
Data Management Branch. 

Pulmonary Branch Support(NHLBI). This project 
involves assisting the Pulmonary Branch to meet its 
computer and data processing needs. CSL has 
helped to maintain the computer portion of the two 
automated pulmonary function analyzers installed 
last year. Cooperation with DMB resulted in a 
proposal accepted by PB for DMB to develop a 
clinical data base. During FY82, we expect the data 
base system to be completed. We will interface both 
the automated pulmonary function analyzers and the 
pulmonary physiology computer (reported separately) 
to it. 

Computer Interfaces for Clinical Laboratory 
Instruments(CC). Efforts were continued this year to 
improve the acquisition and reporting of clinical 
laboratory test results. A second Coulter Model S- 
Plus cell counter was connected to the Clinical 
Laboratory Computer System using the method 
designed in FY80. Development of a multistation 
microcomputer system for white cell differential 
counting, started in FY80, was also continued this 
year. The system will support up to eight stations 
and will be connected directly to the Clinical 
Laboratory Computer System. In addition to 
differential counting, technologists will be able to use 
the system to access and review Coulter results. 
Testing of a single station prototype system is 
complete; the full system should be operational early 
in FY82. 

Distributed Laboratory Data Acquisition and 
Control System(NIADDK). A Distributed Laboratory 
Data Acquisition and Control System (DLDACS) has 
been implemented for NIADDK, in Building 2, as a 
replacement to the Laboratory Computer System 
developed by CSL ten years ago. The new system 
consists of a network of remote microcomputers 
connected in a star configuration through a 
communications processor to a central data 
processing computer. The remote microcomputers 
handle all of the realtime data acquisition 
requirements and provide instrument control 
functions when required. The collected data is 
normalized, buffered, and transmitted at a 
convenient time to the communications processor as 
files over a serial line, using a standard block 
communications protocol. The communications 
processor serves as a store and forward front end 
for the central computer. Currently there are seven 
satellites connected to the system supporting ten 
instruments which includes four added this year. 
Presently the system is configured with two host 



processors, a Honeywell-516 and a DEC PDP-11/70, 
connected to the communications processor allowing 
a staged transition of processing programs from the 
H-516 to the PDP-11/70. 

Molecular Interactions Laboratory Data 

System(NHLBI). This microcomputer (PDP-11/03) 
data system supervises the acquisition and 
processing of information from an ultracentrifuge and 
a circular dichroic spectropolanmeter used in MDB, 
NHLBI to investigate the interactions between 
human lipoprotein subunits. Current capabilities 
include acquisiton, display, and preprocessing of 
data from the ultracentrifuge and transfer of 
preprocessed data files to the DECsystem-10 for 
further analysis. The characterization of interacting 
systems is then carried out under MLAB on the 
DECsystem-10. Computation of molecular weights 
for both associating and non-associating systems are 
also performed under MLAB. An interface to the 
GARY 61 spectropolarimeter was designed, 
fabricated and tested in FY81. Software support 
includes the ability to add, subtract, and average CD 
spectra and to transfer files to the PDP-10 for further 
analysis. 

Californium-252 Plasma Desorption Mass 
Spectrometer Data System(NHLBI). The 
Californium-252 plasma desorption mass 
spectrometer puts unusual and stnngent demands 
on the data system that controls the spectrometer 
and acquires and processes its data output. 
Realtime performance and the ability to access very 
large data arrays in main memory are key 
considerations. A data system design modeled after 
one in use at Texas A & M University has been 
purchased and will be operational soon. Special 
interface electronics and stepping motor controllers 
for automatic tuning are under construction in CSL. 

Combined EDS-WDS X-Ray Analysis Scanning 
Electron Microscope System(NIADDK). This 
Project entailed the development of a scanning 
electron microscope system capable of 
simultaneously analyzing the x-ray emissions of a 
sample under observation by both energy-dispersive 
and wavelength-dispersive detectors. The system 
permits localization and quantitation of both light and 
heavy elements in the sample, storage of raw and 
reduced data within the data system, processing of 
data, and transmission of data to a remote 
DECsystem-10 computer system at DCRT. This 
system is complete and no further development is 
anticipated. 

Radiation Counter Data Recorder. A six-month 
extension was negotiated for the contract awarded in 



21 



FY80 for the manufacture of Radiation Counter Data 
Loggers. Under this extension NIH investigators 
could purchase Data Recorders using a Record of 
Call. NIH purchased 18 Data Recorders under this 
extension. Since the response was small, a new 
contract will not be sought. NIH laboratories can 
continue to order Data Recorders using a standard 
requisition (NIH-402). CSL continues to assist 
laboratories in converting their liquid scintillation and 
gamma counter data outputs to Data Recorder 
systems, and in overcoming related data 
communications problems. 

Measurements of Transepithelial Resistance of 
Kidney Tubules(NHLBI). A microprocessor-based 
instrument was developed in FY80 to facilitate the 
determination of the transepithelial resistance of an 
in vitro preparation of kidney tubule. The instrument 
was upgraded this year to include the online 
calculation of the tubule resistances and other 
parameters of interest using a complex set of 
equations derived from a transmission line model. 
Floating-point hardware was added to the instrument 
along with the software to utilize it. This new function 
provides the investigator with immediate feedback 
about the course of the experiment. 

Electron Microanalysis Facility{DRS). CSL in 
collaboration with DRS/BEIB is developing an 
automated electron microanalysis facility consisting 
of two electron microscopes interfaced to a PDP-11/ 
60 computer system. The facility will be used for 
research into the elemental composition of biological 
specimens, and for the development of new 
techniques in electron microscopy. CSL is designing 
and implementing the computer system, which will 
acquire and display the spectra and images resulting 
from Electron Energy Loss and x-ray spectrometry. 
This year, software was developed for defining 
specimen target areas, for acquiring EELS and EDS 
spectra and electron current signals while scanning 
the target, and for acquiring, calibrating, monitoring, 
and displaying 'housekeeping' parameters. Work was 
begun on software for the DeAnza Color Display 
System and for retrieving empirical x-ray information. 

Metabolic Energy Measurements(NHLBI). This 
project is directed toward the development of new 
and improved instrumentation and techniques for the 
study of energy transfer at the cellular level. In FY80 
a microprocessor based system was designed to 
study energy parameters of respiring membranes 
and this year the system was made operational. 
Electrodes are used to measure concentrations of 
specific ions as well as those of protons and of 
oxygen. The membrane potentials measured by 
these new techniques agree with those determined 

22 



by traditional methods. 

Bioassay Information System(NCI/NTP, NCTR/ 
FDA). Through an interagency agreement, the 
National Center for Toxicological Research (NCTR) 
of the Food and Drug Administration is developing a 
computerized data base system for the NCI Bioassay 
Program using the NCTR Toxicology Data 
Management System (TDMS). The NCI Bioassay 
Program involves the testing of chemicals with 
animals for the detection of carcinogens by contract 
laboratories located at many sites around the 
country. A goal of this project is to automate the 
acquisition of the animal data at the laboratory sites 
and to transmit this data to a central computer so 
the data can be used to monitor the progress of all 
tests and to evaluate test results. Since FY77 CSL 
has served as a consultant to NCI specifying and 
evaluating hardware and software components of 
this system. 

During FY81, the terminal contractor delivered thirty 
of the microcomputer-based programmable data 
acquisition terminals to NCTR in Arkansas for final 
testing before being shipped to the laboratory sites. 
Five of these terminals were shipped to Southern 
Research Institute, Birmingham, Alabama, the first 
bioassay laboratory location chosen for automation. 
CSL has monitored the contractor performance on 
providing the hardware and software components of 
the terminal. Another terminal was shipped to NCI, 
Bethesda, Maryland and was used by CSL to 
establish and test a high speed (4800 baud) 
synchronous communications link with the NCTR 
IBM computer facility in Arkansas using a 
communications protocol emulator software package 
provided by the contractor. This link will be used by 
the bioassay laboratories in the future to transfer 
files with NCTR. CSL also reviewed a report written 
by three outside consultants evaluating the entire 
TDMS. CSL expects to continue its consulting role 
which includes the evaluation of possible new 
configurations for this computer network. 

Small Animal Section Data Base Management 

System(DRS). The Small Animal Section of the 
Veterinary Resources Branch requires a data base 
system to facilitate the data entry, record keeping, 
and report generation associated with animal 
breeding and distribution. CSL has concluded an in- 
depth study of the Small Animal Section operations 
and has composed the functional specifications for a 
system that will satisfy the needs of the SAS. From 
these specifications we anticipate releasing one or 
more Requests for Proposals early in FY82 to cover 
requirements for a total system to manage the data 
associated with breeding, ordering/inventory control. 



quality control, and genetic resources. No known 
system currently exists that will meet all of these 
requirements. 

Voice Output Terminal for the Blind The voice 
output terminal design was first made operational in 
FY79 and has proved to be a valuable asset to the 
blind computer professional. Voice output terminals 
are now available from several vendors. Two of 
these are based on original CSL work. We are 
collaborating with our blind users to develop ways of 
presenting complex text formats-tables, forms, etc- 
in audible form. This should make voice output 
devices useful to a wider segment of the blind 
community. 

Library Automation Project(DRS). This project is 
directed toward automating the major functions of 
the NIH Library. In response to the CSL study 
reported on in FY79, the NIH Library has decided to 
purchase the available elements of an automated 
total Library system. As no currently procurable 
system is expected to meet all NIH Library 
specifications, CSL plans to adapt a purchased 
system to unique Library requirements. CSL 
exhaustively studied Library needs and generated a 
comprehensive 'Request tor Proposals' covehng 
hardware, software, and conversion of Library 
holdings to machine readable form in FY80. The 
Library system procurement was delayed one year 
for administrative reasons. During the last year, CSL 
developed a cost-benefit analysis for the Library 
system and updated the RFP to reflect the 
improvements in the available systems. System 
installation is anticipated for late Summer 1982, at 
which time CSL's development effort should be 
underway. 

Image Processing Facility. This project is intended 
to provide a utility to display and analyze digital 
images. The system will consist of a powerful 32-bit 
computer with a mixture of medium and high 
resolution video displays. Also, the system will 
include a microdensitometer to allow precise 
digitization of x-rays, micrographs, and other images. 
The computer and peripherals have been purchased, 
with delivery expected duhng the next fiscal year, 
and the design for physical space to house the 
system is complete. Construction will begin soon, 
with completion expected during the next eight 
months. The display subsystem specifications are 
complete, and procurement is expected during the 
next year. 

Medical Information Technology Project. This 
project is concerned with developing improved man- 
machine interface methods such that modern 
microprocessor technology can better serve the 



needs of physicians and the practice of medicine. 

This year we have started to field test some of the 
concepts developed in previous years. We have 
created a friendly assistive environment, which 
allows a dermatologist to create and retrieve the full 
spectrum of data contained in a patient record. With 
a minimum of keystrokes the examining physician 
can select and simultaneously record from a set of 
choices for diagnosis, symptoms, signs, procedures, 
tests, and prescriptions. We have chosen the most 
frequent dermatologic disorders, encompassing over 
half of all patient visits, as a basis for this pilot study. 
Advantageous techniques that result from this study 
will be applied to other clinical and research 
recording situations. 

Computers in Cardiology Conference. CSL has 

continued its support of the annual International 
Conference on Computers in Cardiology. The 
conference provides a forum for direct interaction 
and exchange between physicians, computer 
scientists, and engineers who are involved in various 
aspects of clinical systems in the field of cardiology. 
CSL was responsible for planning and organizing the 
1980 Conference in Williamsburg, VA, and helped 
edit the Proceedings. This year, six pre-conference 
tutorials were organized. Each tutorial is designed to 
introduce its subject area to those who are unfamiliar 
with it. Participants who need an introduction to the 
computer aspects of a subject area or those with 
computer experience who need an introduction to 
the medical aspects of a subject area are 
encouraged to attend. This year's Conference was 
attended by 300 people from 14 countries. 



23 



!iss"ja,g'i8:'.srrKf:rs" 


.«««!i^«iuKN'^r*« 


ZOl CT00050-02 CSL 


October 1, 1980 to September 30, 1981 


Computer Support for Flow Mlcrofluorimetry/Cell Sorters (FHF) 


PI: Robert J. Romanoff Computer Specialist CSL 

OTHER: Ronald Fico Electronics Engineer CSL 
Irving Levy Electronics Engineer CSL 
Lee w. Freeman Computer Prograimer CSL 
Eric S. Loiederman Computer Aid CSL 
Arthur R. Schultz, Jr. Chief. Processor Design Section CSL 


DCRT 

DCRT 

DCRT 
DCRT 


't:;:v;:";c.:.b.hc. «»B.«:„.™., 


Computer Systems Labordtory 


""'processor Design Section 


•"'•mf:m:%a^u. ™ 20205 


•™ ■""""' 4.0 ]"""•-'■ 3.7 r- 0.3 


^'•'-~' °" ""'" *'■'-»" 




arious levels for four 

hroughput is the principle 
T-11 operating syste-.i is 
system in order to allow 

ies are being added during 


This project provides PDP-11 computer support at 
Becton-Dickinson FACS II Flow Microfluorimeter (FHF)/C 


WADS FHF. Data displav and analysis for hiqh samole 
system feature. Software currently running under the 

more sophisticated recordkeeping and more effective su 
cipated workloads. New hardware and software capabili 
the conversion effort. 



Research Projects 



■ 



siSEifiSBiE'gssiS'SH*!;?: iT^ai" 


,1'ifsiiif'™!, 


"ZZ 


0056-02 CSL 


'"octo^r™. 1980 to September 3D. 19B1 | 


Distributed Laboratory Data Acquisiti 


n and Control System 




PI: John Powell Elect 

OTHER: Milliam Jennings Physi 
Ronald Fico Elect 

Eugene O'Bryan Elect 
Arthur fi. Schult;. Jr. Chief 


onics Engineer 

ronics Engineer 
°^jcsjngjneer_ ^ ^^^^_^ 


LCP, NIAOOK 


LCP and LMB. NIAOOK 


Computer Systems Laboratory 








|OIH»i 




:(.) MWNMjEm dWhi»«iiss 


« p U) .E.IHER 




A local computer networ 
Building Z at the National I 
laboratory data acquisition 


n5""oces 




as part 


an integrated 




crocompu 




and control over an mstrument/experirr 
Storage medium. The hub of the netwo 


ent. Although 




storage device to the satell 


■" 


age rriedium to appear as a 





Computer Support for Flow Microfluorimetry/Cell 
Sorters (FMF) 

This project provides PDP-1 1 computer support at 
various levels for four Becton-Dickinson FACS II 
Flow Microfluorimeter (FMF)/Cell Sorters and one 
Coulter MDADS FMF. Data display and analysis for 
high sample throughput is the principle system 
feature. Software currently running under the RT-1 1 
operating system is being converted to function 
under the RSX-11M operating system in order to 
allow more sophisticated recordkeeping and more 
effective support of current and anticipated 
workloads. New hardware and software capabilities 
are being added during the conversion effort. 

Background and Objectives: Since FY75, CSL has 
provided engineering, system integration, and 
software support necessary to meet the data 
acquisition, data display, and analysis needs of 
several investigators using Flow Microfluorimeters 
(FMF's) at NIH. 

In FY81, CSL continued to support a FACS-II/PDP- 
11/34 FMF system for I, NCI and a similar system 
for LP, NCI. LP, NCI replaced their Los Alamos 
Scientific Laboratory FMF with a FACS-II in FY80. 
Two new FACS-ll/PDP-11/34 systems were installed 
in FY81 with CSL assistance. These are located in 
EA, H and EEB, NCI. CSL also assisted in installing 
and making software modifications for a Coulter 
MDADS/PDP-11/34 system located at the VA 
Hospital in Washington, D.C. for VA MOB, NIC. All 
CSL supported systems are currently using the RT- 
1 1 single-user operating system. 

Progress in FY81: The major software effort in FY81 
was continuing the conversion of RT-1 1 programs to 
run under the RSX-11M multi-user operating system 
as well as adding functionality to these programs. 
The RSX-1 1 M system is being developed to replace 
RT-1 1 in selected CSL-supported systems in order to 
provide more effective support of current and 
anticipated workloads and more sophisticated data 
acquisition and record keeping functions. In FY81, a 
multi-level report generation program for FMF data 
file information was developed, as was a more 
efficient driver program for ZETA 1 553 incremental 
plotters. Improvements were also made to the data 
display and analysis programs. 

Currently, both the RT-11 and RSX-1 1M systems use 
the Digital Equipment Corporation (DEC) VT-1 1 as 
the graphics display device. The VT-1 1 is no longer 
available for purchase and DEC will not provide 
guaranteed maintenance after three additional years. 



24 



After a thorough evaluation of available devices, CSL 
selected a new graphics terminal, a Tektronix T4025, 
initially for RSX-1 1 M use and eventually for RT-1 1 
use. During the third quarter of FY81, CSL was 
pursuing a contract to develop software packages 
that will respond to the existing VT-1 1 graphics calls 
and drive the T4025 under both RT-1 1 and RSX- 
11M. 

In order to deal with the degradation of data 
acquisition speed under RSX-11M, it was decided to 
develop an LSI-11 microconnputer-based data 
acquisition system that will independently interact 
with the FMF operator during parameter entry and 
send the acquired data to the host PDP-11/34 RSX- 
11M system over an interprocessor link. An 
important feature of the LSI-1 1 will be the ability to 
create a 'laboratory notebook' as a permanent hard 
copy rather than continuing this as a manual task. 
Considerable benchmark testing and design of this 
system was accomplished in FY81 and appropriate 
hardware was ordered. 

Several minor improvements to the RT-11 display 
and analysis programs were made in FY81 in order 
to accommodate immediate needs of our users. 

Also, an efficient program to interface RT-11 to the 
ZETA 1553 incremental plotter was written and 
installed. The ZETA 1553 plotter replaced the 
Houston Instruments plotter on all CSL-supported 
systems in FY81. 

During FY80, a contract was negotiated to provide 
four FMF hardware interfaces. These were delivered, 
tested, and installed during the third quarter of FY81. 

CSL has also responded to many external requests 
and has provided copies of the interface hardware 
schematics, software, and documentation to FMF 
sites in the U.S., Europe, and Australia. 

Proposed Course: In the forthcoming year, CSL 
plans to complete the first RSX-1 IM-based FMF 
system and LSI-11 -based data acquisition system 
and put them into operation at the I, NCI facility. If 
resources permit, the RT-11 software will be 
rewritten to use the T4025 graphics terminal as a 
replacement for the VT-1 1. 

Distributed Laboratory Data Acquisition and 
Control System 

A local computer network has been developed for 
LCP and LMB, NIADDK, in Building 2 at the National 
Institutes of Health (NIH) as part of an integrated 
laboratory data acquisition and processing system. 
This network is configured with satellites connected 
in a star configuration to a host processor. At each 
satellite a dedicated microcomputer system performs 



data acquisition from and control over an 
instrument/experiment. Although acquired data files 
may be stored locally, they are normally transferred 
via the network to a host storage medium. The hub 
of the network, the concentrator, utilizes DMA 
hardware on all communicating links and performs a 
file store and fonward function. The local network 
allows the host storage medium to appear as a 
'virtual' storage device to the satellites. 

Background and Objectives: A system of 
microcomputers capable of independently controlling 
and acquiring data from an instrument/experiment 
was proposed in December 1976 as the best system 
architecture of upgrading laboratory data processing. 
A prototype laboratory data acquisition and control 
(LDACS) computer and the essential elements of the 
communication system were developed. 

Satellites perform the realtime data acquisition and 
instrument control functions. Their configuration 
includes a Digital Equipment Corporation (DEC) LSI- 
1 1 microcomputer, 28K words memory, low density 
random access storage, graphics terminal, and all 
the necessary I/O hardware to interface the 
instrument/experiment. Satellite software runs under 
DEC'S RT-11 realtime operating system. A hardwired 
serial link connects a satellite to the concentrator. 
Although each satellite is capable of stand-alone 
operation, the immediate transfer of the files allows 
data processing to proceed on the host system 
simultaneously with data acquisition on the satellite 
system. The host processor, a DEC PDP11/70, is 
configured with: 128K words of memory, a high 
speed printer/plotter, a 9-track magnetic tape drive, 
and two large capacity disc drives. DEC's multiuser, 
multitasking operating system, RSX-1 1M, is used to 
service the processing needs of the users. User 
access to the host is provided by hardwired links 
between terminals and host timesharing ports. 
Instruments previously reported as being connected 
to the network include: Spectrophotometers; CARY 
118, Perkin Elmer 580B, CARY 14; 
Spectropolarimeter, CARY 60; Electron spin 
resonance spectrometer; Varian; and a stimulus 
response retina experiment. 

Progress in FY81: Seven satellites supporting ten 
instruments are currently connected to the system. 
Four instruments added this year include a CARY 
219 spectrophotometer, a microspectrophotometer 
designed by NIADDK, a Jasco J500A 
spectropolarimeter, and a I.S. Co. Model 1440 liquid 
chromatograph. 

Presently, the system is configured with two host 
processors, a Honeywell-516 and a DEC PDP-11/70. 

25 



The H516 is the data acquisition and processing 
computer from a 1 0-year old centralized system in 
NIADDK. Having both host processors available 
allows a gradual transition of processing functions 
from the H-516 to the PDP-11 /70. 

In addition, a multipurpose counter/timer module 
was developed for use with LDACS and software 
was provided for the GARY 1 1 8 LADACS to 
accommodate a new experimental technique for 
measuring absorbance versus concentration. 

CSL had the first part of a two-part users manual 
written under contract. This 60-page document is 
intended as a guide to introduce scientists to the 
system and the LDACS functions and utilities that 
are common to all instruments. The second part of 
the manual will be specific for each LDACS and will 
describe unique instrument related functions. 

Proposed Course: Support for the system will 
continue. Additional software to further utilize 
systems capabilities will be provided for the Jasco 
J500A, the Perkin Elmer 580B, the CARY 219, and 
the microspectrophotometer. Earlier LDACS software 
will be modified as necessary to incorporate the 
software libraries and modules that have now been 
standardized for LDACS. Documentation of the 
system will be given a higher priority, with the 
objective of completing an LDACS users guide for 
each LDACS and of documenting the common 
software libraries developed for LDACS. 

Publications: 

Powell, J. I., Fico, R., Jennings, W. H., O'Bryan, E. R., Sohultz, Jr., A. R.: A 
Local Network for Distributed Laboratory Microcomputer. Proceedings 
of the Twenty-first iEEE Computer Society International Conference. 
September 1980, pp. 185-190. 



1 



26 



:is:cTi£.s'iS"w - lis :sT' 



Z01 CT00057-OZ CSL 



) SepmKxr 3 0. I?81_ 



I Ho1ecu1«r Interactions Laboratory Oat^ Systcs 



Conputer Systems Laboratory 
Processor Design Section 



, MRT, HIA. fietheiila^ MD _2Q2QS_ 

rtiH. luirMj. TmgiisuWM,. 



This microcomputer [PQP-llVojf <lVta system supervises the acquisition it 
icesstng of Information from an a nalytical ul tracentr lfuge and a circular 






itlng iys 

I by a few simple coinnanas. ai 
•clropolarlmeler signals, clMi 
■s tne ability to add. subtrac 
;o the POP-10 for further anal. 



Ointly developed by CSL an( 
lie perfonned under MLAB usi 



ng predefined pi 
ivelength. Softi 



_Octobcr 1, 1980 to Septewfter 30. 19B1 



C«1iromtum-2S2 Plasma Oesorptlon Mass Spectrometer Data Systea 



OTHER: Henry Fait 



CSL. OCRT 
LC. NHLBI 



CoBputer Systems Laboratory 



Processor Design Section 



DCRT, NIH. Sethesda, W 2020S 



3 to « 



The CaHfomljB-35g plasma , desorptlon mass spectr gTCSpr [ 
stringent demands on the data system that controls the spectre 
acquires and processes its data output. Real-time performance 
to access very large data arrays in main memory are key consK 
evaluated alternatives to meet trie data processing needs of tt 



Molecular Interactions Laboratory Data System 

This microcomputer (PDP-11/03) data system 
supervises the acquisition and processing of 
information from an analytical ultracentrifuge and a 
circular dichroic spectropolarimter used in MDB, 
NHLBI to investigate the interactions between 
human lipoprotein subunits. Current capabilities 
include acquisition, display, and preprocessing of 
data from the ultracentrifuge and transfer of 
preprocessed data files to the DECsystem-10 for 
further analysis. The nonlinear analyses necessary 
for the characterization of interacting systems are 
then carried out under MLAB on the DECsystem-10. 
CLINK, the PDP-1 1/PDP-10 communications 
software package jointly developed by CSL and 
CCB, is used to perform the data transfers. 
Computation of molecular weights for both 
associating and non-associating systems can be 
performed under MLAB using predefined procedures 
invoked by a few simple commands. An interface 
has been constructed to acquire the 
spectropolarimeter signals, ellipticity, and 
wavelength. Software support includes the ability to 
add, subtract, and average CD spectra and to 
transfer files to the PDP-1 for further analysis. 

Californium-252 Plasma Desorption Mass 
Spectrometer Data System 

The Californium-252 plasma desorption mass 
spectrometer puts unusual and stringent demands 
on the data system that controls the spectrometer 
and acquires and processes its data output. 
Realtime performance and the ability to access very 
large data arrays in main memory are key 
considerations. CSL evaluated alternatives to meet 
the data processing needs of the instrument and 
recommended a more recent model of the computer 
used for this purpose at Texas A&M University. This 
data system has been purchased and will be 
operational soon. Special interface electronics and 
stepping motor controllers for automatic tuning are 
under construction in CSL. 



27 



siii;iris«JSTaS'»ifrii?:js?.T' 


.l^iHiii^ffil 


PflDJECr hlWBEH 

ZOl CT00059-02 CSL 


"Octoberi. 1980 to September 30. 1981 


Combined EDS-WDS X-ray Analysis Scanning Electron Microscope System 


PI: Ramon L. Tate. Ph.D. Computer Specialist CSL, DCRT 

OTHER: Htlliam A. Hagins. M.D.. Ph.D. Chief. Meinbrane Biophysics 

Section LCP, N!ADOl 




Computer Systems Laboratory 


Processor Design Section 


'""oCRrTNIHV'e'ethesda. HD 20205 


,ouL-«u.iiL. ^^ |™c.(ss.w.L. ^_^ |omB. 








In 1979. following the reconmendations of CSL. Dr 








and heavy elements in the sample, storage of raw and r 

computer at DCRT. This system is complete and no furt 
pa ted. 


ntitation of both light 
ed>jced data within the dat; 
a remote DECsystem-lO 
her development is antici- 



ZOl CT00051-02 CSL 



October 1. 1980 to September 30. 1981 



Harold OstroH 

Scott Allen 
Steve Bacharach 

Robert Bonow 



I Research Analyst 



■ Medicine, CC, Cardiology Branch, NHLBI 



'"Dt^'f,"l3lWt*ye*thesda. HD 



'tit TiSs 'con£rSJe?f tKe deve'TopSenToT'its Carxliac Scintillation P robe Sysi 
begun in_1977. This non-imaging ECG-gated scintillatton' probe. when used in 



Implemented capability for 
itudy the effects of nephi< 
;epta1 hypertrophy . The pr 





T"Z 


sly 




on frac 


inn 




sjimetr 






c'° 



; effect; 

I Center personnel. Development 1 



Combined EDS-WDS X-ray Analysis Scanning 
Electron Microscope System 

In 1979, following the recommendations of CSL, Dr. 
Hagins procured a combined energy dispersive 
(EDS) and wavelength-dispersive (WDS) X-ray 
analysis system for his scanning electron 
miscroscope. A single vendor did not market a 
combined EDS-WDS system. CSL arranged for the 
EDS vendor to integrate a WDS into their system. 
The system permits: localization and quantitation of 
both light and heavy elements in the sample, storage 
of raw and reduced data within the data system, 
processing of data, and transmission of data to a 
remote DECsystem-10 computer at DCRT. This 
system is complete and no further development is 
anticipated. 

Cardiac Scintillation Probe 

CSL has continued the development of its Cardiac 
Scintillation Probe System begun in 1977. This non- 
imaging ECG-gated scintillation probe, when used in 
conjunction with left ventricular (LV) catheterization, 
permits simultaneous quantification of the variation 
of LV volume and pressure. By simultaneously 
measuring LV volume and LV pressure, parameters 
such as LV compliance can be continuously 
monitored, in addition to such measurements as 
ejection fraction, filling and ejection rates, and 
temporal relationships. This year the previously 
implemented capability for realtime pressure-volume 
measurements were used to study the effects of 
nephidipine and verapermil on patients with 
asymmetric septal hypertrophy. The pressure-volume 
relationships produced by the probe system allowed 
the effects of drugs to be quantitated in a manner 
not possible before. New hardware and software 
have been developed to allow online calculation of 
new parameters and to permit the system to be 
easily used on a routine basis by the Clinical Center 
personnel. Development is continuing on increasing 
the detection efficiency of the probe and in 
quantifying the limitation of the technique. 



I 



28 



.,;iswrxsissr'^,^pyij^ 


im CTIKKIS4-U Ol 


October 1. 1980 te SeptcMber 30. 1981 [ 


Hedlcal Intensive Care Unit Patient Honltorfng Ctwputtr Systm 


Ssn.^s:*ss^>':::^;;vz':;;:;; - -"•' •"•■•■"'-■ - ■" •■- 


PI: Kenneth H. Keitipner Electrvnlct Engineer 
Robert L. Hdrttno Clectronlcs Engineer 


est. K«t 
est. OC«I 



CrUkal Care Hedlclne OipartMnt. CHnlcal Center 



CfiuuUr- 



S yitawt >sl9n Sectloi 



PCWT. NIM. B«th>«d. HP gPgOS 



system for IM 





utoiuted patient ronlt 


orlna 


ch subsystem, 
er subsystem 
etopnent of n 


ord keeping functions 
re already operallona 




ms, compoter- 


ontroUed drug infuse 




«t^ra^e 


f the patient informa 
for retrospective slu 





Medical Intensive Care Unit Patient Monitoring 
Computer System 

This project involves the development of an 
automated patient monitoring system for 
measurement, analysis, control, and recordkeeping 
functions in a nine-bed medical intensive care unit. A 
minicomputer-based patient data management 
system, a cardiovascular research subsystem, a 
software development subsystem, and a medical 
mass spectrometer subsystem are already 
operational. Future project goals involve the 
development of noninvasive measurements of the 
cardiovascular and respiratory systems, computer- 
controlled drug infusion, and automated unne output 
measurement. A data base of the patient information 
obtained with the systems will be created and used 
for retrospective studies by the medical staff. 

Background and Ob/ectives: The IVIedical Instensive 
Care Unit (MICU), which is administered by the 
Department of Cntical Care ivledicine in the NIH 
Clinical Center, receives critically ill patients from 
clinical programs of NIH. The MICU comprises a five- 
bed ward area, a two-bed special study area, a pair 
of isolation beds, and a vascular research laboratory. 
The research goals of this unit include the 
development of techniques for automated patient 
monitoring and noninvasive measurements of the 
cardiovascular and respiratory systems. In addition, 
the medical staff is performing complete cardiac 
cathetehzation studies. 

Over the past four years, CSL has invested six man- 
years in this project. Working with clinical center 
staff, CSL contributed to the engineering design of 
the intensive care unit. CSL also undertook the 
specification, procurement, and installation of the 
bedside patient monitoring equipment and four 
computer systems: 

1 . a Patient Data Management System used for 
automaticaly monitoring patient variables, manually 
entering patient data, retrieving information online, 
and keeping medical records; 

2. a Cardiovascular Research Subsystem used for 
acquiring and processing cardiovascular pressure 
waveforms, measuring cardiac output, displaying 
measured results online, and generating a cardiac 
catheterization report; 

3. a Software Development Subsystem used for 
developing software for the above described 
systems; and 

4. a Medical Mass Spectrometer Subsystem used 
for monitoring both the patient airway gases and the 
gases delivered by the patient's respirator at all nine 
beds. Featuring the same minicomputer, the first 



29 



three systems were purchased from the Hewlett- 
Packard Corporation. The Chemetron Corporation 
manufactures the microprocessor-based mass 
spectrometer system. 

Major Findings: The automation of the MICU has 
aided the medical staff by: managing the large 
amount of data needed for the care of the critically ill 
patient, performing desired calculations, and allowing 
measurments that would not otherwise be possible. 

Progress in FY81: The four computer systems have 
been in operation for over two years. The departure 
of the unit's senior staff at the end of the last 
reporting year resulted in the temporary cessation of 
hardware/software development. With the arrival of 
a new department chief, a reconsideration of system 
goals was undertaken. Particular emphasis was 
placed on upgrading the system's cardiac 
catheterization capabilities. Data collection and 
retrieval functions of the primary patient data 
management system are being reconfigured to 
support anticipated research protocols. 

Significance to Biomedical Researcii: Many 
hospitals around the world are automating various 
functions in their intensive care units. In particular, 
the Hewlett-Packard computerized patient monitoring 
system purchased for this project has been installed 
in many private, university, and government 
hospitals. Therefore, any new developments made 
on this project will benefit many users of automated 
systems, as well as patient care and clinical 
research within the MICU at NIH. 

Proposed Course: Depending on the research 
goals of the new medical staff, possible 
modifications to the primary patient data 
management system include the addition of urine 
output measurement hardware and the 
computerization of fluid infusion therapy with 
microprocessor-controlled infusion pumps. In 
addition, software modifications to the patient data 
management and vascular research computers can 
tailor their functional capabilities to the unit's 
developing research interests. 

Publications: 

Marlino, R. L., Kempner, K. M., McClellan, J. R., and McLees, B. C: 
Automation of a Medical Intensive Care Environment with a Flexible 
Configuration of Computer Systems. Proceedings of the Fourth Annual 
Symposium on Computer Applications in Medical Care. New York, 
Institute of Electrical and Electronics Engineers, Inc., 1980, pp. 1562- 
1658. 



5Sij?inaa!S'r"irrji?:.^S2i!i" 


;2fw»;«;" 


ZOl CT00053-0^ CSL 


"octobeM, 1980 to September 30, 1981 


Cardiac Intensive Care Unit Patient Monitoring Computer System 


PI: Kenneth M. Kempner Electronics Engineer CSL, OCRT 
Andrew G. Morrow, H.D. Chief, Sfi SB, NHLBI 

OTHER: William L. Risso Electronics Engineer CSL, OCRT 
Robert L. Hartino Electronics Engineer CSL. OCRT 
Lee Freenan Computer Prograniner CSL, OCRT 


Surgery Branch, NHLBI 




Systems Design Section 








The computerized patient monitoring system developed in conjunction with 
Surgery Branch. NHLBI, provides realtime, beat-to-beat analysis of physiologic 

a Xerox Sigma-3 Computer, the system monitors the electrocardiogram, arterial 
and venous blood pressures, body temperature, urine output, and blood loss, as 
well as thermal and dye-dilution cardiac output. 

A Morphology Analysis of tne electrocardiogram and arterial blood pressure 
waveforms is performed to detect fiducial markers. The application of cross- 
correlation techniques to the electrocardiogram allows the detection of pre- 


sand electrocardiographic complexes may be recalled at any time in the form of 


minute electrocardiogram memo 


arrhythmias. Vital si 


nrire retrievab"%r''"' 


eight-hour nursing shift surmianes in tabular form. 


r as 12-, 24-, or 72-hour 


vital siqns graphs. 







30 



Cardiac Intensive Care Unit Patient Monitoring 
Computer System 

The computerized patient monitoring system 
developed in conjunction with Surgery Branch, 
NHLBI, provides realtime, beat-to-beat analyis of 
physiologic waveforms from patients within a four- 
bed intensive care unit. Based around a Xerox 
Sigma-3 Computer, the system monitors the 
elctrocardiogram, arterial and venous blood 
pressures, body temperature, urine output, and blood 
loss, as well as thermal and dye-dilution cardiac 
output. 

A Morphology Analysis of the electrocardiogram and 
arterial blood pressure waveforms is performed to 
detect fiducial markers. The application of cross- 
correlation techniques to the electrocardiogram 
allows the detection of premature ventricular 
contractions. Beat-to-beat data from the most recent 
thousand electrocardiographic complexes may be 
recalled at any time in the form of Joint Interval 
Histograms and Correlation Coefficient Scatter 
Diagrams. A five-minute electrocardiogram memory 
is maintained in realtime for use in the detection of 
transient ventricular arrhythmias. Vital signs are 
retrievable as eight-hour nursing shift summaries in 
tabular form, or as 12-, 24-, or 72-hour vital signs 
graphs. 

Background and Objectives: The principle objective 
of this long-term project was to investigate new 
approaches to the problem of providing an 
automated patient monitoring environment. More 
specifically, project goals include: the release of the 
nursing staff from routine clerical tasks, the uniform 
collection of all vital signs at high frequency, the 
automated detection of potentially life-threatening 
ventricular arrhythmias, and the development of an 
online tool for the evaluation of signal processing 
algorithms. 

Methods Employed: A general-purpose process- 
control computer system was chosen as the central 
element of the interactive patient monitoring system. 
The Xerox Sigma-3 computer was interfaced to a 
high-speed 16-channel video generator system. 
Commercially available bedside electronic modules 
were chosen to provide for the monitoring of all 
relevant physiologic parameters. Transducers were 
developed to monitor urine output and blood loss 
(chest drainage). Active analog and digital 
transmission lines were fabricated to connect the 
intensive care unit with the computer room. 

The operating system provided with the Xerox 
Sigma-3 was extensively modified to provide a more 
suitable environment for realtime, beat-to-beat 



patient monitoring tasks. Many signal analysis 
programs, and a sophisticated graphics package, 
were written to accomplish the functional 
specifications selected for this system. 

Major Findings in FY81: No major findings occurred 
during FY81. The system was maintained in 
continuous operation without additional hardware/ 
software development or evaluation. The use of this 
system was terminated at the end of December in 
anticipation of the relocation of the Surgery Branch 
to new facilities within the Clinical Center. This 
system had been in place for ten years, but because 
the mainframe hardware was no longer in 
production, replacement parts and maintenance 
became an increasing burden. Design of the new 
five-bed intensive care unit was carefully planned to 
incorporate all utilities, conduits, and structural 
accommodations necessary for the computerization 
of the unit. 

Significance to Biomedical Research: The 
computerized patient monitoring system can be 
practically applied to the continuous collection of 
data for research protocols. 

The novel approach to the monitoring of cardiac 
arrhythmias implemented on this system is easily 
adaptable to other computer systems with waveform 
capture capability. These arrhythmia monitoring 
techniques should prove useful in many clinical and 
experimental animal protocols, such as drug 
evaluation studies. 

Proposed Course: A final decision must be reached 
concerning the desired approach for replacement of 
the previously developed ICU Monitoring Computer 
System. Some factors to be considered are: medical 
utility of computerization, trade-offs of alternative 
systems, time for implementation, staffing for 
design/implementation, and operational personnel 
required. The selected approach would then be 
carried through the final design phase, including a 
detailed cost estimate. 



31 




1 



SiSJflS.iS'r.W^rKK 


'S!fl" 


'31if:r' 


"zoiTl^-O. CSL 














or Ui>) 




Image Processing Faci 


ity 










MOFESSIOHH. PtHSOHNEL EM.OtD 


M IH£ «OJ£CI 




PI- H Risso 


Electronics Engineer 


CSL, DCRT 


D. Syed 


Chief, Systems Design Sect 


on CSL. DCRT 


OTHER: D. Foxvog 


Computer Specialist 


CSL. DCRT 


A. Pashayan 


Computer Specialist 




B. Trus 


Research Chemist 






Computer Systems Laboratory 








Systems Oesiqn Sectf 


n 










DCRT. NIH. Bethesda, 


HO 20Z05 




10UL«Mt«S. ^^ pCftSSIONAL. ^^ lOIHER. 1 


Tut 'ZTZ'xzT'^ 




(d «E,mR 


a(><) H'NMS Q(.IJ .NIEOyiE 


•s 




UMUr » M*i> (100 rtr^. ,r I 


.. - W.rli«. l..,,.ort.) 




This project Is t 


tended to provide a utility t 




lal imaqes. The syste 


win consist of a powerful 3 




ture of medtun- and high-resolution video displays. 


Also, the system will 






tion of "-rays, micrographs 








delivery expected durt 


g the next fiscal year, and t 














isplay subsystem specffi- 


cations are complete. 


nd procurement is expected du 













Computerized Radiation Tlierapy 

CSL has developed a computer system, now in 
clinical operation in the Radiation Oncology Branch, 
NCI, to use the detailed contour and density 
information available from computer assisted 
tomography to improve radiation treatment planning. 
This system for external beam treatment planning is 
based on a generalized 3-D dose field model that 
covers photon, electron, and neutron beams. 

The computer program and most of its clinical 
implementation has been completed for the photon 
and electron fields available from the local 6 MV and 
12 MV linear accelerators. The current capabilities 
include interactive simulation of most irradiation 
techniques, including the effect of most beam 
modifying devices. The system enables the display 
of dose distributions computed in several transverse 
contours and overlaid on corresponding CT scans. 

Background and Objectives: During the past five 
years, CSL has maintained an active collaborative 
relationship with staff of the Radiation Oncology 
Branch, NCI. The goal of this effort is to develop and 
implement a generalized system for computer 
assisted radiation treatment planning. Initially aimed 
at utilizing CT scans during planning, the scope of 
the project has now broadened. 

Methods Employed: The dose field model 
developed by Jan van de Geijn was implemented in 
RSX-11M FORTRAN and experimentally tested for 
the local radiation facilities. The theoretical model 
was extended to cover irregularly-shaped beams as 
well as irregularly-shaped shielding blocks. Emphasis 
has been placed on optimization of interactive 
operational facilities and accommodations of input 
and hard copy techniques to clinical demands. 

Major Findings: The system is now in routine use 
for clinical treatment planning. It offers high speed 
computation and display of complete dose 
distributions in multiple slices, superimposed on CT 
images, including effects of wedge filters, shielding 
blocks, and diaphragm rotation. 

Significance to Biomedicai Researcti: The 
convenient interactive manipulation of key beam 
parameters in combination with fast response is 
highly valuable in complicated dosimetry problems 
encountered in special protocol studies. 

Proposed Course: 

• Implement the Dose Field Model for regular and 
irregular electron fields. 

• Establish multiple treatment planning stations to 
allow simultaneous use of the computer display 
equipment. 



32 



• Extend the capabilities to compute and display 
dose distributions in sagittal and coronal 
sections of the patient. 

Publications: 

Padikal. T . Lichter, A., Tepper, J,. Glatstein. E-. Schwade. J , Frednckson, 
H., Risso. W,, Roberson, P., Her. v., Van de Gei|n. J., and Kinsella. T , 
Experience with a CT Based Treatment Planning System In O'Neill. J 
T. (Ed ): Proceedings of the Fourth Annual Symposium on Computer 
Applications in Medical Care New York. Institute of Electrical and 
Electronics Engineers, 1980. pp 83-88 

Image Processing Facility 

This project is intended to provide a utility to display 
and analyze digital innages. The system will consist 
of a powerful 32-bit computer with a mixture of 
medium- and high-resolution video displays. Also, the 
system will include a microdensitometer to allow 
precise digitization of x-rays, micrographs and other 
images. The computer and peripherals have been 
purchased, with delivery expected during the next 
fiscal year, and the design for physical space to 
house the system is complete. Construction will 
begin soon, with completion expected during the 
next eight months. The display subsystem 
specifications are complete, and procurement is 
expected during the next year. 

Background and Objectives: This project arose in 
response to a critically overcrowded situation that 
exists on the present DCRT Evans and Sutherland 
Graphics computer. As image processing 
applications at NIH have increased, the limited 
resources of that graphics system have been 
saturated. During FY80, CSL, in collaboration with 
present and potential users designed a new general- 
purpose computer facility to aid the acquisition, 
display, and analysis of images such as electron 
micrographs, CAT scans, and radiographs. This 
facility will be available for use by the NIH 
community. 

Progress in FY81: The system will be based on a 
32-bit, one megabyte computer, with a smaller 16-bit 
processor to handle image acquisition. A multidisplay 
raster scan frame buffer will provide several users 
concurrent access to the central processor. Images 
will be digitized when necessary through a 
microdensitometer or a vidicon camera. Hard copy 
will be provided by a camera system. 

The computers and several related peripherals have 
been purchased, with delivery expected toward the 
middle of next year. A formal 'Request for Proposals' 
for the display sub-systems has been prepared and 
is awaiting funding for purchase next year. The 
design for the physical space of this facility is 
complete, and a contract for the construction should 



be awarded early in FY82. 

Significance to Biomedical Research: Study of 
images obtained in the biomedical laboratory is 
proving more and more fruitful as technology is able 
to supply the proper tools at a reasonable cost. 
Biomedical scientists are employing image analysis 
for a wide variety of research goals, and the use of 
such techniques is expected to grow very rapidly in 
the near future. 

Proposed Course: Pending the availability of funds, 
the outstanding system components will be 
purchased, the physical site will be prepared, and 
the development work to assemble the various parts 
into a working system will begin. 



33 





iluSiiii^'i!!. 


PflOJEOt ««.BER 






ZOl CT00075-02 


CSL 


"TlobeM. 1980 to September 30, 19B1 | 


IITL. » WJtei 1«0 =^.r.c..r, .r I,,.) 




Digital Imaging Applications tn Cardiovascular Research 




«»^.i'C'noNTa'«S'oN*''.lt «i*cJ *'" """ " """'"' '•"""""'" ""' *'' """ 




PI: Douglas Foxvog Computer Specialist CSL, 

Michael Jones Senior Surgeon H Ifi SU, 


NHLBI 


OTHER: Wfl Ham Barrett Staff Feltow CSL, 
James M. DeLeo Computer Systems Analyst CSL, 


°c"t 


Joseph Zulschenberger Clinical Associate H IR SU, 




GOOPIHIIIKC UNITS (if •'.,) 




Clinic of Surgery. NHLBI 




Computer Systems Laboratory 






Systems Desiqn Section 


j 






o.t« iPWOPHuiE eo<U;) 




j(.| mHu.iMXCi'. [j)(»)h.N»«r.sigts a{'=) -timB 




G (.1) KisDOS u (.J) iMCHvliys 






a1 topics in myocard 


jl am 


Digital image analysis has been applied to seve 
valvular research. This project has been renamed fro 


-Morphoraetric analysis of myoc 


ardial hypertrophy. Software has been devel 
of samples of hypertrophied myocardium from 
These packages allow determination of cell 


ped 


to enable interactive analysis 
and from experintental aniinals 






-Bioprosthetic heart valve studies. To investigate t 


e cause of stenosis 
jgraphy. Analysis c 


L 


imaged radlographlcally, and a 


'ter sectiol^nq'! by°pho 


-Reqional myocardial blood flo 


w Quantitative data from experimental anim 


displayed as a map of the hear 
atlons in intensity or color. 


Comparison of epicardia 


'-""°"""'""°"'' 



'So?Sit'6F" 



ZOl CT00O73-O2 CSL 



1980 to Septerrber 30. I9B1 



Cerebral Metabolic InE 



Laboratory of Cerebral Metabolism (LCH), NIMH 



CnnpiitPr <;y;tfm<L l aboratory 



flTRT WIH Bethesda Maryland 20205 



-flJ I [LJ_ 



showing cerebral metabolic activity in micro tornei 
- ^ntal animal brains has been developed to facili 



aligned paralled Inages. Use of the algorithm, which is based on minimizing 



ilgorlthm development on the i 



i NIHH De Anza Imaging 



Digital Imaging Applications in Cardiovascular 
Research 

Digital image analysis has been applied to several 
topics in myocardial and valvular research. This 
project has been renamed from last year to reflect a 
somewhat broader scope. 

• Morphometric analysis of myocardial 
hypertrophy. Software has been developed to 
enable interactive analysis of samples of 
hypertrophied myocardium from humans and 
from experimental animals. These packages 
allow determination of cell diameter, nuclei 
densities, capillary densities, and percent 
interstitium. 

• Bioprosthetic heart valve studies. To investigate 
the cause of stenosis of prosthetic valves, 
diseased valves are removed from experimental 
animals and imaged radiographically, and after 
sectioning, by photography. Analysis can 
quantify sites of calcifications and the amounts 
of fibrous sheath and normal tissue. 

• Regional myocardial blood flow. Quantitative 
data from experimental animals are displayed as 
a map of the heart, with increases in blood flow 
indicated by variations in intensity or color. 
Comparison of epicardial-endocardial flows and 
control vs. test data will be studied. 

Background and Objectives: This project applies 
computer assisted digital imaging techniques to 
areas of interest in cardiovascular research. The 
project began last year with the morphological 
analysis of cardiac tissue, and this year includes a 
preliminary study of bioprosthetic heart valves and 
development of a system to examine regional 
myocardial blood flow. 

Methods Employed: 

• Morphometric Analysis. Specimens are 
microtomed, mounted, stained, and 
photographed under a light microscope. 
Resultant negatives are then digitized by 
microdensitometry, and entered into the Evans 
and Sutherland display computer. The programs 
calculate myocyte cross-sectional areas, 
myocyte diameter and perimeter, capillary and 
nuclei count, and statistics on one or more than 
one sample. 

• Bioprosthetic heart valves. Transverse 
microtome sections taken from the valve leaflets 
and radiographs of the valve are scanned for 
computer input. The amount of calcification, 
fibrous sheath, and normal leaflet tissue can be 
determined quickly and efficiently using an 
interactive contouring and thresholding algorithm. 



34 



• Regional blood flow. Data are obtained from the 
injection and tracking of tagged microspheres. 
The epicardial and endocardial wall of the right 
and left ventricle as well as both sides of the 
septal wall are examined. Numerical data 
consisting of regional blood flow measurements 
are represented graphically as varying shades of 
intensity or color. Values between adjacent 
regions can be interpolated to form a more 
continuous surface, with high to low flow 
variations repsented by a rainbow color 
spectrum. This should provide an effective 
means for the visualization of flow patterns in a 
single animal and of the distribution of flow 
patterns over the entire experimental population. 

Progress in FY81: An interactive computer program 
MORPH has been used to study the left ventricular 
mid-anterior wall of adult foxhounds as part of a 
preliminary study of the effectiveness, accuracy, and 
utility of computer-assisted morphometric analysis. In 
addition to inherent reproducibility, the variance per 
sample analyzed decreases tenfold when compared 
to manual point counting techniques. There is also a 
ten percent reduction in the standard error of the 
estimate for the mean interstitial fraction of a tissue 
evaluated from five separate microscope fields. The 
primary benefit of this method is in facilitating the 
analysis of multiple fields and multiple slides per 
tissue sample. Increasing the number of samples 
decreases the overall variance proportionally. 
Programs that display regional blood flow and 
provide ratio comparisons of endocardial and 
epicardial flow have been completed. 

Significance to Biomedica/ Researcti: These digital 
image analysis techniques provide scientists with 
accurate, reproducible, and efficient tools, 
superceding alternate manual approaches such as 
planimetry or point counting. 

Proposed Course: This project will continue to 
refine the morphometric analysis now in use, and 
extend the techniques to three-dimensional 
examination of cardiac muscle cells by 
reconstruction from thin slices of human and animal 
cells. The prosthetic valve study will resume after 
leaflets are resectioned, to produce a more 
anatomically consistent data base than was 
previously obtained. Regional blood flow studies will 
continue. Methods are to be devised to meaningfully 
compare myocardial blood flow during different tests 
on one animal, as well as to compare different 
animals. Correlation of endocardial with epicardial 
flows will also be provided. 



Cerebral Metabolic Imaging 

An algorithm to align spatially sequential, parallel 
images of autoradiographs showing cerebral 
metabolic activity in microtomed sliced sections of 
experimental animal brains has been developed to 
facilitate arbitrary plane viewing in the three- 
dimensional image space resulting from the stacking 
of aligned parallel images. Use of the algorithm, 
which is based on minimizing non-overlapping areas, 
demonstrates the need for fixed fiducial reference 
points in specimen images. Preliminary attempts in 
prepanng specimens to provide fiducial points 
illustrate the difficulty of this task. LCM will continue 
algorithm development on the newly installed NIMH 
DeAnza Imaging System. CSL will remain available 
for consultation. 

Bacl<ground and Objectives: The Laboratory of 
Cerebral Metabolism, NIMH has been engaged in 
computer-assisted image processing of 
autoradiographs that illustrate cerebral metabolic 
activity in microtome-sliced sections of experimental 
animal brains since 1977. In 1979, the Computer 
Systems Laboratory was asked to collaborate in the 
development of a methodology to enhance and align 
spatially sequential microtome brain slice images for 
the purpose of reconstructing alternate views 
through this three-dimensional image data. 
Algorithms developed by CSL on the DCRT Evans 
and Sutherland Image Processing System were to be 
transportable to a DeAnza image processing system 
on order by NIMH. 

Methods Employed: CSL has implemented software 
to align sequential images by an algorithm which 
minimizes overlapping areas. The limitations of this 
approach are especially evident when aligning 
images in which brain segments are physically 
detached. It was, therefore, agreed to investigate 
new methods of specimen preparation that would 
provide fixed fiducial marker points to enhance 
automated alignment. 

Major Findings: To assure quality three-dimensional 
imaging of cerebral metabolic activity, it is essential 
to have sufficient internal and/or external fixed 
fiducial marker points to facilitate automated 
alignment. Unsuccessful early attempts in specimen 
preparation to provide fiducial points illustrate the 
complexity of this problem. 

Proposed Course: Procedures and algorithms 
developed by CSL have been given to LCM. Now 
that the new NIMH DeAnza image processing 
system has arnved, LCM staff members will continue 
algorithm development on that system. CSL staff will 
continue to be available for consultation in this work. 



35 



'^-^^---^' 


,.3i"!s!»i'I. 


ZOl CT00080-01 


CSL 


""'octobe"l. 1980 to Septenber 30. 1981 1 


(tiki » FHOJtei (BO <h,r.cl.r* or 1...) 






Computer Analysis of Gel Electrophoresis 






itwir HiwrwT .NO ihsllTuu ifniTiTiwisr^soTrru'. o< pfliNcipiL f 






PI: B. L. Trus Rese.rch Chemist 


est. DCBT _ 




OTHER: ». iHoOe. Staff Fellov. 


CE. NIAODK 

NIADOK 

tea. NIAOOK 

Program 


1 
searci, 












! 




k../...«. 




















^^ p»,i.,».L, u^ """■ } 


cici «»onii.ii .o.(tsl 

:l.) »».«!..jttl! Q|.).™ili:nl! -J 


(.) ,n«. 




al.il ...•.! D(.i) i.u...e-s 






sm.i.Y tr g»>. III. ..™. ., 1.. .,„i., ..,-..1 ) 








r^r'the coTOuter'a 


alysls 


qated. 


Dqraphs are beinq i 


1 gels 



«ijEC?'NSBM'f3"lio"«rIhlt JJf?^" 


'ZsiSlilir 


TcToI^-oa CSL 




Automated Analysis of Plaque Formation in Experimental Atherosclerosis 


PI: HilHam Barrett Staff Fellow CSL. OCRT 

OTHER: James H. OeLeo Computer Systems Analyst CSL. XRT 

Donald Fry H.O. Ohio State University 
J. F. Cornhill M.D. Ohio State University 


COOMRHIhC UNITS (it ..,) 

None 


Conputer Systems Laboratory 


Systems Design Section 


'DCRT'."NTH.''Bet'h"esda, Maryland 20205 


lOMi «mEMs, |Pfl<irtssiw*L,^ joiner. 


:(.) .iw..N iuB^cii u(t)w«'".iiiw£s aic)NEimR 


Computer programs have been developed and package 
quantitatinq atherosclerotic disease formation and its 


d for detecting and 








Ifet. and vessel wall permeabi 


ity. Permeability is indicated with Evans" Blue 
^TSudan IV Red dye. Colors are separated using 

dinate system. An iterative edqe-findinq 


algorithm searches the image f 

coincidence of disease at a gl 
longroent distribution of the 

it^tologlc. physical, blochemi 


te diseased areas. Topographic naps, formed bv 
from different animals, show the incidence and 
ven location. The raps describe the extrenely 
disease in a concise and quantitative manner and 
tool for correlation of disease patterns with local 



Computer Analysis of Gel Electrophoresis 

This project, which was transferred frorri NIDR during 
this reporting period, is intended to design hardware 
and software mechanisms for the computer analysis 
of one- and two-dimensional gel electrophoresis. 
Image processing is being used to evaluate and 
compare differences in protein composition. Only 
computer analysis can make such comparisons 
strictly quantitative. Analysis of two-dimensional gels 
is completely automatic, while analysis of one- 
dimensional gels is interactive. Additional 
applications with autoradiographs are being 
Investigated. 

Background and Objectives: The objective of this 
project is to develop computer software that can 
automatically analyze photographs of two- 
dimensional gels or autoradiographs, and to 
optimally utilize available hardware for accuracy and 
reliability. In addition, two-dimensional photographs 
of one-dimensional gels, using the PIC system can 
be easily converted into one-dimensional 
representations for quantitative comparison or 
integration. 

Significance to Biomedical Research: Use of gel 
electrophoresis or autoradiographs is commonplace 
in chemical, biochemical, and biomedical analysis. 
However, the quantization or comparison of such 
gels is tedious, qualitative, and difficult. Use of the 
Evans and Sutherland system to evaluate such 
samples permits scientists greater information with 
increased reliability and simplicity of operation. 

Progress in FY81: This project, which was begun in 
1 980, has produced many successful results. 
However, as new projects contain slightly differing 
needs, there is an ongoing expansion and 
generalization of software to satisfy these needs. In 
addition, the number of laboratories that use the 
methods developed in this project is increasing 
rapidly. Some new laboratories will be using the 
methods collaboratively, while others will be 
independent. 

Metiiods Employed: Samples are scanned on the 
Perkin-Elmer microdensitometer and stored on 
magnetic tape for later processing. Two-dimensional 
gels are processed by a program, CI NT (Computer 
Integration), which corrects for background, 
automatically locates peaks, automatically integrates 
all peaks, and provides a printed summary of results. 
In addition, an image can be obtained either from the 
video frame buffer or from a Calcomp plotter. Two or 
more gels can be compared graphically by means of 
a program OVERLP which correlates two gels and 
provides data to generate overlapping Calcomp plots 



36 



with each gel as a different color. 

Proposed Course: Computer software will be 
expanded to encompass new types of data. A prime 
consideration of software development has been 
machine independence, and the software can thus 
be easily transported to other private facilities. With 
the acquisition of the image processing laboratory 
hardware, the present programs will be converted to 
execute more quickly and efficiently. 

Publications: 

Nikodem. V M., Trus, B. L., and Rail, J. E.: Two.-Dimensional Gel Analysis 
of Rat Liver Nuclear Proteins alter Thyroidectomy and Thyroid Hor- 
mone Treatment. Proceedings of the National Academy of Sciences 
(in press). 

Automated Analysis of Plaque Formation in 
Experimental Atherosclerosis 

Computer programs have been developed and 
packaged for detecting and quantitating 
atherosclerotic disease formation and its location 
along the entire arterial tree, and and for statistical 
correlation of disease patterns with anatomy, diet, 
and vessel wall permeability. Permeability is 
indicated with Evans' Blue dye and disease is 
indicated with Sudan IV Red dye. Colors are 
separated using appropriate filters in the scanning 
process and the resulting digital images are 
transformed to a standard coordinate system. An 
iterative edge-finding algorithm searches the image 
for an optimal threshold of maximum optical intensity 
to identify and isolate diseased areas. Topographic 
maps, formed by overlaying the disease patterns 
from different animals, show the incidence and 
coincidence of disease at a given location. The 
maps describe the extremely congruent distribution 
of the disease in a concise and quantitative manner 
and provide the researcher with a tool for correlation 
of disease patterns with local histologic, physical, 
and biomedical events. 

Background and Objectives: Atherosclerosis was 
produced experimentally in 41 animals by feeding 
them high-calorie, cholesterol-rich diets. Eleven 
control animals were also included in the population 
for comparison with the experimental group. Pnor to 
sachfice, animals were injected with Evans' Blue 
dye, an indicator of vessel wall permeability. A 
systematized necropsy procedure and standardized 
format were developed for arranging and fastening 
the opened arterial tree to reflective panels, after 
which the vessel segments were stained with Sudan 
IV Red dye to indicate areas of disease. 

Although atherosclerosis is a multifaceted disease, 
experimental studies indicate that this disease is very 
discrete and local in occurrence. Moreover, the 
patterns of distribution of sudanophilic lesions tend 



to have a characteristic common topography in a 
given species. Thus, it is important to define the 
spatial occurrence statistically to facilitate correlation 
of disease patterns with local structural, chemical, 
histological, and fluid mechanical events. 

Methods Employed: Thirty-five millimeter slides of 
the preparations are scanned at a resolution of 50 
microns using neutral density, red, and blue filters. 
Digitized images are displayed and analyzed on an 
Evans and Sutherland Video Frame Buffer. Images 
are transformed to a common coordinate system 
through the use of local anatomical landmarks, the 
coordinates of which are identified by a technician 
familiar with the anatomy, using a graph pen and a 
tablet that is interfaced to the computer (PDP-1 1/70) 
driving the display. An iterative edge-finding 
algorithm outlines the image at an initial intensity 
value, and then performs an orthogonal search of 
the outlines to identify an optimal threshold, which is 
used to segment the image into areas of disease 
and non-disease. 

Thresholded images corresponding to a given 
arterial segment are added together to produce a 
topographic map representing the location and 
relative frequency (probability) of lesion occurrence 
in a given population of animals. The map is 
displayed on a video screen by using various 
intensity values or colors to represent levels of 
incidence and coincidence of the disease. The 
intersection of two different maps provides a 
quantitative measure of the correlation between 
disease patterns as they occur in different groups of 
animals, or between different variables within the 
same group. 

Major Findings: Programs for image processing and 
analysis, file management, storage, retrieval, and 
review of results, have all been efficiently integrated 
into a system. Close to 1000 images have now been 
processed, producing topographic maps showing the 
statistical distribution of the disease along major 
arterial pathways. The eight hours previously 
required to manually identify landmarks and 
transform images has been reduced to 
approximately three minutes, making this objective 
and statistical approach practical. In addition, 
automated thresholding of disease patterns 
compares favorably with manual methods, while 
reducing the time involved and assuring 
reproducibility. 

The probability maps illustrate the consistent 
distribution of disease patterns and show a strong 
relationship between these patterns and the 
anatomical detail in the vessel wall. This is striking 



37 



"ISJcfiliSsiS'lSs'toi^S?!!'';!?!! iJST' 


i« i.«!lli''«i«SCT"«rjJoi 


ZOl CTO0O74-O2 CSL 


TTnhr'!' IQW) tr. <;.p,«nh.r W IQfll 


Computerized Three-Dimensional Hodel of the Cat's Brain Stem 


PI: Douglas A. Foxvog Computer Specialist CSL, DCRT 
James H. DeLeo Coinputer Systems Analyst CSL, DCRT 
Joe Adams Senior Staff Fellow LND, NINCDS 


Laboratory of Neuro-otolarngology. LND, NINCDS 




Systems Design Section 












A ihree-dirwnsional. colo 


red. 50^ volume element 
allows rapid selection 

al animals. InvestT^^ 


block model representing 


the neuronal structure of the 




data from different e;(oeriment 


f a viewing surface com- 
ical and physiological 


MrilSng'EhrpIaneTlcribed 


by various electrode penetrations in the experi- 




been developed and used successfully on the Evans 
ation-quality output has been produced. An atlas 







s'f3SSif^iif!i.';.'?.T%aiT5: 



ZOl CT 00081-01 CSL 



- 1, 19B0 to September 30. 19B1 



1 Medicine Department Conputer System 



1 Medicine Department, 



Systems Design Secti 



; the development of c 



H\ti Clinical Center. 



<. begun in Frei ■ 



1 

■»nl\/ ^B 



evidence that the formation of lesions is strongly 
influenced by local factors such as hemodynamic 
forces acting on the surrounding anatomy. Areas of 
coincidence or 'hot spots' indicated a high statistical 
occurrence of sudanophilic lesions and may point to 
key locations in the genesis of the disease. Maps 
have also been created as a function of diet, and 
show a dramatic difference in the propagation of 
disease patterns when factors such as cholesterol 
and thyroid inhibitor are added to the diet. 
Preliminary results show a high correlation (r=.92) 
between maps created from the red and blue stained 
images, indicating a striking relationship between 
sites of lesion formation and vessel wall permeability. 
However, this relationship appears to be most 
pronounced near the periphery of a lesion when 
studied on an individual basis. 

Proposed Course: The present study will culminate 
with the processing of the coronary arteries, a site of 
vital clinical interest. Correlation between Evans' 
Blue and Sudan IV images will be completed in an 
effort to further quantitate the relation between 
vessel wall permeability and disease. It is hoped that 
results of this study will not only help to characterize 
the pattern and distribution of atheromatous plaque 
formation, but will provide the researcher with a tool 
for further investigation into the cause and 
mechanism of the disease. 

Publications: 

Barrett, W. A., DeLeo, J, M., Cornhill, J. F., and Fry, D. L.: A system for 
Automated Analysis of Plaque Formation in Experimental Atherosclero- 
sis. 53rd Scientific Sessions of the Americal Heart Association, Novem- 
ber 19, 1980. 

Barrett, W. A.: An Iterative Algorithm for l^^ultiple Threshold Detection. IEEE 
Conference on Pattern Recognition and Image Processing, August, 
1981. 

Computerized Three-dimensional lUlodel of the 
Cat's Brain Stem 

A three-dimensional, colored, volume element (60- 
cubed) block model representing the neuronal 
structure of the superior olivary complex of a cat's 
brain stem has been developed. The model allows 
rapid selection of a viewing surface computed for 
any selected cutting plane through the model 
volume. The model provides a common frame of 
reference for comparing anatomical and 
physiological data from different experimental 
animals. Investigators should be able to view a 
section of the 'normal' cat on the same plane M 

described by various electrode penetrations in the \ 
experimental animal. The model has been developed 
and used successfully on the Evans and Sutherland 
System. Publication-quality output has been 
produced. An atlas of over 1 50 views is being 
submitted for publication. 



I 
I 



38 



Background and Objectives: The objectives of this 
project were: 

1. To develop and implement a three-dimensional 
block model of a cat's brain stem for laboratory- 
collected data specifying cell types and their spatial 
relationships. 

2. To provide a method for producing publication- 
quality colored and black-and-white two-dimensional 
graphics that illustrate the three-dimensional 
graphics block model. 

Methods Employed: Consecutive 80-micron-thick 
microtomed slices for the superior olivary complex of 
the cat's brain stem were stained and examined by a 
neuroanatomist by means of a Zeiss microscope. 
The stage of the microscope was attached to a 
computer, as was a function box, thus enabling the 
neuroanatomist to log spatial coordinates and cell 
types in a computer data file. This data was 
transferred to the Evans and Sutherland System and 
used as input to the volume element (60-cubed) 
block model program specifically developed for this 
project. The block model software allows the 
investigator to view the surface of any selected 
cutting plane through the model volume. 

Progress in FY81 Publication-quality output has 
been produced. The software developed for this 
project has shown itself to be applicable to several 
other biomedical imaging applications. 

Proposed Course: This project has been 
completed. It is expected that the block model 
software will be extended to other applications. 

Rehabilitation Medicine Department Computer 
System 

This project involves the development of computer 
techniques in rehabilitation medicine in collaboration 
with the Rehabilitation Medicine Department of the 
NIH Clinical Center. CSL has recommended 
computer techniques that can be used to 
automatically acquire anatomical and physiological 
information from patients, perform the required 
calculations on the data obtained, and display the 
necessary results to the medical staff. The 
automated techniques include: the measurement of 
body forces (hand and ground reaction forces), 
muscle activity (monitoring the electromyogram of 
muscles), and body kinematics (the position and 
angles of the limbs and joints in space and time). 
The system will allow the medical staff to easily 
enter patient and staff data into a data base with 
computer-generated forms displayed on a terminal 
screen, and to perform inquiries and generate 
reports using the accumulated data. In FY82 CSL will 



continue the work begun in FY81 including the 
specification of the computer system, the evaluation 
of methods to perform the desired measurements, 
the selection of the necessary transducers and 
instrumentation, and the specification of the required 
software. 

Bac/<ground and Objectives: The Rehabilitation 
Medicine Department provides psychiatric evaluation 
and treatment, physical therapy, occupational 
therapy, and speech therapy for NIH Clinical Center 
patients referred by Institute physicians. In addition, it 
develops various indices to evaluate these services. 
The department supports the efforts of and 
collaborates with Institute physicians engaged in 
research relevant to physical rehabilitation medicine. 
It also initiates both clinical and basic research 
independent of the Institutes in the rehabilitation of 
mentally and physically handicapped individuals. 

In support of these goals, CSL is specifying a 
computer system for the Rehabilitation Medicine 
Department. Initially, the department will use the 
system for the following three projects: 

1. The Physical Therapy Quality Assurance 
System-a data base system that will be used to: 
assess medical staff effectiveness in providing the 
types of patient care needed, determine staff 
workload and scheduling, and identify areas for 
clinical research for the Physical Therapy Service; 

2. The Hand Dynamometer Instrument-a device 
that will be used to measure the magnitude and 
direction of the forces in the hand and to develop 
clinical tests to diagnose the mechanical and 
functional status of the hand, arm, and shoulder; 

3. The Automated Biomechanics Laboratory-a 
system that will be used to automatically measure: 
the position of the limb segments in space, the 
forces in the lower limbs, and the electromyographic 
signals from the muscles in the limbs. The computer 
will perform measurement, analysis, display, and 
recordkeeping functions in these and future 
applications. 

Significance to Biomedical Research: The computer 
system will be used with arthritic, orthopedic, and 
neurological patients, and with amputees in order to 
evaluate drug therapy, orthopedic and prosthetic 
devices, and medical interventions. It also will be 
used as a teaching tool to help these patients learn 
to function with their disability in an efficient manner. 
Many hospitals in the United States are presently 
establishing automated biomechanics and gait 
analysis laboratories. Therefore, any new 
developments made on this project will benefit users 
of these automated systems, as well as patient care 



39 



iffifJSl.S'ff.K'Si'ffi.'S.T* 


!31i!!^":L' 


Z01 CT00082-01 CSL 




Image Processing of Electron Micrographs 


PI; 8. L. Trus Research Chemist CSL. DCRT 

OTHER: A. C. Steven Visiting Associate LPB. NIADDK 
K. A. Piei Chief LB, NIDR 


™:;r™.r;';;:«™ 


Computer Systems Laboratory 


""system Design Section 


DCRT. NIH. Bethesda. HO 20205 


0.6 r "''**^' 0.6 \' 


CHLC« .PPI.OMIUIE m(£S) 

:(.) HUMAN SuWu; a(6)HUI-.N T.SOUIS aU)'(E.IH£fl 


i<L infpnrlpd to dpsian hardware and software mechanisms for the image processinq 
and imaqe reconstruction of electron micrographs. Computer software that has 
been developed is used by a number ot NIH research groups and programs have been 
exported to outside facilities. Under investigation at tliis time as part of 
this project are virus structures and collagen structures; other groups are look- 


ing at numerous other biological specimens such as keratin, membrane structure. 



October 1, 1960 Co September 30, 1981 



1 Tomography (PET) Scan Image Analysis in Aging Studif 



Computer Systems t 



i Design Sectioi 




and clinical research within the Rehabilitation 
Medicine Department at NIH. 

Progress in FY81: CSL's collaboration with the 
Clinical Center's Rehabilitation Medicine Department 
began this year. Areas that could benefit from 
computer support were identified and specifications 
for the required computer hardware and software 
components were developed. CSL established the 
preliminary requirements for the department's 
computer system. 






A large amount of specialized instrumentation is 
needed to perform the required automated 
measurements. The Biomedical Engineering and 
Instrumentation Branch of NIH's Division of 
Research Services built the hand dynamometer. CSL 
acquired information on commercially available force 
plates that are used to measure ground reaction 
forces and portable electromyogram telemetry units 
that transmit the muscle signals from the patient to 
the computer without any impeding cables. CSL also 
evaluated camera systems that can automatically 
measure body kinematics including the United S 

Technologies Research Center infrared system, the^ 
Oxford Medical Systems strobe light system, and the 
Selcom Light Emitting Diode system. 

Proposed Course: CSL will continue to offer advice 
regarding computer hardware and software, 
transducers, and instrumentation. Rehabilitation 
Medicine expects to purchase these components 
during the coming year. 

Image Processing of Electron Micrographs 

This project, which was transferred from NIDR during 
this reporting period, is intended to design hardware 
and software mechanisms for the image processing 
and image reconstruction of electron micrographs. 
Computer software that has been developed is used 
by a number of NIH research groups, and programs 
have been exported to outside facilities. Under 
investigation at this time as part of this project are 
virus structures and collagen structure; other groups 
are looking at numerous biological specimens such 
as keratin, membrane structure, immunoglobulin 
structure, and muscle structure. 

Background and Objectives: The objective of this 
project is to develop a general-purpose software 
package for the analysis of electron micrographs. In 
addition, the computer analysis requires optimal 
utilization of the available hardware. These 
techniques and software have been used in 
independent applications by some researchers, and 
in collaborative projects by others. 

Significance to Biomedical Researcii: Computer 



40 



analysis of electron micrographs is a relatively recent 
addition to the tools available for structural analysis. 
It is an extrennely powerful technique when applied 
to two-dimensional crystalline structures, and can be 
used to correlate and align similar particles that are 
not crystalline. In addition, images can be corrected 
for a number of artifacts and experimental problems. 

Progress in FY81: This project, which was begun in 
NIDR, has involved some growth in software and 
hardware methods, but primarily has grown in the 
utilization of programs and in the number of people 
who are using the software system PIC. In addition, 
use of Brookhaven STEM data is markedly 
increased. PIC has been expanded to interact with 
much of the general purpose image processing 
software in the image processing computer system. 

Computer analysis of electron micrographs of beet 
necrotic yellow vein virus (BNYVV) has been 
completed. The structure is similar to that of tobacco 
mosaic virus (TMV), but important differences exist. 
The spatial arrangement of BYNVV RNA (49 
residues per helical turn, pitch = 2.6nm) is virtually 
identical to that of TMV (49 residues per turn, 
pitch = 2.3nm). However, the helical packing of the 
protein cost is significantly different. The BNYVV 
protein has a helical packing of 49/4 subunits per 
turn and 4 bases per subunit while TMV has 49/3 
subunits per turn and 3 boxes per subunit. 

Proposed Course: This project will continue 
software development as needed, and will be 
converted to utilize the new image processing facility 
when it is available. In addition, because new 
biological structures are regularly available, these will 
be examined. 

Publications: 

Steven, A. C, Trus, B. L.. Putz. C, and Wurtz. M.: The Molecular Organiza- 
tion of Beet Necrotic Yellow Vein Virus. Virology (in press) 

Trus. B. L. and Steven, A. C: Digital Image Processing of Electron li^icro- 
graphs-The PIC System. Ullramicroscopy (in press). 



Positron Emission Tomography (PET) Scan 
Images Analysis in Aging Studies 

Exploration of the use of the combined power of 
Computer Assisted Tomography (CAT) scanning and 
Positron Emission Tomography (PET) scanning to 
study brain cell metabolism in disease associated 
with aging is the purpose of this project. The initial 
goal is to provide a computerized image analysis 
procedure to delineate brain substructures 
represented in spatially, sequenced CAT scan 
images and to determine metabolic activity in these 
substructures from corresponding sequenced PET 
scan images. 

Bacl<ground and Objectives: Positron Emission 
Tomography (PET) scanning performed in the 
Nuclear Medicine Department of the NIH Clinical 
Center pi^ovides a spatially sequenced series of 
images of regional cerebral glucose metabolism in 
man. The Laboratory of Neurosciences of the 
National Institute of Aging wishes to incorporate PET 
scanning technology in the study of disease 
associated with aging. The initial goal of this project 
is to delineate brain substructures represented in 
spatially sequenced Computer Assisted Tomography 
(CAT) scan images and to determine metabolic 
activity in these substructures from corresponding 
spatially sequenced PET scan images. 

Met/iods Employed: Methods for establishing 
external coordinates to align corresponding PET and 
CAT scans are being investigated. Equations for 
converting radiation units to glucose utilization units 
are being finalized. Procedures for transporting PET 
scan and CAT scan images from the NIH Clinical 
Center to the DCRT Image Processing Facility have 
been established. Software is being written to 
determine statistical characteristics of PET scan 
regions of interest as delineated manually on 
corresponding CAT scan images. 

Significance to Biomedical Research: It is 
anticipated that this work will provide a basis for 
evaluating the utility of PET scanning in studying 
diseases associated with aging. Successful 
implementation of an external coordinate system 
should provide for accurate anatomical region 
designation via higher resolution CAT scan images 
to measure physiological processes from 
corresponding lower resolution PET scan images. 

Proposed Course: Upon implementation of an 
external coordinate device and finalization of 
radiation-to-glucose conversion equations, analysis 
will commence. 



41 




s!SinM'p.K:'!if::s!S" 


!™!SiLSr' 


ZOl CT00084-0, CSL 




nut Of WJICI (80 ch>r>ctT> «r Utt) 

Computer Analysis of Autoradiographic Images of Recombinant ONA Colonies 


PI: James M. DeLeo Computer Systems Analyst CSL. DCRT 
Floyd Taub Research Associate LB, C 

OTHER: Brad Thompson Section Chief LB, C 


Laboratory of Biology (LB), C 


* Computer Systems Laboratory 


Systems Design Laboratory 




L «..»«., ^^ pOftSSIWAL, ^^ |OTH£H. 




A computerized methodology for analyzing autoradiographic spot images asso- 
:iated with recombinant UNA bacterial colonies has been developed in collaboratio i 
<ith scientists in Nil. Ihis system represents a unique refinement in a method 

ievelopmentally or homonally induced. Spot density measurements are computed 
from dipitized images produced via microdensitometry. These measurements are 
:orrected for variability in exposure and local background, calibrated to 
lybridization standards, and normalized for comparison purposes. The system 
irovides a variety of graphical and tabulation output that effectively sunwarizes 









"'"'"'■'"'*'•'"""■'''■'''" 


.l«i^";iii'fifcT 


ZOl CT00085-01 


CSL 




igHi 




Cataract Grading via Computerize 


d Slit-lamp Image Analysis 




PI: James M. DeLeo 

Robert D. Sperduto 
Leo T. Chylack, Jr. 


Computer Systems Analyst 

Chief, Division of Ophthalmolrtgy 


" 


Office of Biometry & Epidemiology (BE), NEI 
Division of Ophthalmology, Harvard Medical School 


Computer Systems Laboratory 


Systems Design Section 






OCRT. NIH^'se'thesda. HD 20Z05 | 


,«.«..«»«:., ^^^ |™W£SSI«H. 


0. ■""■ 




""■'"-"»•- ,»...»,.„ ,,.,„«. 


The aim of this project is t 




sch«e- 




SjeTCped 




of 




aj(ls for purposes of comparing 
Early results demonstrate the 

ducted to further detomiine th 


for improved quali 
the methodology. Ex 


en lens ha; been d 
f this approach. 



Computer Analysis of Autoradiographic Images 
of Recombinant DNA Colonies 

A computerized methodology for analyzing 
autoradiographic spot images associated with 
recombinant DNA bacterial colonies has been 
developed in collaboration with scientists in NCI. 
This system represents a unique refinement in a 
method to directly identify cloned seqences 
complementary to messenger RNA that are 
developmentally or hormonally induced. 

Spot density measurements are computed from 
digitized images produced via microdensitometry. 
These measurments are corrected for variability in 
exposure and local background, calibrated to 
hybridization standards, and normalized for 
comparison purposes. The system provides a variety 
of graphical and tabulation output that effectively 
summarizes experimental results and identifies 
significant induced hybridization events. 

Background and Objectives: NCI scientists have 
been refining techniques to directly identify cloned 
sequences complementary to messenger RNA that 
are developmentally or hormonally regulated. This 
refinement has led to a methodology which produces 
autoradiographic spot images representative of cell 
colony hybridization. The objective of this project is 
to provide an automated procedure for a quantitative 
analysis of understanding of these images. 

Methods Employed: Radioactively-induced cell 
colonies are placed into microliter wells, grown on 
agar, transferred to filter paper.and hybridized to 
end-labeled mRNA or cDNA probes. 
Autoradiographs of the filters are digitized and the 
density of each spot relative to background is 
established by means of CSL-developed image 
processing software operational on the DCRT Evans 
and Sutherland PDP-1 1 /70 computer system. 
Compensation for variations in background, film 
exposure conditions, and hybridization are included 
in the methodology. A variety of graphical output 
including scatter diagrams, histograms, and listings is 
provided. 



Significance to Biomedical Research: Classical 
solution hybridization techniques are too 
cumbersome to be performed en masse. The 
methodology developed allows quantitative 
hybridization studies on a large number of 
sequences. Earlier qualitative assessment of 
autoradiographic spot images is now superseded 
with automated procedures yielding more accurate, 
more reproducible data. Computer graphic 
presentation of results greatly facilitates identification 
of significant experimental events. 



I 



42 



I 



Proposed Course: Production image analysis has 
started and will continue. Further software 
refinements, to include more complete automated 
analysis and new graphics software, are planned. 

Cataract Grading via Computerized Slit-lamp 
Image Analysis 

The aim of this project is to develop an objective 
cataract grading scheme--a goal having high priority 
among cataract researchers. Slit-lamp camera 
images of the human eye lens are entered into the 
DCRT Image Processing System via 
microdensitometry methods. A method for 
quantitating opacity along the visual axis for 
purposes of comparing similar images of a given 
lens has been developed. Early results demonstrate 
the need for improved quality control in image 
preparation to assure the success of the 
methodology. Experiments are being conducted to 
further determine the accuracy limitations of this 
approach. 

Background and Objectives: A major problem for 
cataract researchers has been the lack of an 
objective, reproducible in vivo cataract classification 
scheme. Subjective classification methods are 
currently depended upon. With the tremendous 
variability in the morphology of cataracts, it is difficult 
to rely on such methodology either in survey work or 
in longitudinal studies. Development of an objective 
cataract grading scheme is seen as a high priority 
item among cataract researchers. 

Methods Employed: Images of the human eye lens 
obtained from Topcon and Zeiss slit-lamp cameras 
are digitized via microdensitometry and the resultant 
qualified image representations are entered into the 
DCRT Image Processing System for analysis. 
Analysis consists of computing a histogram of the 
density(opacity) values within a narrow band 
centered about the visual axis. The resultant 
histogram is partitioned into five equal bands and a 
corresponding color isodensitometric image is 
displayed. A quantitative distribution of percent 
involvement for each of the five bands is also 
provided. Time-spaced images of the same eye lens 
may be compared with these results to track disease 
progression or regression. 

Major Findings: The interactive computer system 
developed for quantitating visual axis opacity is 
operational and easy to use. Analysis of initial 
prototype images suggests the need for improved 
quality control in all aspects of image preparation. 

Significance to Biomedical Research: Development 
of an objective cataract grading scheme is seen as a 
high priority item among cataract researchers. 



Proposed Course: Studies on a larger population of 
images are planned. These studies should result in 
procedures to facilitate image comparison. The 
studies should also provide accuracy measurements 
for variables introduced by factors such as camera 
type, photographic procedures, microdensitometry 
procedures, and computer analysis. 



43 



' " ™ '*"■ '■oflck'i* 



ZOl CT00076-02 



October I, 1980 to Septentier 30, 1981 



linage Analysis In Coirputerijed Tomography (CT) Contrast Kate 



PI: Jares M. DeLeo 
Michael Verness 


Computer Systeir 


IS Anali^st 




Diagnostic Radiology (DR). CC 


Coonuter Systenis Laboratory 


Systems Design Section 








DCflT, NIH. Bethesda. Marylan 


20305 






10m«Hn«„ 1'^'"" 


*;" 1*"""' 








The intent of this project i 

efficacy of a ne« experiment 

the liver. Software for oprf 
Inaqe Processinq System. Hea 


to provide a tool for 
F coitputerized tomogra 


the quantita 
m (Cf) to * 
at was develo 


tl.e analysis 
ed at the NIH 


the demonstration of e 
classical image te*tu 




s are being 


injection tomograms. Several 
tried as discriminating featu 
effectiveness. 



44 



Image Analysis In Computerized Tomography 
(CT) Contrast Material Evaluation 

The intent of this project is to provide a tool for the 
quantitative analysis of liver and spleen section of 
computerized tomograms (CT) to demonstrate 
efficacy of a new experimental contrast material that 
was developed at the NIH Clinical Center as an aid 
in the demonstration of early metastatic disease of 
the liver. Software for performing this analysis is fully 
operational on the Image Processing System. 
Measurements have been made on pre- and post- 
injection tomograms. Several classical image texture 
measurements are being tried as discriminating 
features in the demonstration of contrast material 
effectiveness. 

Background and Objectives: The objective of this 
project Is to provide a procedure for performing 
statistical analysis of attenuation values lying within 
any specified closed amorphous contour of 
computerized tomogram (CT) images. This 
procedure is to be used for quantitative analysis of 
liver and spleen sections of CT images in order to 
demonstrate the efficacy of a new experimental 
contrast material that was developed at the NIH 
Clinical Center to aid the demonstration of early 
metastatic disease of the liver. 

Methods Employed: CT images are transported by 
means of magnetic tape from the Diagnostic 
Radiology Department (DR) of the Clinical Center to 
the Image Processing Facility in DCRT. Interactive 
software has been written and implemented to read 
the CT images from magnetic tape, draw selected 
contours to indicate areas of interest, and perform 
statistical analysis on the attenuation values within ' 
the chosen areas. Liver and spleen sections of pre- 
and post-injection images are processed and 
compared. 

Significance to Biomedical Research: Most current 
CT computer systems provide limited statistical 
analysis within small fixed-shape contours (circles or 
squares). The approach presented here offers 
unlimited analysis within any shape contour. The 
software may be easily extended to measure other 
features based on groupings of attenuation values; 
such features could be useful in further 
differentiating image elements and diseased areas. 
The software has obvious extension to the analysis 
of other types of biomedical images. 

Proposed Course: Analysis of liver and spleen 
sections on pre- and post-injection tomograms are 
continually being performed as images are made 
available. It is intended to explore the usefulness o 
several classical image texture measurements in th 
demonstration of contrast material effectiveness. 



1. 1960 to Si>pteiii>er 30. 1981 



zoi cmmi-m csi 



-^..u, .««. 1« 1...) 




Ial9« «n.lysH In AuWmut Hjdlolhtrapj treawent PUralBg 


ai me uiTma ' 






Jmt H. DrLro Computer Syjtwii toi.l, it 
Ell Ol.llteln Chlel 


est. OC«T 
HO . «C1 


Itodlatlon Ootol09y Brjnch (BOl. NCI 
Coflputer Syttcm Laboratory 
Systems DtsWn Section 










0.0 1 0.0 1 








K.I .u»ua.<IU ii.|-*» ■..;.,. 


1 


llL!,'»°tk~»'!.''.'.!^'!'!nf. . ,. ;;7^^, 


— 


This project u directed to»ardi finding iimroved ci»lhodi o 


analyting 


coavuterlied tonnqraphy (CT) images in order to provide opt 


mate dati^ 
Irvlenented and 


automated radio therapy treatment plannina. various alqori 
enhancement, contour detection, e.tractlon. follo-lno. coor 
coioresslon. and three-dimensional representation have heen 


tested. The project has been Inactive during the past year 



«'tK*«irs!n»* i£t^' 



""'^Td!fk\^*^"' 



201 CT 0005S-02 CSL 



JJOflter. L .1960 to Swlciftcr 30. 1981 



OmLto^tenKsL, 
Automated Pulnonary Phystoloqy Testing 



Ldwrence D. N<de1 . 



PDS. CSL. DCRT 
PB. IR NHLBI 



PulMMry Branch, NHLBI 

Coaput er Systens Uboratory 

Project Development Section 
0(.((]_. NIH. Bethestfa. HO 20205 




Image Analysis in Automated Radiotherapy 
Treatment Planning 

This project is directed towards finding improved 
methods of analyzing computerized tomography (CT) 
images in order to provide optimal methods of 
automated radio therapy treatment planning. Various 
algorithms for contrast enhancement, contour 
detection, extraction, following, coordinate data 
compression, and three-dimensional representation 
have been implemented and tested. The project has 
been inactive during the past year. 

Background and Objectives: The objective of this 
project is to find improved computerized methods for 
isolating specific organs and diseased areas in 
computerized tomography (CT) images for the 
purpose of improving upon computer-assisted 
methods of radiotherapy treatment planning. 

Methods Employed: CT scans are transported by 
magnetic tape to the Evans and Sutherland System 
in DCRT, where they are entered and analyzed. 
Analysis consists of applying a variety of classical 
and experimental algorithms for contrast 
enhancement, density slicing, texture analysis, 
contour detection, contour extraction, contour 
following, contour coordinate data compression, and 
three-dimensional reconstruction. 

Significance to Biomedical Research: 

Improvements in computer-assisted radiotherapy 
treatment planning should result in improved patient 
care. 

Proposed Course: Plans for future work have not 
been specified at the time of this writing. 

Automated Pulmonary Physiology Testing 

Procedures such as exercise testing, pulmonary 
compliance, and work of breathing have been found 
successful for evaluating pulmonary function. By 
exercising a patient on a treadmill and gradually 
increasing the workload (i.e., speed and incline), the 
physician can better assess cardio-pulmonary 
disease, which in its early stages generally does not 
manifest itself except under physical exertion. In 
order to help the physician perform these procedures 
more effectively, a microcomputer system has been 
developed to enable automated realtime collection, 
analysis, and display of pulmonary compliance data. 
Work is in progress to complete an automated 
exercise procedure as well. Data is stored in a local 
disk data base for future reference. 

Background and Objectives: Physicians monitor 
pulmonary parameters during exercise to better 
assess pulmonary function and to diagnose 
pulmonary dysfunction that only manifests itself 

45 



under physical exertion. Procedures such as 
pulmonary compliance and inspiratory muscle 
strength also give insight into respiratory function. 

Until recently, pulmonary treadmill exercise testing 
was performed manually at NIH. Data were written 
down and later entered into a programmable 
calculator for determination of results. Additional 
summary statistics and a final report were prepared 
by hand. Work of breathing and pulmonary 
compliance measurements, done in the same lab, 
were likewise performed manually. 

In order to speed both exam and data analysis time, 
and to improve accuracy, these procedures are 
being automated with a microcomputer system. 

Methods Employed: The microcomputer system is 
a DEC MINC-11/03 (Modular Instrument Computer) 
containing an LSI-1 1 microprocessor, 32K words of 
memory, auxiliary disk storage, and analog-to-digital 
and digital-to-analog conversion capability. There is 
also a video graphics display, a keyboard console, a 
hard copy unit for printing the video display, and a 
line printer. 

In determining pulmonary compliance, 
transpulmonary pressure (the difference between 
alveolar pressure, i.e., mouth pressure with mouth 
shutter closed, and esophageal pressure, as 
measured by a balloon transducer swallowed by the 
patient) and lung volume (measured with a wedge 
spirometer) are determined by the computer as the 
physician repeatedly closes a mouth shutter 
throughout a patient's inhalation or exhalation. A 
graphical plot of the data is then produced to aid in 
evaluating the 'stretchability' of the patient's lungs. 

In a similar manner, a patient's relative inspiratory 
muscle strength is determined by measuring the 
most negative pressure developed when inspiring 
against a closed mouth shutter. 

During the treadmill exercise procedure, the 
computer monitors expired volume and flow via a 
Tissot spirometer and pneumotach, respectively. 
Inspired and expired oxygen, carbon dioxide, and 
nitrogen concentrations are monitored via a Perkin 
Elmer mass spectrometer gas analyzer. Acid/base 
and gas concentrations are determined offline from a 
sample of the patient's arterial blood, and entered at 
the keyboard. Pulmonary volumes, flows, and oxygen 
consumption-a measure of how hard the patient 
actually works to perform a given level of exercise- 
are then calculated. 

Progress in FYS 1: TherMINC-11 microcomputer 
and related peripherals were installed and 
functioning by November 1980. The manual data 

46 



er on V 



entry exercise analysis package developed earlier 
the CSL LSI-1 1 /03 development system was directly 
transferred to the MINC, immediately enabling the 
automatic processing of exercise data. Due to a 
change in clinical priorities, it was decided to first 
automate the compliance and inspiratory muscle 
strength tests, and then proceed further with the 
exercise protocol. 

Utilizing the MINC's realtime data acquisition and 
graphics capabilities, subsequently, software was 
developed to enable a full automation of the 
compliance and inspiratory muscle strength 
procedures. Only minimal modifications to the 
existing pulmonary hardware were required. In 
addition to the capabilities for data acquisition, 
analysis, and display, a scheme for storing data 
locally on disk allowing subsequent retrieval was 
developed. Later on, the stored data will be 
incorporated with the central pulmonary data base 
being developed by DMB. 

Proposed Course: First, the existing exercise 
protocol will be completed. Then, having automated 
all existing pulmonary lab procedures, we will use the 
computer for work that could not be done by manual 
methods. For example, we plan to include closed 
loop computer control of treadmill speed. By 
monitoring heart rate and dynamically varying the 
treadmill speed, it should be possible to apply a 
more constant workload to the patient, thus leading 
to more stable results. In addition to performing 
more complicated laboratory procedures, additional 
forms of mathematical analysis will be applied to the 
data in order to gain further insight into the patient's 
pulmonary function. Having developed a general- 
purpose tool for pulmonary data collection, it will be 
this latter portion of the project that has the potential 
for advancing the state-of-the-art in pulmonary 
medicine. 

Publications: 

Nadel, L. D.: Automated Pulmonary Analysis by an On-line Microcomputer. 
In Nair, 8. (Ed.): Proceedings of the Conference on Computers in 
Criticai Care and Puimonary Medicine. New York, Plenum Press (in 
press). 



SiSSfltiS'tS'-?™!™ !ST' 


"ijSiii^zr 


MMUt m 
201 C 


r 00060-02 CSl 


Octolier 1. IMO to Septtoter 


30. 1981 




Hlasurewnt of Tranieplthel u 


1 (teslsunu of Kfdney Tubule 




«wu, LMMiiom *w iaiiir<ni irfiii*! 


ad, iM iinii ff MiKim i 




a i» ui cnu 



ict £fl9ln«er CSL. OCRT 

FpIIom LKEH. NHLBl 



I jhnratftry af XldnCX and EttttfO ljff HeUboll««. NHJl 

fniipMt»r Sy^t^an Laboratory 

Project D ewl opnw t_ Sec 1 1 on 

ncRT NlH, H^thg^d*. KirvUnd 20M5 



3 






: Mas upgraded In FV8I 



Measurement of Transepithelial Resistance of 
Kidney Tubule 

A microcomputer-based instrument was developed in 
FY80 to facilitate determination of transepithelial 
resistance of an in vitro perparation of kidney tubule. 
The instrument controls the onset, intensity, and 
duration of a series of electric current pulses through 
the preparation; measures the induced voltage 
changes; and prints those values. The instrument 
was upgraded in FY81 to include online calculation 
of the transepithelial resistance and other 
parameters of interest. 

Background and Objectives: The objective is to 
automatically measure the transepithelial resistance 
of in vitro preparations of kidney tubule. A 
microprocessor-based instrument was developed to 
facilitate these measurements. Based on operational 
directives entered by the user, the instrument 
controls the polarity, duration, and intensity of each 
of a series of current pulses through the tubule 
preparation, and determines and records the steady 
state voltages induced at either end of the 
preparation. Transepithelial resistance and other 
parameters are calculated from these values. 

Significance to Biomedical Research: Previous 
manual methods for obtaining these measurements 
were time-consuming and prone to error. With the 
control instrument, a complete set of measurements 
can be made in a few moments. The experiment can 
easily be rerun under the same or a different set of 
conditions to test repeatability and consistency. 

Progress in FY81: The capability to calculate the 
resistances and other parameters using a complex 
set of equations derived from a transmission line 
model was added to the instrument. This was 
accomplished by adding floating point math 
hardware and software to utilize it. The calculated 
values for every current pulse are now available to 
the investigator immediately following execution of 
the pulse sequence. Of particular interest is the 
immediate display of the tubule diameter. A close 
correpondence of the calculated diameter with the 
actual diameter indicates that the preparation is 
intact and that other calculations are valid. 

Proposed Course: We may make a few minor 
software changes to tailor the instrument's user- 
interaction features to this experiment. No major 
effort is anticipated for the future. 

Publications: 

Hauser, S E and Alfneida. A : A Control and Data Processing Instrument 
lor Kidney Tubule Research. Biomedical Sciences Instrumentation 17: 
13-19. 1981. 



47 



_Qctolier U isao to September 3Q. 1981 
Electron Microanalysis Facility 



s'tsTiir !?!*«?; ."Wiiu.H^^cDw^^^^ 



Electron Microanalysis Facility 



Physical Scientis 



KHLBI. NIADOK, NIMH. NINCE 

Comp uter Systems Laborator 

_Ecaleci ()evelop<nent Sectic 

-^nrRi^^tmt.. fletnesda. fiD ; 

5.2 






1 6E1B is developing i 



I Energy loss Spectrometry 
I Dispersive X-ray Spectra 



tranetry . 

ZOI HS 10058-03 
ZOl RS 10059-03 
ZOl CT 00042-03 



SSScTiiS'BWrH!! SW 


.l'zJ";il!«iiil 


ZOl CTO0053-O2 


CSL 




Computer Assistance for Blind Computer Users 1 


PI: Scott [. Allen Medical Besearch Analyst CSL, OCRT ' 
David C. Songco Electronics Engineer CSL. DCRT 
Perry S. Plenico Chief. Project Development Section CSL, DCRT 
Oavld M, Stoffel Computer Programmer CSL. OCRT 

; 


cco*iuT>iw mm (u .-,) 
None 








1 1 n » 1 




M* have developed a voice output terminal for us 


by Mind computer 
igram running in a 


'by 


cortlning a voice lynt^esl^e^ with a te.fto-speech p 
for presenting visual data In audible form. 



^ 



CSL in collaboration with BEIB is developing an 
automated electron nnicroanalysis facility consisting 
of two electron microscopes interfaced to a 
computer system. The facility will be used for 
research into the elemental composition of biological 
specimens, and for the development of new 
techniques in electron microscopy. CSL is designing 
and implementing the computer system, which will 
acquire and display the spectra and images 
produced by Electron Energy Loss Spectrometry, 
Energy-Dispersive X-ray Spectrometry, and 
Wavelength-Dispersive X-ray Spectrometry. 

See also: 
Z01 RS 10058-03 
ZOl RS 10059-03 
ZOl CT 00042-03 

Background and Objectives: The Computer 
Systems Laboratory is designing and implementing a 
computer system as part of the BEIB Microanalysis 
Facility. The facility consists of two electron 
microscopes, and will be used for research into the 
elemental composition of biological specimens and 
for developing new techniques in electron 
microscopy. Work on this project is now in its third 
year. 

One of the electron microscopes is an Hitachi H- 
700H 200 keV Scanning Transmission Electron 
Microscope (STEM) equipped with: 

• a lithium-drafted silicon [Si(Li)] detector 
connected to a Kevex 7000 Analytical 
Spectrometer for performing Energy-Dispersive 
X-ray Spectrometry (EDS); 

• an electron spectrometer for performing 
Electron Energy Loss Spectrometry (EELS); and 

• detectors for forward-scattered, back-scattered, 
secondary, and sample electron current signals. 

The other electron microscope is a Cameca 50 keV 
Electron Microprobe equipped with: 

• an Si(Li) detector for performing EDS; 

• three Wavelength Dispersive X-ray (WDS) 
spectrometers; and, 

• detectors for forward-scattered, back-scattered, 
secondary, and sample electron current signals. 

A PDP-1 1 /60 computer system ultimately will be 
interfaced to both microscopes to perform the 
following functions: 

• control electron beam position, stage position, 
and the various detectors; 

• acquire spectra and image data from all 
detectors; 



48 



• process and display the spectra and Innage data; 

• monitor and display a wide variety of 
'housekeeping' parameters, including: lens 
currents, lens temperatures, beam current, beam 
energy, pump temperatures, coolant flow rates, 
vacuum pressures, water leak detectors, floor 
vibrations, ambient AC fields, power supply 
voltages, room temperature, and room humidity. 

Progress in FY81: CSL's software efforts this year 
have been concentrated on four aspects of data 
acquisition from the STEM: 

1. acquisition of EELS spectral data and control of 
STEM beam position, 

2. acquisition, calibration, monitoring, and display 
of housekeeping parameters, 

3. retrieval of empirical X-ray information, and 

4. installation and programming of the color 
display systems. 

EELS data acquisition and control of the STEM 
beam position is done by a satellite processor 
connected to the PDP-11/60 by a high-speed link. 
Software has been wntten that allows the STEM 
operator to define areas of a specimen as targets for 
data acquisition and to collect EDS, EELS, and 
electron current signal data from the target areas. 
The data is acquired in SPECTRUM mode, which 
produces a single X-ray and/or EEL spectrum from 
the target area along with as many as four electron 
current signal images. 

EDS data acquisition is done by the Kevex 7000, 
which is connected directly to the computer. 
Software has been developed to: allow programs on 
the 1 1 /60 to directly communicate with and control 
the Kevex 7000; save or restore spectra and 
associated information to or from user library files; 
allow programs on the 1 1/60 to access, insert, or 
delete spectra contained in user library files. This 
software is currently being used by BEIB scientists 
for research into methods of processing EDS and 
EELS spectra to remove background and resolve 
overlapping peaks/edges. 

Housekeeping parameters are acquired by the 
computer by means of an Analogies AN5400 data 
acquisition subsystem. ,Software has been developed 
to acquire, monitor, and display these parameters. 

Calibration parameters for housekeeping and other 
signals are managed by a Calibration Utility that was 
developed this year. This utility maintains calibration 
information on a disk file and allows it to be listed, 
updated, and restored to memory-resident tables at 
system boot. 
To simplify the operation of the data acquisition and 



display software that is being developed, a menu 
selection scheme is used. The menu selection 
software is completely table-dnven so that it is easy 
to add new functions as they become available. 
Currently, the housekeeping parameter display and 
specimen target definition functions can be activated 
through menu selection. 

Work has begun on software to retrieve information 
associated with X-ray emitting electron energy 
transitions within atoms. This software will allow an 
operator or another program to specify an element 
and the transition(s) of interest using a convenient 
notation. It will then look up the associated transition 
energies and relative peak amplitudes. Conversely, 
an energy range may be specified, in which case a 
list of the transitions within that range will be 
retrieved. This work is being done under contract by 
Systex, Inc. 

A DeAnza ID5400 color display system was 
delivered in October 1980 and was successfully 
interfaced to the PDP-11/60. Work has begun on 
software to support image display and processing. 

A second satellite processor for interfacing the 
microprobe to the computer was delivered in FY81 
and is currently connected to CSL's PDP-11/70 
system where it is used for development of software 
for the STEM satellite. 

Proposed Course: Next year, we expect to: 

• implement additional modes of acquisition for 
EELS and EDS spectra and images and the four 
electron current signals from the STEM; 

• have a package to facilitate the display and 
processing of EDS, EEL, and electron current 
signal images on the DeAnza display system; 

• begin work on the Cameca microprobe interface. 

Computer Assistance for Blind Computer Users 

We have developed a voice output terminal for use 
by blind computer professionals. Full word unlimited 
vocabulary speech output is made possible by 
combining a voice synthesizer with a text-to-speech 
program running in a microcomputer. We are working 
with our blind users to develop additional techniques 
for presenting visual data in audible form. 

Background and Objectives: In previous years we 
have developed a voice output terminal that 
combines unl.mited vocabulary with extensive text 
review capabilities. Three terminals were provided to 
different Government agencies and proved to be a 
valuable asset to the visually impaired computer 
professionals who used them. Similar voice output 
terminals are now marketed by the private sector. At 
least two of these are based on CSL work.* 

49 




?^Ti&t 


'ts'.irr!iR !W «.;;:• rr/."',,?.,.,. 


201 CT00OS2-02 CSL 


October 


. 1980 to September 30, 198) 


Httaboll 
OTHER: 


Energy Neasurements 

D. C. Songco Electronics Engineer CSt, DCRT 
». ». Henai.r Chler, HES tC, IR, NHLBI | 

1 


Non« 


_Co»pmer 
Project 


S/stems Laboratory 


Developwnt Section 


'"DCRT.*NiH?Bethesda. HD 20205 I 




0.5 P'"'**^' Q_5 !<""'"' 




11 Bo«{i:) 

a (.0 .««»..« 




rocomputer- based instrument was developed t 


B Study eneroy transduc- 


lion pnenomena or respiring raemhrSnes. tiectrodes are interfaced to the mfcro- 
eo«iputer to measure concentrations of specific ions, of protons, and of oxygen. 
Program! are being developed to assay the change in ph across the membrane and 
the ratios of ion movements to oxygen uptake. 



siScrits'iS'.iS'rxffisrir' 


"" 


fil 


ilf'iiii.', 


!01 n000»5-02 


CSL 


October 1. 1980 toSeoteober 30. 


1981 





















InforHtion Technology Project 



Crawford S. BrOMi, H.D, 
Oavld M. stoffel 



relopnent CSL, DCRT 




Progress in FY81: This year saw rapid 
developments of voice synthesis technology which 
resulted in the availability of integrated circuits, 
costing less than one hundred dollars, for the 
synthetic generation of English phonemes. We have 
incorporated these devices into our voice output 
terminal. This has resulted in not only reduced size 
and cost of the units but in enhanced capabilities in 
terms of rate of speech and intelligibility. 

Proposed Course: Prior work, both by CSL and the 
private sector, has concentrated on the presentation 
of sequential, linear text in audible form. We plan 
additional studies in cooperation with our blind user- 
colleagues aimed at presenting text in other formats 
commonly employed on computers. Examples 
include tables, forms, and lists. 

Metabolic Energy Measurements 

A microcomputer-based instrument was developed 
to study energy transduction phenomena of respiring 
membranes. Electrodes are interfaced to the 
microcomputer to measure concentrations of specific 
ions, of protons, and of oxygen. Programs are being 
developed to assay the change in ph across the 
membrane and the ratios of ion movements to 
oxygen uptake. 

Background and Objectives: CSL has collaborated 
with the LC, IR, NHLBI for a number of years in the 
development of new techniques for the study of 
bioenergetics. This effort began with the 
development of the Potentiometric Titration 
Controller, which continues to be used in the 
investigation of energy transfer phenomena at the 
energy parameters of respiring membranes. To be 
included are determinations of membrane potential 
and the concentrations gradient for protons across 
the membrane. New methods are also being 
developed to measure the rate of oxygen uptake and 
the ratio of protons and other ion movements to the 
atoms of oxygen consumed. 

Progress in FY81: The microcomputer system 
designed in FY80 has been installed. A single board 
microcomputer was used in conjunction with a 6- 
channel D/A converter and a 16-channel A/D 
converter with programmable gain. A high speed 
mathematics module was interfaced to the system to 
enhance realtime calculations. As proposed last 
year, the data acquisition and control software were 
developed on the disk-based Potentiometric Titration 
Controller system and down-loaded to the target 
system. By utilizing the existing development 
facilities, only the hardware specific to the new 
experiments needed to be purchased. 



50 



An electrode was constructed specifically for the 
nnembrane-soluble probe, tetraphenylphosponium 
(TPP), and this electrode has been shown to exhibit 
Nernstian behavior in the measurment of low 
concentrations of TPP in solution. A program has 
been developed which calculates in realtime the 
electric potential across a respiring membrane based 
on the internal volume of membrane vesicles in the 
suspension being analyzed, the external volume of 
the solution, and the voltage sensed by the TPP 
specific electrode. The membrane potential 
measured by this procedure has been found to 
agree with that determined by the traditional 
technique of flow dialysis, which is much more 
cumbersome and time-consuming. 

Proposed Course: Development of the system to 
determine energy parameters of respiring 
membranes by use of electrodes will continue. 
Emphasis will be placed on using the pH and oxygen 
electrodes to assay the change in pH across the 
membrane and the ratios of ion movements to 
oxygen uptake. 

Medical Information Technology Project 

This project involves the application of 
microprocessor technology and improved man- 
machine interface methods to permit physicians and 
their associates to more directly communicate with 
computer record systems. This year we have begun 
implementation of concepts developed in previous 
years. A pilot study involving medical recordkeeping 
by direct input of examining physicians and staff is 
underway. 

Background and Objectives: The use of computers 
within the biomedical community is increasing as the 
cost of systems is decreasing due to technological 
innovation. Enhancements in the area of man- 
machine interfaces must keep pace with the rapid 
advance of computer hardware and software 
technology. With this in mind, we have investigated 
devices, methods, and structures that could provide 
a more human-oriented interface while maintaining 
an acceptable level of flexibility and efficiency. 

Progress in FY81: This year we have identified an 
area in which we can apply concepts developed in 
previous years. In collaboration with a practicing 
dermatologist, we have begun the development of a 
clinical care system which will allow the physician to 
store and retneve the data contained in a medical 
record. This data includes histories, physical 
examinations, progress notes, treatments, and 
procedures. 

The immediate objective of the pilot study underway 



is to provide the physician with rapid and simple 
access to a dedicated microcomputer system. 
Disease-specific and problem-specific forms and 
protocols are used to prompt the user through the 
hierarchy of programs available. Much of the 
software is table-driven to allow the physician to add 
and modify not only the data base but the logic of 
the presentation. This approach also provides a 
convenient means of adapting the programs to other 
clinical care and research protocols. To support this 
effort, we are working in two main areas. The 
clinicans are designing the disease-specific protocols 
and formats as well as a general workup logic. The 
computer programmers are developing a generalized 
software system to provide a convenient, assistive 
interface for physicians to use. This software will aid 
the physician in accessing and selecting data from 
complex tree-structured files, and in entering data via 
a CRT terminal by simple menu selection for form 
fill-out methods. 

Proposed Course: Both the clinical forms and 
clinical data access software are being developed on 
the CSL time-shared computer system. Online 
program trials from the dermatologist's office are in 
progress. We expect to be able to transfer 
operations of the software to a dedicated 
microcomputer system situated in the physician's 
office by the first quarter of FY82. At a later time, we 
plan to perform experiments using a touch screen, 
bar codes, and special cursor controls for data 
selection and entry. 



51 



52 



Laboratory of 
Applied Studies 



Eugene K. Harris, Chief 



Summary of Activities 

Computer-aided analysis of electrocardiograms. 

J. Bailey, M. Horton (LAS); cardiologists and 
biomedical engineers in the U.S.A., and abroad. To 
evaluate the utility of leading computer programs for 
ECG interpretation, and to search for optimal 
computer-based methods of extracting medically 
significant ECG patterns. A Gompanson of IBM and 
GRI (Glasgow Royal Infirmary) ECG programs, 
including clinical documentation and semantic 
equivalences of output statements, has been 
published. A study of serial ECG's has begun jointly 
with staff of the Framingham Heart Study, NHLBI. 

Computer systems for nuclear medicine. J. 

Bailey, M. Douglas, and others (LAS); H. Ostrow 
(CSL); M. Green, et al. (CC, Nuclear Medicine). 
Development and application of computer systems 
to such diagnostic imaging activities as ECG-gated 
radionuclide angiocardiography, functional mapping, 
and other scintigraphic studies of the kidney, brain, 
heart, and lung. In collaboration with Nuclear 
Medicine (CC) and the Cardiology Branch (NHLBI), 
various parameters measunng regional heart wall 
mobility are being studied for their discriminating 
ability in normal volunteers and heart patients. Data 
collection in the renal scintigraphy study to detect 
arterial stenosis in dogs has been completed. 
Results are now being analyzed and written for 
publication. 

Computer-based studies of pulmonary 
pathophysiology and respiratory disease. J. 

Bailey, R. Burgess, and others (LAS); R. Crystal, A. 
Nienhuis (NHLBI); A. Jones (CC, Nuclear Medicine). 
To achieve better understanding of pulmonary 
pathophysiology through use of computer-based 
models of pulmonary gas exchange and respiratory 
mechanics, comparing predicted values with real 
patient data. A joint study of gas exchange in normal 
volunteers and patients at rest or exercising has 
advanced with development of a reliable gas 
analysis system and receipt of computer-controlled 
exercise test equipment. 



Statistical research in clinical pathology. E Harris, 
M. Horton, A. Albert (LAS); G. Shakarji (DMB); 
clinical chemists and others in the U.S.A., Europe, 
and Japan. Application of variance component, time 
series and other analyses to descnption of reference 
distributions of clinical laboratory tests, to serial 
studies of blood chemistries in health and disease, 
and to the design of criteria for recommended 
precision and accuracy of laboratory methods. A 
comparative study of the sensitivity and specificity of 
univariate and multivanate time series models, using 
real and simulated data, is nearing completion. A 
chapter on statistical aspects of reference values in 
clinical pathology has been published. New research 
has begun on statistical methods for dynamic 
assessment of risk in acute illness. 

Computer-based studies in ultrasonography. R. 

Burgess, M. Douglas, J. Bailey, E. Pottala (LAS); B. 
Maron. Ultrasonography allows non-invasive 
visualization of many organs without the hazard of 
ionizing radiation. This project involves development 
of minicomputer systems for image enhancement, 
pattern recognition, and three-dimensional 
reconstruction from ultrasound data sources, 
principally wide-angle phased array 
echocardiography. Bone structures opaque to sound 
necessitate development of an esophageal 
transducer interfaced to a minicomputer. Lack of 
staff time during FY81 forced deferral of this effort 
until FY82. 

Mathematical Modeling of biological processes. 

J. Fletcher (LAS); R. Schubert (Louisiana Tech. 
University). Development and application of 
mathematical models in studies of substrate 
transport in the microcirculation, in diffusion 
processes in physiology, and in macromolecule- 
ligand binding equilibria. A theoretical reanalysis of 
concurrent flow models for organ perfusion 
experiments has been completed. Modeling and 
expenmental work on microcirculatory processes in 
the autoregulation of oxygen supply within an organ 
is continuing. 



53 



Mechanisms of active transport/biochemical 
kinetics. B. Bunow (LAS); A. Kaplan (NCI); D. 
Mikulecky (Medical College of Virginia; J. Kernevez 
(University of Tecii., Compiegne, France). 
Experimental and mathematical studies of the energy 
mechanisms for active transport and of multi-state 
biochemical kinetics in cells and membranes. 
Theoretical studies last year revealed the 
insufficiency of current, widely accepted methods 
and hypotheses to explain the energizing or 
localizing of active transprot mechanisms. 
Collaborative work with NIH scientists has begun on 
the use of newly implemented network simulation 
programs to improve the understanding of active 
transport mechanisms in membranes. 

Hybrid computing to analyze physiologic signals 
and construct simulation models. E. Pottala, J. 
Wolpert (LAS); various NIH and FDA scientists. 
Using LAS minicomputer system (MAC-16) for 
hardware simulation of physiologic functions and for 
analysis of analog signals (myogram, EEG, etc.). An 
operating system developed and implemented last 
year has been used extensively for A/D conversion 
and spectral density analysis of EEG's and EMG's of 
patients in various studies with the Medical 
Neurology Branch, NINCDS, and of ECG's in a 
cardiac drug toxicity study sponsored by the FDA. 

Image processing in electron-loss spectroscopy. 

M. Douglas, E. Pottala (LAS); J. Costa (NIMH); 
Development and implementation of mathematical 
models and image enchancement techniques to 
analyze computer-acquired information from 
electron-loss and X-ray spectra indicating the 
location of extremely small quantities of important 
chemical elements and active protein molecules 
within cells. Image processing capabilities developed 
on the newly expanded DeAnza system have been 
used to determine fluorine distributions within the 
dense bodies of blood platelets as a possible model 
for intracellular patient monitoring of therapeutic 
drugs. 

Mathematical and computational methods for 
nonlinear equations. R. Shrager, R. Hendler 
(NHLBI); A. Schechter (NIADDK). Study of methods 
of fitting nonlinear models and mathematical 
methods of spectral analysis. A decomposition 
procedure for the analysis of the spectra of mixtures 
of chemical reactants to determine individual 
physico-chemical characteristics and reactive 
mechanisms has been completed and is being 
tested in several experimental applications. 



54 



giis^inaiaiS'fss^M'riiKfw^ 


INI«A«wi''«SH«rMQJECI 


m CToono7-i3 las 


October 1. 1980 to Septenher 3D, 1981 


St«,sUca,»..a.hi„c„„,ca,P«.o,.,, 


PI: E.K. HarHs "^ """Thief. Lflb. of Applied Sturties LAS OCRT 
OTHERS: A. Albert Fogarty International LAS DCRT 

G. Shakarji Sgpv. Systems Analyst DHB DCRI 
M.R. Morton Computer Systems Analyst LAS OCRT 
L.H. Norton Research Hathenatician LSI DCRT 
R. Eltn Clinical Pathology CC 
G.Z. Williams Institute for Health 

San Francisco, CA 
T. Yasaka PL Medical Service nept. 

Osaka, Japan 
G. Siest and Centre du Medccine 
R. Gueguen Preventive, Uancy, France 






Laboratory of Applied Studies 


flCRT. HillTBethesda, MD 20205 


'"'*' '"'"""o.a |™*'»i""L' ^_g |<»H«. 


Dl..)-,No« G(.0..,«m-s 


tUUIAHT U MM (iOO .=„!• or Utl - untf.rlin. fc.^oM.) 

technifiues are heing applied to various data base 


_s and discriminant 
consisting of short 


patients with myocardial infarction. The purpose 




experience in the use of these statistical predic 
detect changes and trends within individuals, tak 




series varies from daily to weekly, 6-inonth, and 
between oTiservat ions. Parallel conputer-based si 

Mathematical investigations into the properties o 
nodel of linear change are continuing. 


astlng methods. 



Numerical methods for the solution of 
mathematical models describing reaction- 
diffusion and other processes in biological 
systems. M. Bieterman, J. Fletcher, B. Bunow (LAS); 
I. Babuska (University of Maryland). Study, 
development, and implementation of efficient, flexible 
numerical methods for the solution of nonlinear 
ordinary and partial differential equations involved in 
modeling dynamic physiological processes. Research 
and testing continued during FY82 on the use of 
finite element numerical methods for the solution of 
time-dependent reaction-diffusion equations. 
Mathematical theory has been delevoped for the 
adaptive solution of coupled systems of kinetic 
equations. Applications to laser light scattering in 
gels and some hydrodynamic models of blood flow 
are currently in progress. 

Research Projects 

statistical Research in Clinical Pathology 

Univariate and multivariate time series models and 
discriminant techniques are being applied to various 
data bases consisting of short series of 
measurements of serum biochemistries in healthy 
subjects and patients with myocardial infarction. The 
purpose is to gain practical experience in the use of 
these statistical predictive techniques to detect 
changes and trends within individuals, taking into 
account biological variation and measurement error. 
The time scale of these series varies from daily to 
weekly, 6-month, and 12-month intervals between 
observations. Parallel computer-based simulation 
studies are also underway, particularly to estimate 
the relative sensitivities and specificities of 
multivariate and univariate forecasting methods. 
Mathematical investigations into the properties of a 
new stochastic model of linear change are 
continuing. 

Objectives: To investigate applications of statistical 
theory, particularly the use of variance components, 
discriminant analysis, and the theory of discrete and 
continuous time series, to the interpretation of serial 
clinical laboratory measurements in healthy subjects 
and patients with acute and chronic disease. 

Recent Background: During the past several years, 
the major effort in this project has been devoted to 
the application of recently developed mathematical- 
statistical models to various data bases of serial 
biochemistries in healthy subjects. The smallest of 
these data collections, weekly measurements of 10 
common constituents in 37 male volunteers over a 
5-month period, arose from a cooperatively designed 
study with the Clinical Research Centre, Harrow, 



England. Larger studies, in San Francisco and in 
Osaka, Japan, involve hundreds of men and women 
undergoing annual or semi-annual examination 
including clinical chemistry and hematology 
measurements. Running parallel with these studies 
has been a rapidly growing interest among many 
clinical laboratory workers and clinicians in the 
statistical bases of reference values commonly used 
in diagnosis and the evaluation of therapy. In 
particular, the need to take account of within-person 
biological variability over time is becoming more 
widely recognized among pathologists and others 
who interpret clinical laboratory reports. 

Progress during FY81: The application of 
multivariate and univariate time series models to 
selected groups of semi-annual measurements from 
a large-scale health maintenance program in Japan 
made considerable progress. A preliminary report 
was presented at an international meeting on the 
subject of automated health testing in Tokyo, 
October 1980. Current findings, based on both these 
and computer-generated observations, indicate that 
multivariate subject-specific reference regions, like 
their counterparts derived from cross-sectional 
population-based data, are much more conservative 
in their interpretation of clinical laboratory results 
than are univariate reference ranges. At the same 
time, the multivanate region will occasionally show 
sensitivity to a combination of results, each of which 
appears normal when tested separately. Work is 
continuing to implement and test the multivariate 
random walk model, whose application to short 
series is much more difficult than the stationary 
homeostatic model. 

Mathematical study continues of the univariate 
stochastic linear growth model able to detect true 
changes in slope, free of effects of measurement 
errors and random 'ups and downs.' In cooperation 
with the Laboratory of Statistical and Mathematical 
Methodology, research into the asymptotic properties 
of this model has been completed and efforts have 
been directed towards obtaining more efficient 
estimates of the parameters than are now available 
from current methods based on the variances of 
second and third differences. 

In a new area of research, methods involving 
discretized response curves are being developed to 
deal with rapidly changing biochemical variables, as 
in patients under intensive care for myocardial 
infarction. The object here is to discriminate as early 
as possible the probable outcome of the patient's 
condition using only the most appropriate 
measurements at each stage. 



Studies with collaborating scientists outside NIH on 



55 



I iMTitiwuL >U(UM Mojtci I ZOI CTOOOOS-U 



' 1, 1980 CO Septerrber 30, 1981 



; of Binding EquH 



, Or. Chief, Appti( 



. Erifchmann, Dr. 



Applied (tetheraf 
DCRt". NrH/Bethc! 



The objective of 

are examined for 



le study of mathematical models of ligand- 
inding studies at equilibrium. The moteTs 
ill as for conceptual validity and are 
ity for fitting to experimentally obtained 





™'"'I"c™ 


00«-03 LAS 


QctobEr.l>-19aO-tii SpptPnhPr in, i 


981 




Hathensttcal Modeling of Substrate 
Envlronnents 


Transport in Physiological 


1 


w(L"»r"«o;ru «"«""« wi'"i"!; "^ 


llUi « «.I«IML .««STrC«MS m .LL 


"'"'" 


"'= J.f. Fletcher, nr. 


Chief, Applied Mathematics L 
Section 


" ""T 


OTlltR; R. M. Schuhert, Hr. 


Assoc. Prof. 


L 


NiiUn^ i 



Laboratory of Appl i ed Studi es 

_flpPl1_edJ1atheraat Ics Section 

^DCfiT^JUH^JethesdajLjID 20205 



U(.tl«iM*i r;(.0 iMiniui 








tiatheraattcal models 


of mt-ro 


cil,U„. 




partial differential equ 




'nto syste<T 


s 0' couple 




are dev 





iilatory physiology i 



> physiologic challi 



;h1s project '■ 
i between v^r\ 



56 



the transferability of reference values in clinical 
pathology are still in the exploratory stage with 
appropriate methods under development. In another 
area, practical procedures for studying the heritability 
of quantitative blood levels of certain constituents 
such as cholesterol or of such characteristics as 
blood pressure have been researched and proposed 
for use in analyzing data from family health 
maintenance programs. 

Medical Significance: The development, testing, and 
routine use of univariate stochastic models to 
describe and forecast sequential results of laboratory 
tests in individual cases have proven useful when 
applied to periodic monitoring of healthy individuals 
as part of a general program of preventive medicine, 
introduction of multivariate models for this purpose 
may prove even more valuable since many 
laboratory tests are interpreted as part of a multi-test 
organ battery, or in concert with other 
physiologically-related measurements (e.g., calcium, 
total protein, albumin). However, the sensitivity and 
specificity of multivariate, as compared with 
univariate, methods-whether for diagnostic or 
monitoring purposes-needs careful assessment 
based on simulation studies and real patient data. 
The perfection of practical, yet mathematically 
sound, methods for reliable prediction of patient 
outcome based on dynamic risk assessments as 
new data become available has great potential for 
improving the efficiency and efficacy of medical care 
in both acute and chronic illness. 

Future Course: Current studies on the statistical 
properties of univariate and multivariate time series 
models should be completed by the end of FY81, 
except for mathematical-statistical research on the 
linear growth model which will continue during next 
year. Development and testing of methods for 
judging the transferability of reference values in 
clinical chemistry will proceed more rapidly as the 
time series studies are concluded and prepared for 
publication. Research in dynamic risk assessment 
will continue and may extend to studies of chronic 
disease in cooperation with the Arthritis and 
Rheumatism Branch, NIADDK. 

Publications: 

Harris, E.K.: Further applications of time series analysis to short series of 

biochemical measurements. Proceedings of Workshop on Reference 

Values in Clinical Pathology. Helsinki, 1981 (in press). 
Harris, E.K.: Regression, least squares and correlation. In Seligson, D., M.D. 

(Ed.): Handbook of Clinical Chemistry. CRC Press (in press). 
Harris, E.K.: Statistical aspects of reference values in clinical pathology. In 

Stefanini, M., and Benson, E. (Eds.); Progress in Clinical Pathology 

New York, Grune & Stratton, Inc., 1981, Vol. VIII, pp. 45-66. 
Harris, E.K.: Use of statistical models to detect subject-specific changes. 

Proceedings of International Conference on Automated f\Aultiphasic 

Health Testing & Services. Tokyo, 1980 (in press). 



Mathematical Models of Binding Equilibria 

The objective of this project is the study of 
mathematical models of ligand-receptor or ligand- 
macromolecule binding studies at equilibrium. The 
models are examined for mathematical as well as for 
conceptual validity and are studied to determine their 
suitability for fitting to experimentally obtained 
laboratory data. The appropriateness of various 
model fitting critena are studied and general 
guidelines and computational algorithms are 
designed for computer-aided interactive model fitting. 

Background: Mathematical models of 
macromolecule-ligand binding equilibria have been 
investigated since 1966. This continuing effort has 
revised many of the concepts related to the binding 
of ligands to macromolecules, particularly small ions 
to proteins. This project has produced an interactive 
methodology for the fitting of binding models to data 
and has developed other computer oriented tools for 
the analysis of data from laboratory experiments. 

Progress in FY81: In FY81 numerous requests for 
copies of exportable computer algorithms were 
honored and some consultation was provided. 
Preliminary studies of cooperative binding in 
Aspartate Transcarbamylase in collaboration with the 
Clinical Pharmacology Branch of NCI were not 
experimentally reproducible. Activity in this area has 
been suspended until the experimental procedure 
can be improved. 

Significance to Biomedical Research: The fitting of 
models to experimentally obtained data is a 
procedure used to estimate unknown parameters in 
mathematical models. The proper choice of a model, 
a choice of goodness of fit criteria, and the ability to 
estimate the unknown parameters is a basic need of 
biomedical research. The estimation of unknown 
biochemical parameters contributes to new 
biomedical insight and adds to basic scientific 
knowledge only if the fitted models represent the 
underlying biological process and the unknown 
parameters can be readily and accurately estimated. 
A thorough and continuing critique of such models 
and their appropriateness for the interpretation of 
current laboratory and clinical experiments is 
therefore essential to the growth of fundamental 
knowledge in these areas. 

The cumulated findings of the previous AMS 
research in this area has been collected in the form 
of a summary report. This report details the various 
alternative models, graphical presentations of data, 
and algorithms for fitting models to data. Fitting 
algorithms are also available for fitting with other 
than the least squares criteria. The development of 



these criteria is detailed elsewhere. Publication of 
this summary report was delayed again due to 
extensive revision requirements caused by the 
computer center conversion to the new version of 
WYLBUR. 

Some mathematical considerations of new 
experimental designs are being explored in 
collaboration with Dr. Roy of NIADDK. The questions 
involve the ultrafiltration techniques and the 
measurement of 8-amino-adenosine binding. The 
experimental procedures have not yet been 
validated. 

Proposed Course: Applications of existing and new 
methodology to data analysis will continue to be 
made as they are requested by collaborating 
laboratories. Computer programs, reprints and 
reports continue to be provided to requesting 
consultees. Publication of the summary report, which 
was again delayed due to conversion problems in 
the new version of WYLBUR special train print, is 
expected in FY82. Analytical development of new 
models and continued research in this area will 
emphasize validation of experimental techniques, 
multi-receptor models, and conformational changes 
in macromolecules due to binding of ions. 

Publications and Abstracts: none 

Mathematical Modeling of Substrate Transport in 
Physiological Environments 

Mathematical models of microcirculatory structure 
and function are developed from conceptual models 
into systems of coupled ordinary and/or partial 
differential equations. Methods of solution of these 
nonclassical formulations are developed and tested 
and satisfactory cost effective methods are used to 
explore the properties of these models. The results 
are interpreted in terms of microcirculatory 
physiology and are published in the scientific 
literature. 

One objective of this project is to study whole organ 
and organ tissue level phenomena by means of 
mathematical models in an effort to determine 
relationships between variables that govern the 
organ response to physiologic challenges. 

Background and Objectives: The objectives of this 
project are to develop mathematical models that can 
be used to simulate microcirculatory physiology and 
to explain, interpret and/or predict physiologic 
behavior and limits. Such models should lead to a 
better understanding of the biological control 
processes and should suggest improved clinical 
approaches to microcirculatory disorders. 

57 



iNiiAiwuL MiURCH nojjci joi CT00033-05 LAS 




October 1. 1960 to September 30. 1981 




Analysis of Coupled Transport and Biochemical Kinetics 




PI: B. BunoH, Dr. Expert LAS DCRT 

OTHEfiS: J. Kernevez, Dr. Professor Univ. of Tech. Compiegne, 

A. Kaplan, Dr. Research Biochemist DCCP HCl 

0. Mikulecky. Dr. Professor Medical College of Virginia 




None 




Laboratory of Applied Studies 




A^,,l,.d^^the™nci Section 




MRT. NiH.'Bethesda. Maryland ^0^05 




'""'"i'o" ]-«"-*- -[<"""■ 




„ (.) »«» ZMJlCi D (,J mou, I.SSUtS a (=1 NHIWR 




This project 1nve"t^gat« thTee'fundVmental problems in biology: (1) the role 
of dynamic patterns in embryoloqiy and evolution, {2) the kinetics of enjvmes 




normal cells in culture. The first area involves investigation of the role 
which sitrultaneoos reaction and diffusion might play in the spatial and temporal 

special emphasis on limitations in inferring molecular mechanism solely from 

for data acquisition, display, and transmission, as well as mathematical 
analysis of kinetic Studies on lactate dehydrogenase, an enzyme whose function 
and molecular form is altered in hepatocytes in tissue culture subject to 
chemical transformation to a malignant state. Oiqital computer simulation. 




ijai LiLularly by means or network mooeiinq [anquaqes. numerical solution of 








these investigations. 





58 



The mathematical modeling of organ substrate 
supply by the microcirculation has been under study 
since FY69. The substrate of primary interest is 
oxygen. Such modeling studies are aimed at the 
prediction of threshold and critical limits of substrate 
supply necessary to sustain cell function under a 
variety of physiologic conditions. The responses of 
models to varying blood flow, blood hemoglobin 
characteristics, tissue metabloic rate, tissue binding 
proteins, and other physiologic parameters have 
been examined. The complex interaction of 
microcirculatory geometry, nonlinear oxygen 
hemoglobin dissociation properties, intracellular 
binding proteins, and substrate dependent metabolic 
rates requires a detailed description to achieve 
physiologic validity. These models require the 
mathematical and numerical computer solution of a 
system of coupled equations from a distributed 
parameter model which are of an unusual nonlinear 
type. The objective of the computations is to identify 
mathematical models of organ microcirculation 
having characteristics that correlate with 
experimentally obtained measurements. Such 
models can then be used to examine probable organ 
response to physiologic challenge. 

Significance to Biomedical Researcii: Such modeling 
is necessary to predict the state of local tissue 
conditions since direct measuremenfs are generally 
not possible and must be inferred from boundary 
observations. Studies of this type have the potential 
of predicting tissue oxygenation and reoxygenation in 
ischemia, hypoxia, anemia, coronary obstructions, 
sickle cell anemia, shock, and other conditions of 
substrate normal and abnormal physiology. 

Mathematical models offer the only technique 
available for the quantitative study of possible 
autoregulatory mechanisms. The qualitative and 
quantitative nature of such mechanisms can be 
explored by means of appropriate models. 

Proposed Course: It is anticipated that the research 
course of this project will have the following stages. 

(a) Reexamine the Krogh cylinder model and its 
adequacy for the representation of perfused organ 
microcirculation. 

(b) Develop exact mathematical solutions for the 
Krogh model that exhibit tissue axial diffusion and 
capillary axial diffusion for the steady state constant 
metabolic rate experiments with perfused organs. 

(c) Develop or modify numerical algorithms that 
will compute substrate levels for nonconstant 
metabolic rates and other nonlinear effects. 

(d) Develop algorithms for the direct comparison of 
distributed substrate level computations with 



experimentally obtained micro-electrode 
measurements. 

(e) Identify those critical ranges of parameters that 
control organ response to physiologic challenge. 

Progress in FY81: During FY81 attempts were made 
to use mathematical solutions in the existing 
literature to validate an experimental design for 
perfused organ studies. The numerical results 
revealed that these solutions, for the cell-free 
perfused Krogh cylinder model, were mathematically 
incorrect. Our efforts were subsequently redirected 
to the development of new, intricate, mathematically 
correct, solutions of the Krogh cylinder model. The 
analytical construction and examination of these 
mathematical solutions is being completed, and 
these solutions are being explored parametrically for 
critical ranges and limits. Two preliminary reports on 
these properties have been presented at 
international meetings, and a detailed theoretical 
publication has been submitted. 

The following stages of the proposed course have 
been accomplished. 

(a) Reexamine the Krogh cylinder model and its 
adequacy for the representation of perfused organ 
microcirculation. 

(b) Develop exact mathematical solutions for the 
Krogh model that exhibit tissue axial diffusion and 
capillary axial diffusion for the steady state constant 
metabolic rate experiments with perfused organs. 
Publications and Abstracts: 

Fletcher, J.E,: Simulation; Procedures and Pitfalls. Proceedings of the 34th 
ACEMB meeting. Houston. Texas. 1981 (abstract), 

Fletcher, J.E,. and Jolly. M,; The Computation of Substrate Levels in Per- 
fused Tissues, Proceedings of the Annual Siam Meeting. Houston, 
Texas, 1980 (abstract), 

Fletcher, J,E., and Schubert, R,W,: Substrate Level Prediction and Histo- 
grams in Perfused Tissues, Proceedings of the 34th ACEMB meeting 
Houston, Texas. October 1981 (abstract), 

Fletcher, J,E,, and Schubert, R.W,: The Theoretical Prediction of Substrate 
Levels and Their Histograms in Cell Free Perfused Tissues, Proceed- 
ings of the International Meeting of OTT Society. Detroit, Michigan, 
August 1981. Plenum Press (in press). 



Analysis of Coupled Transport and Biochemical 
Kinetics 

This project investigates three fundamental in 
biology: (1) the role of dynamic patterns in 
embryology, (2) the kinetics of enzymes in cell 
membranes, and (3) the kinetics of enzymes from 
malignant and normal cells. The first area involves 
the role which simultaneous reaction and diffusion 
might play in the spatial and temporal organization of 
organ shapes and surface markings. The second 
area involves enzyme kinetics with special emphasis 
on limitations in inferring molecular mechanism 
solely from gross observations. The third area 
involves the integration of instrumentation for data 
acquisition, display, transmission, and mathematical 
analysis of kinetic studies on lactate dehydrogenase, 
an enzyme whose function is altered in hepatocytes 
a malignant state. Digital computer simulation, 
particularly by means of network modeling 
languages, numerical solution of differential 
equations, and nonlinear regression analysis are the 
main tools in these investigations. 

• Dynamic Patterns 
Bac/<ground and Objectives: Temporal and spatial 
organization is universal in living organisms. What 
are the mechanisms which produce it? The 
equations describing diffusion of chemically reacting 
molecules possess solutions which spontaneously 
develop spatial and temporal patterns of 
concentration. Because reaction and molecular 
motion are the predominant physico-chemical 
processes of living organisms, it is tempting to 
suppose that the one is the mechanism of the other. 
The objectives of this project are to develop 
numerical methods to solve reaction-diffusion 
equations so as to investigate the principles 
controlling pattern formation. 

Progress in FY81: The bifurcation diagram of a 
reaction-diffusion equation summarizes the 
multiplicity and stability of solutions. This year the 
bifurcation diagram for the model equations 
developed previously was successfully computed. 
Techniques from functional analysis have been 
applied to determine the multiplicity of solutions. One 
technical manuscnpt, long in press, has appeared. A 
popular presentation is in press, and another 
popularization has been solicited. The technique for 
computing bifurcation diagrams has appeared as an 
internal report at a collaborating institution and will 
shortly be submitted for publication. 
Significance to Biomedical Research: Patterns 
generated by reaction and diffusion have been 
hypothesized to play a role in a variety of 
developmental processes in biology. The validity of 

59 



these hypotheses has never been critically tested. 
Several critical questions about such patterns are 
addressed: What is the relation between pattern 
form and the shape of the medium in which the 
pattern grows? Is there a relation between the 
patterns seen and the molecular mechanism of the 
reaction? Are the patterns unique and stable? 

Future Course: A formalism for dealing with 
bifurcation to temporally periodic solutions has been 
acquired. During FY82 this theory will be used to 
direct experimental research into spatially distributed 
oscillations in a membrane containing the 
immobilized enzyme phosphofructokinase. 
• Kinetics of Enzymes in Membranes 

Background and Objectives: Studies of the 
mechanism of membrane transport and excitable 
membranes are generally less precise than studies 
of the mechanisms of enzymes in solution. This 
uncertainty arises because it is difficult to manipulate 
the environment of the interior of a biological 
membrane and also difficult to measure responses 
there. There are two objectives to this project: to 
determine the extent to which the actual organization 
of models for membrane-associated processes can 
be correctly inferred from the kinds of experimental 
measurements currently made, and to develop a 
formalism within which complex kinetics can be 
entered rapidly and accurately as data into a 
simulation program. 

Progress in FY81: During FY81 the network 
simulation languages SPICE and NET-2 were 
installed on the IBM 370 central computers. Several 
collaborative projects have shown the utility of these 
languages for rapidly building and testing models for 
a variety of biological phenomena. In the area of 
membrane transport, SPICE modeling of several 
molecularly distinct schemes showed them to be 
experimentally indistinguishable. This material was 
presented at the Polish Winter School of Membrane 
Transport and has been submitted for publication. To 
demonstrate the utility of network modeling to the 
NIH community, a course is conducted through 
DCRT; this year a dozen investigators attended. A 
book on the application of network methods to 
physiological simulation is in preparation, as is a 
manuscript on use of network methods for 
physiological models based upon partial differential 
equations. Examples from the areas of 
microcirculation and axon physiology are illustrated. 
Recently, a second type of network, called a state 
transition network, has been employed successfully 
in the characterization of several types of membrane 
processes. The means of translating such networks 
into SPICE, gaining considerably in the size and 

60 



\ 



generality of models that can be considered, has 
been demonstrated. An invited illustrative paper was 
delivered at a recent conference on NET-2 
conducted by the Navy. 

Significance for Biomedical Research): The choice of 
a model for a biological process strongly conditions 
the design of experiments to confirm and extend it. 
By making the analysis of models sufficiently simple, 
an investigator is given the freedom to consider 
many models. From comparisons among the models 
using simulation, it should be possible to develop 
incisive experiments which permit scientifically valid, 
rather than arbitrary, selection among the models. 
The network languages nicely complement the 
MLAB system in permitting users to model 
phenomena too complex to be conveniently 
described in MLAB. 

Future Course: During FY82, an effort will continue 
to proselytize the NIH community on the utility of 
network modeling. A compendium of transition state 
networks, currently in preparation, will provide the 
basis for development of the concept of 
distinguishable classes of models in a number of H 
areas of membrane biology. The extent to which B 
current experimental techniques actually provide 
confirming data for the models which they employ 
can then be determined. 

• Kinetics of Lactate Dehydrogenase(LDH) from 
Normal and Malignant Hepatocytes ^ 

Bacliground and Objectives: When the enzyme LDH ™ 
is isolated from normal and malignant hepatocytes 

grown in tissue culture, there are differences both in h 

the activity of the enzyme and its sensitivity to H 

inhibition by a reaction product. The objective of this ™ 
project is to characterize the differences in LDH from 

various sources, emphasizing particularly the h 

differences between normal and transformed cells. ■ 

Progress in FY81: The method of kinetic analysis 
was applied to LDH obtained from several sources, 
including normal and transformed hepatocytes. 
Differences detected in this way were also reflected 
in iso-electric focusing studies on the enzyme from 
the two sources. The apparatus for performing the 
kinetic studies was installed at NIH. A microcomputer 
was purchased and installed by the principal 
investigator that permits data from stopped-flow 
spectrophotometry of LDH kinetics to be stored, 
edited, and transfered to NIH computers for further 
analysis. Software for this purpose has been 
acquired. A manuscript describing the work is about 
to be submitted. 

Significance to Biomedical Research Differences in 
the enzyme not evident from other types of analysis 



I 



can be shown through kinetic studies. The method of 
kinetic analysis employed here distinguishes 
between LDH from normal and chemically 
transformed hepatocytes, whereas the enzymes from 
these two sources are indistingusihable by isozyme 
analysis. The technique is probably applicable to 
LDH from other organ sources, although this remains 
to be shown. 

Future Course: During FY82 the new computer will 
be employed to extend this type of analysis to cells 
from other types of malignancies. 

Publications: 

Bunow. B-: Cellular Enzymology: Edect of compartmentation on steady 
slate kinetics, J. Theor Biol. 84 611-628. 1980, 

Bunow. B,: Turing and the physico-chemical basis of biological patterns. In 
Prewitt. J, (Ed,): IEEE Turing memorial 1980 (in press), 

Bunow. B,, Kernevez. J P , Duban. MC. Jolly. G,. and Thomas. D,: Pattern 
formation by reaction-diffusion instabilities: Application to morphogane- 
sis in Drosophila, J. Theor. Biol 84: 629-649. 1980, 

Bunow. B, and Kernevez. J P . Reaction-diffusion patterns as a basis for 
biological form-some discouraging results. Abstracts ol American 
Math. Soc 1-521. 1980. 

Bunow. B,. and Mikulecky, D.C: On the feasibility of using flux meas- 
urements to distinguish among active transport models. Polish Winter 
School ol Membrane Transport 1981 (in press), 

Bunow. B . and fvlikulecky. D C: Where does metabolic energy couple into 
the active transport process? J Theor. Biol (in press), 

Kernevez. J,P . Jolly. G . Thomas. D , and Bunow. B,: Pattern formation and 
wave propagation in the S-A system In Bardos. C , Lasry. J M., and 
Schatzman, 1^, (Eds.): Lecture Notes in Mathematics 782: 201-221. 
Springer-Verlag. 1980 

Sharan, M,. Kernevez. J P , and Bunow. B : On numerical exploration of 
bifurcating branches of solutions in reaction diffusion equations model- 
ing enzymatically active artificial membranes. Research report ol the 
Department ol Applied Mathematics. University of Technology of Com- 
piegne, Compiegne. France. 50 pp 



61 



"»""-"'i5"""""""" •" 


;.T' 


TJSiili- 


if" 


'io, aooo.0-0 


LAS 












Nonlinear Equations 


PI: R.I. ShraGer 
O.E. Fletcher 


Hjthen 

Cotnput 
Chief 
Bloche 


er SpccUllsl 
MS 




LAS DCRT 

IAS DCRT 
LCR nilLOl 
LCR NIAOOK 






Laboratory of Applied Studies 


Applted Itothenatics Section 


DCflT. Hill. Bethesda. HD 


20205 










lom MAMtws, par 


"Til' 


1"'""' 








at..)->«« a 1.2) .MMv.us 


o..,™» 


„„.„ 


° 


..... 










r equations frequently 

h as asymptotic error 
systems, are also 




Methods are developed For so1 


ja'it Kill. 




T? 


It 


irohlems, sfic 


differential equations. 
analysis and the efficle 



3iSEJrJS«!S'8S^ifS:*!iftJi.^^ 


'I'^iSiilL^nilir 


zoi CTnnoi5-o3 las 


'"october°i. 1980 to September 30, 1981 


Numerical Appronination T«:finiques for the Solution of 
Systems in Biology 


Reaction- Diffusion 


OTHER: J.E. Fletcher Chief, AMS 

B. Bunow Bioinatheniatician. AMS 
1. fiabuska Professor, Institute of Physi 
Science and Technology 


LAS nCRT 

LAS DCRT 
LAS nCRT 


„.. 


Laboratory of Applied Studies 


Applied Mathematics Section 


nCRI. «IH. Pethesda. TO 20205 


O.a 0.8 


31.) ««.!«.«,! a(>l««.»i.».is ow.si™ 


""'ZIZZZIZ^ 


'Z27L.,.. .eactio 


n-diffusion 


systems and other bfoloqica 
reliable, computationally e 
Of nonlinearly co-ipled part 


processes requires the m 
al differential equations 


pWmentation of 
computer solution 
New numerical 
are is developed 


approximation techniques ar 
for the solution of model r 




biological problems. 


ans are altereri and applie 



Nonlinear Equations 

Methods are developed for solving nonlinear 
equations frequently encountered in mathematical 
modeling at NIH, usually in the context of 
constrained nonlinear least squares or in the solution 
to nonlinear differential equations. Related problems, 
such as asymptotic error analysis and the efficient 
treatment of sparse systems, are also considered. 

Background and Objectives: The objective of the m 
project is to develop methods for solving the ■ 

nonlinear equations frequently encountered in 
descriptions of biomedical problems at NIH. General 
categories of problems for which solutions were 
developed here (partially or wholly) are nonlinear 
least squares with linear constraints, nonstiff and stiff 
ordinary differential equations, and nonlinear curve 
fitting in norms other than least squares. In addition, 
every project with which this one interfaces (i.e., 
collaborations) presents its own set of special 
equations which must be solved, in either the 
analytic or numerical sense. Methods which prove to 
be of general utility are developed into accessible 
computer programs or routines, e.g., MODELAIDE 
and MLAB. 

Progress in FY81: 

• MLAB Projects 

Gary Knott, (DCRT). The root-finder has been 
revised with considerable improvement in efficiency 
and the techniques are being described in a 
manuscript in progress. An extension of the curve- 
fitter is being considered which could include not 
only least-pth power fits (added this year) but also 
maximum likelihood and M-estimation in which the 
likelihood function itself may contain parameters. 
The approach thus far has been to use the current 
Levenberg-Marquardt algorithm but alter the manner 
in which the function to be fit is presented to the 
algorithm. 

• GABA Metabolism 

E. Anthony Jones, Dan Schafer, Peter Ferenci, 
(NIADDK). GABA is a neurotransmitter which is 
stored only in brain cells. In normal animals, GABA 
concentration is higher in plasma than it is in 
cerebro-spinal fluid, and higher still in the portal vein. 
Recent experiments indicate that gut bacteria 
produce GABA which is then mostly degraded in the 
liver. When the liver fails, the blood-brain barrier 
breaks down, and the role of GABA in this process is 
not well understood. Data is now being processed 
on the metabolism of 3H-GABA from normal rabbits 
and rabbits in liver failure to determine the resulting 
alteration in GABA metabolism. Other experiments 



62 



(e.g., impairment of GABA metabolism without liver 
failure) will follow. 

• Equilibrium Studies of Magnesium Phosphate 
Lev Jacobson (NIADDK). The physiologically 
important reactions between Mg and P04 are not 
well understood. Various models are being applied to 
the fitting of NMR and pH data, with considerable 
improvement in data expected when a divalent metal 
electrode is used to detect magnesium ions. 

• Analysis of Experimental Spectra 

R.W. Hendler, D.Y. Setty, Dan Robertson (NHLBI). A 
paper on the use of singular value decomposition 
(SVD) in data matrices to detect chemical transitions 
and their associated spectra is in the final stages of 
revision. New data from beef heart mitochondria 
indicate that cytochromes in mammalian cells have 
essentially the same mechanism for passing 
electrons as that observed in the bacteria. Spectra 
to verify this will soon be forthcoming for processing 
by SVD or a related method. A paper on oxygen 
uptake and proton release by cells and reconstituted 
vesicles has been submitted. 

• Phytic Acid Titration. 

William Evans (Dept. of Agriculture, New Orleans). 
Phytic acid is a chelating agent which may be a 
factor in metal-deficiency diseases. A manuscript, on 
the liganding mechanism, has been revised and 
resubmitted to The Journal of Agriculture and Food 
Chemistry. 

• Hemoglobin production 

A.N, Schechter, Ann Dean, Francois Erard 
(NIADDK). An abnormal human cell line, called k562 
cells, can be induced by addition of hemin to 
produce embryonic and fetal hemoglobins. Inhibitors 
of cell division cause these cells to accumulate 
hemoglobin, and the combined effect is a hundred- 
fold increase in Hb concentration. Can all the effects 
observed be explained by a simple kinetic model in 
which only hemin controls production rates? The 
data are now being processed. 

Significance to Biomedical Research: These 
methods are essential for the resolution of problems 
of data analysis in metabolism, cell growth, chemical 
kinetics, and spectral analysis (UV, IR, CD, ORD, 
NMR, ESR). 

Proposed Course: The new root-finder will be 
compared with the current techniques in the existing 
literature and the results will be published. The 
maximum-likelihood and M-estimation curve-fitter will 
be developed and tested. If it proves feasible, the 
syntax of MLAB will be modified to permit easy use. 
The new data from beef-heart mitochondna will be 
processed by the SVD technique. Simultaneous 



measurements from pH and divalent-metal 
electrodes are expected to clarify the modeling of 
the magnesium phosphate system. 

Publications and Abstracts: 

Berk, P D . Blaschke, T F , Shrager. R I . Waggoner. J G : Phenobarbitol 
does not increase hepaticheme turnover in Man Gaslroenlerology 79: 
1004. 1980 

Numerical Approximation Techniques for the 
Solution of Reaction Diffusion Systems in 
Biology 

The use of mathematical models to describe 
reaction-diffusion systems and other biological 
processes requires the implementation of reliable, 
computationally efficient algorithms for the computer 
solution of nonlinearly coupled partial differential 
equations. New numerical approximation techniques 
are studied and related software is developed for the 
solution of model reaction-diffusion systems. Existing 
methods and previously written programs are altered 
and applied to other biological problems. 

Background and Objectives: Mathematical models 
describing many biological and physicochemical 
processes involve systems of coupled ordinary and 
partial differential equations that first must be 
formulated and then solved via computer. It often 
happens that the necessary numencal algorithms do 
not exist in the literature, the related computer 
programs are unavailable, or that applicable software 
requires a large amount of computer time, rendenng 
its implementation impractical. The two objectives of 
this project are the development of new, efficient 
numerical techniques for the solution of equations 
describing reaction and diffusion processes and the 
modification and implementation of existing 
programs to solve other specific biological problems. 

Progress in FY81 The investigation of an adaptive 
'method of lines' solution approach and the 
implementation of known techniques for the 
approximate numerical solution of time-dependent 
partial differential equation systems ansing in biology 
have continued in FY81 . 

Significance to Biomedical Research: A variety of 
reaction, diffusion, and transport processes occurnng 
in biomedical application areas are modeled by 
equations which can be solved by the techniques 
being developed. Among these processes are 
oxygen transport in the microcirculation, embryologic 
pattern formation, nerve impulse transmission, and 
population dynamics of ecological systems. Effective 
modeling of such nonlinear distributed parameter 
systems involving both reaction and diffusion 
requires the use of accurate, computationally 

63 



"•™'" 


ffct iwowi»n» tiCHiaci u.s. otMHiMtNi or 


ZOl CTO0O34-05 LAS 


""""'"" October 1. 1980 to Septerter 30. 1981 j 




Conputer-based Studies in Pulmonary Pathophysiology 
and Respiratory Disease 


PI: 

OTHEBS: 


R. C. Burgess Senior Staff Fellow LAS DCRT 1 
A. H. Nienhuls Chief CHB NHLBI 1 
J. J. Bailey Chief. MAS LAS DCRT 
R. G. Crystal Chief PB NHLBI 

E. K. Harris Chief LAS DCRT \ 
M. R. Horton Comiuter Systems Analyst LAS DCRT 
E. W, Pottala Electronics Engineer LAS DCRT 


Clinical HeiiBlolo^. Pulmonary Branch. NHLBI, Nuclear Medicine Dept.. CC 1 
Branch. NHLBI | 


Laboratory 


of Applied Studies 


Applied fc 


theratics Section, Medical Applications Section 


KRT. NIH. 


Bethesda. Maryland 20205 


2.0 


|..«.si.«.L, B..l„ 


J I.I «««.» 
al..)-i><« 


J»JtH □ 1.) muH iiiju. u 10 .[„„„ 


Tills project— through a collaborative effort of LAS with the Nuclear Medicine 
Departaient, CC and the Clinical Hematology and Pulmonary Branches, NHLBI— is 
directed toward a deeper understanding of pulmonary pathophysiology through 
the construction of computer-based rrodels of pulmonary gas exchange and 
respiratory mechanics and comparisons of model predictions with real patient 









I 



I 



efficient numerical algorithms. These methods and 
other standard techniques are available to the NIH 
biomedical community as research tools and can be 
modified or used directly to solve problems of 
physiologic interest. 

In the 'method of lines' approach the finite element 
technique is used to approximate the spatial 
variables. This yields a system of ordinary differential 
equations which can be solved, for example, by one 
of the many high quality software packages 
available. The scheme is adaptive in that decisions 
concerning mesh selection and refinement are made^ 
by the computer during the problem solution. Such HI 
decisions are based upon reliable a posteriori VI 

estimates of the error between the exact solution 
and the computed approximate solution. Changes in 
the mesh structure are made in order to control the 
error and increase the efficiency of the solution 
process. Such schemes are especially applicable to 
dynamic biological systems whose behavior is 
localized in space. 

Work has been completed which extends the full 
mathematical framework of the a posteriori error 
analysis developed for single, steady-state equations 
by I. Babuska and W. Rheinboldt to coupled, time- 
dependent linear systems of partial differential 
equations. Adaptive mesh selection strategies based 
upon these estimates have been developed and are 
being implemented in computer programs on the NIH 
IBM System 370. The accuracy and efficiency of the 
adaptive procedure are to be tested on a collection 
of nonlinear problems of biological interest. 

In addition to the above investigation, recent 
collaborative efforts have been initiated with two NIH 
researchers: 

1. A preliminary study with Dr. Robert Lutz, DPS/ 
BEI, has been completed. The study examined the 
solution of the Navier-Stokes equations for a two- 
dimensional model of blood flow through a section of 
an inelastic canine aorta model. This collaboration 
will involve implementing a fluid dynamics finite 
element code, which will either be obtained and 
modified, or designed and written by LAS. Computed 
wall shear stresses and blood velocity profiles are to 
be compared to laboratory experimental data and 
used in a parametric study of factors relevant to the 
onset of atherosclerosis. 

2. Work with Dr. Ralph Nossal of DCRT/PSL is 
underway to determine if modified versions of the 
Bathe-Wilson finite element code could prove useful 
in deducing the elastic properties of gels from light 
scattering measurements. In a mechanically excited 
gel, the frequencies of standing waves are related tr 
the elastic properties of the gel. In most 



64 



experiments, these frequencies can be only 
approximated crudely. The objective here is to 
develop a scheme for approximating these 
frequencies by a finite element technique. 

These collaborative projects will continue at levels 
determined by LAS/AMS priorities and manpower. 

Publications and Abstracts: None 

Computer-based Studies in Pulmonary 
Pathophysiology and Respiratory Disease 

This project-through a collaborative effort of LAS 
with the Nuclear Medicine Department, CC and the 
Clinical Hematology and Pulmonary Branches, 
NHLBI--is directed toward a deeper understanding of 
pulmonary pathophysiology through the construction 
of computer-based models of pulmonary gas 
exchange and respiratory mechanics and 
comparisons of model predictions with real patient 
data. 

Bacliground and Objectives: Numerous attempts 
have been made in the past to quantify pulmonary 
function. Inhomogeneities in the lung required certain 
simplifying assumptions to be made tending to 
obscure the true nature of lung function. 
Furthermore, certain nonlinearities inherent in the 
lung system allowed only partial quantitative models 
and sometimes these could only be expressed in the 
form of nomograms or graphs. 

Within recent years it has been possible to apply 
computer technology to numerous diagnostic tools, 
viz, spirometry, dynamic compliance studies, multiple 
inert gas studies, pulmonary scintigraphy, cardiac 
catheter studies, and blood gas studies. 

The objectives of this program include the use of 
computer technology to refine diagnostic methods 
and to construct models for pulmonary gas 
exchange and respiratory mechanics. 

Progress in FY81: The existing system for analysis 
of gas exchange was tested extensively. It was 
found that instabilities in the gas analysis lines and in 
the analog circuitry prevented any hope of obtaining 
reliable data. Therefore, the system was completely 
re-designed. Literature on exercise-testing 
laboratories was reviewed; the state-of-the-art in 
commercial equipment was analyzed; and one of the 
most advanced exercise-testing laboratories 
(Wasserman in California) was visited. LAS 
developed specifications and directed purchase of 
the equipment, much of which has been delivered. 
The basic components of the new system include: 
an LSI-1 1 based computer for rapid data acquisition 
and processing; a pneumotachometer and mass 
spectrometer to provide flow and concentration data; 



and a bicycle ergometer and treadmill whose work 
loads can be varied under direct computer control. 

Significance: In certain diseases of the lungs 
(restrictive, obstructive, etc.), blood 
(hemoglobinopathy sickle cell anemia, polycythmia). 
and cardiovascular system, the assessment of the 
patient's condition at various points in the course of 
disease may be no better than a subjective 
impression given by the patient, his family, or his 
physician. An alternative, and potentially more 
objective, method of assessment involves the use of 
continuous exercise to quantitatively evaluate the 
overall ability of the patient to meet the demands of 
exercise. This is accomplished by monitoring his 
ECG, blood pressure, blood gases and lactate, 
oxygen consumption, etc., in a reliable and 
reproducible manner. The detection of 'anaerobic 
threshold' may be of particular clinical importance 
when considering response to therapy or disease 
progression. 

The sensation of 'shortness of breath' is poorly 
understood. This phenomenon clearly has 
components of neural as well as clinical origin. Non- 
steady state exercise measurements should give 
additional insight into the origin of this complaint in 
patients with vanous disease processes. This data, 
in conjunction with other parameters, should improve 
prognostic accuracy and aid the theraupeutic 
decision-making process. 

Proposed Course: Configuration and interfacing of 
the system is in progress and will continue into 
FY82. Many components from the old system (e.g., 
bicycle ergometer, Tektronix display) will be 
incorporated. When assembled, the new 
configuration will be tested extensively for reliability. 
Subsequently, a normal data base using volunteers 
will be acquired as a baseline against which the 
tests in pathological conditions can be compared. 

Publications: None 



65 






October 1. 1980 to Scpteinbc 



Mybriri Cwipirtlng for t 



PI: r.M. Pottala 

OnrERS: J.J. BaHey 
I. iaar 
H.C. Van Ars-ldle 



Phar(naca1o<]fs 
Tug. Aid 
Vlstttng Scic 



lanoralory of Applied ' 



Jl£aL_tlllU_afiii£ilA^ 



Id 1 yze physiologic 



doviilop physioloqic simutati 


n models 


use hybrid canputinci techmq 
electrocardioirain, eloctroen 


ephaloqrain 



si^irisiais'^^of SSI"** it^" 


.l^iFsilimfcT 




ZOl 




October I. 1980 to September 30. 1981 


Computer Systems for Nuclear Medicine 


PI: J.J. Railey Chief, Med. Appl. Sec. 
S.L. flacharach Physicist 
M.A. Douglas Comp. Syst. Analyst 
fi.C. Burgess Senior Staff Fellow 

OTHERS: ».G. Ostrow Engineer 

M.v, Greun Ch. Appl. Physics Sec. 
A.E. Jones Chief. Diagnostic Imaging 
G.S. Johnston Chief 
R.O. 8onow Clinical Associate 


Cf! 


DCRT 
CC 


Nuclear Medicine nepartiront. CC, NIH 
Conputer Systems Laboratory, OCRT, HIH 




Medical Applications Section 




1.1 1 3.0 1 o'.l 


3 10 K«« «•*«» □ ,., H^« ,.».tS □ (. 


"'■'" 




This project involves computer-based qathenatical 


*n.lv 


,^ 


patt<^rn recognition, and ,maqe processlm in support 


i] 


ii 


Center 


and colUhorating Institutes. Applications Include 


sclninraphy. renal dynamics 




reMtlonsMps. 

Wssi 



investigation of l-iybrid Computing for the 
Construction of Simulation Models and for the 
Analysis of Physiologic Signals 

This project was undertaken to develop physiologic 
simulation nnodels using hybrid computing and also 
to use hybrid computing techniques to analyze 
physiologic signals such as electrocardiogram, 
electroencephalogram, and electromyogram. 



I 

i 



Background and Objectives: This was extensively 
described in last year's Annual Report (FY80) pages 
144-146. 

Progress During FY81: The Medical Neurology 
Branch, NINCDS, has studied electromyograms with 
intra-muscular probes to determine muscle fiber 
conduction velocities in various disease states 
(references 1 and 2). Analog-to-digital conversion 
and spectral analysis were performed on the MAC- Jj 
16 system. The Medical Neurology Branch also ™! 
analyzed EEG's on patients being treated for hepatic 
coma, using the MAC-16 system. 

The Division of Cardio-Renal Drug Products, FDA, is 
investigating the early detection of cardiac toxicity 
resulting from drug therapy. Rat electrocardiograms 
are being used to determine the sensitivity of 
detection. Analog-to-digital conversion of the data 
has been completed to provide data for the analysis 
programs. 



4 



I 



Significance: In some simulation models, certain 
pieces or functions can be split off and implemented 
in hardware circuitry or in a set of microprocessors. Jl 
This has several advantages. First, parallel ™ 

processing is allowed, which can shorten computing 
time and make interactive model testing feasible. 
Second, the hardware circuitry or microprocessors 
are usually quite inexpensive. Third, some models 
are so complicated and extensive that their 
implementation on a large scale digital computer is 
not feasible; however, with hybrid computing, such 
models may be achieved. An example was the 
model of the Purkinje network in the alligator 
cerebellum, which required a system of 35 cells 
connected by nonlinear differential equations 
(completed in FY76). Another example was the 
simulation of retinal cone cells (described FY79). 

Currently, the effect on cardiac behavior of various 
drugs~in particular, cancer chemotherapy agents-is 
monitored by a single lead electrocardiogram (ECG) 
in animals. The end point for cardiac toxicity is 
terminal ventricular tachycardia. The current study is 
investigating multiple lead ECG's and the computer 
analysis of this data to provide more sensitive and 
accurate end points for drug effects. 



66 



Proposed Course: No new projects involving 
simulation models are being undertaken at this time. 
However, the capability remains, should a 
collaborative project be proposed in the future. For 
example, the Laboratory of Neurotoxicology, 
NINCDS, has studied a rat preparation with bipolar 
electrodes in the sensory cortex and in two layers of 
the hippocampus. The theta activity generators in 
the hippocampus produce certain correlated 
spindling activities in the sensory cortex, varying 
according to degree of a drug-induced, immobility 
state. Data already collected will be analyzed, using 
the fy/IAC-16 system. Whether additional study is 
needed using a simulation model remains to be 
investigated. 

The MAC-16 system will have continued use for 
ECG processing from the Framingham Heart Study 
(see project report on electrocardiography). 

Analysis of FDA data on cardiotoxicity has begun 
and will continue in FY82. 

Publications and Abstracts: 

Yaar, I.. Shapiro, MB. Mitz. AH,, and Poltala. E,W,: A computer assisted 
monitoring of muscle fiber conduction in full interference patterns: ALS 
versus normal subiects, A preliminary report, Ttie Amencan Association 
of Electromyography and Electrodiagnosis Meeting, Philadelphia. Penn- 
sylvania. 1980. Eleclroencephol- Clin. Neurohysiol. 50:245P. 1980 (ab- 
stract) 

Yaar. I,, Shapiro, M.B.. Mitz, A,R,, and Pottala. E,W.; Introducing a new 
computer assisted technique for measunng muscle fiber conduction 
velocity at full interference pattern, American Neurological Association, 
Boston, Massachusetts. 1980, Ann. Neurology 8:124. 1980. and Trans- 
actions of the Amencan Neurological Association 1980 (abstract), 

Yaar, I , Shapiro, MB. and Pottala, E W,: An EEG power spectral analysis 
of dopaminergic mechanisms in patients with hepatic coma, Amencan 
Electroencephalographic Society meeting Boston. Massachusetts. 
1980. Electroenceph. C/in. Neurophysiol. 51:31P, 1981 (abstract). 



Computer Systems for Nuclear Medicine 

This project involves computer-based mathematical 
analysis, pattern recognition, and image processing 
in support of diagnostic activities in the Nuclear 
Medicine Department of the Clinical Center and 
collaborating Institutes. Applications include 
computerized ECG-gated radionuclide 
angiocardiography and myocardial perfusion 
scintigraphy, renal dynamics, and pulmonary 
ventilation-perfusion relationships. 

Bacl<ground and Objectives: Since FY72, LAS and 
CSL have collaborated with the Nuclear Medicine 
Department, CC in the acquisition and development 
of several minicomputer systems that gather and 
process data from the scintillation cameras in the 
Nuclear Medicine Department. 

The objective of this program is to continue 
development of computer-based algorithms, which 
have already found wide-ranging applications, 
including: fitting mathematical models; mapping the 
parameters of such models over time and in different 
regions of an organ; image processing; interpolation, 
expansion, and contraction of image arrays; and 
pattern recognition. 

Progress during FYS 1: 

Renal Scintigraphy-V^ ork in FY76 showed a 
significant enhancement of radionuclide renography 
by the use of functional maps, and, since FY77, 
functional maps have come into routine clinical use 
in the Nuclear Medicine Department. In FY79-80 a 
pilot study of partial renal artery ligation was carried 
out on four dogs. The results of this study were 
reported at the Society of Nuclear Medicine, and a 
manuscript is now in preparation. In FY81 the 
techniques of the pilot study were improved in 
several ways: 

• The studies are uniformly performed with 1-123 
iodohippuran, improving the signal noise ratio 
about 16-20 fold; 

• A radio-transparent table was designed so that 
positioning of the animal could be carefully 
controlled; 

• Precise positioning of the kidneys under the 
gamma camera was guided by technecium-99m 
DPTA given in minute doses; and 

• The software has been entirely rewritten to work 
in conjunction with the current Nuclear Medicine 
Operating System and to decrease image 
variability. 

Data on nine dogs including renal scintigraphy and 



67 



KS"js.s'i!;'rr!i!" 


;S!.T' 


,Zil!:L 


ZOl CT00002-11 LAS 


""'» ""■" October 1. 


1980 to September 30. 1981. 




Cofrpiiter-Aided 


Analysis of Electrocardiograms 




PI: J. J. Bailey 

OTHERS: E. K. Harris 
H. E. Wonbte 
P. tecFarlane 

D. Savage 
S. Allen 


Chief, HUS 

Computer Systems Analyst 

Chief 

Kedica! Cardiology 

Framingham Heart Study 
Medical Research Analyst: 


LAS DCRT 
LAS DCRT 

LAS OCRT 

Glasgow Royal Infirmary 
Scotland 

CSL DCRT 


co«p<iu.i« u.iii 0. ..,1 


HBrtMNO. 






Medical Appl 






'"'""" ""krt'.'^IH. Bethesda, Karyland E0E05 1 




fflOFessiMAL, \atm, ^ 


1 


J(.) «l»»Sl»JU15 


a(b)«u»v.ii5!ULS : 


,.,«,.». 


These studies continu 
computer analysis fo 


ing sin 


ce 1970 have been direc 


ed toward the 









contrast angiography before and after renal artery 
ligation has been collected. Preliminary results 
suggest a vast improvement in technique over the 
pilot study. 

Cardiac Scintigraphy-\n collaboration with Nuclear 
Medicine and the Cardiology Branch, LAS has begun 
investigation of several parameters reflecting mobility 
of the heart wall including ejection fraction, time-to- 
end-systole, phase (of a fitted cosine curve), and 
ratios of areas over/under the time-activity curve of 
a blood pool scan. Programs have been written to 
compute these parameters globally or for any of four 
sectors of the heart image. The test data base 
includes rest and exercise studies on 40 normal 
volunteers, 10 patients with chest pain but 
completely normal cardiac studies (including ECG 
and coronary angiogram), 10 patients before 
beginning adriamycin therapy, and 15 patients with 
coronary disease and known resting apical 
abnormalities (hypokinesis, akinesis, or dyskinesis). 
Preliminary results suggest that ejection fraction and 
phase are the best parameters for separating normal 
from abnormal cases. 

Image Processing Deve/opments--Jhe DECsystem- 
10 based image processing packages, IMAGE and 
PSTACK, continue to be expanded and used both 
for dynamic scintigraphic images and for electron 
microscopy images. 

In the past year, the LAS DeAnza image processing 
system has undergone rapid expansion. Two major 
interactive software packages have been developed. 
The first, PICTUR, currently includes a variety of 
edge detection and tracking routines and alignment 
procedures in addition to display and image 
enhancement options. 

The second, MOVIE, is aimed at time varying image 
sequences. During the past year this package has 
been enlarged to include generation of flow/volume 
loops, phase/amplitude maps and a variety of 
enhanced dynamic display procedures. These two 
packages are designed to be compatible with one 
another; transition from the use of one package to 
the other is invisible to the user. 
A magnetic tape drive to facilitate data interchange 
and a disk drive to improve storage capabilities have 
been purchased and are being interfaced into the 
system. 

Proposed Course: 

Renal Scintigraphy--A project planned in FY80 would 
have allowed the use of renal scintigraphy to study 
patients with renal vascular hypertension in 
collaboration with the Hypertension Branch, NHLBI; 
it is still awaiting FDA approval of 1-123 iodohippuran 



68 



as a diagnostic agent. Extension of the canine 
studies with different lesions is under discussion. 

Cardiac Scintigraphy-fK statistical analysis of the 
data collected will be pursued, one possible outcome 
of which might be a discriminant function of ejection 
fraction and phase to achieve optimal separation of 
normals from abnormals. Another interesting study 
will involve those patients with myopathy secondary 
to adriamycin therapy, using each patient before 
therapy as his own control. Those patients who have 
normal contractility at rest but abnormalities upon 
exercise form another interesting data base. 

Image Processing--lhe DeAnza image processing 
system is to be expanded from its current 256x256 
image size to a 512x480 image. This expansion, 
together with the new magnetic tape drive and the 
disk drive, will facilitate study of paired myocardial 
(Thallium) and blood pool image sequences. Refined 
edge detection, assessment of wall motion 
abnormalities and perfusion, and determination of 
'absolute' volumes, if possible, are planned. 

A model is to be developed to demonstrate the 
effect of known amounts of additive noise on the 
detectability of regional wall motion abnormalities. 

Publications: None. 

Computer-Aided Analysis of Electrocardiograms 

These studies continuing since 1970 have been 
directed toward the evaluation of accuracy, clinical 
utility, and cost effectiveness of various computer 
systems for analysis of routine electrocardiograms 
(ECG's). Further studies will involve new methods of 
feature extraction and design of criteria by computer 
techniques and their use in epidemiological studies. 

Background and Objectives: This has been 
extensively described in last year's Annual Report 
(FY80), pages 111-114. 

Progress during FY81: A study of 300 ECG's from 
the Royal Infirmary in Scotland was published 
(bibliography follows). This study resulted in the 
comparison of two EGG programs using guidelines 
and definitions that were developed by LAS. This 
method of evaluation, when applied to two or more 
programs, can reveal their relative strength and 
weaknesses. The reason for a program failure can 
often be pinpointed to a specific defect in pattern 
recognition, measurement algorithm, or design of 
criteria. 

In FY80, the Framlngham Heart Study proposed 
longitudinal studies of routine ECG's in that 
population. In FY81, LAS collected ECG's on 200 
subjects, including 100 blacks, and 25 normals 
repeated 4 times. 



Significance: See Annual Report (FY80), pages 
111-114. 

Proposed Course: Testing of ECG programs 
developed by other organizations was largely 
completed in FY79. However the development of 12- 
or 15- lead ECG data acquistion devices and the use 
of sophisticated feature extraction methods (e.g., 
Karhunen-Loeve expansion) opens the possibility for 
further investigation of ECG diagnostic systems. 

Meanwhile LAS continues to study the epidemiologic 
significance of the routine ECG in collaboration with 
the investigators of the Framingham Heart Study. 
The ECG correlates of such heart diseases as 
coronary disease, mitral prolapse, and asymmetnc 
septal hypertrophy in a free-living population are of 
particular interest. 

Studies of statistical variance within and between 
individuals already begun will extend into FY82. A 
separate study to test a black population for 
differences is planned. 

Publications and Abstracts: 

Bailey. J J : The future of gold standards and computenzed eieclrocardio- 
grapfiy In Tolan. G,D and Pryor. T A (Eds.): Computenzed Inlerprela- 
tion ot the ECG V New York. Engineering Foundation. 1980. pp. 229- 
233, 

Bailey JJ.. Berson. AS,. Jacl^son, L,K,, fylilliken. J, A,, Stevens. Jfyl,, Totan. 
G,D,. and Wolf. H,K,: Evaluation Methodologies lor ECG diagnostic 
systems In Bonner, RE and Pryor. T A, (Eds,); Computerized Interpre- 
tation of ttie ECG VI. New York. Engineering Foundation. 1981 (in 
press), 

Bailey, J.J., and Harns. E K,: Evaluation o( ECG interpretation: Truth versus 
beauty In Tolan,G.D. and Pryor, T,A. (Eds): Computenzed Interpreta- 
tion of Itie ECG V. New York. Engineering Foundation, 1980, pp 179- 
182, 

Bailey, J,J., and Horlon, M,R,: Type A electrocardiogram data bases: Pur- 
pose and development. In Wolf. H,K., and Macfarlane, P W. (Eds): 
Optimization of Computer-ECG Processing New York, North-Holland 
Publishing Company, 1980. pp, 189-195, 

Macfarlane, PW. Chen, C,Y , and Bailey. J, J,: A companson of point 
scoring techniques for the diagnosis of LVH, In Macfarlane, PW, (Ed.): 
New Frontiers in Electrocardiology. London. Pitman Medical Publ. Co., 
1981 (in press) 

Macfarlane, PW,, Melville, D,l,, Horlon, M.R,. and Bailey. J J, Comparative 
evaluation of the IBM (12 lead) and Glasgow Royal Infirmary (3 ortho- 
gonal lead) ECG computer programs Circulation 63: 354, 1981, 



69 



■KOJtCI «««« (S« ■« «. IM. .p.«^ 


,l"rl?°ii:'k'"!!i'l 


ZOl aO0043-O 


LAS 




nut Of PNOJtCT (W ch.r«t.r. «r l.») 

Computer-based studies in ultrasonography 






Pi: E.W. Pottala 
OTHERS: H.A. Douglas 


UAS DCRT 
CG NIILBI 
LAS OCR! 
LAS nCRT 




Cardiology Branch, NHL8I 


Li8/0R"W 








r;:r:r:r"' o, .,...»..».„ p, .,«,.. 


This project involves collaboration of LAS, with 
Br^nrh NHI Rl . It IS directed toward computer-based 


the Cardiology 
orocessinq for 
ensional 




wide-angle, phased array echo-cardiography. 



sist?.s.s'rrr«!:sy' 


"ZSiiSir 


ZOl CTOOOflZ-03 LAS 


October 1. 1980 to Septenber 30, 1981 


Computer Based Analysis and lnage Processing in Electron 
Microscopy and X-ray and Electron-Loss Spectroscopy 


PI: H. A. Douglas Computer Systems Analyst 
J. L. Costa Medical Officer 

OTHERS: E. W. Pottala Elec. Eng. 


LAS DCRT 
CN BIHH 

CSL DCRT 


Clinical Neuropharmacology, NIMH, Laboratory of Chemistry, NIADDK 


Laboratory of Applied Studies 


Hedical Applications Section 


DCRT. NIH. Bethesda. Maryland 20205 


«.p.. |««,M.ON.L. jOIMDl. ^ ^ 


^ ""-«--» DU,««Hn.o. ni0.m.. 


This project tnvoWes collabo 


t of conputer-based mathematical and 


tatistica! 


analyses, pattern recognition 




data, princi 


ally x-ra,s 
studieTTn 


and electron energy loss spec 







Computer-based studies in ultrasonography 

This project involves collaboration of LAS, with the 
Cardiology Branch, NHLBI. It is directed toward 
computer-based processing for image enhancement, 
pattern recognition, and three-dimensional 
reconstruction from ultrasound data. The principal 
sources of data are wide-angle, phased array echo- 
cardiography. 

Background and Objectives: Ultrasonography allows 
non-invasive visualization of many organs without the 
hazard of ionizing radiation. Due to its safe nature 
and little or no patient discomfort, it is an excellent 
tool for screening and multiple repeat follow-up 
studies. Unfortunately, the presence of bone, which 
is completely opaque to sound waves, and certain 
processing practices have limited this technique. 
Using the computer, it is possible to overcome or 
circumvent many of these limitations. This project is 
directed toward ultrasound studies of the heart but 
could be expanded later to other organs. 

Progress in FY80: The acquisition of real data 
requires the development of an esophageal 
transducer with interfacing to a minicomputer 
system. Lack of staff has resulted in deferment of 
this objective until FY82. 

Meanwhile development of hardware (DeAnza 
System) and software (PICTUR) for general image 
processing has continued (see Nuclear Medicine and 
Electron Microscopy Sections). 

Significance: Patients with hypertrophic 
cardiomyopathy have an increased risk of sudden 
death. Unfortunately, many of these persons are not 
diagnosed ante-mortum because they are 
asymptomatic. A reliable technique to screen those 
persons with a family history of hypertrophic 
cardiomyopathy would be of great use, because 
prophylactic drug therapy is probably feasible. 
Unfortunately, many different patterns of hypertrophy 
appear to exist in the population afflicted. 
Techniques are needed to assess the distribution of 
hypertrophy in those patients with uncommon 
distributions, especially those missed by traditional 
M-mode echocardiographic techniques. Additionally, 
it is likely that the prognosis may differ among the 
various patterns of hypertrophy. The use of this 
technique to determine regional wall motion 
abnormalities and other parameters of left ventricular 
function in patients with coronary artery disease 
could prove superior to the techniques of nuclear 
cardiology. 

Proposed Course: To overcome the above 
mentioned problem, a phased array transducer 



70 



capable of use in the esophagus will be developed. 
This transducer will scan in a transverse plane. 
Multiple parallel images can be obtained by moving 
the transducer up or down in the esophagus. The 
transducer must be small enough to be easily 
tolerated by patients and will need to interface to the 
Varian electronics. Modifications to the existing 
reconstruction software will be made to permit use of 
the new images. Additional work is needed to 
capture the Varian data for image analysis and 
processing by the DeAnza system. 

Groups of normal volunteers and patients with 
hypertrophic cardiomyopathy or coronary artery 
disease will be studied using this new method of 
echocardiography. 

Publications and Abstracts: None 



Computer Based Analysis and Image Processing 
in Electron Microscopy and X-ray and Electron- 
Loss Spectroscopy 

This project involves collaboration of LAS and 
several NIH Institutes. It is directed toward the 
development of computer-based mathematical and 
statistical analyses, pattern recognition, and image 
processing of data, principally X-rays and electron 
energy loss spectra, derived from biological 
specimens studied in an analytical electron 
microscope. 

Background and Objectives: Microanalysis using 
electron and/or x-ray beams is a relatively new tool 
in biology which promises to reveal correlations 
between structure and function on sub-cellular and 
molecular levels. However, the extent to which this 
potential can be realized depends critically upon the 
use of computer methods, both in the acquisition of 
raw data and in the subsequent analysis of the data. 

The Microscopic Analysis Section/BEIB in 
collaboration with CSL/DCRT is constructing a 
facility where digital acquisition of raw data can be 
performed. In the meantime several NIH 
investigators currently obtain raw data at other sites. 
The data is brought back to the NIH campus for 
analysis. 

Of particular interest are the electron-energy loss 
spectra (inelastically scattered electrons) that 
contain information about the chemical composition 
of the specimen, the back-scattered electrons, and 
the elastically scattered electrons (both related to 
the mass density of various specimen regions). The 
physics of these various types of electrons, as well 
as the measured specimen current and secondary 
electrons emitted, needs to be elucidated in order to 
formulate the proper mathematical or statistical 
models that can combine this information into a 
'true' or corrected elemental map on a pixel-by-pixel 
basis. These models should adjust for contributions 
of neighboring pixels as well as a Poisson process in 
some cases. The relationship between elemental 
(energy-loss) peaks, zero-loss (transmission) peaks, 
and the plasmon peaks, as well as the background 
contribution, requires further exploration and 
quantification. 

The potential resolution for chemical analyses of 
specimens is in the range of 10-6 to 10-20 gms. and 
the spatial resolution is in the order of 100 to 1000 
square Angstroms. In addition, maps up to 1024 x 
1024 pixels can be constructed. The further 
development of this tool will require a new kind of 
image processing which will differ radically from the 
usual sort applied to nuclear medicine, x-ray devices, 



71 



and ultrasound. 

An array processor (CSPI MAP200) has been 
acquired by LAS. The speed of the array processor 
should make several procedures fast enough to 
become part of normal analyses. For example, a 
complex FFT (1024) can be accomplished in only 20 
milliseconds; a Gaussian curve fitting can be 
shortened from 8 hours to 10 minutes. Hence, many 
image processing algorithms, which can be 
developed offline on the DECsystem-10, can 
subsequently be implemented on the array processor 
and allowed to operate almost in realtime. 

Progress During FY81: In FY80 the array processor 
was tested extensively and was returned to the 
vendor due to faults detected in a number of printed 
circuits and in the power supply. In FY81 the array 
processor was sent back to NIH, interfaced with the 
PDP 1 160 in the BEIB Facility, and tested fully. 
Development of analysis programs for it is now 
proceeding as originally anticipated. 

The DeAnza image processing system has been 
enlarged by the acquisition of additional hardware (a 
1600 bpi tape drive and an RL02 disk drive) and by 
the further development of two major interactive 
software packages. (See project report on Nuclear 
Medicine.) 

The PICTUR package, particularly, has been used 
extensively for the investigation of the characteristics 
of dense bodies in digitally acquired electron 
micrographs of platelets. Assumptions of their 
geometry and composition based on their two- 
dimensional grey scale images have been 
investigated. A subsection of PICTUR, ALIGN, has 
been used to spatially align pairs of digitized images; 
for example, to produce composite elemental maps 
from paired pre-edge and fluorine K-edge filtered 
images. 

Significance: The ultimate biological goal is to relate 
structure and function at the subcellular and 
molecular levels. Certain active molecules (e.g., 
enzymes, neurotransmitters, hormones, antibodies, 
etc.) can be tagged with appropriate labels (e.g., 
fluorine) and then localized and quantified within 
cells by means of this tool. It should also be possible 
to determine the distribution of double bonds within 
membranes, microtubules, and cytoplasmic 
organelles. The distribution of elements of great 
biological importance (viz, calcium, magnesium, 
nitrogen, sulfur, and oxygen) also can be determined. 

Thus, the research potential of this tool has 
widespread applications in all areas of biology 
concerned with ultrastructure, much as the 
development of the imaging capability of the electron 
72 



microscope itself has provided important insights in 
almost every area of biology. 

Proposed Course: The study of the basic physics 
and the formulation of appropriate mathematical/ 
statistical models needed to achieve the analytical 
capabilities will require extensive work with 
phantoms, i.e., specimens of known composition that 
are very thin, prepared by such means as vacuum 
evaporation. There will need to be extensive studies 
of the signal/noise ratio in phantoms and in 
biological specimens. Potential problems with 
contamination and with specimen destruction by the 
high energy beam also need to be studied. 
Sophisticated algorithms for element recognition and 
location, image enhancement, etc., need to be 
designed and, where practicable, implemented on 
the array processor for rapid turnaround. 

LAS proposes to undertake some of these objectives 
in collaboration with participating wet laboratories. 
The DeAnza system is to be upgraded from a 
maximum image size of 256 x 256 to 512 x 480. 
Images acquired at Brookhaven are 512 x 512; 
hence this expansion plus the new magnetic tape 
drive will allow more rapid processing of the images 
obviating the need for data compression or 
partitioning. 

Publications and Abstracts: 

Douglas, M.A., Hui, S.W., Costa, J.L., and Bailey, J, J.: Computerized proc- 
essing and subtraction of energy-filtered electron images as an aid to 
elemental analysis. In Bailey, G.W. (Ed): Proceedings of the Thirty- 
Eighth Meeting EMSA. Baton Rouge, Claitor's Publisfiing Division, 
1980, pp 128-129. 



73 



Physical Sciences 
Laboratory 



George H. Weiss, Chief 



Summary of Activities 

Consulting Services. George H. Weiss (PSL); 
James E. Kiefer (PSL); J. Shapiro (CC); A. Pikus 
(CC); D. F. Dillon (Walter Reed); M. Brodsky (NIDA). 
A statistical analysis of data comparing auditory 
deficits in patients with osteogenesis imperfecta, 
their families, and normal volunteers has been 
completed. Several types of auditory abnormalities 
have been shown to characterize both patients and 
close family members. These results suggest 
audiologic measurements as a useful diagnostic tool. 
Analysis of posttraumatic epilepsy in head-injured 
Vietnam veterans has led to a simple formula 
relating the time to first occurrence of postinjury fits 
to a cluster of symptoms. The formula, suggested in 
two earlier analyses, will be valuable in estimating 
benefits due to victims of head injury. We have 
developed a mathematical formalism to describe 
data on the rate of entrance of drug addicts to 
treatment facilities. The simplest one of a class of 
models suffices to describe the rate data for a 
number of years with excellent accuracy. 

Theory of Biochemical Separation Techniques. 

George H. Weiss (PSL). This project applies 
mathematical methods to the interpretation of data 
from and the design of experiments involving 
ultracentrifugation, electrophoresis, and 
chromatography. Little, outside of planning for an 
experiment to determine optimal methods for 
determining molecular weight distributions, was done 
this year. On the completion of equipment by Marc 
Lewis (BEIB) the experiments will be started. 

Actin in Nonmuscle Cells-Biophysical and 
Biochemical Studies. Stephen L. Brenner (PSL). 
Studies on the mechanism of polymerization of actin 
have led to the discovery of monomer actin ATPase 
activity. Further kinetic studies are being pursued to 
verify a new species of monomeric actin. 

Theory and Application of Nuclear Magnetic 
Resonance Spectroscopy. James A. Ferretti (PSL); 
G. H. Weiss (PSL); J. E. Kiefer (PSL); R. J. Highet 
(NHLBI). Experiments utilizing two dimensional 
Fourier transform spectroscopy have been 



performed on a number of molecules. The technique 
allows an order of magnitude of greater accuracy 
than the comparable one-dimensional methods. An 
investigation of the accuracy of measuring chemical 
shifts has been completed and a number of 
improvements to present practice have been 
suggested. 

Correlation Function Spectroscopy/Laser Light 
Scattering. Ralph J. Nossal (PSL). Techniques are 
being developed for the application of laser Doppler 
techniques to measure blood flow in tissue 
microvasculature. The theory that has been worked 
out has been found in excellent agreement with 
experiments on artificial blood cells. A theory is 
presently under development for interpreting 
experiments on blood flow in muscle. 

Cell Motility and Chemotaxis. Ralph J. Nossal 
(PSL). Little has been done on this project in the 
past year. 

Theory and Measurement of Intermolecular 
Forces. V. Adrian Parsegian (PSL); M. Prouty (PSL); 
B. K. Lee (PSL); A. N. Schechter (NIADDK); R. P. 
Rand (Brock University). Measurements of 
intermolecular forces in proteins or nucleic acids 
using the osmotic stress methods developed in this 
Laboratory are now underway. Because sickle cell 
hemoglobin has been gelled, the investigators will 
undertake a systematic measurement of 
thermodynamic data on gelation and crystallization. 
B. K. Lee has written a number of programs to 
integrate crystallographic data on proteins into a 
project that will explore the way that proteins 
assemble into more complicated structures. 

Studies in Mathematics and Statistics. George H. 

Weiss (PSL); J. E. Kiefer (PSL). A review article on 
random walks in chemical physics was completed. A 
study of order statistics of diffusion processes is 
presently underway. 

Diffusion of Molecules on Cell Surfaces and Light 
Scattering from Fluids. Nahum Gershon (PSL); B. 
Aizenbud (MIT). The investigators have found that 

75 



:l RIMU {On Ml !••• thl( aMiiJ HULIH 



701 CT 00022-14 PSL 



Consulting Services 






PI: George H. Weiss. Chief. PSL 
James E. Kiefer. PSL, DCRT 


DCRT 




A. Plkus. M.Sc. CC. OPD: W. f. Caven 
Halter fieed; H. Srodsky. NIDA; G. Kno 
N.M.; E. Laska. Ph.D., Rockland State 


tt.'LSH.'oCflT, D. 


ed): David Dillon, H.D., 


Physical Sclwces Laboratory 


HCI.ON 


Division of Computer Research and Technology 


loUL-ANUMS. |«(ress,«u. |DmH. ^^ 


^'""~" 0.0 H... 


-" 


c) ^E.lHER 


Members of the PSL provide consu 

We have cwrcleted a study on the phan 
finding that it is effective at dimin 
facts. A considerable amount of work 


was spent on the 
imperfecta. An 


several areas of applied 

esV^ble streak art!-'' ''" 
study of audioloqic 
nalysis ot post traumatic 

as been developed. 

lable data. Several 
solved to yield asympto- 


epilepsv in head-iniured Vietnam vetc 


ans IS being com 


uler storage wer 


tic formulae. 



si^iriSeg's^irrji?:.^^* 


""'"SS'"""'"' 


ZOl CT 00011- 


4 PSL 


Oc'tobeT'Tr 1980 to September 30. 1981 


Theory of Biochemical Separation Techniques I 


PI; George H. Heiss, Chief. PSL. OCRT I 

i 


H. S. Lewis, Ph.D., BEI6, 1. G. Darvey, Ph.D., University of Sydney 


Physical Sciences Laboratory 




Division of ConDuter Research and Technology 


io.«.««um. (""'"'o^o' f'""" 0.0 


D.»« imronuu aai((s) 


This project explores the 
interpretation of experiments 


n biochemistry. Prepar 


lions to collec 


progress 

i 


data collection equtpment is 1 
Is possible on the conlenolate 


mm are being made by D 
stalled on Or, Le»is' u 
study. 



the effect of surface curvature does not greatly 
affect diffusion constants calculated using 
fluorescence photobleaching recovery methods. The 
light scattering spectrum from viscoelastic fluids was 
derived. 

Quantitative Analysis of the Electronmicroscopy 
of Cells and their Plasma Membrane. Nahum 
Gershon (PSL); P. Gorden (NIADDK); K. Porter (U. 
Colorado); L. Jarett (U. of Pennsylvania). Work is in 
progress on the analysis of hormone binding to their 
receptors on cell membranes using statistical 
analysis of digitized electronmicrographs. In 
particular, a study was completed on the spatial 
distribution of binding sites by cytochalasins B and 
D. 

Computerized Typesetting of Scientific Papers. 

V. Adrian Parsegian (PSL); N. Crawford (PSL); M. 
Douglas (LAS); M. Norton (LAS); M. McNeel (PSL); 
P. Miller (OD); J. Prewitt (OD); R. Fajman (CCB); J. 
Fajman (CCB). This project is intended to produce 
magnetic tape versions of material for publication, for 
direct typesetting. A program is presently being 
written to write tapes on the IBM System 370 using 
WYLBUR. The tape writing program is also being 
generalized to prepare manuscripts for several 
journals. 



1 



76 



Research Projects 

Consulting Services 

Members of the PSL provide consulting services in 
several areas of applied mathennatics and tfie 
phiysical sciences to researchers at NIH and 
elsewhere. We have completed a study on the 
phantom view method in computerized tomography, 
finding that it is effective at diminishing certain 
undesirable streak artifacts. A considerable amount 
of work was spent on the study of audiologic defects 
in patients with osteogenesis imperfecta. An analysis 
of post traumatic epilepsy in head-injured Vietnam 
veterans is being completed and a model that fits 
data very well on the rate of occurrence of fits has 
been developed. A mathematical model for the 
entrance of drug addicts into treatment facilities was 
devised and shown to give an excellent fit to 
available data. Several combinatorial problems 
related to computer storage were solved to yield 
asymptotic formulae. 

A study of the phantom view method for diminishing 
image artifacts due to interpolation in computerized 
tomography has been completed with Dr. Rodney 
Brooks and a paper has been submitted for 
publication. 

Study of the occurrence of posttraumatic epilepsy in 
a group of head-injured veterans has shown that the 
occurrence of epileptic fits in different injury 
categories follows a negative exponential distribution 
to a good approximation. The incidence varies with 
severity of injury but the average time to first fit 
appears to be independent of injury. Further work 
has gone into planning a second phase of the study 
of these veterans, in which as many as possible will 
be called in for an extensive examination. Several 
questions can then be answered that are presently 
obscured by a non-uniform follow up time. 

A study of audiologic defects in patients with 
osteogenesis imperfecta has shown that there is a 
marked excess of audiologic defects both in 
diseased patients and in otherwise unaffected close 
relatives. Further investigations are planned. Dr. 
Susan Hauser, CSL, DCRT, has coupled a digitizer 
to the Clinical Center otoadmittance meter allowing 
the collection of data with a previously unattainable 
accuracy. A study of the features of normal 
tympanograms will be undertaken with the new 
equipment. 

Together with M. Brodsky we have developed a 
model for the rate of entrance of drug addicts into 
treatment facilities. The model gives a good fit to the 
data and allows early prediction of the number of 



addicts in a given cohort to use these facilities. 

Several combinatorial problems were solved for Dr. 
Gary Knott, LSM, DCRT, which will be used in the 
analysis of computer data storage algorithms. 

Keyword Descriptors: Image reconstruction, 
interpolation, computerized tomography, audiologic 
defects, osteogenesis imperfecta, mathematical 
models. 

Publications: 

Meirowsky, A M,, Caveness, W. F.. Rish. B. L. Dillon, J. D. Mohr. J. P.. 

Kisller, J. P . and Weiss, G. H.: Cerebrospinal fluid complicating nnissile 

wounds ol the brain. J. Neurosurgery 54:44-47. 1981. 
Mohr, J, P., Weiss. G. H , Caveness, W. F., Dillon. J. D., Meirowsky. A. M., 

and Rish, B L : Language and motor deficits following penetrating head 

injunes in Vietnam. Neurology 30^273-1279. 1980. 
Rish, B, L . Dillon. J D , Caveness. W F . Mohr, J P , Kistler. J. P.. and 

Weiss. G. H. The evolution of craniotomy as a debndement technique 

for penetrating craniocerebral injuries. J. Neurosurgery 53:772-775, 

1980. 

Theory of Biochemical Separation Techniques 

This project explores the use of mathematical 
techniques applied to the interpretation of 
experiments in biochemistry. Preparations to collect 
data on fractionated serum albumin are being made 
by Dr. Marc Lewis. Until data collection equipment is 
installed on Dr. Lewis' ultracentrifuge no progress is 
possible on the contemplated study. 

Keyword Descriptors: Biochemical separation, 
ultracentrifugation, enzyme kinetics. 

Publications: 

Weiss. G H : Can one measure rate constants using chromatographic 
melhods'' Separation Sci. 16:75-80, 1981 

Weiss, G. H., and Darvey. I. G.: A note on the choice of substrate concen- 
tration in enzyme kinetics. J. Theoretical Biology (in press) 



77 



-"-«•"'-'--'" 


.Zi^HJii'f!!! 


ZOl CT 000,0-03 PSl 




Actin in Nonmuscle Cells - BtODhysical and Bioch 


mical Studies 


PI: Stephen L. Brenner, Research Chemis 
Other: E. 0. Korn. Chief. LCB, NHL8I 




Laboratory of Cell Biology. NHLBI 


lWiw.«h 






,!ffil!Mr — «psAsl:c ^L„, ^^ 




a,.,»™. 


actin. a major component of the motile apparatus 

effects of the drug cytochalasin, which accelera 
led to the postulafioJ of a oe. species of monom 


olynterization of the protein 
ot all eukaryotic cells. 

ed. Kinetic analysis of the 
es this ATPase activity, have 



^^■^^^Z'VS,,. ^^-^-^^^^^^^ 


ZOl CT 00025-06 PSL 


October 1, 1980 to Septerrtier 1, 1981 


Theory and Application of Nuclear Magnetic Resonance Spectroscopy 


PI: James A. Ferretti, Ph.D., Research Cheipist, PSL. DCRT 

G. R. Marshall, Ph.D., Professor of Physiology, Department of Physiology | 

and Biophysics, Washington University School of Medicine, i 

St. Louis. MO 
flollie J. Myers, Ph.D., Professor of Chemistry, Department of Chemistry, 

University Of California. Berkeley, CA 
Oaraes A. Kiefer, and G. H. Weiss, Ph.D.. Chief. PSL. DCRT 
R. J. Highet. Ph.D.. Laboratory of Chemistry, NBLBI 
Lance R. Pohl . Ph.D.; Terrence A. Marks. Ph.D.; and Jack S. Henson, Ph.O. 

Laboratory of Clinical Pharmacology. NHLBI 
William M. Egan. Ph.D.. Bureau of Biologies. FDA. Bethesda, MD 








Division of Convuter Research and Technology 


,L,ML.«,t«i. |Pi.cFEMiiiN.L, p^o.l 


^ -'"-'"-"•" a,.,™»,....s 0,0 .i™ 


The purpose of this project is to develop new me 


thods in nuclear magnetic 
:ility of existing techni- 


and conformational problems in small peptides and pro 
is the development of two-dimensional Fourier transfo 


blems in organic molecules 
rm NHR spectroscopy. 


Ihe importance of this technique is that it permits one to restrict one nuclear 
parameter to one dimension and a second parameter to the other dimension. 
A second part of the project Is to determine best data processing techniques 


for precise deteririi nation of chemical shifts. 



Actin in Nonmuscle Cells-Biophysical and 
Biochemical Studies 

Studies were continued on the mechanism of 
polymerization of the protein actin, a major 
component of the motile apparatus of all eukaryotic 
cells. 

A monomer actin ATPase activity was discovered. 
Kinetic analysis of the effects of the drug 
cytochalasin, which accelerates this ATPase activity, 
have led to the postulation of a new species of 
monomeric actin that may be an essential 
intermediate in actin polymerization. 

Actin is one of the major proteins of the cytoskeleton 
of all eukaryotic cells. As such, it is involved in many 
different motile processes and in the regulation of 
cell shape and cell organization. Actin is a globular 
molecule of molecular weight 42,000 that 
polymerizes into double helical filaments under ionic 
conditions similar to those of the cytoplasm of cells. 

It is in this polymerized form (microfilaments) that 
actin presumably functions in non-muscle cells just 
as it is the polymerized form of actin (the thin 
filaments) that functions in muscle contraction. In 
contrast to the situation in muscle, most of the actin 
in non-muscle cells is unpolymerized and its 
polymerization is spatially and temporally regulated 
so that the microfilaments occur in the cell when and 
where they are needed for specific motile events. 
The regulation of actin polymerization has been the 
focus of our research this year. 

1 . Identification of a new species of monomeric 
actin. We have shown in the past two years that the 
cytochalasins, a group of natural products that are 
potent inhibitors of a variety of motile processes in 
eukaryotic cells, have pronounced effects on actin 
polymerization in vitro. One of these effects is the 
uncoupling of actin ATPase activity from actin 
polymerization. Normally, when polymerization 
occurs, a monomer actin bound ATP is hydrolyzed to 
polymer bound ADP. Cytochalasins greatly increase 
the rate of ATP hydrolysis even while inhibiting actin 
polymerization. We have now shown that this effect 
is due to the acceleration of a monomeric actin 
ATPase cycle, which can exist independent of the 
ATPase associated with monomer addition to 
polymer. Kinetic analysis of the effects of four 
cytochalasins on actin from two sources (rabbit 
muscle and Acanthamoeba, a soil amoeba) requires 
the existence of an until now unknown species of 
actin monomer. Formation of this intermediate is the 
rate determining step in the monomer ATPase cycle 
when cytochalasins are present. In the absence of 
cytochalasins, disappearance of this intermediate is 



78 



rate limiting. The coupling of the monomer ATPase 
cycle to the polymerization of actin, and the possible 
role of the newly discovered actin species as an 
obligatory intermediate in actin polymerization, are 
under investigation. 

2. Equilibrium Studies of Actin Polymerization. It is 
now clear that the polymerization of actin ATP is 
extremely complex (see above). Actin with bound 
ADP will also polymerize. The result is an equilibrium 
system of actin filaments (containing bound ADP) 
and monomeric ADP actin. The ADP actin system 
may provide a useful tool for understanding actin 
polymerization since it uncouples the polymerization 
process from the hydrolysis of nucleotide in an 
equilibrium, rather than a steady-state system. We 
have begun a detailed study of ADP actin 
polymerization using the techniques of 
ultracentrifugation, high-shear viscometry, and light 
scattering. 

Because ATP hydrolysis is not obligatory for actin to 
polymerize, and because cellular energy stores are 
depleted when ATP is used, we continue to search 
for a role for nucleotide hydrolysis in actin 
polymerization. It is possible that it is at the level of 
the monomeric actin, and not the hydrolysis 
associated with polymerization per se, at which 
regulation is affected. Efforts will be made to 
examine this idea through the use of 
nonhydrolyzable ATP analogs, cytochalasins, and 
inhibitors or promoters of actin ATPase activity as 
well as purified proteins known to have direct effects 
on the state of actin assembly. 

Keyword Descriptors: polymerization, actin, 
cytochalasin, cytoskeleton, eukaryotic. 

Publications: 

Brenner. S- L-, and Korn. E. D: Stimulation of actin ATPase activity by 
cytochalasins provides evidence for a new species ol monomenc actin. 
J. Biol. Chem. (in press). 



Theory and Application of Nuclear Magnetic 
Resonance Spectroscopy 

The purpose of this project is to develop new 
methods in nuclear magnetic resonance 
spectroscopy, to extend and determine the utility of 
existing techniques, and to apply these techniques to 
structural problems in organic molecules and 
conformational problems in small peptides and 
proteins. Of current interest is the development of 
two-dimensional Fourier transform NMR 
spectroscopy. The importance of this technique is 
that it permits one to restrict one nuclear parameter 
to one dimension and a second parameter to the 
other dimension. A second part of the project is to 
determine best data processing techniques for 
precise determination of chemical shifts. 

Since the acquisition of the NT-360 NMR 
spectrometer 18 months ago, we have been carrying 
out various experiments on complex organic and 
biological molecules, which were heretofore 
impossible. The major emphasis in this project has 
been the application of multiple pulse and two- 
dimensional Fourier transform NMR spectroscopy. 
The principal idea behind multiple pulse experiments 
is to map out the behavior of the nuclear 
magnetization during an evolution period as well as 
during a detection period after the last radio 
frequency pulse. Because this permits many choices 
for the second frequency axis, the advantages of 
two-dimensional spectroscopy are numerous and 
include: correlating frequencies of pairs of nuclei to 
simplify assignments, improving spectral resolution, 
measuring cross relaxation processes for distance 
determinations, and studying multiple quantum 
coherence phenomena. 

In one study, we have determined the structures of 
two stereoisomeric metabolites isolated by HPLC 
from the bile of rats treated intraperitoneally with 
bromobenzene. The 2D J spectra of both 
metabolites permitted complete resolution of all the 
transitions and allowed us to carry out the 
assignments of lines. These results demonstrated 
that the metabolites were formed by the opening of 
aryl epoxide intermediates by glutathione. This study 
also enabled us to demonstrate the feasibility of 
studying spectra in the presence of a strong water 
peak. 

We are applying a three-pulse experiment followed 
by a two-dimensional Fourier transformation to 
investigate cross relaxation pathways in rigid organic 
molecules. Cross relaxation analysis allows the 
estimation of internuclear distances, yields rotation 
correlation times, and produces information on 
molecular complexes. We have carried out such 



79 



S!SS!"lS.I."RJ'i"SI'KS SS?,T' 


""™Si'l'""'' 

IHIRUUIU. RMtMCH MOJt 


ZOl CT 00021-10 PSL 




Correlation Functiwi Spectroscopy/Laser Light Sc 


ttering 


PI: R. J. Hossal, Ph.D.. Research Physicis 


■ "'•• °"^ 


J. Gladner, Ph.D.. LBC. NIAODK 
fl. Bopner. Ph.D.. BEIB, DPS 














0{.)«iii.i. 


E.per™enta, and the.ret.ca 


1 Studies have been 


ioloqical qeJs. viscoelastic 


p.l,»er s.1„?io»s. a.P silla 


s also are being pertonned 


blood flp» in tissue r.icro«asculatore. 






ZOI CT 00017-09 PSL 



Cell Motility and C 



I Septenter 30. 1981 



, J. Nossal, Research Physicist, PSL. DCRT 



DW<sif»n of Conputer Research and Technoloi 



>loped. Mathematical i 



studies on veratraldehyde and a series of Enkephalin 
analogs. Specific interactions are demonstrated in 
the Enkephalen derivatives. We are attempting to 
quantitate the results on veratraldehyde in order to 
evaluate the competitive inter- and intramolecular 
dipolar contributions to the nuclear Overhauser 
enhancement factor. 

We are continuing to study molecular motions in 
oligopeptides using spin-lattice relaxation times, spin- 
spin relaxation times, and NOE factors as vi^ell as 
using the two-dimensional techniques for structural 
investigations on these systems. We are attempting 
a complete assignment of the proton spectrum of 
Bleomycin. We have carried out a complete 
temperature and field dependence study on 
Bradykynin and evaluated the motional parameters. 
We have also carried out a study on liver alcohol 
dehydrogenase, by investigating the cadmium-113 
chemical shifts where the cadmium-113 replaced 
Zinc at both the catalytic and non-catalytic sites. We 
are attempting to learn more about the nature of the 
metal ion coordination at the catalytic site. 

We have investigated, by a simulation study, the 
error in measured chemical shifts when various 
strategies are used for filtering and smoothing the 
data and for estimating peak position. We have 
found that curve-fitting the data in a neighborhood of 
the observed maximum is always considerably more 
accurate than simply choosing the position of the 
peak maximum as the true maximum. Further, the 
use of a matched filter leads to a considerable 
improvement over the use of no filter, and the 
improvement is not sensitive to the choice of filter 
constant. 

Keyword Descriptors: Nuclear Magnetic Resonance, 
two-dimensional, Fourier transform, bile metabolite, 
liner alcohol dehydrogenase. Enkephalin. 

Publications: 

Egan. W., Ferretti, J. A., and Marshall, G. R.: Relaxation parameters and 
motional properties in biological macromolecules. Bui/. Magnet. Reso- 
nance 2■.^5-^^ , 1981. 

Ferretti, James A., Highet, Robert J., Pohl, Lance R., Marks, Terrence R., 
and Hinson, Jack A.: Application of 2D J-resolved spectroscopy in the 
structural investigation of bile metabolites. Abstract, 22nd Experimental 
NMR Conference, 1981. 

Gupta, R. K., Ferretti, J. A., Becker, E. D., and Weiss, G. H.: A modified fast 
inversion recovery technique for spin-lattice relaxation measurements. 
J. Magnet. Resonance 38:447-452, 1981. 

Weiss, G. H., Ferretti, J. A., and Kiefer, J. E.: A study of precision in the 
measurement of chemical shifts. J. Magnet Resonance (in press). 



80 



Correlation Function Spectroscopy/Laser Light 
Scattering 

Experimental and theoretical studies have been 
performed to develop laser inelastic light scattering 
methods for studying biological gels, viscoelastic 
polymer solutions, and similar materials. Studies also 
are being performed in order to understand how 
laser Doppler techniques can be used to measure 
blood flow in tissue microvasculature. 

Quasielastic light scattering has been used to study 
dynamical properties of various biological substances 
and model systems. Emphasis has been on 
developing novel applications of this relatively new 
technology and on devising physical and 
mathematical theories in support of experimental 
protocols. Current work continues on schemes for 1) 
non-invasive surveillance of blood flow and 2) non- 
perturbative probes of the mechanical properties of 
soft polymer gels similar to those found in biological 
cells and within tissue interstitia. 

A collaboration with R. Bonner (BEIB, DRS) and P. 
Bowen (NHLBI) has resulted in publication of a 
theory which relates quasielastic light scattering 
measurements to blood flow in tissue 
microvasculature. This theory accounts for diffuse 
scattering by immobile tissue constituents and for 
multiple interactions of photons with moving blood 
cells. The theory has been tested with an 
experimental analog in which particles flow through a 
fiber capillary bundle imbedded in a composite gel, 
the latter containing polymer microspheres which 
impart optical properties characteristic of those of 
biological tissue. Dr. Bonner and Mr. Bowen have 
been using their laser Doppler flowmeter in various 
clinical studies and we have continued to collaborate 
on related theoretical and experimental problems. 
Attempts currently are being made to provide a 
theory for interpreting measurements of tissue blood 
flow in moving heart muscle. 

Efforts to apply quasielastic light scattering to 
studies of dilute polymer networks recently have 
involved improvements in instrumentation and 
development of methodology for measuring lattice 
damping parameters ('internal viscosity'). An 
apparatus has been constructed to facilitate torsional 
excitation of a sample by an external mechanical 
field. Also, a fiber optics version of the spectrometer 
has been developed which promises to be 
particularly useful for studying small specimens. Data 
have been acquired to characterize the dependence 
of internal dissipation on such variables as polymer 
concentration, crosslink density, and solvent 
viscosity; also, we have begun work on a theory to 
relate those measurements to microscopic physical 



properties of a gel. Our collaboration with Dr. J. 
Gladner (NIADDK), on studies of the biophysical 
chemistry of polymer networks which arise during 
blood coagulation, has been reinitiated. 

Keyword Descriptors: Laser light scattering 
macromolecules, diffusion coefficients, correlation 
functions, gels, blood flow, Doppler flowmeter. 

Publications: 

Bonner. R . and Nossal. R : A model (or laser Doppler measurements o( 
blood (low in tissues Applied Optics 20:2097-2108. 1981 

Nossal, R : Quasielastic light scaltenng from polymer gels. In Chen. S H . 
Chu, B, and Nossal. R. (Eds.); Scallenng Techniques Applied to Supra- 
molecular and Nonequilibnum Systems. New York. Plenum Publ. Corp., 
(in press), 

Nossal, R . and Jolly. M Shear waves m cylindrical gels J AppI Phys (in 
press) 

Cell Motility and Chemotaxis 

This project has been undertaken to study vanous 
aspects of cell locomotion, including the 
mathematical basis of macroscopic assays for 
leukocyte chemotaxis. New procedures for 
measuring parameters of cell migration, including 
computer assisted tracking techniques, are being 
developed. Mathematical relationships between 
microscopic cell motion and macroscopic response 
are derived. 

This study relates to cell locomotion and chemotaxis. 
Recent emphasis has been on examining certain 
immunologic aspects of leukocyte migration and on 
constructing physical models of signal transduction 
occurring at the surfaces of chemoresponsive cells. 

During the past year this project has been carried on 
with reduced effort while laboratory personnel have 
been otherwise occupied (cf. project Z01-CT00021- 
10 PSL). Literature searches have been performed in 
order to identify in vitro models of polymerizing 
cytoskeletal structures which might be amenable to 
study by recently developed quasielastic laser light 
scattering techniques. Attention also has been given 
to assimilating information on physical theories 
concerning detection of chemotoxins by receptors at 
the surfaces of motile cells. A previously prepared 
review of mathematical theories of chemotactic 
responses was published. 

Keyword Descriptors: Cell locomotion, leukocyte 
chemotaxis, capillary migration assays. 

Publications: 

Nossal, R.: Mathematical theones of topotaxiS- In Jager. W,. Rost. H., and 
Taulu, P (Eds): Biological Growth and Spread Mathematical Theories 
and Applications Heidelberg. Spnnger-Verlag. 1980. pp 410-440. 



81 






ZOl CT 00026-06 ( 



October 1. IMP to September 



in. Ph.O.. PSL. DCRT 

.0.. PSL. DCRT 

1.0.. Brock Universitj 



. N. Schechler, H.D., LCB. NIflDDK 



Physical Sciences laboratory 



Division of Cotrputer Research ; 




*ojEcrilSatfi'B"iioir'S!If*'hhI l^^" 


ilfilss^ffiiicT 


201 CT 00024-06 PSL 


October 1, 1980 to September 30. 1981 


Studies in Mathematics and Statistics 


PI: Georqe H, Weiss, Chief. PSL. DCRT 
Other: James E. Kiefer. PSL. OCHT 


Ph.D., Univ. of Ca1 ifnrnia-San Diego; K. Lindenberq. Ph.D., Univ. of California- 
San Dieqo; J. Rice. Ph.D., Univ. of Ca1ifornia-San Oiego; M. F. Shlesinger, 
Ph.D.. Univ. of Maryland. College Park. 


Physical Sciences Laboratory 


lECIlC* 


Oivtsion of Computer Research and Technology 


'o.a 1 ' ols \' 0.3 


:j(.J «u««;wBJtctS a i») HUKU. nssut! □(O-eiiwh 


A co<nirehens(ye reylew 
logles tn chemical physics i 


rticle on the applications of random walk methodo- 
in the orocess of completion. A study of. some 




oroblems related to the 


matching of DNA sequences 



Theory and Measurement of Intermolecular 
Forces 

Our capacity to determine intermolecular forces in 
aqueous or physiological milieux now covers proteins 
and nucleic acids as well as phospholipid bilayer 
membranes. There is good evidence that hydration 
forces dominate interactions near contact. These 
forces are due to the disturbance of water by 
charges on the exterior of large molecules or 
aggregates. 

Exponentially decaying hydration forces, first 
characterized quantitatively between bilayer 
membranes, are typically repulsive. They depend on 
the chemical identity and packing of water polar 
soluble groups on the membrane surface. They 
vanish when an opposing body sticks more strongly 
to these water soluble groups than does water. 

Measurements of intermolecular forces in proteins or 
nucleic acid are now underway using osmotic stress 
methods developed for membranes. Evidence is 
strong for the action of hydration forces between 
parallel DNA double helices. 

Programs written on the DCRT molecular display 
system are being developed to examine contacts 
between proteins in crystals and between nucleic 
acids in condensed arrays. 

Our activities this year have been to examine the 
consequences on membrane processes of earlier 
measurements of forces between bilayer membranes 
and to develop several methods for identifying and 
measuring forces experienced by proteins and 
nucleic acids. 

Beginning from the systematic determination of 
electrostatic, dispersion, and hydration forces 
between a large range of phospholipid bilayer 
membranes, we have been able to correlate the 
intermembrane forces with the ability or inability of 
vesicular membranes to fuse. Exponentially growing 
hydration repulsive forces dominate below 30 
angstroms separation. Their strength varies with the 
density and identity of water soluble groups on the 
membrane surface. This repulsive force collapses 
upon addition of substances, such as calcium, with 
phosphatidylserine membranes that bind more 
strongly than does water to these water soluble 
groups. 

Electrostatic interactions, seen beyond 20 to 30 
angstroms membrane separation, almost follow the 
behavior expected from double layer theory. A 
systematic deviation from that theory gives evidence 
for perturbation of water structure around ions in 
solution. 



82 



Methods of measurement analogous to those used 
to determine membrane-membrane forces are now 
successfully being applied to the aggregation, 
gelation and crystallization of proteins and to the 
packing of parallel double helical strands of DNA. 
Sickle cell hemoglobin has been successfully gelled 
(Prouty, Schechter) under conditions where one can 
measure the work of removal of water from the 
protein and water mixture. Systematic collection of 
thermodynamic data on gelation and crystallization 
now appears feasible. Similarly measurements on 
the repulsive force versus separation curve for 
parallel DNA helices are now in progress. 

These molecular studies are complemented by a 
major study of protein and nucleic acid contact using 
the DCRT molecular graphics facility. To this end a 
computer program has now been written (Lee) to 
construct protein crystal structures by combining 
atomic coordinates with the crystal symmetry. As in 
previous studies of contacts stabilizing protein 
dimers and tetramers, the stable association of 
peptides involves close approach and tight fits. We 
expect that the lessons learned from these studies 
will be helpful in work anticipated on the self 
assembly of elaborate protein structures. 

Keyword Descriptors: phospholipid bilayer 
membranes, hydration forces, electrostatic forces, 
dispersion forces, ionic solutions, protein gelation, 
nucleic acids. 

Publications: 

Lis, L. J.. Lis. W T.. Parsegian. V A., and Rand. R P.: Adsorption of 
divalent cations to a variety of phosphatidylcholine bilayers. Biochemis- 
try 20^77^■^777. 1981 

Lis, L, J,. McAlister, M,. Fuller, N.. Rand, R. P., and Parsegian. V A.; 
Interactions between neutral phospholipid bilayer membranes. Biophys 
J. (in press) 

Lis, L. J.. McAlister, M . Fuller. N,. Rand, R, P, and Parsegian. V.A.: 
Measurement ol the lateral compressibility ol several phospholipid bi- 
layers. Biophys J (in press). 

Lis. L J,. Parsegian, V A , and Rand. R P.: Binding ot divalent cations to 
dipalmitoylphosphatidylcholine bilayers and its effect on bilayer interac- 
tion, S/oc/7ems/ry 20:1761-1770. 1981 

Parsegian, V. A,: Forces between membranes approaching contact. Scandi- 
navian J. ol Clinical Investigation A^ MS. 39-47,1981. 

Parsegian. V, A,. Rand. R. P . and Slamatoff. J : Perturbation ot membrane 
structure by uranyl acetate labeling. Biophys- J. 33:475-478, 1981 

Parsegian, V. A., and Weiss. G H.: Spectroscopic parameters for computa- 
tion of van der Waals forces. J, Colloid interlace Sci. 81:285-289. 1981. 



Studies in Mathematics and Statistics 

A comprehensive review article on the applications 
of random walk methodologies in chemical physics is 
in the process of completion. A study of some order 
statistics related to random walks and diffusion 
processes has been completed. Several statistical 
problems related to the matching of DNA sequences 
are under investigation. 

A review article on random walks in chemical 
physics is being completed and will be submitted to 
Advances in Ciiemical P/iysics. In addition we have 
applied some rigorous results on the number of 
distinct sites visited in a finite random walk to the 
study of materials with trapping sites. Together with 
R. J. Rubin we have developed statistical results on 
the probability of visiting a set of points by a lattice 
random walk and with M. F. Shiesinger we have 
studied properties of the expected number of distinct 
sites visited during an n-step lattice random walk. 
We have studied order statistics of diffusing 
particles. Together with K. E. Shuler we are 
analyzing combinatorial problems related to the 
matching of DNA sequences. These problems are 
currently studied by simulation techniques, leading to 
results of questionable generality, but some analytic 
progress is possible and some results on matching 
probabilities have been obtained. 

Keyword Descriptors: Random walks, diffusion 
processes, trapping sites, first passage time 
problems, sequence matching. 

Publications: 

Kiefer, J E., and Weiss, G H,: A companson ol two methods lor accelerat- 
ing the convergence of Fourier series Computers and Mathematics (in 

press). 
Lindenberg. K., Seshadn, V E., Shuler. K E,, and Weiss. G. H.: Lattice 

random walks for sets of random walkers J. Statistical Physics 23:11- 

25, 1980 
Oppenheim. I . Shuler. K E . and Weiss. G H.: Stochastic processes. In 

Lerner. R G . and Trigg. G L (Eds ): Encyclopedia of Physics. New 

York, Addison-Wesley. 1980. pp. 964-967 
Weiss. G. H : Asymptotic form for random walk survival probabilities on 3-D 

lattices with traps Proc Natl Acad. Sa. USA 77:1273-1274, 1980. 
Weiss. G H First passage times for one dimensional random walks. J 

Statistical Physics 24:581-589. 1981 
Weiss. G H Note on lattice random walks with an excluded point. J. 

Mathematical Physics 22 562-563, 1 981 . 
Weiss. G, H., and Shiesinger. M, F : On the expected number of distinct 

points in a subset visited by an N-slep random walk. J. Statistical 

Physics (in press). 



83 



sssris. 


:ipa IWMMIIW tlCH^EE 


.uu'\"oX»l.?..ci. 


«,o.Ec. H^m 








ZOl CT OOOai-03 PSL 


'"«"»;'" 


. 1980 to Septerrber 30. 1981 






QMntita 
Hsabnne 


l,e An.l,s s of the ElectronmKroscoor of Cells and their Plas«a 




"f" ^ ^ ,„,s'eiTpr"rr'""°'"'" 


Other: 


n •; hool of Medicine, Philadelphia, 




P to NIADDK 




r' ''"'' Sv't^^'-l^nd ^ 




' P C Ph n . s !y Of Colorado. Bo.lder. Colorado i 




Ht ssberaer Ph PSL DCRT 


Physical 


Sciences Laboratory 






of Computer Research and Technolociy 






1.8 1 0.7 1 0.1 i 




,»( BO«[£S) 


.(.) ««•» 


WB«rS aWHWU-KSSUtS DIOMIIHEB 


Gl.D "..-.ufi 


(.J) iNitR.^f.:, 






af r.l". 




Ss^rmernbra^irS^e^OTfd^q°mon5meric^"^ | 
















etJ on rat adipocyte eel! surfaces while thev are uniformly distributed 1 


groups a 


Se's^cr/parVjr:. 






'es and coalef pits on 


ad'poc?te-likI fibroblasts. 


of oo'rw.lr1zan™ JeliLr' ""^cSleclilS' .n'ceik 




has heen pursued usinq 














of mcrtr 


'ins^drcells°anrits 


qanizalion and oolvmeruation of cytoskeletal 


relation to transmembrane siqnalling. 



I Septeniber 30. 1981 



Diffusion of Molecules on Cell Surfaces and Light Scattering from F 



T!!!!S 



At-^or 


etical study was....- 


,...,, 


photob leach 


nq recovery, f^,, 






mentranes. u «.i. 
not affect the _. :. 




In the 


second part hydrodyriamic 
astic fluids were derive 


equations 


viscoelastl 


fluids were studied. 





Quantitative Analysis of the Eiectronmicroscopy 
of cells and their plasma membranes 

In this study, the distributions of insulin binding sites 
on the surface of rat adipocytes and liver plasma 
membranes were studied using monomeric ferritin- 
insulin viewed by eiectronmicroscopy and analyzed 
by newly developed quantitative methods. It shows 
that insulin binding sites are non-randomly distributed 
on rat adipocyte cell surfaces while they are 
uniformly distributed on liver plasma membranes. 
The quantitative characteristics of the receptor 
groups and the mode of action of cytochalasin B and 
D have been elucidated. 

In the second part we have been looking at the 
organization of intramembraneous particles on 
lymphocytes and coated pits on adipocyte-like 
fibroblasts. 

In the third part of this investigation, the three 
dimensional structure of polymerization centers for 
macromolecules in cells has been pursued using 
three-dimensional reconstruction and image analysis 
techniques. 

The fourth part is an experimental study of the 
effects of aggregation of membrane proteins on the 
organization and polymerization of cytoskeletal 
elements inside cells and its relation to 
transmembrane signalling. 

We have developed quantitative methods to analyze 
electronmicrographs of biological systems. The 
methods include digitization of micrographs and 
computational analysis of their contents (e.g., protein 
particles on membranes). 

The binding of hormones to their receptors on cell 
membranes is believed to be the first step in their 
biological action. Insulin binding sites appear in 
groups on the plasma membranes of rat adipocytes 
and as separated sites on rat liver plasma 
membranes. These configurations are seen on 
electronmicrographs of these membranes using 
insulin bound to monomeric ferritin particles. The 
present study used computer analysis of the spatial 
distribution of these insulin binding sites and of the 
effects on this distribution by the disrupting agent 
cytochalasin B and also by cytochalasin D. This 
study points out that the distance between adjacent 
sites in a group does not seem to exceed 400 
angstroms. Also, the relative change by Cytochalasin 
B in distribution of this insulin preparation in groups 
of 1 , 2, and 5 or greater was not affected by the 
defined separation distance (at least up to 600 
angstroms) used to define a group. Cytochalasin B 
appears to disrupt portions of groups of insulin 
without regard to distances between adjacent 



I 



84 



receptor sites within a group. Computational analysis 
of nnorphological studies on insulin and other 
receptors should provide more information than 
visual analysis. 

To further the understanding of insulin action, 
experiments are done on fibroblasts transformed into 
adipocyte-like cells. These cells look and respond to 
insulin as do adipocyte cells. Coated pits were 
inferred to participate in the mechanism of insulin 
action in some cells. We would like to study the 
reorganization of coated pits on membranes of ceils 
undergoing this type of transformation and examine 
what the changes are in spatial distributions on the 
membranes during this kind of differentiation. This is 
part of a broader study of the membrane's structure 
in the undifferentiated and in adipocyte phenotype of 
the 3T301 cells. 

A three-dimensional arrangement of some cell 
organelles is important to cell function. At the 
present time electronmicroscopy is the only method 
available to visualize microfilaments, microtubules, 
and cell membranes at sufficient resolution to study 
their arrangement and interrelationships. Using 
electron micrographs produced by the high voltage 
electron microscope (a national resource sponsored 
by NIH) in Boulder, Colorado; the new digitizing 
system being constructed at PSL; and the 
computational facilities at DCRT; the three- 
dimensional location of organelles in cells is being 
measured and their structure is going to be 
elucidated. This will shed light on what determines 
and maintains cell form in normal and malignant 
cells. This work is done with R. Nossal, K. Porter, B. 
Bowers and M. Weissberger. The method has been 
developed and the analysis of the data has been 
initiated. 

Membrane proteins can interact with various 
components inside the cell, e.g., cytoskeletal 
elements. We have looked for possible physical 
mechanisms to account for the attachment of 
membrane proteins to cytoskeletal filaments, e.g., by 
entanglement or by polymerization of cytoskeletal 
elements around aggregated membrane proteins. 

There are some indications, on the visible light 
microscopic level, that these cytoskeletal structures 
are affected by changes in the location of cell 
surface proteins. We have developed an electron 
microscopic method that makes it possible to 
observe these structures inside the cell. In order to 
enhance the visibility of the cytoskeletal 
microfilaments we use the Evans & Sutherland 
facilities for image display and analysis. With this 
method we plan to quantitate the amount and 
morphology of cytoskeletal elements. In doing so. 



changes in the cytoskeleton will be detected and 
correlated with changes in cell surface proteins. 
These studies might shed light on how signals are 
transferred through membranes to cell interiors. 

Keyword Descriptors: Insulin binding sites, insulin 
receptors, cytochalasin B, membrane proteins, 
electronmicroscopy, coated pits, high voltage 
electronmicroscopy, digitization of 
electronmicrographs, cell surfaces, microfilaments, 
cytoskeleton. 

Publications: 

Gershon. N D . Smith, R M . and Jaretl, L : Computer assisted analysis o( 
fernlininsulin receptor sites on adipocytes and the effect of cytochala- 
sin B on groups of insulin receptor sites. J. Membr. Biol 58:155-160, 
1981 

Diffusion of Molecules on Cell Surfaces and Light 
Scattering from Fluids 

A theoretical study was carried out to determine the 
rate of cell surface curvature on the observed 
diffusion rate of membrane proteins using 
fluorescence photobleaching recovery. For with a 
similar geometry to natural microvillous membranes, 
it was found that the assumption that the membrane 
is curved does not affect the calculated diffusion 
constant to a large extent. 

In the second part, hydrodynamic equations and the 
light scattering spectrum from viscoelastic fluids 
were derived. Two cases, fluid-like and solid-like 
viscoelastic fluids, were studied. 

A. Diffusion of Molecules on Cell Surfaces. 
Fluorescence photobleaching recovery (FPR) 
techniques have been used to study lateral 
movement of molecules on membranes. Fluorescent 
molecules within a circular spot are bleached and 
the time dependence of the fluorescence recovery in 
the spot is measured. The physical interpretation of 
these results in terms of diffusion along the 
membranes is based on the assumption that the 
surface is planar. However, biological membranes 
may be nonplanar, e.g., they may have blebs and 
microvilli. To study the effect of nonplananty on the 
diffusion rate, the diffusion equation for motion on 
curved surfaces was derived. This equation was 
employed in studying the diffusion along a 'wavy' 
surface of the form Acos(kx)cos(ky). The numerical 
calculations show that for k = 10 (micrometers)-l and 
a bleached spot of 1 micrometers in diameter, the 
time dependence of the intensity of fluorescence in 
the bleached spot depends on A at low values of A 
(0, .1, and .5 micrometers) while the dependence on 
A at higher values of A (.5 compared with 1 and 2 
micrometers) is weak. Assuming that the membrane 
is planar, the interpretation of FPR measurements 



85 






riled Typesetting of Sct« 



ZOl CT 00066-OZ PSL 



lalyst, LAS. OCRT 



lioohysical Journal, Goverr 



1 nf Coirputpr Rpsparch and TechnolDgs 



The object of this project is to be able to produce magneti 
laterial intended for publication. Those tapes can be "hung" 
copy editing/typesetting computer systems of the publisher. 



ii writing progrd 



way of retyping maten 



in page charges I 



might yield a diffusion constant that is 40 percent 
(for A=1) of the real diffusion coefficient along the 
curved surface. These results suggest that the 
transition fronn a plane to a surface with small 
microvilli slows the diffusion process while the 
transition from small to large microvilli practically 
does not affect the diffusion rate of molecules in the 
bleached area. 

B. Light Scattering from Microelastic Fluids. 
Viscoelastic systems, as all other macroscopic 
systems, can be studied by phenomenological and 
statistical approaches. The common 
phenomenological approach consists of three 
stages: 

1 . The finding of the stress-deformative structure 
of the system 

2. The construction of differential equations of 
motion. [This construction usually involves some 
phenomenological (transport) coefficients.] 

3. The carrying out of some, usually mechanical, 
experiments estimating values for these 
coefficients. 

However, this scheme may not always be applied. 
The problem is that there is a wide class of fluid 
systems, namely, overcooled liquids, which do not 
permit (because of their instability) usual rheological 
experiments in order to define their structure. For 
such systems, nonmechanical experiments became 
very important. 

In this work we continued the phenomenological 
study of the two simplest viscoelastic systems: 
Kelvin body (the solid-like system) and Maxwell body 
(the fluid-like system). We generalized and corrected 
the dynamical equations for these systems and 
calculated the vertical-horizontal (VH) spectrum of 
light scattered by these systems. We found that the 
form of these spectra can sometimes uniquely define 
the structure of the system. 

For the solid-like system we have derived the 
hydrodynamic equations and the VH light scattering 
spectrum using a molecular approach. 

Keyword Descriptors: Membrane proteins mobility, 
lateral diffusion, microvilli, fluorescence 
photobleaching recovery, light scattering, 
hydrodynamic equations, viscoelastic fluids. 

Publications: 

Aizenbud, B. M.. and Gershon. N.D.: Hydrodynamic equations and VH light 
scattering from viscoelastic (solid-like and fluid-like) systems. Pfieno- 
menological approach. Physica A (in press). 



86 



Computerized Typesetting of Scientific Papers 

The object of this project is to be able to produce 
magnetic tape versions of material intended for 
publication. Those tapes can be 'hung' directly on 
the copy editing/typesetting computer systems of 
the publisher. 

Our method is to write programs using WYLBUR files 
as source material that will produce magnetic tapes 
to the specifications of each publisher. The 
execution of these tape writing programs should be 
as simple as requesting a paper copy of the same 
text. 

Such electronic conversion of texts has been shown 
to be cheaper, faster, and more accurate than the 
old way of retyping material by the publisher. 
Typesetting costs can be halved. Already one journal 
is offering a major discount in page charges to 
authors submitting 'compuscripts.' Others should 
follow. The ultimate savings to NIH are expected to 
be significant. 

It has been shown that the texts of many if not most 
of the scientific papers produced on magnetic 
memory typewriters ('word processors') at NIH can 
be transferred into the WYLBUR system. Many 
papers are also keyed directly into WYLBUR. A 
program by Bonnie Douglas and Martha Norton is 
being used and improved by Nancy Crawford to 
produce magnetic tape records of such texts. These 
tapes are then suitable for text conversion 
electronically to typeset galleys without retyping 
manuscripts. 

Our early success in producing a paper (by John 
Fletcher, LAS/DCRT) in this way has lead to an 
agreement with the Biophysical Society, its publisher 
Rockefeller University Press, and its printer Science 
Press, to supply us with magnetic discs, cards or 
tapes of papers for experimentation. The object here 
is to obtain material in order to learn to transfer 
various forms of magnetic record into the WYLBUR 
system and then to produce typesetter-ready 
magnetic tapes from WYLBUR. 

The original tape writing program is being transferred 
by M. McNeel from the PDP-10 system to the IBM 
370 system to allow direct writing of tapes using 
WYLBUR command procedures. Features of the 
new WYLBUR system are being incorporated to 
allow far more efficient tape writing. The tape writing 
program is being generalized to prepare material for 
other publishers. Initial experiments have been 
performed by Waverly Press under the auspices of 
the Journal of Biological Chemistry and a new 
collaboration is being established with the Computer 
Society. 



With the automatic conversion of texts that would 
otherwise be laboriously retyped and with the 
likelihood that publishers' copy editing can now be 
done at a terminal rather than on paper copy, the 
new system provides speed, accuracy and cost 
savings. Consequent reduction in page charges is 
already a fact with the Biophysical Journal. Other 
journals appear ready to grant similar reductions 
when similar procedures are available. 

Keyword Descriptors: computer typesetting, 
WYLBUR, computerized composition, magnetic tape, 
floppy disc, magnetic card, compuscripts. 

Publications: None. 



87 



Data Management 
Branch 



J. Emmett Ward, Chief 



Summary of Activities 

Clinical Information Utility (CIU). Clinical Support 
Section (DMB). This ongoing major effort nnaintains a 
data base for research and patient care in the NIH 
Clinical Center. During this past fiscal year a number 
of improvements were made: the Surgical Pathology 
subsystem and database were fully implemented 
using the SNOP System to encode the diagnoses; 
the general design of an integrated database was 
completed; a subsystem was defined, designed, and 
implemented for purging cumulative laboratory data 
from the weekly production runs; software was 
developed for producing final cumulative laboratory 
summaries; several modifications were made to the 
CIU that reduce the run and connect times for 
producing the weekly laboratory summaries and 
updating the databases; and a subsystem was 
designed to preprocess clinical laboratory data for 
the integrated database. 

Combined Cardiology/Heart Surgery Data 
System. Larry Martin (DMB/ASPS); Roger Dailey 
(DMB/DBAS); C. Mcintosh (NHLBI); D. Rosing 
(NHLBI). This combined system provides a 
chronological record of the medical activity of NHLBI 
Cardiology and Heart Surgery Branch patients. In 
FY81 effort was directed toward meeting the routine 
and ad hoc reporting requirements and new 
statistical needs of the NHLBI physicians and 
researchers and the system was expanded to 
include nuclear angiogram information. An online 
private disk was assigned to the project during the 
year to improve data query and analysis response 
time. 

Pulmonary Function Data System. Judy Mahaffey 
(DMB/ASPS); Ronald Crystal (NHLBI/IRPB); Larry 
Nadel (DCRT/CSL). The Pulmonary Branch of 
NHLBI has requested the development of a 
combined computerized data base for pulmonary 
function and exercise testing data to replace existing 
separate ones. It is planned that the system will 
interface a NOVA and an LSI-11 minicomputer. This 
year analysis was completed and a design proposal 
was prepared during the last year. 



Analysis of SLE Nephritis Patients. George 
Shakarji (DMB/OC); John H. Klippel (NIADDK); John 
Decker (NIADDK). The storage phase of the 
development and implementation of this ongoing 
project is now completed with major modifications 
included. The system has the capability of storing 
and retrieving chemistry and therapy data on all SLE 
(Systemic Lupus Erythematorus) nephritis patients. 
Data on a subset of SLE patients, participating in the 
immunosuppressive trials and assigned to receive 
either prednisone only or the combination of 
prednisone and cyclophosphamide, are now being 
studied to evaluate certain chemistry constituents. 

Multivariate and Univariate Forecasts for Blood 
Constituents. George Shakarji (DMB/OC); Eugene 
K. Harris (DCRT/LAS). This study, which is part of 
continuing studies on variability of blood chemicals in 
normal people, uses data compiled through the 
health maintenance program in Japan. The database 
includes 15 to 18 semiannual values for 6 
biochemical tests in over 16,000 men and women 
between the ages of 20 and 70. Programming for 
this study involved univariate and multivariate time 
series systems. Programs were completed to 
compute homeostatic and random walk (non- 
stationary) models for both the univariate and 
multivariate analyses. All of the methodologies were 
applied to the database to compute and compare 
results and predictions. 

Psychobiology Patient Information System. 

Dennis George (DMB/ASPS); Steve Soroka (DMB/ 
ASPS); Frank Putnam (NIMH/BP). The purpose of 
this project is to condense a large amount of data 
for a small number of patients studied by the 
Biochemical Psychiatry Branch into a format that is 
useful for research analysis. During the last year 
analysis was completed and a design proposal was 
submitted for a system that extracts and reformats 
data from the Clinical Information Utility for analysis 
by existing statistical packages. 

Dyslipidemia Computerized Recordkeeping 
System. George Roberts (DMB/SAS); Ernst 



89 



Schaefer (NHLBI/MDB). This system keeps records 
on clinical laboratory data for normal and 
dyslipidemic subjects and provides for routine 
reporting as well as for ad hoc queries and 
preparation of selected subfiles for statistical 
analysis. An individual history report program was 
supplied during FY81. For current reporting requests, 
the new SAS online graphics package was used. 

BRIGHT Augmentation. Brian Cole (DMB/SAS); 
David Rodbard (NICHD/BES); Jay Shapiro (CC). A 
computer system is being developed on the 
DECsystem-10 that will enable Clinical Center 
investigators to analyze their own clinical data. 
Available thus far are a t-test module and a plotting 
module. Also to be included are descriptive statistics, 
chi-square test, linear regression, ANOVA, normality 
test, non-parametric tests, and life table analysis. 
Modules are to be added as requested by 
investigators. 

Diet Composition/Menus. Diane Feskanich (DMB/ 
SAS); Dennis Sprecher (NHLBI). This is a system for 
determining the nutritional profiles of patient menus. 
With appropriate modifications, the 'MR FIT' 
nutritional coding tape and diet composition 
programs were made operational at NIH. An input 
module for Dyslipidemia patients' dietary records has 
been supplied. 

Survival System. Diane Feskanich (DMB/SAS); 
Ardyce Asire (NCI). This life table analysis system 
was originally developed in the 1960's to support the 
End Results in Cancer studies of NCI. Maintenance 
and improvement of the system is now the primary 
goal. During FY81 the system was sent to: Rhode 
Island Health Services Research, Inc.; Department of 
Health, San Juan, Puerto Rico; and Rosewell Park 
Memorial Institute, Dept. of Health, Buffalo, NY. 

Prevalence of Major Neurological Diseases- 
Nigeria. Joe Huston, Mary Lee Dante (DMB/SAS); 
Bruce Schoenberg (NINCDS/NS); Dr. Osuntokun 
(University of Ibadan). This WHO-sponsored study 
consists of four parts: census and health screen, 
evaluation of risk factors, neurological exam results, 
and follow-up. A pilot study was done to determine 
validity and usefulness of the questions and worth of 
the questionnaire. Processing of these pilot forms 
led to many suggestions for improving the study 
protocol and questionnaire format. 

Neurological Screening Summary. Brian Cole 
(DMB/SAS); B. Schoenberg, D. Anderson (NINCDS/ 
NS). A survey of neurological disorders was made in 
a Mississippi county; this study examines the 
epilepsy, stroke, psychomotor delay/cerebral palsy, 
transient ischemic attacks, and Parkinson's Disease 

90 



data. Extensive validity checks and consistency 
editing were required. Preliminary analysis of 
Parkinson's Disease, including frequency tables and 
bar graphs, has been supplied. 

Seroepidemiology Data Processing System. Judy 
Mahaffey (DMB/ASPS); Paul Levine (NCI). The 
Clinical Studies Section, NCI Laboratory of Viral 
Carcinogenesis, is trying to find characteristics of 
serum samples that can be used to predict cancer. 
To this end, a computer system has been designed 
to manage all data necessary for efficient inventory 
control, test results feedback, and statistical analysis. 
The system is now operational and reports from the 
system are being sent to collaborating scientists in 
the U.S., Ghana, Greenland, and Singapore. During 
the past year a new contractor took over the the 
running of this system, and was provided with 
assistance in setting up to correctly run the system. 

Idiopathic Hypereosinophilic Syndrome Protocol. 

Brian Cole (DMB/SAS); John Harley (NIAID). A data 
base is being set up that will allow easy storage, 
retrieval, and analysis of a large amount of data that 
has been paper-collected on hypereosinophilic 
patients since 1967. Information drawn from the CIU 
will be included. In FY81 drug therapy information 
was extracted from the CIU and patient response 
patterns studied. 

Physiologic and Behavioral Responses to 
Apomorphine. Diane Feskanich (DMB/SAS); Neal 
Cutler (NIMH). The effects of apomorphine on 
physiological and psychological variables are being 
studied in groups of patients and volunteers, male 
and female. Graphs have been run on time of peak 
hypothermal response, mean duration of the 
response, and time of rebound by age, sex, and 
other variables. Also being examined is the 
development of brain tolerance to the drug. 

Smithsonian Tick Collection Query/Retrieval 
System. Diane Feskanich (DMB/SAS); Carleton 
Clifford, Jim Kearins (NIAID/RML). The Rocky 
Mountain Lab has catalogued its tick collection on 
tape and sent the data to the Smithsonian Institution. 
DMB is supplying the ability to query this file from 
Montana using the DCRT central computer facilities. 
During FY81 Ms. Feskanich assisted RML in the 
selection of a DataPoint word processor for 
installation in Montana, and ensured that the 
software would interface with DCRT software. Ms. 
Feskanich has provided interactive DataPoint 
programs for data entry and query/report and will be 
training RML personnel in use of the word processor 
when it is installed. 

Monkey Management System. Diane Feskanich 



(DMB/SAS); Robert Williams (NICHD/ERRB). A data 
base is being developed of bibliographic and medical 
information, plus the experimental history of each 
monkey in a colony of 500 to 1,000 monkeys, with a 
turnover of 300 each year. The system will be used 
to select appropriate individuals for specific 
experiments, and to prepare daily work assignments 
for caretakers and technicians. 

Musculoskeletal Model. Sig Knisley (DMB/SAS); 
Richard Lymn (NIADDK/ABSDP). This is a study of 
the structure of live muscle fibers as they contract. 
From a model, diffraction patterns will be computed 
and compared with real diffraction patterns produced 
by living muscle fibers. During FY81 myosin data has 
been processed with the PDP-10, linked with the 
PDP-11, and put up on the Evans and Sutherland 
picture system and the frame buffer for 3-D 
conformational analysis and manipulation. 

Wild Mouse Breeding Colony Data Processing 
System. Vivian Pelham (DMB/ASPS); Ernest Plata 
(NCI). The Laboratory of Viral Carcinogenesis, NCI 
Division of Cancer Cause and Prevention, breeds 
and raises a rare and valuable strain of wild Asian 
mice originally acquired from Vietnam, A system that 
will maintain all data collected on these animals and 
aid in selective breeding, carcinogenesis studies, 
aging studies, etc., was completed during the past 
year. 

Canine Breeding Colony Data Processing 
System. Peter Basa (DMB/DBAS); Dennis George 
(DfvlB/ASPS); T. Wolfle (DRS/VRB/ACS). The goal 
of this project was to develop a system to assist the 
Veterinary Resources Branch, DRS, with its record 
keeping and work scheduling. The system is 
complete. DRS is now in the process of installing a 
word processing system (CADO) in Poolesville to 
handle all data entry, maintenance, etc. When this is 
complete, DMB will work on interfacing the two 
systems. 

Strain Specificities Reference System. Steve 
Soroka (DMB/ASPS); David Sachs (NCI). A 
computer system is being developed for the Division 
of Cancer Biology and Diagnosis, NCI Immunology 
Branch, to assist in transplantation biology research. 
The system will be used to help locate existing 
cogeneric mouse strain products and/or to design 
mouse strain products with specific antigens that are 
used in experiments relative to the development of 
sera. Analysis was completed and a design proposal 
was prepared during the past year. 

Ectromelia Epidemiologic Survey. Dennis George 
(DMB/ASPS); Gordon Wallace (NIAID). This project 
provides the data processing services necessary to 



determine the environmental factors that most likely 
contribute to the spread of Ectromelia (Mouse Pox). 
The project was completed during the past year. 
Using the results from the study, Dr. Wallace wrote a 
paper, which he presented at a National Conference 
on Lab Animal Diseases in October, 1980. 

Estimating Q Matrix in the Kalman Recursion. 

George Shakarji (DMB/OC); Eugene Harns (DCRT/ 
LAS). A package has been implemented that would 
estimate the matrix of 'shift' variances and 
covariances, or the Q matrix. The package assumes 
either that Q is unknown, in which case the weight 
function and the matrix of predicted values can be 
calculated directly, or that Q is unknown but that the 
series of observations is long. 

Multivariate Time Series Packages. George 
Shakarji (DMB/OC); Eugene Harris (DCRT/LAS). 
Two packages that would input multivariate 
observation vectors of related tests were completed 
and tested. The homeostatic approach assumes the 
existence of a constant set point about which the 
sehal measurements fluctuate; successive 
observation vectors are presumed to be mutually 
independent. The non-stationary, random walk 
approach does not assume any constant 
homeostatic subpoint. It postulates instead that the 
true value at any time is shifted randomly from a 
previous time. In this procedure the predicted value 
is a weighted (exponentially smoothed) average of 
past results. 

Materiel Management System (MMS). Marvin 
Katz, Ron Wicks (DMB). This ongoing administrative 
project utilizes data base technology in support of 
NIH-wide procurement, receiving, and payment 
activities. As the MMS entered its fourth year of 
development and operation, much time was spent in 
enhancing existing software. During FY81 some 50 
change control items successfully went into 
production. New developmental efforts implemented 
were: 

1. Provision of a delegated interface to MMS for 
the B/I/D's Administrative Offices. This interface 
includes: procurement entry at point of origin, receipt 
of entry, and online review and control of all 
pertinent data by the Administrative Office. This 
feature will be phased into the B/l/D's as terminals 
become available. 

2. Development of a new numbering system for 
the delegated procurement actions that originate in 
the B/I/D's. 

3. Streamlining of DFM Accounts Payable 
functions by automatic voucher generation as part of 
the Treasury schedule preparation. 

91 



4. Development and implementation of an invoice 
entry subsystem for DFM. 

5. Completion of subsystem design for a stock 
inventory subsystem. 

6. Cutover to production of a source subsystem 
thiat allows blanket purchiase agreements and 
indefinite delivery contracts to control telephone 
charge order and record of call validity. 

Requirements Analysis for Financial Management 
Data Base. Clare Hoover, Jeff Schriver (DMB/ 
DBECS); Harry Hsu (SIMCOM/USAF). As an adjunct 
to the full statement of Division of Financial 
Management accounting requirements, the study 
participants: developed a complete structured flow 
chart of the existing Central Accounting System 
(CAS); defined the detailed flow of key CAS 
transactions; and created a matrix of all transactions, 
master files, and their processing relationships. 
These documents will be used as the basis for 
evaluating vendor accounting packages. 

General Support for Central Accounting System. 

Clare Hoover, John Price, Jeff Schriver (DMB/ 
DBECS). During the fiscal year, the Data Base 
Enhancement and Control Section developed 
enhancements and new programs, and conducted 
studies for, the Central Accounting System. Because 
this support involved almost 80 separate projects of 
varying size, they will not be enumerated here. 

Performance Index/Grant Awardees. Mary Lee 
Dante (DMB/SAS); William Parker (NINCDS/EAP). 
Under study is a system for tracking K07, K04, and 
F32 applicants. To do this, information must be 
acquired from the IMPAC and CRISP files, organized 
(by category, principal investigator, and grant), and 
printed in matrix format. If pursued, this can be 
generalized for other requests. 

DRR Grants Subproject System. Vivian Pelham 
(DMB/ASPS); Jean Babb (DRR). The existing DRR 
Grants Subproject System, which uses Conversation 
Programming System, is being evaluated. A proposal 
was made for the redesign of this system to make 
use of more current, supportable technology. The 
proposal was accepted and the system currently is 
being developed. 

NIH Nutrition Grants Monitoring System. Judy 
Mahaffey (DMB/ASPS); Thomas VogI (OD). A 
system has been designed for the NIH Nutrition 
Coordinating Committee to assist them in monitoring 
and reporting data on biomedical and behavioral 
nutrition research at NIH and at other agencies 
within DHHS. The system is operational and Dr. 

92 



Vogl's office is currently using it to answer inquiries-- 
from NIH Directors' offices, the White House, 
Congress, and the public--that relate to dollar 
amounts and percentages of grant money being 
spent in the area of nutrition. 

Review and Evaluation Branch Grants 
Information System. Penny Brogan (DMB/ASPS); 
Harry Canter (NCI). The computerized Research 
Analysis and Evaluation Branch Grants Information 
System, a highly specialized grants management 
system, was designed and implemented for the 
Division of Cancer Grants, NCI. Enhancements are 
currently being made to the Intramural and Funded 
Grants subsystems, and the system is being 
extended to include contract data. In the future, a 
Training Grants system will be developed and history 
file maintenance will be added to the Intramural 
projects and unfunded grants systems. 

NIH International Activities and Personnel 
Monitoring System. Penny Brogan (DMB/ASPS); 
Libby Low (FIC). A system has been developed to 
process financial and visa data on the foreign guest 
workers and foreign visitors at NIH. The Fogarty 
International Center is currently using this system to 
provide reports for the Visiting Program, for foreign 
nationals, and for foreign embassies. The growth of 
the current data files will necessitate a future 
revision of the system to separate current data from 
historic data. During the past year an analysis of all 
FIC requirements was completed and a design 
proposal with recommendations for improving their 
system was prepared. 

Committee on Academic Science and 
Engineering (CASE) Reports. Darius Georg (DMB/ 
ASPS); J. Bailey (OD/OPPE). This project involves a 
broad spectrum of data processing support required 
for the collection and reporting of DHHS obligations 
to institutes of higher education, research and 
development centers, and non-profit institutions. This 
is an ongoing project. 

MMS Query and Reports. Jane Blessley (DMB/ 
ASPS); Joe Campbell (DMB/DBAS). This project 
provides an economical method for the selection 
and reporting of data from the NIH Administrative 
Data Base. Ms. Blessley provides recurring and ad 
hoc reports from the data base for all segments of 
the NIH community. 

System for Statistical Complaints of 
Discrimination at NIH. Darius Georg (DMB/ASPS); 
G. Yee (OD/DEO); M. Williams (OD/DEO). This 
project establishes and maintains a file that provides 
statistical data, on a case by case basis, of formal 
and informal complaints of discrimination at NIH. In 



the past year Mr. Georg revised and simplified the 
retrieval process. 

ARMS/TDCS Interface. Dennis George (DMB/ 
ASPS); B. Hughes (OPA/P); A. Amatucci (OA/M). 
This project is intended to create an NIH Personnel 
System that is a composite of the current NIH 
personnel system (ARMS) and the DHHS Personnel 
System (TDCS). In the past year the analysis and 
design of the proposed system was completed. The 
ARMS Steering Committee has approved the 
proposal and the system is presently being 
implemented. 

Radiation Safety Control System. Charles Twigg 
(DMB/ASPS); R. Zoon (DRS/RSB). This system is 
designed to monitor the use and users of radioactive 
isotopes at NIH. When complete, this system will 
include five subsystems: inventory and bioassay, lab 
survey and airborne release, waste processed and 
activity balance, training, and film badges. In the past 
year extensions were made to the inventory and 
bioassay subsystems to satisfy Nuclear Regulatory 
Commission requirements. Development of the lab 
survey and airborne release subsystem was begun. 
All subsystems have been completed except the 
waste processed and lab survey. 

Electrical Safety Program System. Larry Martin 
(DMB/ASPS); Steve Soroka (DMB/ASPS); Howard 
Metz (DRS/BEIB). The chief of Scientific Equipment 
Services of the Biomedical Engineering and 
Instrumentation Branch has requested a system to 
help monitor maintenance of equipment at the 
Clinical Center. A system is being designed to 
computerize the results of routine electrical safety 
checks and preventive maintenance performed on 
hospital equipment. The system will be used by DRS 
to schedule equipment checks, to provide reviews on 
instruments checked by contractors and by the CC, 
and to provide statistical information on different 
types and repair histories of equipment. A design 
proposal has been accepted, and the system is 
currently being implemented. 

Design Billing System. Peter Basa (DMB/ASPS); 
Robert Weymouth (DRS/OD). This project converts 
the manual accounting system for the Design Unit in 
MAPB/DRS to a computerized system. Analysis, 
design, and development of the system were 
completed during this year. 

Information System of Extramural Scientists. 

Darius Georg (DMB/ASPS); William Rhode (00/ 
OPPE). This system creates a data base of 
information drawn from various sources in order to 
perform analysis of various patterns of involvement 
in NIH science review activities by extramural 



scientists. The data base was created during the 
year and reports are being run as requested. 

Medical Records Auditing System. Judy Mahaffey 
(DMB/ASPS); Gloria Burich (CC/MRD). The purpose 
of this system is to assist Medical Records in the 
monitoring and reporting of the status of medical 
records from the time they enter the department until 
they leave. When the system is developed, it should 
replace four manual systems now being used by the 
Medical Records Department. During the past year 
analysis was completed and a proposal was 
submitted. The proposal has been accepted, and the 
system is currently being implemented. 

Correspondence Control System. Steve Soroka 
(DMB/ASPS); Dennis George (DMB/ASPS); Zaven 
Khachaturian (NIA/OD). The objective of this system 
is to monitor the status of staff assignments and 
correspondence assigned to NIA for processing and 
review, and to provide a query facility that will allow 
NIA personnel to locate correspondence (old and 
current) pertaining to a current issue or problem. 
During the last year analysis was completed and a 
design proposal was submitted for approval. 

AIRS Personnel System. Vivian Pelham (DMB/ 
ASPS); L. Lee Manuel (DCRT/OD). This project 
involved a complete revision of this system due to 
the discontinuation by CCB of the Conversation 
Programming System. Analysis, design, and 
implementation were completed during this year. 

HMO Label Programs. George Roberts (DMB/ 
SAS); Lois Eberhart (OHMO/OPS). The Office of 
Health Maintenance Organizations maintains several 
address lists. They were provided with an interactive 
updating capability, a number of reports, and a 
gummed label option. 

Space Management System. George Roberts 
(DMB/SAS). This system provides a method 
whereby the Office of Research Services can keep 
track of all space in all buildings occupied by NIH. 
The Fort Dietrick facility was added to this system 
during FY81. 

Chinese Personalities and Institutions in 
Biomedicine. Judy Mahaffey (DMB/ASPS); Joseph 
Quinn (FIC); Joseph Lee (FIC). International 
exchanges in the field of biomedicine between the 
U.S. and the People's Republic of China have 
increased rapidly. The Fogarty International Center 
has requested DMB services to design a system for 
the computerization of data on biomedical scientists 
and institutions in the PRC. The system will be used 
by the FIC officials in bnefing NIH and non-NIH 
scientists interested in biomedical research in China. 



93 



During the year, analysis was completed and a 
design proposal was submitted to FIC. 

Selective Dissemination of Information. Sig 

Knisley (DMB/SAS). SAS has continued its support 
of the current awareness search for both Chemical 
Biological Activities (CBAC) and Biosciences 
Information System (BIOSIS). Retrospective 
searches are referred to the NIH Library staff. 

Sickle Cell Disease, Health/Science Seminar 
Evaluation. George Roberts (DMB/SAS); Katrina 
Johnson (NHLBI/SCDB). A battery of tests probing 
general knowledge of Sickle Cell Disease is given to 
groups of teachers prior to a two-day seminar on the 
disease. Then a post-test is given to measure the 
value of the seminar. Some analysis was done on a 
set of tests already scored, mainly for the purpose of 
evaluating the test questions. Advice also was 
provided during the process of requesting proposals 
from contractors for administering the program. 

SLANG (Structured Language) Compiler. Bob 

Magnuson (DMB/OC). SLANG is designed to assist 
programmers to generate block structured assembly 
language code on the IBM system 370. The features 
added include automatic indentation, generic 
statement numbering, boxed comments, and 
program structure display. 

Voice Input and Synthesis Support. Bob 

Magnuson (DMB/OC). As a follow-up to DMB's initial 
voice input project, this section is examining the 
feasibility of developing voice input/output 
applications by way of microprocessors, which can 
act alone or as front end processors to larger 
computers. This year was spent modifying the design 
of a microsystem with plug-in peripherals, purchasing 
some hardware, and providing appropriate software. 

SFOR (Structured FORTRAN) Compiler. Bob 

Magnuson (DMB/OC). Designed to assist 
programmers writing structured programs, the SFOR 
compiler generates block-structured IBM FORTRAN 
source code. There are six different kinds of blocks 
available to the FORTRAN programmer- 
CASENTRY, FOR, IF, LOOP, REPEAT, and WHILE. 

RMAG Products Support. R. Magnuson (DMB/ 
OC). Necessary support is provided for RMAG, SLR, 
Logic Subroutines, Arithmetic Subroutine, SLANG, 
REFORMATGEN, REPORTGEN, TRANSACTGEN, 
Standardized Update, Voice Input, and SFOR. This 
ongoing support includes software maintenance, 
customer assistance, and the teaching of formal 
DCRT courses on these products. In particular, a 
special effort had to be mounted to change over to 
the new WYLBUR format data sets. 

94 



Computer Center 
Branch 



Joseph D. Naughton, Chief 



Summary of Activities 

New Languages. An entirely new version of 
WYLBUR, designed and implennented by the Center 
over the past several years, was made operational in 
a full production nnode in early January. In addition to 
retaining all the capabilities of the previous version, 
the new version provides many new functions 
including document formatting, pattern matching, 
catalogue and PDS support, session recovery, and 
command procedures (EXEC files), plus many 
enhancements to existing facilities. The new version 
uses much less CPU time than its predecessor at 
the same level of user load and can accommodate 
many more simultaneous users. 

SPEAKEASY, a computing language for scientific 
and mathematical problem solving, became available 
under TSO on the System 370. Developed by a 
physicist at the Argonne National Laboratory, 
SPEAKEASY provides a quick and simple means of 
formulating mathematical problems and obtaining 
results. The language contains over 500 functions 
and commands that perform matrix and array 
operations, numerical differentiation and integration, 
statistics, and character processing. Because its 
notation and syntax are similar to those used in 
mathematics, it provides extensive new scientific and 
mathematical capabilities without burdening the user 
with the details often inherent in computer 
programming. SPEAKEASY may be self-taught at 
NIH through two online facilities; TUTORIAL and 
HELP. 

The IBM PASCAL/VS Compiler was tested and 
installed on the System 370, and is available for 
interactive use under TSO or for batch processing. 
PASCAL is a relatively small but very powerful 
language that has become widely used throughout 
the data processing industry. It is a very concise 
language with few defaults or implicit conversions, 
making it easy to use. The PASCAL/VS compiler 
contains many extensions to the International 
Standards Organization proposed PASCAL standard. 
It provides the ability to divide programs into 
separately compilable sections, a debugging facility, 



and numerous other features. 

The Conversational Programming System (CPS), the 
first interactive programming facility offered at NIH, 
was replaced by the more modern VS BASIC, an 
improved version offering many features not 
previously available. Guidelines were developed to 
assist users in converting CPS-PL/I programs to use 
the PL/I Optimizing Compiler under TSO. The 
command procedures facility of WYLBUR provided a 
viable mechanism for rewriting some programs. 

New Software. Two powerful new graphics 
packages were made available on the IBM System 
370. TELL-A-GRAF, an interactive system that uses 
conversational commands, may be used to create a 
variety of graphs and charts. The system is easy to 
learn and does not require programming experience. 
The other program, DISSPLA, is a library of 
subroutines that can be called by the user's 
program. Although some programming experience is 
necessary, DISSPLA is also easy to use and enables 
the user to integrate data analysis and graphical 
display into a single program. Both TELL-A-GRAF 
and DISSPLA are capable of driving a wide variety of 
graphics devices. 

A new software package was installed to assist 
users in the diagnosis of program abends that cause 
dumps. ABEND-AID analyzes program abends, 
extracts diagnostic information, and presents the 
results in a few easy to understand pages. This 
enables users to analyze abends without having to 
go through the painstaking process of interpreting a 
dump. 

Output Facilities. The OMNIGRAPH package was 
enhanced by the implementation of a library of 1,377 
different alphabetic and graphic characters known as 
Hershey's fonts. These high quality, esthetically 
pleasing characters and symbols, manually digitized 
by Dr. Allen V. Hershey of the Naval Weapons 
Laboratory, greatly enhanced the appearance of 
graphs and plots generated by OMNIGRAPH and 
MLAB users. 

All standard output forms provided by the Computer 

95 



ZOl CT00077-01 CCB 



Hueltlc Ad d S tructure Synthesis and Display 

P. I.: Richard J. Feldttann. CCB, DCfiT, Computer Specti 



None 


UNMi (r» •- 


) 




Cwrpu 


er Center Branch 










DCRT. 


NIH. Bidg. 


12A, Room 3009. Bethesda. Maryland 20205 




r.o 


.«, 


iMOFtStintL. jOIHlNi 




^^■'~ °'"""""""- 











;xported to other lnsi 



1 sequences. Because crystal 



Center were made directly available through JES2, 
the component of the Operating System that handles 
printer/punch output. This change eliminated the 
need for creating and processing a SPOUT tape, 
thereby reducing the overhead and cost for creating 
output at the NIH Computer Utility, and provided the 
ability to examine, reroute, locate the output of a job 
during any stage of its processing. By reducing 
operator involvement, the possibility of error was 
also decreased and turnaround was improved. An 
easier mechanism for users who require printed 
labels was developed. Each of the 21 different types 
of standard labels is now identified with a simple 
four-character name and an appropriate carriage 
control loop is automatically requested by JES2. 

Communications Improvements. The link between 
the DECsystem-10 and IBM System 370 was 
strengthened by the addition of a second 
communication line between the two systems. This 
doubled the capacity of the link and ensured that 
jobs get through quickly. In addition, the effective 
speed of the link was further increased by several 
software improvements, nearly doubling the speed of 
transmission on each communication line. 

Documentation/Publications. Publications by the 
Computer Center are oriented toward familiarizing 
users of the Computer Utility with the computer 
services, languages, and training available. Ten new 
publications and ten revised or updated titles were 
released this year. 

Over 1 ,200 pages of documentation were prepared 
for the new version of WYLBUR, including a 
Fundamentals manual and four additional manuals 
on document formatting, command procedures, 
batch processing, and general editing. Other 
documentation included a Master Index; two 
reference handbooks to give the syntax of the 
language in concise form; and a special edition of 
INTERFACE, the Computer Center's technical notes, 
which introduced the new version in considerable 
detail and described specific differences between the 
new version and the old. 

Seven editions of INTERFACE including the special 
WYLBUR edition and the Annual Index were issued 
this year. Two new features were added, 'WYLBUR 
Wisdom' and 'MLAB NOTEBOOK.' INTERFACE a\so 
ran a five part series written by Dr. Benes L. Trus, 
Computer Systems Laboratory, describing various 
techniques of scientific image processing using the 
Center's Surface Display System. 



96 



User Training and Assistance. Seven new lecture 
courses were added to the Computer Center training 
program this year, including 'Managing and 
Processing Data Sets at NIH,' 'Dynamic Biological 
Stimulation,' and 'BRIGHT--A System for the 
Creation and Use of Data Tables.' The popular 
WYBLUR introductory course was divided into two 
courses, one focusing on secretarial applications and 
the other on data processing applications. New 
seminars this year included 'Introduction to Scientific 
Data Analysis at NIH,' 'IMSL: International 
Mathematical and Statistical Library,' and 'Graphical 
Representation of Multivariate Data.' There were 71 
sessions of 41 different classroom courses given to 
over 1,600 students during the year. 

Self-study courses, involving either programmed 
instruction, workbook, audiovisual, or computer- 
assisted learning, continued to be popular. 'PL/I 
Programming,' a new independent study program 
written by IBM, was added to the roster this year. 

Programmer Trouble Reports (PTR's) researched 
and answered dunng the year numbered 2,745. User 
Services applied over 4,000 system software fixes 
during the year and installed 14 new releases of 
current software packages. There were 23,000 calls 
or visits by users for assistance during the years. 



developed by others for experimentally determining 
the pairing of nucleic acid bases. This project has 
developed a technique for displaying the two- 
dimensional structure of general nucleic acid 
sequences. Two-dimensional diagrams for the 16S 
fragment of the ribosome (1.6 kilo bases) have been 
generated. A complete language has been 
developed for input, manipulation and display of 
nucleic acid two-dimensional structures. Copies of 
the program package have been exported to other 
institutions. 

An attempt has been made to synthesize the three- 
dimensional structure of general nucleic acid 
sequences. Because crystallography has been done 
on only a few nucleic acid structures (straight helix 
DNA and RNA, and two tRNA's) the cntical insights 
needed for general structure synthesis are still 
missing. Modeling techniques are being developed to 
solve this problem. 



Research Projects 

In addition to the many activities, services, and 
facilities for NIH, the Computer Center Branch 
serves biomedical computing with its research work. 
The DECsystem-10 and the molecular graphics 
systems have been used to develop techniques for 
creating and displaying two- and three-dimensions of 
genetic control molecules. It is now clear that both 
DNA and RNA fold up into complex structures. 
Algohthms have been developed to create 
reasonable folding patterns for rather long nucleic 
acid sequences. A two-dimensional display algorithm 
has been developed to layout many of these 
complex structures. Using the layout algorithm it is 
possible to study the dynamics of the folding 
process by making movies. By understanding the 
dynamics of folding, it may be possible to 
understand how processing and gene control 
function. 

Nucleic Acid Structure Synthesis and Display 

Rapid advances in nucleic acid determination have 
led to questions about the secondary and tertiary 
structure of DNA and RNA. It is clear now that the 
sequence alone of a nucleic acid is not sufficient to 
determine many of the processing and control 
functions. Computer techniques have been 



97 



Office of the 
Director 



Arnold W. Pratt, M.D. 



Summary of Activities 

Library Automation. E. Chu; J. Mahaffey (DMB); J. 
Knight (CSL). In conjunction with other DCRT staff, 
the DCRT Librarian applies computer techniques to 
DCRT needs, advises other libraries, and maintains 
knowledge of work done outside NIH. (Details of 
activity in FY81 appear in Volume I of the Annual 
Report.) 

DCRT Publication File. P. O. Miller; R. Baxter 
(DMB). In FY79 the Information Office began to 
create a file of citations for all papers published by 
DCRT authors. In FY81 additional work was done to 
correct errors in the file. 

Text-to-tape Copy Preparation. P. O. Miller. This 
project is an offshoot of work begun in 1 979 as part 
of a joint PSL/LAS/OD effort. In FY81, a JCL was 
written for creating a WYLBUR tape to drive GPO 
typesetting equipment Documentation was begun in 
an effort to make the technique available to all NIH 
Information Offices. 

DCRT Communications Program. P. O. Miller; W. 
C. Mohler. Previously called the DCRT Information 
Program, this is an ongoing project to develop 
improved and coordinated communication 
techniques to support DCRT activities. It has four 
parts: Analyzing Needs, Creating and Evaluating 
Products, Developing Resources, and Education. In 
FY81 work continued on developing and distributing 
information products, including a videotape about 
computer terminals to aid handicapped programmers 
and a slide show about DCRT work. 

Who at NIH uses DCRT for what? W. C. Mohler; 
L. L. Manuel. DCRT services support activities 
throughout NIH for more than 3,000 registered users 
on some 1,100 accounts. A survey in FY79 and 
FY80 asked users to categorize the types of NIH 
activities supported by each account in the previous 
year. The FY81 survey in conjunction with the PAS 
account update confirmed previously observed 
difficulties in obtaining consistent and usable 
information by this collection mechanism. The 
project was stopped. 



Clinical Data Management and Analysis. W. C. 

Mohler; B. Cole (DMB); D. Rodbard (NICHD); J. R. 
Shapiro (Clinical Center). In spite of the rapid growth 
in use of data management and statistical packages 
provided for NIH scientists on DCRT computers, 
there is a perceived need for facilities that would be 
easier to learn and use in NIH clinical research 
projects. In FY81 work began using BRIGHT, a 
table-oriented data management/analysis package 
on the DECsystem-10, developing added data 
analysis and display programs and exploring their 
usefulness for a few NIH clinicians on data sets from 
the Clinical Information Utility. 

Multi-function Microprocessor Interface. A. W. 

Pratt; D. Songco (CSL). The project begun in FYBG 
seeks to adapt a variety of information acquisition 
techniques on a single microcomputer as a versatile 
data input/output interface for biomedical scientists 
and clinicians. 

Medical Linguistics. A. W. Pratt, et al. This is a 
long-term project to define a set of semantic and 
syntactic forms that can aid in the analysis and 
interpretation of written medical statements. 

Image Processing, Decision Analysis, and 
Computer Architectures. J. M. S. Prewitt. 
(Summaries for projects in this group of activities 
were not available in FY81.) 



98 



Research Projects 



October t. 1980 to Septcnter 30. 1981 



^"'" '""" "■" "* "" "*" tiiiiwjajj^^^ 

"""October I, 1980 to Sept 
[ 1 Iext-to-t«pe Copy Prepa 



ZOI CT00078*Ot 00 



1 Orftcrr. CD. OCRI 



















Government 


Print 


tn9 Offlc 


ei Co«np«ter Center Branch 






Lulm^ 


JfflM of tlie Of 
Informtlon Off 


ce. KRT 
uter Rese 






















OWisfon 


CMV 


arch and Ttchnolosy 








0.1 


r 


0.1 


lOtHOII 








Ql..)-IW« 


LI (.'1 I"" 




o,.,~- 


— 


a,.,«™. 






Governmenl public af 

Once text is records 
conputer tape reeU) 
be furnished to » tjf 

produce camera-ready 
qalley proofreading 


ii 


documen 
for cor 

magnetic 


^, These syste 
t f.snion. Ed 


ms enable i 
te. ijving 

typesetter 


ocesstng 

jntJSe: 
s equtpm 


eclal 
Jtes the 







: U.S. GOVERNMENT PRINTING OFFICE: 1981- 720-020, 6586 REGION 3-t 



99 



DrVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



FISCAL ANNUAL VOLUME 1 

YEAR REPORT 

1982 



rVHT T 

KDCRTRT 



Foreword 



The Division of Computer Research and Technology DCRT progranns focus on three primary activities: 
has primary responsibility for incorporating the power conducting research, developing computer systems, 
of modern computers into the biomedical programs and providing computer facilities, 
and administrative procedures of NIH. DCRT serves jhe fiscal year 1982 annual report descnbes our 
as a scientific and technological resource for other yvork in two volumes: 

parts of PHS, and for other Federal organizations Volume 1 gives an overview of the work of 

with biomedical and statistical computing needs. each group, highlighting the year's 

accomplishments; 

Volume 2 gives details about the projects and 
activities of each group. 



From the Director 



I am pleased to submit this Annual Report for the 
Division of Computer Research and Technology. 
It serves to remind all of us of the extent and 
breadth of subject matter involved in the conduct 
and management of the NIH biomedical research 
program. Biomedical computing has matured and 
become another essential element of the scien- 
tific excellence of NIH. 

One has to be aware of increasing involvement 
and contributions of applied mathematics, statis- 
tics, engineering, and computer science across 
all biomedical research. The following examples, 
limited only to clinical medicine, serve to il- 
lustrate how DCRT mal<es these several discip- 
lines productive throughout NIH. 

The Computer Systems Laboratory is broadly in- 
volved in both the laboratory and clinical research 
programs in many Institutes and several Depart- 
ments of the Clinical Center. Current clinical 
work includes; 

• The Radiation Therapy System with the 
Radiation Oncology Branch of the National 
Cancer Institute, 

• Automated systems for the Pulmonary 
Branch of the National Heart, Lung, and 
Blood Institute, 

• The Positron Emission Tomography Facility, 
the focus of exciting collaborative projects 
among staff of the National Institute on Ag- 
ing and the National Institute of Neuro- 
logical and Communicative Disorders and 
Stroke, and 

• Other projects in the Medical Intensive Care 
Unit, Electrocardiography/Heart Station, and 
Anesthesiology Service of the Clinical 
Center and its Departments of Clinical 
Pathology, Nuclear Medicine, and Rehabilita- 
tion Medicine. 



The Laboratory of Applied Studies has been ac- 
tive in a variety of clinical research projects. 
Many of these have been collaborations with 
users of the systems designed by CSL. Areas of 
activity currently include laboratory medicine, 
electrocardiography, and pulmonary medicine. 

The Laboratory of Statistical and Mathematical 
Methodology provides mathematical tools and 
consultation for clinicians from all Institutes. In 
addition, it too has collaborative projects with 
medical scientists. Even the Physical Sciences 
Laboratory, with its major focus on physics and 
chemistry, consults on a few clinical research 
projects. 

The Data Management Branch has created scores 
of computer programs for clinical scientists in 
the Institutes and for departments in the Clinical 
Center. Its Clinical Information Utility project pro- 
vides the archival data base for information col- 
lected by the Clinical Center medical information 
system. DMB is working with members of the In- 
stitutes and Clinical Center to develop better 
systems to retrieve and analyze this archived 
data. 

Finally, the Computer Center Branch provides the 
reliable, accessible, modern central computing 
facilities that support hundreds of clinical 
research projects and virtually all of the adminis- 
trative activities without which research and pa- 
tient care could not go forward at NIH. 



rtytyf^ci iJjncufJf ^ 



Arnold W. Pratt, M.D. 
DCRT Director 



Contents 



Computer Systems Laboratory 1 

Provides consultation and collaboration in the design 
and implementation of specialized computer systems 
for laboratory and clinical applications. 

Laboratory of Applied Studies 7 

Relates mathematics, statistics, and computer 
science to such biomedical problems as ECG 
analysis, evaluation of physiological systems in 
health and disease, modeling of the microcirculation, 
and estimation problems in laboratory medicine. 

Physical Sciences Laboratory 1 1 

Conducts research in mathematical theory and 
practical instrumentation to explain biological 
phenomena in terms of chemistry and physics at 
subcellular molecular levels. 

Laboratory of Statistical and Mathematical 
Methodology 15 

Provides statistical and mathematical help in the 
computer analysis of biomedical data; offers 
statistical and mathematical packages for users; 
develops methodology in multivariate analysis, curve 
fitting, biological shape and pattern theory. 

Data Management Branch 21 

Serves as a central resource of systems analysis, 
design, and programming for data processing 
projects relating to scientific, technical, management, 
and administrative data. 

Computer Center Branch 25 

Designs, implements, and operates the NIH 
Computer Center; provides assistance, training, and 
technical communications to the more than 8,000 
users of the Central Utility. 



Office of ADR Policy Coordination 31 

Coordinates the complex Federal policies and 
procedures that govern getting and using computers 
at NIH. 

Office of Administrative Management 33 

Provides general administrative management support 
for the Division's work. 

Office of Scientific and Technical 
Communication 35 

Serves as a central source of information about 
DCRT activities and about computer-related 
disciplines. 



Computer Systems 
Laboratory 



Alan M. Demmerle, Chief 



Function and Scope of Work 

The Computer Systems Laboratory--26 professionals 
representing the disciplines of engineering, computer 
science, medicine, and chemistry--is the major 
source of expertise at NIH for minicomputer and 
microcomputer technology. CSL engineers and 
scientists, in collaboration with NIH intramural 
laboratory and clinical investigators, apply this 
technology in the areas of laboratory automation and 
patient care. Some projects are occasionally 
undertaken with NIH extramural program staff and 
with other Federal agencies. CSL's multidisciplinary 
approach aids both the recognition of problem areas 
that will benefit from automation and the 
interpretation of research needs in terms of 
computer methods. 

Computers may be used only in an adjunctive 
manner--for example, as a more convenient means 
to acquire laboratory and clinical data--or they may 
be integral parts of an elaborate instrumentation 
system, such as a computer-controlled mass 
spectrometer. Advances in large scale circuit 
integration (LSI)-the microprocessor revolution-have 
brought about the miniaturization of computer 
components and a dramatic decline in their prices 
and power requirements. CSL engineers are now 
able to use microprocessors to deal with problems 
that once defied solution because of cost, size, or 
manpower constraints. 

CSL projects range in size from consulting activities 
of a few days' or weeks' duration to large-scale 
efforts taking many manyears. Much CSL work 
involves the development of new methods or 
technology or is influenced strongly by the changing 
needs of research. Thus, it is often difficult to predict 
the long-term scope at the outset of a project. 



FY82 Highlights 

This year, CSL engineers and scientists worked on 
32 projects, representing collaboration with almost 



all of the NIH Institutes. Some of these projects 
require only limited resources, while others take 
many manyears. The latter deserve particular 
emphasis because of both their sheer magnitude 
and the importance of the patient care or research 
activity they support. 

One of these large scale efforts is the Radiation 
Therapy Project, a collaborative effort with the 
Radiation Oncology Branch, NCI. This project began 
in late 1975. Since then, it has experienced 
modifications to both the original goals and 
implementation techniques. At project inception, 
short-term priorities focused on integrating scans 
from a recently acquired CAT scanner into a 
computerized treatment planning system. More 
ambitious goals included three-dimensional treatment 
planning, treatment plan optimization, and dynamic 
treatment planning. Short-term goals were quickly 
achieved. Despite changes in program emphasis and 
significant technical problems, a limited three- 
dimensional capability has been implemented. 
Optimization requirements, however, have been 
reduced and are being satisfied through pseudo- 
optimization techniques that use interactive beam 
manipulation. 

Aspirations for achieving dynamic therapy have 
remained dormant. Instead, efforts were expended 
upon supplementing traditional isodose curves with 
point dose calculations, upon producing implanted 
seed calculations, and upon generating sophisticated 
treatment plan displays that emphasize particular 
structures or magnify special features. 

The original treatment planning system consisted of 
a Digital Equipment Corporation PDP-11/70 
computer with treatment planning software 
developed by J. R. Cunningham and marketed by 
Atomic Energy of Canada Limited (AECL). Clinical 
use of this system began at NIH in 1977 in a single 
user mode. Two years ago, the AECL software was 
replaced by algorithms developed by J. Van de 
Geijn. Sophisticated displays featuring up to eight 



beams, CAT or ultrasound scan, and point dose or 
isodose curves on one screen also have been 
operational for some time. 

The core hardware configuration remains essentially 
the same as originally purchased except for the 
addition of new display equipment. However, the 
original single user software implementation has 
been replaced by new software that supports 
multiple treatment planning stations. The success of 
the system can be measured by its high use by the 
Radiation Oncology Branch, and by the fact that a 
number of commercial firms have expressed interest 
in including it in their product line. 

Another major project in the Clinical Center that was 
completed this year is the computerized ECG 
Management System. The Heart Station of the 
Clinical Center is responsible for obtaining and 
interpreting electrocardiograms (ECG's) from 
hundreds of patients each week, and for maintaining 
a filing system for all previous ECG records. 

With the impending completion and occupancy of 
the new Ambulatory Care and Research Facility 
(ACRF), the Clinical Center wanted a new 
comprehensive computer-based system to facilitate 
the processing of the projected ECG workload and 
to provide a practical method for long-term storage 
and retrieval. 

The project's requirements included analog-to-digital 
conversions of ECG waveforms, computer-assisted 
interpretation of the ECG data, physician approval of 
all interpretations, automatic generation of final 
diagnostic reports, and generation and maintenance 
of a data base of all tracings and interpretations in 
machine readable form. An additional requirement 
was the ability to search the data base using 
selection factors to be specified by researchers 
engaged in the wide class of retrospective clinical 
studies envisioned for the ACRF. 

CSL developed the Request for Proposals for the 
system in 1979. The procurement process 
culminated with the purchase of a Hewlett-Packard 
ECG Management System that was specifically 
configured and programmed to conform to the 
Clinical Center's operational procedures, reliability 
needs, and report format requirements. Relatively 
rapid installation of the computer system was 
achieved by using an existing computer site in 
Building 10. The Clinical Center's telephone network 
was adapted so that the system's portable ECG 
machines could be used to transmit ECG's from all 
patient care areas or patient rooms to the centrally 
located ECG computer facility. 

Routine clinical use of the ECG diagnostic computer 
system began in January 1 982 after a short 



orientation session for the heart station's ECG 
technicians. CSL expects to provide software 
modifications to support future research protocols. 

Other major clinical activities on which CSL worked 
during FY82 are as follows: 

• Intensive Care Units (ICU's): CSL is involved in 
the automation of three ICU's. The general 
goals are to capture patient vital signs, generate 
comprehensive displays of patient status and 
trends on demand, and substantially simplify 
medical recordkeeping. Data archival is 
expected to facilitate research in areas such as 
noninvasive intervention. 

• Nuclear Medicine: In this project, the design 
and implementation of a stand-alone computer 
system facilitates the analysis of digitized patient 
brain scans obtained on floppy discs from the 
Nuclear Medicine Department's PET Scanner. 
Three Institutes currently use this system for 
diagnostic purposes and for pursuing basic 
research in the areas of schizophrenia, epilepsy, 
and aging. 

• Gait Laboratory: The Automated Biomechanics 
Laboratory System involves the measurement of 
limb and spine position in space, forces in the 
hand and between foot and ground, and 
electrical activity of limb muscles. Designed for 
use with arthritic, orthopedic, and neurological 
patients, as well as amputees, it is anticipated 
that the data collected will assist in the 
evaluation of drug therapy and orthotic and 
prosthetic devices. 

• Clinical Pathology: Work with the Clinical 
Center's Clinical Pathology Department centers 
on the automation of laboratory procedures that 
have resisted computerization by conventional 
methods. A major innovation is the development 
of a microcomputer-based system to facilitate 
the recording of differential white cell counts. 

• Pulmonary Laboratory: Automated methods for 
evaluating pulmonary function using such 
procedures as measurement of pulmonary 
compliance and work of breathing and exercise 
testing on a treadmill have been developed. The 
goals are improved speed and accuracy in test 
performance and evaluation by the NHLBI staff 
who provide this service in the Clinical Center. 

In the laboratory automation area, CSL work on the 
Distributed Laboratory Data Acquisition and 
Control System (DLDACS) for the laboratories of 
NIADDK in Building 2 was largely completed this 
year. Planning and design of this system began in 
1 976; it replaces a centralized data acquistion and 
processing system developed by CSL over a decade 



I 



ago. The system is designed as a local computer 
network consisting of a group of microcomputers 
that communicate with a host processor by way of a 
front-end communications processor that, in turn, 
performs a file store-and-forward function. Each 
satellite microcomputer performs data acquisition 
and control for a single instrument or experiment. 
Acquired data files may be stored locally, however, 
they are normally transferred via the network to the 
host processor. 

Development of DLDACS was phased over a period 
of several years to avoid interfering with the 
research of users of the old, centralized data 
acquisition system. In fact, the communications 
processor initially was connected to the old 
Honeywell-516 central computer, and one satellite 
microcomputer was put into operation early in 1979. 
Since then, additional satellites have been added, 
one at a time, in place of the hardwired instrument 
interfaces used in the old system. There are 
currently eight satellite processors in use. This year, 
the final step of replacing the H-516 processor with 
a multiuser DEC PDP-11/70 was made. Because of 
the design of the local network, this major 
accomplishment was completed without requiring 
any changes to the satellite microcomputer software. 

Other large CSL projects in NIH Research 
laboratories during FY82 were: 

• Electron Microanalysis: CSL is collaborating 
with the Biomedical Engineering and 
Instrumentation Branch, DRS, in the 
development of an automated electron 
microanalysis facility for use by NIH scientists. It 
will be used for research into the elemental 
composition of biological specimens and for 
developing new techniques in electron 
microscopy. Some of these, such as electron 
energy loss imaging, can only be done with the 
aid of a computer. 

• FMF Cell Sorters: CSL pioneered data 
management systems for FMF cell sorters at 
NIH. Support for data acquisition, display, and 
analysis is provided for four Becton-Dickinson 
FACS II FMF/Cell Sorters. Two additional FMF/ 
Cell Sorters used at the Naval Medical Center in 
collaborative programs with NCI are also 
supported. 

• Molecular Interactions System: A 
microcomputer-based data system supervises 
the acquisition and processing of information 
from an analytical ultracentrifuge and a circular 
dichroic spectropolarimeter used in NHLBI to 
investigate the interactions between human 
lipoprotein subunits. Capabilities include 
acquisition, display, and preprocessing of data 



from the ultracentrifuge and spectropolarimeter. 
After limited local processing, files are 
transferred to the central facilities' PDP-10 for 
further processing. 
• Animal Management: The Small Animal Section 
of DRS supplies large numbers of rodents to 
NIH investigators and also serves as an 
international genetic resource. CSL is assisting 
the Section in the design and procurement of a 
computer system to improve their colony 
management and breeding research programs. 
CSL also invested considerable effort this year in the 
development of new computer-related technology to 
support biomedical and clinical research. 

The Medical Information Technology Project, now 

in its third year, achieved a major milestone this year 
with the installation of a computer in a physician's 
office for field trials. This project involves research 
into source data automation techniques for patient- 
physician encounters in primary care settings. 

One outgrowth of the project is a system that (under 
the physician's supervision) automatically produces 
prescriptions and advice on followup home care, 
both of which are keyed to diagnosis. It is this 
prescription and advice system, initially tailored for 
use in a dermatologic practice, that is undergoing 
field trials. 

In operation, the physician gives the system a 
diagnosis for the patient examined, followed by the 
selection of one or more drugs appropriate for that 
diagnosis. The system's drug formulary contains 
medications the physician would normally prescribe. 
Using this formulary, the system displays a default 
formulation for each drug prescribed based on the 
diagnosis and, when necessary, on other factors 
such as the patient's age, sex, or weight. The 
physician may simply approve the default selections 
or he may alter them to fit particular circumstances. 
Once approved, the computer prints the 
prescriptions on blank prescription forms. 

In addition to the drug formulary, the computer 
system's data base also contains numerous generic 
advice 'modules' containing advice on followup 
home care, additional information about a drug and 
its use, and disease information. Based on the 
diagnosis and the medication regimen prescribed, 
the computer, again subject to the physician's 
supervision, selects from among these advice 
'modules' and tailors them to the situation at hand. 
As is the case with prescriptions, the computer then 
prints them for the patient. 

The computer system just described benefits both 
the physician and the patient. The physician's time is 



saved because the computer produces accurate, 
legible prescriptions, thus eliminating the need to 
write them in longhand. Furthermore, summaries of 
the prescriptions and instructions to the patient can 
be produced for inclusion in the patient's record. The 
patient benefits by receiving detailed printed 
instructions about his disorder and any drugs 
prescribed for him. Thus, problems resulting from the 
patient forgetting or misunderstanding instructions 
given to him orally by the doctor may be avoided. 

The Image Processing System, another area of 
advanced technology, involves the implementation of 
both general-purpose and special-purpose hardware 
and software to meet the growing image processing 
requirements of the NIH community. Until recently, 
this effort was conducted using the Evans and 
Sutherland facility at DCRT, however, a new system 
using state-of-the-art equipment has been designed 
and is being procured. It is expected that a special 
application of this technology will be to provide the 
NIH Clinical Center with a medical imaging network. 

Future Plans/Trends 

FY83 can be expected to present an increased 
demand for computers in laboratory and patient care 
settings. More complex research goals of biomedical 
research investigators point to a greater need for 
automation in the laboratory. Technological 
developments in large-scale circuit integration 
continue to lead to lower costs and smaller sizes for 
computers. The current popularity of 'personal' 
computers is resulting in greater awareness on the 
part of NIH scientists of the potential benefits of 
computers. Concurrently, in common with many 
other organizations, CSL is faced with contracting 
personnel and budgetary resources. 

In response to the challenge imposed by this 
conflicting set of circumstances, CSL expects to 
maintain high quality engineering and laboratory 
computer support to NIH programs by continuing 
policies developed in the past for managing resource 
issues. For example, CSL staff will continue to be 
deployed on projects promising maximum impact to 
the NIH community-those that serve a significant 
number of scientists, affect the quality of patient 
care, or represent general-purpose developments. 



Finally, a trend toward greater emphasis on software 
engineering, begun several years ago with the 
introduction of microprocessors, is expected to 
continue with concomitant improvements in 
productivity. Many of CSL's engineering design 
functions-instrument interfaces, data acquisition 
devices, special signal processors-are now 
accomplished with microcomputers. Because of the 
adaptability of software, these new design concepts 
readily meet the changing needs of research 
programs. 



Publications and Presentations 



Barrett, W.: An Iterative Algorithm lor Multiple Threshold Detection. Pro- 
ceedings ol the IEEE Computer Society Conference on Pattern Recog- 
nition and Image Processing. 1981, pp. 273-278. 

Bonow, R., Ostrow, H., Rosing, D. Lipson, L., Kent. K.. and Allen, S.: 
Verapamil-induced Changes in Ventncular Volume and Diastolic Func- 
tion in Hypertrophic Cardiomyopathy. Mechanisms for Reduced Subval- 
vular Gradient and Improved Symptoms. Circulation 64;IV-11, 1981. 

Bonow, R., Ostrow, H., Rosing, D., Lipson, L., Kent. K.. Allen, S., Bachrach, 
S., Green, M., and Epstein. S.: Scintillation Probe Measurement ol Left 
Ventncular Cardiomyopathy. Circulation &4A\/-35. 1981. 

Fiori, C, Gorlen, K., and Gibson, D : Comments on the Computenzation ol 
an Analytical Microscopy Microscope, Proceedings of the Thirty-ninth 
Annual Meeting of the Electron Microscopy Society of America. Baton 
Rouge, Claitor's Publishing Division, 1981, pp 246-249 

Fion, C. E.. Myklebust, R. L., and Gorlen, K. E.: Sequential Simplex: A 
Procedure for Resolving Spectral Interlerence in Energy Dispersive X- 
ray Spectrometry. Energy Dispersive X-ray Spectrometry Gaithers- 
burg, MD, National Bureau of Standards, 1979, pp. 233-272, 

Fiori, C. E., Swyt, C. R.. and Gorlen, K E.: Application of the Top-Hat Digital 
Filter to a Nonlinear Spectral Unraveling Procedure in Energy-Disper- 
sive X-ray Microanalysis. Microbeam Analysis San Francisco, San 
Francisco Press, Inc., 1981, pp. 320-324. 

Gershon, N., Porter, K., and Trus, B.: The Microtrabecular Lattice and the 
Cytoskeleton. Their Volume, Surface Area and the Diffusion of Mole- 
cules Through It. The Aharon Katizir-Katchalsky Symposium on Biologi- 
cal Structure and Coupled Flows. Rehovot, Israel, June 6-11, 1982. 

Hagins, W., Foster, M., George, J., and Trus, B : Combined X-ray Microana- 
lysis and Radioautography of Diffusible Elements in Aqueous Suspen- 
sions of Cells and Cell Fragments. Proceedings of the Microbeam 
Analysis Society August 1982 

Keogh. B., Gadek, J., Pnce, D., Nadel, L., and Crystal, R.: Remarkable 
Similanties in Exercise Gas Exchange Parameters in Markedly Dispa- 
rate Diseases: Compansons Between Idiopathic Pulmonary Fibrosis 
and 1 -Antitrypsin Deficiency American Review of Respiratory Disease 
125:257, 1982. 

McGee, P , Trus, B., and Steven, A.: Techniques to Evaluate the Perform- 
ance of Scanning Microdensitometers in the Digitization ol Electron 
Micrographs. Micron Journal May 1982. 



Nadel, L. D.: Automated Pulmonary Analysis by an Online Microcomputer. 
In Nair, S. (Ed.): Computers in Cntical Care and Pulmonary Medicine 
(in press). 

Nadel, L D Breath-By-Brealh Pulmonary Exercise Testing Using an Online 
Microcomputer First IEEE Computer Society International Conference 
on Medical Computer Science, Computational Medicine (MEDCOMP 
82) Philadelphia, September 25. 1982 

Nikodem, V., Trus, B . and Rail, J : Two-dimensional Gel Analysis of Rat 
Liver Nuclear Proteins after Thyroidectomy and Thyroid Hormone 
Treatment. Proceedings of the National Academy of Sciences. USA 
78:4411-4415, 1981 

Piez. K . and Trus. B.: A New Model for the Packing ol Type-I Collagen 
Molecules in the Native Fibril. Bioscience Reports 1 Great Bntain. 
1981. pp. 801-810 

Steven. A , Senwer, P., and Trus, B : Molecular Packing in Bacteriophase T7 
Capsid Visualized at 2 5 nm Resolution in Computer-Processed Elec- 
tron Micrographs Eighth Biennial Conference on Bactenophase As- 
sembly Falls Creek Park, Tennessee, May 9-14, 1982 

Steven, A , Trus, B , Pultz, C ., and V>/urtz, M.: The Molecular Organization 
of Beet Necrotic Yellow Vein Virus Journal of Virology New York, 
Academic Press. 1981, pp 428-438 

Tate, R , Schullz, A . and Osborne, J Computer-Assisted Analysis of apo- 
lipoprotein subunit interactions Federation Proceedings 41:874, 1982. 

Trus, B , and Elzinga, M : Computer Modeling of a 17.000 Dalton Fragment 
of Myosin Structural Aspects of Recognition and Assembly in Biologi- 
cal Macromolecules. Rehovot. Israel, 1981, p 361 

Trus, B , Serwer, P , and Steven, A Capsid Fine Structure ol Bacteriophase 
T7 Determined by Image Processing ol Electron Micrographs Tenth 
International Congress on Electron Microscopy Hamburg, Germany. 
August 17-24, 1982. 

Trus, B . and Steven, A : Computer Processing ol Electron Micrographs of 
Periodic Biological Specimens Washington Society of Electron Micros- 
copy Annual Picture Meeting Bethesda, MD. May 6, 1982, 

Trus, B , and Steven, A : Digital Image Processing ol Electron Micrographs- 
The PIC System. Journal of Ultramicroscopy 6:383-386. 1 981 



Laboratory of 
Applied Studies 



Eugene K. Harris, Chief 



Functions 

The Laboratory of Applied Studies (LAS) has three 
main purposes: 

1 . in collaboration with biomedical scientists, to 
apply mathematical theory and computing science to 
the construction, testing, and improvement of 
mathematical models of physiological processes-- 
particularly reaction-diffusion kinetics, transport of 
substrate to tissues, and the control of metabolism 
within cells and tissues; 

2. in collaboration with clinicians, to develop and 
apply mathematical or statistical theory and special- 
purpose computing procedures (analog or digital as 
required) to facilitate research projects aimed at 
improving the diagnosis of disease and assessment 
of treatment; 

3. to engage in independent research in applied 
mathematics, statistics, and computer systems 
necessary to provide a sound theoretical basis for 
collaborative studies, and to insure that state-of-the- 
art mathematical and computational methods are 
available as research tools at NIH. 

Two sections carry out these primary LAS functions: 

Applied Mathematics Section"AMS"(John E. 
Fletcher, Ph.D., Chief). This staff of five includes 
specialists in applied mathematics, computer 
science, biomathematics, and medicine. 

Medical Applications Section-MAS"(James J. 
Bailey, M.D., Chief). This five-member staff includes 
physician-scientists, electronic engineers, and 
computer systems analysts. 
The Chief, LAS, is a biostatistician with training in 
public health and the basic medical sciences. 



Scope of Work 

The Laboratory of Applied Studies works on projects 
in basic and clinical biomedical science. Largely, 
these involve collaboration with other groups at NIH, 
elsewhere in the U.S.A., and abroad. The 
collaborating investigators this year included: 



• biochemists and ptiarmacologists at NIH, at the 
Medical College of Virginia, and at other 
universities in the U.S.A. and in France working 
on models for receptors of drugs or other 
ligands, on the kinetics of enzymes in 
membranes, and on other problems in tissue 
metabolism 

• physiologists and chemical engineers in the 
U.S. A and Europe studying the transport of 
substrate within the microcirculation and the 
regulation of tissue perfusion 

• clinicians in the cardiology, pulmonary, and 
hematology branches of NHLBI; in the arthritis 
and rheumatism branch of NIADDK; and in the 
medical intensive care unit and the departments 
of diagnostic radiology and diagnostic imaging of 
the Clinical Center 

• clinical chemists and pathologists at NIH 
(Clinical Pathology Department, Clinical Center) 
and elsewhere in the U.S.A., in Europe, and in 
Japan engaged in the collection and study of 
reference values in laboratory medicine 

• electrocardiologists and biomedical engineers in 
the U.S.A., Canada, and Europe concerned with 
improved algorithms for computer-based 
interpretation of ECG's and evaluation of ECG 
interpretative programs. 

Highlights of the Year's Activities 

Work continued in computer-based studies of 
pulmonary gas exchange during exercise in patients 
with respiratory disease. Despite unfortunate delays 
due to problems with the vendor-supplied software, 
which have now been resolved, considerable 
progress can be reported in this collaborative project 
with the Pulmonary Branch and the Clinical 
Hematology Branch, NHLBI. Interfaces to the 
exercise equipment to enable online computer 
control have been designed and built by Drs. R. 
Burgess and E. Pottala, while system programming 
for realtime control of these devices and for analysis 



of accompanying data on C02 and 02 pressures 
and content is being completed by IV1. Norton. 
Preliminary studies of healthy subjects have begun. 

Dr. A. Albert, Fogarty International Research Fellow 
In LAS, has contributed to the theory of dynamic risk 
assessment in acute disease by creating very 
general, yet practical, methods for sequential 
analysis of time-dependent multivariate measurement 
vectors obtained during the course of the patient's 
illness. Combining these methods with discriminant 
function techniques enables daily reassessment of 
probable patient outcome. The procedures have 
been applied sucessfully to patients under intensive 
care following myocardial infarction. In addition. Dr. 
Albert has published a generalized theory and 
method of computing multivariate likelihood ratios for 
combinations of discrete and continuous variables. 
This work represents a substantial advance in the 
calculation of diagnostic probabilities (predictive 
values of specified illnesses). 

Theoretical work by Dr. M. Bieterman on the 
adaptive finite element method for the solution of 
reaction-diffusion equations resulted in new software 
routines for the efficient cost-effective solution of 
many of the complex systems of partial differential 
equations that arise from biological models. These 
routines, known as Femoll are now available on the 
NIH central computer systems. 

Studies by B. Bunow and E. Pottala of network 
modeling languages have demonstrated that network 
models are feasible for use as biological simulators. 
Although computer times for these models on the 
NIH central system are presently excessive, their use 
on dedicated systems such as the VAX or similar 
machines has established their utility. Presently, 
interested NIH scientists are being instructed in 
network methods, and exploratory applications are 
underway in collaboration with NIH researchers on 
problems of nerve conduction and facilitated 
diffusion in tissues. One advantage of these 
simulations is that a functional rather than a 
mathematical description of the biological process 
suffices as a requirement to initiate study of its 
stimulus-response characteristics. 

A comprehensive report summarizing Dr. J. 
Fletcher's past decade of research on the analysis 
and interpretation of equilibrium binding data through 
mathematical models has been completed and is 
being distributed to interested scientists worldwide. 

During FY82 LAS staff members participated in 
various teaching and consulting, or advisory, 
activities. 



J. Fletcher continued to serve as Chairman of the 
Mathematics and Computer Science Departments, 
Foundation for Advanced Education in the Sciences. 

J. Bailey continued as a member of an NHLBI site- 
visiting team concerned with computer analysis of 
exercise ECG's. He also serves as consultant on 
common standards for quantitative 
electrocardiography for a program in medicine and 
public health, sponsored by the European Economic 
Community. 

E. Harris continued to be a consultant in applied 
statistics to the Food and Drug Administration's 
Division of Medical Devices and Diagnostic Products. 
Dr. Harris also serves as consultant statistician to 
the College of American Pathologists and to the 
International Federation of Clinical Chemistry (Expert 
Panel on the Theory of Reference Values), and is a 
member of the Board of Editors of Clinical 
Chemistry. 

Future Plans 

Testing of the computer-controlled system for 
measuring pulmonary gas exchange in exercise will 
continue on healthy volunteer subjects. Studies to 
evaluate cardiorespiratory abilities in patients and 
controls will be specified in cooperation with the 
Pulmonary and Clinical Hematology Branches of 
NHLBI. A new project in cooperation with the 
Department of Critical Care Medicine, CC, to 
investigate dysfunction in neurologically impaired 
patients will be pursued through development of 
microcomputer-based methods for analysis and 
display of evoked potentials. 

Utilizing the newly upgraded DeAnza image 
processing system, a joint study with the clinical 
neuropharmacology laboratory, NIMH, will continue 
to develop theory and methods for interpreting 
electron energy loss spectra from ultracellular 
biological specimens, particularly in the study of 
dense bodies in electron micrographs of blood 
platelets. 

The analysis of the signal/noise characteristics of 
various parameters of regional ventricular wall 
motion will continue jointly with the Nuclear Medicine 
Department, CC, and the Cardiology Branch, NHLBI, 
in an effort to improve differential diagnosis of 
coronary artery disease and other cardiac 
myopathies. 

A major effort will be made to facilitate the 
conversion of network models simulating biological 
processes into forms compatible with languages 
such as MLAB to enable the use of powerful data- 



fitting algorithms. Training of and collaboration with 
NIH scientists using network simulation modeling will 
continue. In the area of numerical analysis, the 
current program for approximate solution of partial 
differential equations (FEM0L1) will be extended to 
sets of 3 or more equations, enabling it to be of 
more general use to NIH scientists using 
biomathematical models. 

Statistical theory developed to support the 
calculation of reference values for the differences 
between two or three successive measurements of 
blood constituents will be published and its 
application extended through collaborative studies of 
selected patient groups. Investigation of relative 
sensitivities of multivariate and univariate reference 
ranges will progress using clinical chemistry data and 
followup diagnostic information from a large health 
maintenance program. 



Publications and Presentations List 

Albert, A.: Atypicalily indices as reference values (or laboratory data Amer. 
J Clin. Pathol 76: 421-425. 1981 

Albert, A : On the use and computation of likelihood ratios in Clinical 
Chemistry Clin- Chem. 5: 1113-1119, 1 982 

Albert, A,, Chapelle, JR., Heusghem, C, Kulbertus, HE, and Hams, E.K.: 
Evaluation of risi< using serial laboratory data in acute myocardial in- 
farction In Heusghem, C. Albert, A., and Benson, E S. (Eds): Ad- 
vanced Interpretation of Clinical Laboratory Data. Nev» York, Marcel 
Dekker (in press). 

Bailey J J , Berson, AS. Jackson, L K , Milliken, J A., Stevens, J.M.. Toian, 
G D, and Wolf, H K : Evaluation Methodologies for ECG diagnostic 
systems In Bonner, RE., and Pryor, T A (Eds): Computerized Inter- 
pretation of the ECG VI New York. Engineering Foundation (in press). 

Bunow, B: All things flow and change-some thoughts on the role of 
reaction and transport in biology J Wash Acad Sci (in press). 

Bunow, 8 : Turing and the physico-chemical basis of biological patterns. In 
Prewitt, J (Ed): IEEE Turing Memorial (in press) 

Bunow, B , and Mikulecky, DC : On the feasibility of using flux meas- 
urements to distinguish among active transport models. Polish Winter 
School ol Membrane Transport (in press). 

Burgess, R C: An instrument lo add evoked potential capability to the 
standard electroencephalograph. EEC and Clin Neurophysiol. 53: 33, 
1981. 

Evans, W.J., McCourtney, J.E., and Shrager, R I : Titration Studies of Phytic 
Acid. J. Am Oil Chemists' Soc. 59: 189-191, 1982. 

Fletcher, J E The Analysis ol Equilibrium Binding Data by the Fitting of 
Models. May 1982, 102 pp. 

Fletcher, J.E , and Schubert, R,W.: The Theoretical Prediction of Substrate 
Levels and Their Histograms in Cell Free Perfused Tissues Proceed- 
ings of the International Meeting of the Society of Oxygen Transport to 
Tissue (ISOTTJ (in press). 

Harns, E K : Further applications of lime series analysis to short senes of 
biochemical measurements. In Grasbeck. R., and Alstrom, T. (Eds): 
Reference Values in Laboratory Medicine. Chichester, U.K., John 
Wiley & Sons, 1981, pp 167-176. 

Harris, E K : Regression, least squares, and correlation. In Seligson, D., 
M.D, (Ed): Handbook of Clinical Chemistry (in press). 

Harris, E.K.: Use of statistical models lo delect sub|ect-specific changes. In 
Yasaka, T. (Ed): Proceedings of International Conference on Automat- 
ed Multiphasic Health Testing & Services Amsterdam, Excerpta 
Medica, 1981, pp. 35-44. 

Harns, E.K., Yasaka, T , Horton, MR., and Shakarji, G.: Companng multivar- 
iate and univariate subject-specific reference regions for blood constitu- 
ents in healthy persons. Clinical Chemistry 26: 422-426, 1982 

Macfarlane, P.W., Chen, C.Y., and Bailey, J.J: A companson ol point 
sconng techniques for the diagnosis of LVH. In Macfarlane, P.W. (Ed): 
New Frontiers in Eteclrocardiology (in press). 

Shrager, R.I , and Hendler. R.W.: Titration ol Individual Components in a 
Mixture with Resolution ol Difference Spectra, pKs, and Resox Transi- 
tions. Anal Chem. 544: 1147-1152, 1982. 



Physical Sciences 
Laboratory 



George H. Weiss, Chief 



Function and Scope of Work 

The Physical Sciences Laboratory carries out 
research in support of NIH programs and in pursuit 
of its mennbers' interests in the areas of physics, 
physical chemistry, applied mathematics, and applied 
computer technology, it has an active research 
program in its areas of expertise and provides 
consulting services to other NIH scientists in 
theoretical physics, chemistry, and applied 
mathematics. Members of PSL develop theory and 
often develop instrumentation for biomedical 
experiments. 

The PSL staff consists of seven professionals who 
work in the areas of: 

• biophysics 

• light scattering, as applied to problems in 
determining structure and function of biologically 
interesting gels and other forms of matter 

• nuclear magnetic resonance, as applied to 
kinetic and configurational properties of 
molecules 

• the use of image processing techniques to 
interpret electronmicrographs of membrane 
structure 

• the physical chemistry of actin, and 

• applied mathematics in areas suggested by 
investigations at NIH. 

All of the members of the Laboratory collaborate 
with scientists both on and off the NIH campus. For 
example, crystallographic data bearing on 
intermolecular forces is generated at Brock 
University, Canada, in collaboration with members of 
PSL in a joint theoretical and experimental project in 
that general area. A new project has been initiated 
together with members of the Clinical Center and the 
Computer Systems Laboratory, DCRT, on the use 
and interpretation of ultrasonic data in the study of 
tongue motion in different parts of speech. 



FY82 Accomplishments 

A joint project between members of PSL and Brock 
University has succeeded, for the first time, in 
directly measuring the force between biological 
macromolecules. This has provided a good picture of 
how the force between parallel DNA helices behaves 
in response to changes in distance, as well as 
elucidating the effect of structure on this force. An 
extension of this work will allow the determination of 
important thermodynamic parameters of proteins. 

The applicability of two-dimensional Fourier 
Transform Spectroscopy to the determination of 
kinetic parameters for enzyme reactions was 
demonstrated for the first time, in a study of the 
isomerization of glucose-6-phosphate. This work 
required not only the techniques of physical 
chemistry, but also the development of a suitable 
theory as well as the study of the effects of 
instrumental noise on the calculation of parameters. 

Work has continued on light scattenng techniques as 
applied to the determination of mechanical rigidity 
and internal viscosity of polymer gels. The current 
emphasis is on the structure and properties of 
polyacrylamide gels because of their technological 
importance, but studies are also underway on gels 
formed from glycoproteins and clots formed from 
reconstituted human plasma. This research requires 
expertise both in the development of a considerable 
body of theory to interpret experimental data as well 
as experimental skill in the modification of available 
equipment to perform the relevant measurements. 

A theory of errors of parameter estimates in positron 
emission tomography has been developed and will 
be incorporated into programs in use on the PET 
scanner. This theory is part of a continuing study of 
the optimal design of experiments that includes 
present analyses of NMR and chromatography 
experiments in addition to the PET study. 



11 



Dr. Adrian Parsegian has been elected President of 
the Biophysical Society. Dr. George Weiss has been 
appointed Biostatistics Editor of Cancer 
Investigations. 

PSL partially sponsored an international meeting on 
Random Walks in Physics and Biology, which was 
held at the National Bureau of Standards from June 
28 to July 1, 1982. Dr. George Weiss was Chairman 
of the organizing committee for the meeting. 



Future Plans/Trends 

Research in PSL will continue along lines already 
initiated. A new joint project with the Speech 
Pathology Department of the Clinical Center and 
CSL will study the combination of ultrasound and 
image processing techniques on the study of tongue 
position in speech. 

With the recent access to an x-ray spectrometer on 
the NIH campus, Drs. Parsegian and Lee will be able 
to extend their measurements of intermolecular force 
constants to study nucleic acids. A large project is 
being planned on the use of molecular graphics 
programs to investigate molecular contacts in protein 
crystals. 

Further studies will be undertaken in the relation of 
instrumental noise to the precision of parameter 
estimates in NMR measurements. These studies 
should point the way to optimization of such 
experiments. Dr. Ferretti will be joining NHLBI in the 
near future but collaboration on these problems will 
continue. 



12 



Publications 



Aizenbud, B., and Gershon, N. D.: Diffusion of molecules on biological 
membranes of nonplanar form--a theoretical study. Biophys. J. (in 
press). 

Aizenbud B., and Gershon, N. D.: Diffusion of Molecules on Microvillous 
Biological Membranes. In Perelson, A. C, DeLisi, C, and Wiegel. F. W. 
(Eds.): Cell Surface Phenomena. New York, Marcel Dekker (in press). 

Aizenbud, B. M., and Gershon, N. D.: Hydrodynamic equations and VH light 
scattenng from viscoelastic (solid-like and fluid-like) systems. Pheno- 
menological approach. Physica A 107:126-142, 1981, 

Aizenbud. B , and Gershon. N. D,: Hydrodynamic equations and VH light 
scattering from viscoelastic (solid like) systems. II. Molecular approach. 
Physica A (in press) 

Brenner, S. L., and Korn, E. D.: Stimulation of actin ATPase activity by 
cytochalasins provides evidence for a new species of monomeric actin. 
J. Biol Chem. 256:8663-8670. 1981. 

Chen. S-H., Chu. B.. and Nossal. R. (Eds): Scattenng Techniques Applied 
to Supramolecules and Nonequilibnum Systems NATO ASI Series B. 
Vol. 73. New York. Plenum Press. 1981. 

Ciarkowski. J. E., FerrettI, J. A., and Marshall. G. R.: Comparative conforma- 
tional studies of angiotensin II and two sterically constrained analogs 
by 600 MHz proton spectroscopy. J. Am. Chem. Soc. (in press). 

Jacobson. L., and Ferretti. J. A.: The determination of a phosphorus-phos- 
phorus nuclear Overhauser enhancement by two-dimensional magne- 
tization exchange spectroscopy. J. Amer. Chem. Soc. (in press). 

Kiefer. J. E.. and Weiss. G. H.: A comparison of two methods for accelerat- 
ing the convergence of Fourier series Comp. and Math. 7:327-336. 
1981. 

Lis, L. J., McAlister, M., Fuller, N , Rand, R. P, and Parsegian, V A.: 
Interactions between neutral phospholipid bilayer membranes. Biophys. 
J. 37:657-666. 1982. 

Lis, L. J., McAlister, M.. Fuller, N., Rand, R. P., and Parsegian, V. A.: 
Measurement of the lateral compressibility of several phospholipid bi- 
layers Biophys J. 37:667-672. 1982. 

Marks. T. J.. Pohl. L. R., Gillette. J. R.. Hong. M.. Highet. R. J., Ferretti, J. 
A., and Henson. J. A.: Stereoselective formation of bromobenzene 
glutathione conjugates. Chem. Bio. Interactions (in press). 

Nossal, R.: Laser Light Scattering. In Methods of Experimental Physics. 
New York, Academic Press, 1982, pp. 299-336. 

Nossal. R.: Ouasielastic Light Scattenng from Polymer Gels. In Chen, S-H., 
Chu. B.. and Nossal. R. (Eds.): Scattering Techniques Applied to Supra- 
molecules and Nonequilibnum Systems. New York. Plenum Press, 
1981. pp. 301-320. 

Nossal. R.: Stochastic aspects of biological locomotion. J. Stat Phys. (in 
press). 

Nossal. R., and Jolly, M.: Shear waves and 'internal viscosity' in cylindncal 
gels. J. AppI Phys. (in press). 



Parsegian, V. A. (Ed): Protein-Lipid Interactions in Membranes. The Rocke- 
feller University Press, 1982, 401 pp 

Rish, B. L., Caveness, W. L., Dillon. J. D.. Kistler. J. P.. Mohr. J. P.. and 
Weiss. G. H : Analysis of brain abscess following penetrating craniocer- 
ebral injuries in Vietnam /Veorosurge/y 9:535-541. 1981 

Rubin, R J., and Weiss. G. H : Random walks on lattices: The problem ol 
visits to a set of points revisited J. Math Phys. 23:250-253. 1982. 

Shapiro. J R . Pikus. A.. Weiss. G H , and Rowe. D W.: Hearing and 
middle ear function in osteogenesis imperfecta. J. Am. Med. Assoc 
247:2120-2126. 1982. 

Weiss. G H : Optimal parameters for the measurement of the half-width of 
a Gaussian peak. Sep Sci <S Tech (in press) 

Weiss. G H.: Random walks and their applications. Amer. Sci (in press). 

Weiss. G H.. Caveness. W F . Einsiedel-Lechtape. H., and McNeel, M L.: 
Life expectancy and causes of death in a group of head-in|ured veter- 
ans Arch. Neurol (in press). 

Weiss, G H , Feeney. D M., Caveness, W. F., Oillon, J. 0.. Kistler. J. P„ 
Mohr, J P , and Rish, B L : Prognostic factors for the occurrence of 
posttraumatic epilepsy. Arch. Neurol (in press). 

Weiss, G. H., Ferretti, J. A., Kiefer. J A.: A study o( precision in the 
measurement of chemical shifts. J. Mag Res. 46:69-83. 1982 

Weiss. G. H.. and Rice. J.: A combinatonal problem in pharmacology J. 
Math Biol (in press) 

Weiss. G. H.. and Rubin. R. J.: Random walks: Theory and selected appli- 
cations. Adv Chem. Phys. (in press). 

Weiss, G. H., and Shiesinger. M. F : On the expected number of distinct 
points in a subset visited by an N-step random walk. J. Slat Phys. 
27:355-363. 1982. 

Weiss. G H . Talben. A., and Brooks. R A.: The use of phantom views to 
reduce CT streaks due to insufficient sampling. Phys in Biol and Med 
(in press). 



13 



Laboratory of Statistical and 
l\/lathematical l\/lethodology 



James E. Mosimann, Chief 



Function and Scope of Work 

The Laboratory of Statistical and Mathematical 
Methodology (LSM) combines research in 
mathematical statistics, mathematics, computer, and 
information science with collaboration and service in 
these areas to NIH researchers and administrators. 
LSM staff interact with all NIH Institutes, with other 
Federal agencies outside DHHS, and with biomedical 
researchers worldwide. 

In addition to the position of chief, the Laboratory 
has fourteen full-time professional positions 
distributed among four sections: 

• The Statistical Software Section (SSS) 
provides consultation to and collaboration with 
NIH researchers and administrators in all 
computational aspects of biomedical data 
analysis, including selection and support of large 
systems/packages. Three specialists in scientific 
programming are led by a computer systems 
analyst whose specialty is statistics. 

• The Biomathematics and Computer Science 
Section (BCS), directed by a mathematician, 
performs independent research and provides 
consultation and collaboration in the specialties 
of its four computer and mathematical scientists. 

• The Statistical Methodology Section (SMS) 
works closely with the Statistical Software 
Section. Two professionals in mathematical 
statistics provide biostatistical consultation and 
do independent research. 

• The Medical Information Science Section 
(MIS) investigates and develops methods for 
application of information and computer science 
to medical language data processing. Two 
computer specialists work under the direction of 
a computer systems analyst who is an expert in 
computational linguistics. 

A major part of LSM activity is the offering of 
statistical and mathematical systems/packages to 
the NIH user community. LSM accepts responsibility 



for evaluation of new systems/packages and their 
suitability for NIH. When it offers a system/package 
to the NIH community, LSM makes three basic 
commitments: 

1 . Maintenance of the package, with adequate 
documentation, through NIH computer system 
changes, system/package updates, and corrections. 

2. Rapid response to queries concerning user 
access to a system/package program, including job 
control language and program parameters. 

3. Assistance in interpretation of results. 
During this year, as in the past year, the Statistical 
Software Section of LSM maintained the following 
systems/packages and programs on the IBM 370 
system of the DCRT Computer Center: 

• BMD and BMDP, Biomedical Computer 
Programs, UCLA. 

• SPSS and SCSS, Statistical Package for the 
Social Sciences, SPSS, Inc. 

• SAS, SAS/GRAPH and SAS/ETS, Statistical 
Analysis System, SAS Institute, Inc. 

• P-STAT Statistical Package, P-STAT, Inc. 

• IMSL, International Mathematical and Statistical 
Libraries, IMSL, Inc. 

• MSTAT1, Collection of Mathematical and 
Statistical Programs, DCRT. 

In FY82 alone the SSS staff responded to over 
5,500 quehes concerning use of these packages. 
Also during this year, every system/package went 
through at least one major update. 

The Biomathematics and Computer Science Section 
maintains several systems/packages and specialized 
systems on the DECsystem-10 of the Computer 
Center. Foremost in use is the interpretive system 
MLAB, designed (by LSM scientists) for 
biomathematical modeling. C-LAB, previously an 
independent package for pattern recognition and 
cluster analysis, has now been incorporated into 
MLAB. The Unified Generator Package, written and 
maintained by a BCS staff member, is on DCRT's 
IBM System 370. 



15 



Uses per month of 

Statistical packages supported by LSIVI' 



58800 
54600 - 
50400 - 
46200 - 
42000 - 
37800 - 
33600 - 
29400 - 
25200 - 
21000 - 
16800 - 
12600 

8400 

4200 H 




JUN 75 JUN 76 JUN 77 JUN 78 JUN 79 JUN '80 JUN '81 JUN '82 



'Packages supported by the Statistical Software Section only. Does not include 
packages supported by the Biomathematics and Computer Sciences Section. 



As a result of LSM's policy of not only supporting the 
use of these systems/packages but also aiding in 
the interpretation of their output, the statisticians of 
the Statistical Methodology Section provide 
consultation over a wide range of scientific fields. 
Some very brief consultations are very successful 
because there is a known answer to the question at 
hand. Other consultations involve extensive time and 
statistical/mathematical/computer science research 
as well. 



Research projects in LSM vary widely, from studies 
of natural language processing for medical 
information systems and studies of efficient 
algorithms for information retrieval to studies in 
mathematics and statistical methodologies for 
biomedical applications. 



16 



FY82 Accomplishments 

FY82 was LSM's eighth year as a separate entity 
within DCRT. The volume of its computational and 
consultative services continued to expand; its 
research activities decreased slightly, with one 
project terminated. 

Computation 

In FY82, LSM continued to expand teaching and 
documentation for supported systems/packages. 
LSM taught eight introductory courses for SAS, two 
for SPSS, and two for BMDP. In addition, two 
introductory courses and one advanced course were 
taught for MLAB, plus an introductory course for 
BRIGHT (a package supported by DMB) and two 
courses on computer graphics at NIH. The tenth 
edition of the MLAB Reference Manual is now being 
printed. The DCRT Mathematical and Statistical 
Program Manual was updated in FY82. 

BCS staff contributed to the implementation of 
overlay facilities on the DECsystem-10, and MLAB 
was redesigned in a segmented form to use overlay, 
so that software for seldom-used operations is 
loaded into computer memory only when needed. 
MLAB software was substantially enlarged by 
incorporation of C-LAB operators, of new 
OMNIGRAPH character fonts and codes for display 
of mathematical formulas, and of new, more 
informative error messages. 

Consultation, Collaboration, and Research 

As in FY81, LSM consultation and research in FY82 
was closely tied to the use of the computer. Most 
consultations (55 percent) involved statistical advice 
combined with considerable computer use. Others 
(40 percent) involved computer use alone and a 
small fraction (5 percent) involved mathematical or 
statistical advice with only limited computer use. 

In FY82, LSM research, collaborative, and 
consultative efforts merged more closely and were 
less distinguishable among themselves. In a number 
of studies, statistical methodologies were developed 
for, or modified to suit, specific biomedical problems. 

The results of LSM research on simultaneous 
confidence intervals for ratios appeared in the 
Journal of the American Statistical Association in 
1982. A study was completed on the connection 
between statistical and algebraic independence, 
applicable to the sample covariance matrix of 
multivariate data. 

In statistical discriminant analyses--a subject of LSM 
research in collaboration with Dr. J. Darroch, Flinders 
University, South Australia, and Dr. H. Hoffman, 
DRS--methods adapted for size and shape variables 



are being used to study genetic variation in 
laboratory mice with reference to purity of breeding 
stocks. A separate study of independence of size 
and shape variables before and after scale change 
appeared in FY82, along with other LSM studies of 
statistical distributions. A collaborative study was 
undertaken with Dr. P. Turkeltaub (BB/DPB) on 
clinical symptoms and allergic reactions to pollen. 
LSM also participated actively in a study of Chagas 
disease (Dr. F. Neva, NIAID/LPD) and continued 
collaborative work in various studies of 
schistosomiasis (Dr. A. Cheever, NIAID/LPD). 

New studies in linear models were initiated in FY82, 
and two of these have been completed. The first 
gives optimal linear model estimates of variance 
components, while the second presents a solution to 
the multivariate analysis of variance with unbalanced 
data. A collaborative study on the spatial distribution 
of blue cones in the retina with Drs. S. Schein (NEI/ 
CB) and F. de Monasterio (NEI/LVR) is near 
completion. An algorithm related to this work has 
been published. A study of patients with systemic 
lupus erythematosus in collaboration with Dr. T. 
Chused (NIAID/LMI) is near completion. 

In computer science, a study of hashing with 
coalescing lists for information storage and retrieval 
was completed and submitted for publication. A 
method of resolving overlapping spots on two- 
dimensional electrophoretic gels was studied; 
programs implementing the method are under 
development. Studies of equivalence of module 
theories and of classification of Riemannian 
geometries by N-algebras (with possible applications 
to size and shape analysis) continued. A pilot study 
of computer 'reading' of technical text was 
completed. This involved computer translation of 
syntactic and semantic content of a simplified 
chapter of a computer programming textbook into 
operational structures in the procedural language 
PROLOG. 

In medical linguistics, research studies on the 
morphosemantic structuring of medical terms derived 
from Greek and Latin were continued. Previous work 
on computer parsing of medical words into 
morphemes via suffix analysis was extended to more 
than 1,500 terms representing 6 kinds of surgical 
procedure. The results are applicable to the 
construction of dictionaries suitable for information 
retrieval by computer. Also studied were rules for 
morphosemantic transforms useful for computer 
substitution of terms in the automatic encoding of 
medical text. The MIS-developed encoder program 
for automatic assignment of SNOP (Systematic 



17 



Nomenclature of Pathology) codes to surgical 
pathology diagnoses was used on a routine basis 
(Dr. J. Costa, NCI/LP). Collaborative studies on the 
improvement and modification of the SNOP 
dictionary with Dr. Donald E. Henson (NCI/BCPC) 
continued in FY82. An LSM linguist also served as a 
consultant on a machine translation project at 
Georgetown University. 

LSM research on the 'symmetric axis' method of 
describing biological shapes was discontinued this 
year, due to the retirement of the principal 
investigator. Computer software for symmetric axis 
analysis and reconstruction of figures will be 
maintained. Programs and documentation are sent to 
biomedical researchers on request. 

Future Plans/Trends 

No major shift in laboratory service or research is 
anticipated in the coming year. Current levels of 
statistical and mathematical systems/packages 
support, consultation, and user assistance will be 
maintained. Research projects will be continuations 
of those already initiated and reported here. 



Publications 



Buck, J., Buck, E., Hanson, F. E., Case, J. F., Mets, L., and Atta, G.: Control 
of flashing in fireflies IV. Free run pacemaking in a synchronic pterop- 
tyx. Journal of Comparative Physiology 144: 277-286, 1981. 

DeBlas, A. L., Ratnaparkhi, M. V., and (ulosimann, J. E.: Estimation of the 
number of monoclonal hybridomas in a cell fusion experiment. In Vuna- 
kis, H. v., and Lagone, J. J. (Eds.): Immuriocfiemical Techniques (a 
volume of Methods in Enzymology). New York, N. Y., Academic Press 
(in press). 

DeBlas, A. L, Ratnaparkhi, M. V., and Mosimann, J. E.: Estimation of the 
number of monoclonal hybridomas in a cell fusion experiment. Journal 
of immunological Methods ^5: 109-115, 1981. 

Grimes, A. M., Mueller, H. G., and Malley, J. D.: Examination of binaural 
amplification in children. Ear and Hearing 2: 208-210, 1981. 

Hutchinson, G.: A complete logic for n-permutable congruence lattices. 
Algebra Universalis 1 3: 206-224, 1 981 . 

Hutchinson, G.: Exact embedding functors between categories of modules. 
Journal of Pure and Applied Algebra 25: 1 07-1 1 1 , 1 982. 

Knott, G. D.: Graphics Facilities in MLAB. In Sproull, R. (Ed.): Computer 
Graphics, a section of Chang, S. (Ed.): Handbook of Computer and 
Electrical Engineering (in press). 

Knott, G. D.: Fixed-bucket binary storage trees. J. of Algorithms (in press). 

Malley, J. 0.: Simultaneous confidence intervals for ratios of normal means. 
Journal of The American Statistical Association 77: 170-176, 1982. 

Mosimann, J. E., and Malley, J. D.: The Independence of Size and Shape 
Before and After Scale Change. In Taillie, C, Patil, G. P., and Baldes- 
sari, B. (Eds.): Statistical Distributions in Scientific Work. Vol. 4, Models, 
Structures and Characterizations. Dordrecht, Holland, D. Reidel Pub- 
lishing Co., 1981, pp. 137-145. 

O'Connor, M. A.: Invariant metrics on cones. Proc. of the Conference on 
Invariant Metrics and Holomorphic Maps. Rome, Italy, Istituto di Alta 
Matematica F. Severi di C.N.R. (in press). 

Ratnaparkhi, M. V.: On splitting model and related characterization of some 
statistical distributions. In Taillie, C, Patil, G. P., and Baldessari, B. 
(Eds.): Statistical Distributions in Scientific Work, Vol. 4, Models, Struc- 
tures and Characterizations. Dordrecht, Holland, D. Reidel Publishing 
Co., 1981, pp. 357-363. 

Ratnaparkhi, M. V.: Some bivariate distributions of (X,Y) where the condi- 
tional distribution of Y, given X is either beta or unit-gamma. In Taillie, 
C, Patil, G. P., and Baldessari, B. (Eds.): Statistical Distributions in 
Scientific Work, Vol. 4, Models, Structures and Characterizations. Dor- 
drecht, Holland, D. Reidel Publishing Co., 1981, pp. 389-400. 

Roux, J. J. J., and Ratnaparkhi, M. V.: On matrix-variate beta type I distribu- 
tion and related characterization of Wishart distribution. In Taillie, C, 
Patil, G. P., and Baldessari, B. (Eds.): Statistical Distributions in Scientif- 
ic Work, Vol. 4, Models, Structures and Characterizations. Dordrecht, 
Holland, D. Reidel Publishing Co., 1981. pp. 375-378. 

Shapiro, M.: A note on Lee and Schacter's algorithm for Delaunay triangula- 
tion. International Journal of Computer and Information Sciences (in 
press). 

Yaar, I., Shapiro, M., and Pottala, E.: Spectral analysis of the EEG in 
hepatic encephalopathy treated with levodopa. Electroencaphalography 
and Clinical Neurophysiology 52: 61 7-625, 1 981 . 



18 



Data Management 
Branch 



J. Emmett Ward, Chief 



Functions and Scope of Work 

The Data Management Branch (DMB) provides 
advice and assistance to research investigators, 
program officials, and administrators throughout NIH 
in planning for and obtaining computer data 
processing services. In this role the branch is a 
central NIH resource for systems analysis, design, 
and programming. There are currently 47 permanent 
full time employees whose disciplines include 
computer science, mathematics, and statistics. 

DMB staff design and create computer-based data 
management systems that provide practical solutions 
to the unique mix of administrative, scientific, and 
management data processing problems encountered 
at NIH. Each new computer system user is provided 
comprehensive training in all system facilities and 
functions of the system provided by DMB. In addition 
DMB staff teach courses about programming tools; 
provide advice on data management techniques to 
NIH programmers; serve as consultants to the B/l/ 
D's for obtaining and monitoring contracting services 
for computer systems development; and create and 
maintain general purpose, user-oriented 
programming tools to speed building and improve 
operation of applications systems. 

DMB comprises five sections. The Applied Systems 
Programming Section (ASPS) and the Scientific 
Applications Section (SAS) provide general 
computer systems analysis and programming 
services for all of the B/I/D's. The ASPS supports 
general data management, and the SAS handles 
those projects that require scientific data analysis. 

The Data Base Applications Section and the Data 
Base Enchancement and Control Section develop 
and maintain the central administrative data base for 
NIH materiel and financial management. The Clinical 
Support Section develops and maintains the 
Clinical Information Utility as a data base for 
research and patient care in the Clinical Center. 



FY82 Accomplishments 

The Clinical Information Utility is a long term effort 
that, when completed, will provide a unique archive 
of integrated data for use in patient care and 
research. Efforts to date have involved the 
development of software to acquire and to make 
available data from the Medical Information System 
and the individual clinical service activities. The 
integration of these individual data bases has now 
begun. The design of this effort will provide users 
with a single source for most of the clinical service 
data and all of the medical information system data. 
It will allow authorized users to make requests for 
information and will provide either reports or sub-files 
of the data, which will be produced in a format 
acceptable to BRIGHT for further perusal and 
analysis. 

BRIGHT is a user-friendly interactive program 
designed to make data entry, correction, updating, 
manipulation, retrieval, formatting, and printing of 
tabular data bases easier. The first DCRT training 
course for BRIGHT was taught in early December 
1981 and a formal seminar introducing the system to 
the clinical associates was conducted in March 1982 
in the Clinical Center. BRIGHT has had several new 
features added during the fiscal year and it is a 
much more versatile and easily used package for 
data and graphics display, table-making, 
computation, and analysis. 

To assist the Records Processing Section in the 
Medical Records Department of the Clinical Center, 
a tracking system for the medical records audit 
process was designed and implemented. This 
system provides a mechanism for monitoring the 
status of medical records in process and for 
following up with physicians concerning delinquent 
reports. This system has had a significant impact in 
that it has not only reduced the manual effort 
involved in the audit process but also improved the 
organization of the followup procedure. 



21 



The NIH Administrative Data Base is an ongoing 
development project that uses data base technology 
in support of NIH-wide materiel and financial 
management. This project expanded on several 
fronts. A vendor credit system was added to the 
accounts payable procedure; an alphabetic search 
capability was made available for the vendor data 
base; all miscellaneous obligations such as training 
orders, utilities, tuition, work requests, etc., were 
programmed to flow through the ADB; and a facility 
to produce Purchase Orders (SF-147's) online was 
added for reprints and professional services. 

Extending the availability of ADB functions to the B/ 
l/D's is also well under way. in collaboration wth the 
Office of Research Services and the Training 
Assistance Branch, DPM, B/i/D terminal locations 
have been identified and contracts have been 
negotiated for conducting training on delegated 
procurement and receiving. Current plans call for all 
B/I/D's to be making full use of these functions by 
early Fall, 1 982. 

From July 1 982 through February 1 983 the stock 
requisitioning and the central and self service stores' 
inventory systems will be phased in. Full conversion 
to this new set of functions should be in place by the 
end of the first quarter of calendar year 1983. 

Development of the new financial management 
system for the ADB is under way and a detailed 
design document should be available in November 
1982. Plans call for implementing the fund 
certification, fund control, and general ledger 
modules of this system by utilizing off-the-shelf 
software, modified to fit NIH requirements. 

For a detailed review of the many other important 
projects in which the Data Management Branch has 
been involved, please refer to the project reports in 
Volume 2. These projects are too numerous to 
highlight in the summary. 

In the area of general support for NIH activities, 
DMB continued to maintain and teach courses on 
the Inquiry and Reporting System (IRS) and 
MARKIV; to support NIH use of Chemical Biological 
Activities (CBAC) and Biosciences Information 
System (BIOSIS) current awareness searches on a 
biweekly and semimonthly basis, respectively; to 
maintain and distribute the NCI Survival System; and 
to consult with and assist NIH programmers and 
contractors, enabling facile use of DCRT computer 
facilities. 



Future Plans/Trends 

During the next year, the Administrative Data Base 
will expand to include all central inventory functions 
and work will have begun to bring other inventory 
systems (e.g.. Planning and Control Branch, 
Biomedical Engineering and Instrumentation Branch, 
and Pharmacy) under this umbrella. The financial 
management system fund certification, fund control, 
and general ledger modules should be well on their 
way to completion by the end of FY83. Stock 
requisitioning should be installed in all of the B/I/D's 
and open market requisitioning should be ready for a 
pilot test by the end of the next fiscal year. 

The Clinical Information Utility will have a completely 
integrated data base that will provide an effective 
link among the user, the clinical data and the 
analysis. 

Its role as a central resource for computer 
applications development throughout the B/I/D's will 
continue to receive primary support by DMB. 



22 



Publications 

George, D.: Epizootiology of an Outbreak of Mousepox at the National 
Institutes of Health. Laboratory Animal Science 31 :609-61 5, 1 981 . 

Harris, E.K., Yasaka, T., Norton, MR,, and Shakarji, G,: Connpanng Multivar- 
iate and Univariate Subiect-specific Reference Regions for Blood Con- 
stituents in Healthy Persons. Clinfcal Chemistry 28: 422-426, 1982. 

Rodbard, D.. Cole. B., and Munson. P.J.: The Need for Innovative Ap- 
proaches to Radioimmunoassay Quality Control. In Wilson. D. W. (Ed.): 
Quality Control of Radioimmunoassays. 1982. pp. 251-260. 



23 



Computer Center 
Branch 



Joseph D. Naughton, Chief 



Function 

The Computer Center Branch (CCB), the largest 
component of DCRT, designs and operates the NIH 
Central Computer Utility and its associated 
telecommunications facilities; conducts a formal 
computer training program; and provides technical 
documentation, programming support, and 
consultation on the use of computers in support of 
scientific and administrative programs throughout 
NIH. 

Because the Computer Center receives no direct 
appropriation, all services are provided strictly on a 
fee-for-service/cost recovery basis. 

Two large, interconnected, multicomputer facilities-- 
the IBM System 370 and the DECsystem-10--form 
the nucleus of the NIH Computer Utility. Each facility 
is linked by telephone lines to hundreds of remote 
interactive terminals and computers located 
throughout NIH and many other Federal agencies. 
System software is either designed and implemented 
by Computer Center personnel or acquired from 
other sources and adapted to meet the unique 
needs of the NIH biomedical research and 
administrative user community. 

A specialized staff of professional, technical, and 
administrative personnel keep the NIH Computer 
Utility functioning smoothly 24 hours a day. 
Experienced computer systems programmers and 
analysts develop and maintain operating system 
software. They also offer technical consultation, 
design and teach training courses, and write 
technical documentation on the use of the Utility. A 
staff of experienced computer systems technicians 
operates the Computer Utility's hardware and 
telecommunications network, and provides data 
entry services. Systems management professionals 
establish long term program goals and ensure the 
design integrity of the Utility. 

A number of research and development projects are 
also conducted by the Computer Center. These 



include the design and installation of data security 
facilities for over 300,000 online data sets, the 
design and implementation of communications 
networking facilities to make possible the rapid 
exchange of information among research 
investigators, the development of improved graphic 
output facilities, and the exploration of new training 
methods. 

Scope of Work 

The NIH Computer Center plans, designs, 
implements, and operates a large, general-purpose 
Central Computer Utility. This Utility provides a 
variety of computational services in support of a 
dynamic and diverse user community of over 10,000 
research scientists, administrators, secretaries, and 
programmers throughout the Federal Government. 

The primary component of the NIH Computer Utility 
is a uniquely configured, 'loosely coupled' system 
designed around 5 IBM 3081 processors with 80 
million bytes of directly addressable memory. A 
peripheral complex of 115 tape drives, 300 disk 
drives, 2 mass storage systems, 1 1 high speed 
printers, card reader/punches, microfiche output 
units, and teleprocessing facilities for over 1 ,000 
communications lines assure adequate data storage, 
input/output, and communication capabilities. Over 
2,000 interactive terminals and 143 Remote Job 
Entry (RJE) computers located in users' offices and 
laboratories make the power of the Utility available 
at the source of the problem whenever needed. 

Operating in a multiprogramming mode, this facility 
provides timesharing and batch processing, graphic 
services, and data base management facilities. The 
IBM System 370 facility operates 24 hours a day, 
and it currently processes over 13,000 batch jobs 
and 7,000 interactive sessions daily. 

The other major component of the NIH Computer 
Utility is the DECsystem-10 timesharing facility. This 
facility is designed around one DK and two KL-10 



25 



processors with five million bytes of directly 
addressable memory, and it provides timesharing 
services and data communications support to over 
2,000 laboratory research investigators throughout 
NIH. Ten tape drives, 31 disk drives, and a variety of 
teleprocessing facilities make up the peripheral 
complex. 

The NIH Computer Utilty provides a variety of 
programming languages-including FORTRAN, 
COBOL, PASCAL, BASIC, Assembler, PL/I, and 
SAIL--as well as a data base/data management 
system (IMS), the TELL-A-GRAF interactive graphics 
package, and a comprehensive library of statistical 
and utility programs. Online computing and batch job 
submission are available interactively on the IBM 
System 370 through WYLBUR and TSO, and 
through timesharing services on the DECsystem-10. 
Several facilities for job output on paper and 
microfiche are available, and there are programs for 
creating two-dimensional or three-dimensional 
graphic displays for advanced research projects. 

The users of the Computer Utility's IBM System 370 
are informed of current programming standards and 
available facilities in a comprehensive manual, the 
Computer Center Users Guide. The DECsystem-10 
Timesharing Guide provides similar information for 
users of this system. Changes in the Utility are 
announced to users through INTERFACE, a 
periodic technical newsletter. An in-house training 
program offers 40 courses four times a year, to help 
users develop expertise in the use of the Utility. 



NIH COMPUTER UTILITY 
System 370 Services 




i^ 6S 69 70 

Calendar Year 



Through the years, the workload of the NIH 
Computer Utility has increased steadily. 



Highlights of FY82 
Accomplishments 

As in past years, the user demand and workload of 
the NIH Computer Utility showed steady growth in 
1982. A record number of jobs-over 6.5 million- 
were processed on the IBM System 370 during the 
year, and over 93 percent of these jobs were 
completed and available to the user in less than two 
hours. The DECsystem-1 facility also ran at a 
record pace, processing approximately 120,000 
interactive timesharing sessions during the year. 
Overall, the Computer Utility processed an average 
of 26,000 job-sessions per day, a seven percent 
increase over last year. 

A major highlight of the past year was the 
completion of an eight-month transition plan for 
installing the newly acquired IBM System 370 
hardware. Over 1 00 hundred new tape drives, 250 
disk drives, and 5 new Central Processing Units 
were put into productive use without interrupting 
normal service. The new IBM hardware has 
increased power, capacity, and reliability to meet the 
growing needs of Computer Utility users. Major new 
IBM System 370 components installed during the 
year include: 

• Five 3081 Dyadic Processor Complexes-IBM's 
most powerful processor, the 3081 , supports 
two processors in one physical unit. Each 3081 
processor provides more than twice the 
processing power of its predecessor, the 3033 
processor, in a unit smaller than a single 3033. 
Installation of the 3081 complexes resulted in a 
dramatic improvement in batch job turnaround 
time and interactive response time, and has 
increased the capability for processing complex 
applications in the most cost effective and 
efficient manner possible. 

• One hundred fifteen 3420-6 drives-These 
advanced tape drives increase the speed and 
reliability of tape processing through the use of 
6250 bpi reading and recording density. 

• The 3380 Direct Access Storage Device-The 
3380, IBM's newest disk drive, can store three 
times as much data as the 3330 disk it replaces. 
The 3380 also has a data transfer rate that is 
twice the speed of the older 3330 units. Online 
data storage was restructured to permit storage 
of data sets that range in size from a few bytes 
up to many millions of bytes. Over 300,000 
online user data sets were successfully 
transferred to the new 3380 disks. The 
increased availability of high speed online 
storage, a direct result of the conversion to 
3380 disks, made it possible to phase out the 
technologically obsolete 3330 private mountable 
volumes. Use of 3380 disks for online data 



26 



storage substantially improves the capacity and 
cost-effectiveness of the NIH Computer Utility. 
• IBM 3850 Mass Storage System (MSS)--The 
MSS became fully operational this year, resulting 
in a significant improvement in turnaround time. 
This also increased cost effectiveness, 
expanded online storage capability, and 
permitted substantial amounts of offline data to 
be stored online inexpensively. Data space 
equivalent to more than 29 billion bytes has 
already been allocated on MSS volumes (at 
about one-ninth the cost of online direct access 
data storage space). Implementation of the 3850 
MSS has allowed the maxium size of online data 
sets for batch jobs to be quadrupled. Users 
have found that the low charge for MSS 
storage, in addition to the savings that arise 
from fewer tape and disk mounts and faster 
processing time, make the MSS a very attractive 
alternative to other data storage media. The 
3850 MSS has also been made available, 
through the MERCURY intersystem file mover, 
to users of the DECsystem-10 for storage of 
large rarely-accessed files. 
The installation of the new processors and storage 
devices expanded the computational capability of the 
NIH Computer Utility dramatically. As a result, the 
limits on CPU time, scratch disk space, region size, 
and terminal idle time were all increased. This 
permits user batch jobs and interactive sessions to 
use significantly more of each of these resources. In 
addition, a new job class (class E) was established 
to serve the needs of the larger job that does not 
require volume mounting. The new larger limits 
improve job turnaround time, permit greater flexibility 
in job class selection, and allow the running of larger 
jobs during the prime shift. 

DECsystem-10 users also received significant 
enhancements in timesharing services during the 
year. The installation of the latest version of the 
Symmetric Multiprocessing (SMP) operating system 
both increased reliability and extended the 
availability of timesharing services for users. In 
addition, maximum real memory size was increased 
by 25 percent, and timesharing hours were extended 
to allow unattended service during the night. 

As in the past, software was enhanced to meet the 
needs of the user community. This year's additions 
include TVEDIT, a new text editor on the 
DECsystem-10 that uses no command words, and 
POSTER, an easy-to-use program on the 
DECsystem-10 that generates customized posters 
and slides for textual material. 

The building renovation program that began two 
years ago is nearly completed. A new, larger, 
modern user terminal room with a variety of 
interactive terminals and a Remote Job Entry (RJE) 



terminal was put into operation on the first floor of 
Building 12A. Efficient and comfortable, the User 
Area provides an effective work environment and is 
equipped with adequate power and air conditioning 
facilities to support the growing needs of users for 
years to come. 

The custom-designed training facility-equipped with 
sound-insulated walls, carpeting, large work tables, 
and cushioned chairs to provide a comfortable 
learning environment-is now being used for 
Computer Center training classes. Comfort and 
safety are enhanced by vanable lighting that allows 
illumination for note-taking while the screen area is 
dimmed, by dust-free chalkless marker boards, and 
by built-in projection facilities that eliminate the 
hazard of trailing electrical cords. 

This year, the Molecular Graphics System operated 
by CCB was used to solve structure problems where 
very little information was available. An algorithm 
was devloped to determine from first principles the 
shape of RNA molecules in three dimensions. The 
algorithm successfully reconstructed PHE- and SER- 
tRNA molecules. Models of the limulus hemocyanin 
complex were developed from image-enhanced 
electron microscope data. Using symmetry 
constraints, the space-filling representation of 
macromolecular structure was extended to represent 
structures of several million daltons. The models of 
several viruses-including adenovirus, the Semliki 
Forest virus, and polyoma virus-were also 
constructed. 



27 



Future Plans 

The coming year will see the next steps in the 
Computer Center's continuing plans to provide better 
services at equivalent or reduced costs. 

The installation of 64 IBM 3380 disk actuators during 
early FY83 will complete the total replacement of all 
hardware comprising the IBM System 370 facility. 
During the year, the disk drives on the DECsystem- 
10 facility will be replaced by RP07 drives that have 
a capacity two and one half times that of the current 
RP06 drives. 

Following the completion of the installation, the 
major emphasis of the Computer Center staff for 
FY83 will be directed toward the design and 
implementation of facilities and procedures to 
increase system availability. A number of new, 
realtime performance monitoring tools will be 
incorporated into the system to permit the early 
detection of faults or performance deterioration. New 
problem diagnosis methods and procedures to 
permit the rapid identification and resolution of 
problems will be investigated. Steps will be taken to 
minimize scheduled system shutdowns for dumps, 
data migration, fix application, equipment installs and 
maintenance, and to isolate services for improved 
reliability and availability. Fail soft, back-up, and 
recovery procedures will be improved to minimize 
recovery time when failures do occur. 

The teleprocessing facilities of the NIH Computer 
Utility will be expanded and improved to provide 
additional capacity to support more remote terminal 
users and to offer new functions and higher speed 
services. Facilities will be investigated to support a 
greater variety of terminal types, including word 
processors, on all services provided by the 
Computer Utility and to support more convenient 
methods of switching among services. During the 
year, the Computer Center will replace its interactive 
CRT and hardcopy terminals with new, more modern 
units and a procurement will be initiated to acquire 
new Remote Job Entry (RJE) workstations. 



The Computer Center will invest significant effort in 
the installation of improved physical and technical 
security facilities to insure the privacy of user data. 
Access to the physical plant and computer output 
boxes will be controlled by an electronic access 
control system. New software security facilities will 
permit the 'owner' of sensitive data and programs 
stored on the Utility to exert more specific levels of 
control over access to them. 

The coming year also will see an emphasis on the 
development of new self-study training courses for 
users and the completion of the audiovisual facilities 
in the new training rooms. In addition to these 
rooms, the ongoing building renovation program will 
include construction of offices for the Systems Team 
and Office of the Chief, CCB, on the second floor of 
Building 12; the Plotter and Microfiche Units will 
move to new quarters on the first floor of Building 
12A. 



28 



Publications 



Berzofsky. J. A.. Buckenmeyer. G.K., Hicks. G., Gurd, F.R.N., Feldmann. 
R.J.. and Mina, J.: Topographic antigenic determinants, recognized by 
monoclonal antibodies to sperm whale myoglobin. J. Blot. Chem. 257: 
3189-3198. 1982. 

Feldmann, R.J.. Potter, M.. and Glaudemans, C.P.J.: A hypottietical space- 
filling model of the V-regions of the galactan-binding myeloma immun- 
oglobulin J539. Mol. Immun. 18:683-698.1981 

Furie. B.. Bing. D.H., Feldmann. R.J.. Robinson, D.J.. Burnier. J. P.. and 
Fune. B.C.: Computer-generated models of blood coagulation Factor 
Xa, Factor IXa. and Thrombin based upon structural homology with 
other serine proteases. J. Biol Chem. 257: 3875-3883. 1982. 

Pawlita. M.. I^ushinski. E.. Feldmann. R.J.. and Potter. M.: A monoclonal 
antibody that defines an idiotope with two subsiles in galactan-binding 
myeloma proteins. J. Exp. Med. 1981: 1946-1956. 1981. 



29 



Office of the 
Director 



Arnold W. Pratt, M.D., Director 



Three separate units make up the Office of the 
Director, DCRT: the Office of ADP Policy 
Coordination, the Office of Administrative 
Management, and the Office of Scientific and 
Technical Communication. 



These offices complement the work of the six 
Laboratories and Branches by: 

• coordinating the complex Federal policies and 
procedures that govern getting and using 
computers at NIH 

• providing general administrative management 
support for the Division's work 

• serving as a central source of information about 
DCRT activities and about computer-related 
disciplines. 



Office of ADP Policy Coordination 



Henry J. Juenemann, Chief 



Function and Scope of Work 

The Office of ADP Policy Coordination, under the 
direction of the Assistant Director of the Division, has 
four closely related functions: 

Lit serves as a focus for NIH-wide coordination of 
Automatic Data Processing (ADP) policy matters. 

2. It serves as a central NIH point of contact with 
the Public Health Service, the Department of Health 
and Human Services, other HHS Agencies, and the 
General Services Administration on policy and 
regulatory questions. 

3. It provides the point of contact for those NIH 
procurement and contracting matters that must be 
cleared by DCRT prior to procurement or contracting 
action. 

4.lt also provides advice and assistance to NIH 
staff and others concerning the internal operations of 
DCRT in matters of ADP policy and procurement. 



The activities of the Office include: 

• advising the Director of DCRT and, through him, 
the Director of NIH on ADP policy matters 

• reviewing and evaluating proposals from NIH B/ 
l/D's for procurements and contracts related to 
computing and ADP 

• directing the development of the annual NIH 
ADP Plan 

• assisting the NIH Division of Management Policy 
on questions relating to its responsibility for 
administrative and management computer 
applications 

• representing NIH in PHS and DHHS policy 
formulation efforts 

• working with HHS and GSA staff to obtain 
necessary approvals for NIH on procurements 
and contracts, and 

• answering inquiries from scientists and 
administrators who are confused by the whole 
process. 



31 



FY82 Accomplishments 

A major set of tasks accomplished during FY82 
involved obtaining PHS, HHS, and GSA approval for 
full recompetition of three contracts under which 
DCRT supplies the three types of NIH Standard 
Terminals to users. This required extending the 
present contracts in order to provide adequate lead 
time for the reprocurements. It also required six sets 
of interlocking clearances and three minor and three 
major solicitation packages. The tasks were 
accomplished during FY82 in a way that provided 
continuous terminal support to users of the DCRT- 
operated NIH computer utility. 

During the year, this office reviewed nearly 400 
proposals for acquisition of ADP equipment and/or 
services and commented on approximately 100 
research contracts involving ADP. Each was 
reviewed to ensure that it was justified and was in 
conformance with PHS, HHS, GSA, and 0MB 
guidelines. Suggestions and assistance were 
provided to the NIH Procurement Branch and to the 
various Research Contract Branches as to the most 
expeditious procurement route to follow. In many 
cases one or more of the Laboratories and Branches 
of DCRT assisted by providing expertise to help in 
the review of technical aspects of the proposals. 

In addition, the office was heavily involved in the 
development of specifications, RFP's, and 
clearances for a variety of specialized automated 
systems, including many minicomputers for 
laboratory use. A contract was awarded for the 
automation of the NIH Library; implementation of the 
initial phases of that project is underway. A 
delegation of procurement authority was obtained 
from GSA for a long term project to automate the 
data handling associated with the Division of 
Research Services' animal breeding program. 
Preparation of the necessary solicitation package 
and conduct of the procurement was underway 
during the latter part of FY82. Likewise, GSA 
delegations of authority were obtained for the total 
reprogramming of the National Library of Medicine's 
MEDLARS system and for the upgrading of the 
hardware that will be required to develop, test, and 
operate the new and improved MEDLARS System. 

l\\e Annual ADP Plan that combines projections of 
new ADP initiatives and required ADP expenditures 
for all Bureaus, Institutes, Divisions, and Offices of 
NIH was completed. It details an NIH ADP program 
projected to be 75.5 million dollars and 772 work 
years in FY83 growing to 80.5 million dollars and 



788 work years in FY84 and to 83.9 million dollars 
and 853 work years in FY88. Although the accuracy 
of the out-year projection must be regarded with 
caution, the trend of ADP and computing 
involvement in the scientific and managerial life of 
NIH is unmistakable. 

Future Plans/Trends 

FY83 will be marked by major changes in the 
structure, staffing, and focus of NIH's ADP Policy 
Coordination function. These changes, the nature of 
which are not predictable at this writing, should be 
accomplished during the last quarter of FY82. 



32 



Office of Administrative Management 

L. Lee Manuel, Chief 



Function and Scope of Work 

The Office of Administrative Management, under the 
direction of the Executive Officer, consists of 1 5 
people, organized functionally into three sections: 
finance, personnel, and general administration. The 
office serves as liaison between these functions and 
the NIH Office of Administration, Office of Research 
Services and with other NIH, PHS, and DHHS 
offices. It handles a broad range of administrative 
managerial functions for an NIH research division of 
almost 300 people. 



Fiscal Year 1982 
Accomplishments 

The Administrative Office processed a variety of 
procurements and acquired approximately 30 million 
dollars in supplies and equipment during FY82. Day- 
to-day management activities conducted by this staff 
included: procurement purchases and contracts; 
processing of travel and training requests; 
administration of property, space, and 
communications; payroll; and mail/messenger 
services. In October 1981 the Administrative staff 
began using the Delegated Procurement System 
(DELPRO), a subsystem of the NIH Materiel 
Management System (MMS). This process allows B/ 
l/D personnel to enter and control their own 
delegated procurement data such as ICO's and 
ROC's, Repair Orders, reprint orders, FACS book 
orders, and professional services requests. 

Based on a decision by the Director, DCRT, the 
Budget Office developed budgets at the laboratory 
level and implemented a reporting system to allow 
Lab/Branch Chiefs to track actual obligations against 
their budgets. This has resulted in greater awareness 
of spending patterns and more accurate 
recordkeeping. The need for even more detail by the 
Lab/Branch Chiefs will undoubtedly result in some 
modifications, but the reporting system has proved to 
be a valuable management tool. The office also 



participated in setting and reviewing rates for Service 
and Supply Fund data processing activities. This was 
especially challenging in a year of major equipment 
transition in the Computer Center. 

The Project Control Office continued to process 
requests for new accounts, register new users, and 
prepare billing data for the NIH computer facility. The 
office also successfully completed its major annual 
update of information on over 10,000 users and 
2,000 accounts. Several modifications, including the 
elimination of some reports, were made to the 
Project Accounting System. 

The Personnel Office processed approximately 400 
personnel actions that included promotions, 
reassignments, temporary appointments, excepted 
appointments, transfers, and career-conditional 
appointments. The Department continued to operate 
under a partial hiring resthction most of the year, 
making outside recruitment and selection difficult. 

A new Departmental performance appraisal system 
was implemented this year, effective October 1, 1981, 
as a result of the Civil Service Reform Act of 1978. 
Personnel Office staff conducted training and 
orientation for all supervisors and employees on the 
implementation of the new Employee Performance 
Management System (EPMS). This massive effort 
involved 12 separate sessions, each conducted by 
either the Personnel Officer or the Personnel 
Management Specialist. The Personnel Officer 
serves as the coordinator for EPMS and provides 
assistance to employees and supervisors on a 
continuing basis. 

The writing of new factor evaluation position 
descriptions in the GS-334-0 series for all computer 
specializations in DCRT was coordinated by the 
Personnel Office with the assistance of contract 
personnel who conducted workshops on the Factor 
Evaluation System for the 334 series. The Personnel 
Office staff carried out the classification of these 
positions. 



33 



The Executive Officer and Assistant Director served 
on a work group witli the Director and staff of the 
NIH Division of Management Policy to prepare a 
position paper with alternative recommendations for 
the Director, DCRT, and for the Associate Director 
for Administration, NIH, concerning the ADP policy 
functions at NIH. The paper described the functions 
associated with this process and addressed certain 
issues that should be resolved. 

The Executive Officer also coordinated the 
preparation of the Annual Research Plan and 
represented the Division within OD, NIH, as Program 
Planning Officer. He also served the Division as its 
international representative to the Fogarty 
International Center and as its legislative contact 
with the Division of Legislative Analysis, OPPE, NIH. 
This required keeping abreast of issues in these 



areas and advising the Division's staff as necessary 
on items that might have an impact on DCRT 
programs. 

A new Full Time Equivalency (FTE) system was 
implemented to track DCRT employment against 
NIH-mandated ceilings. 

Future Plans/Trends 

The 0AM will continue its efforts in providing 
administrative management support to the Division's 
programs. A new 0MB Circular, A-123, titled 'Internal 
Control Systems' will probably be implemented at 
NIH during the coming year. This Circular prescribes 
policies and standards to be followed in establishing 
and maintaining internal controls in program and 
administrative activities. 



34 



Office of Scientific and Technical 
Communications 

William C. Mohler, M.D., Chief 



Functions 

The DCRT Office of Scientific and Technical 
Communications (OSTC), under the direction of the 
Associate Director, DCRT, includes: 

• the DCRT Library, which maintains a high quality 
collection of materials for use by DCRT and NIH 
staff and serves as a resource for other libraries; 

• the DCRT Information Office, which serves as 
the focus for providing the NIH community and 
the professional and lay public with information 
about DCRT's activities and related applications 
of computing in biomedical research; 

• scientific work in related areas of pattern 
recognition, multidimensional information 
processing, and applications to medical decision 
making. 

Scope of Activities 

The DCRT Information Office handles the full range 
of activities of an NIH Information Office. The 
Information Officer, assisted by a Public Affairs 
Specialist, answers inquiries, produces and 
distributes print and audiovisual materials, and 
arranges briefings for visitors. The office coordinates 
special events, works with members of the media, 
and provides advice, assistance, and educational 
resources on communications for the DCRT staff. It 
also responds to all Freedom of Information requests 
coming to DCRT. 

A major part of the Information Office program is 
directed within NIH toward improving an 
understanding of the Division's work and the 
application of computing to biomedical research. But 
the scope of its communications includes Federal 
agencies, schools, libraries, private industry, medical 
organizations, representatives of the media, and a 
wide variety of individual scientists, engineers, other 
professionals, and lay persons. 



The DCRT Library is a fully independent, special 
library, staffed by the Librarian and a Library 
Technician. The Library provides a full range of 
services and has access to a wide variety of online 
information services and data bases. The collection 
of monographs, periodicals, and other documents 
covers subjects related to the work of DCRT. These 
include computer science, mathematics, statistics, 
electronic engineering, information science, and 
management. 

The Library supports the work of the DCRT staff and 
serves as a resource for employees in the rest of 
NIH. It is an integral part of the Washington area 
network of special libraries and cooperates with 
libraries outside the area to share resources. It does 
this through organizations such as the Interlibrary 
Users Association of the Washington/Baltimore 
Area, the Metropolitan Washington Library Council, 
FEDLINK (a Federal library consortium), and the 
national OCLC (Online Computer Library Center) 
network. 

The other activities of OSTC derive from the 
interests of its scientific professionals. They work 
with other professionals at NIH and with medical and 
technical groups, both government and private, 
outside of NIH. 

Highlights of FY82 

Dr. Prewitt was very active in FY82. She was the 
leader in organizing two major scientific meetings: 
MEDCOMP '82, the First International Conference 
on Medical Computer Science sponsored by the 
IEEE Computer Society, and the First International 
Symposium on Medical Imaging and Image 
Interpretation. Her collaborations with physicians, 
mathematicians, engineers, and computer 
professionals led to papers and presentations on 
medical decision making and on pattern recognition 
applied to biomedical images. She began an 
appointment as Stocker Professor of Electrical 



35 



Engineering at Ohio University under the aegis of the 
Intergovernmental Personnel Act. 

The DCRT Information Office underwent several 
changes. In the first quarter, Mrs. Hodges moved to 
work full time on the needs of the Data Management 
Branch for informational materials describing their 
computer systems. Despite this, the Office had 
another productive year. This is even more 
noteworthy in light of other assignments undertaken 
by its staff and new restrictions placed upon public 
information activities in the general effort to cut costs 
and improve efficiency throughout the Federal 
Government. 

As part of this effort, the Information Officer, Mrs. 
Miller, went on detail for four months to the Office of 
the Assistant Secretary for Public Affairs, DHHS, to 
assist with several projects involving public affairs 
and computer expertise: 

1. A study in response to 0MB Circular M-81-14, 
'Federal Information Centers,' for which the 
Department must develop an implementation plan to 
consolidate, eliminate, and improve its existing 
clearinghouses and resource centers. 

2. The systems review process mandated by 
Public Law 96-511, the Paperwork Reduction Act. 

3. The Department's response to H.R. 4758, which 
proposes legislation on the government sale of data 
processing and telecommunications services, and 
cost recovery of same. 

4. A project to design and implement a data 
processing system to keep track of Department-wide 
public affairs budgets and publications, as part of a 
DHHS Control Plan mandated by 0MB Bulletin Bi- 
le, 'Elimination and Consolidation of Government 
Periodicals and Recurring Pamphlets.' 

During her absence, Mr. Hall handled the office, 
within the limits of his temporary appointment. 

In spite of these constraints, the staff covered the 
many routine day-to-day activities and provided 
consultations on a variety of communications 
services. These included arranging publicity for 
MLAB courses, assisting in slide production for a 
presentation of BRIGHT, obtaining contract editorial 
services for Division staff members, and assisting 
the DCRT director in responding to administrative 
requests. 

In addition to coordinating, editing, and producing the 
DCRT Annual Report, the Office carried out the 
following major efforts during the year: 
• Printed materials about DCRT activities went to 
over 6,000 people. To provide current 
information on DCRT activities, the Office 
produced complete revisions of the old 
brochures describing the Laboratory of Applied 



Studies and the DCRT Library, and it obtained 
reprints of Computers at NIH: Tools for the 
Advancement of Medicine and of the DCRT 
Fiscal Year 1981 Annual Report, Volume 2. 

• The Office wrote articles for the NIH Record and 
helped other DHHS groups report DCRT 
activities. The PHS covered WYLBUR in its 
publication Future Office; the Clinical Center and 
the NIH Audiovisual Branch prepared NIH 
Record articles citing work in DCRT. 

• The Division received mention in several private 
media throughout the year. These included the 
SIGBIO Newsletter, VoiceNews (a newsletter on 
voice technology). Current Health (a magazine 
for high school students). Sky (Delta Airlines' in- 
flight magazine), and Omni (a half-hour 
television science magazine). Film footage of 
molecular graphics was taken for future use on 
a CBS television program. The Body Human. 
The Information Office arranged for regional 
computer voice technology experts to be 
interviewed on the Illinois public television 
network. 

• The short sound/slide show describing the 
Division and highlighting each laboratory and 
branch was completely revised to help brief 
visitors and orient new employees. 

• Work on the DCRT bibliography involved 
correcting and updating records on over 1 ,000 
scientific papers written by DCRT researchers 
over the last 1 2 years. 

The Information Officer continued to explore the 
economics of electronic typesetting. As Chair of the 
NIH Printing Committee, she advised others on these 
methods. She also served on the Board of Directors 
of the Washington chapter of Women in 
Communications, Inc., a national organization for 
public relations, journalism, broadcasting, and 
communications professionals. 

The DCRT Library also had a busy year improving 
service while managing fixed space, budget cuts, 
and changing technology. 

A complete redesign of the Library brochure now 
helps users more readily locate monographs, 
documents, journals, and reference materials. New 
signs made by the Library staff identify shelf areas, 
sections of the card catalog, and locations for return 
of materials. The Librarian, Mrs. Chu, compiled the 
prototype for a series of bibliographies covering 
specific subject areas of the library's collection. 

A fourth online bibliographic reference service 
provided better coverage for the wide range of 
technical and scientific needs of users of the Library. 
The hidden cost to this breadth of coverage is the 



36 



time that the Librarian and the Library Technician, 
Ms. Florentine, must spend mastering system 
updates and the idiosyncracies of the data files and 
indexing strategies. The Librarian attended the 
District of Columbia Online Users group meetings to 
share experiences with such systems and to keep 
current with new developments in this emerging 
technology. 

Budget restrictions caused a reduction of monograph 
purchases and a complete review of the list of serial 
acquisitions. In spite of this, space limitations 
required continued attention to weeding out less 
useful materials and to conversions of essential 
journal holdings to microforms. With the aid of the 
DCRT Library Committee, a review of the collection 
led to removal of some 50 monographs, 400 
documents, and 200 journal volumes. 

A multistage, multiyear project continued to cope 
with the new Anglo-American Cataloging Rules 
(AACR2). This effort is complicated by the high cost 
of computer-based library systems. None yet meets 
the needs and budget of our small special library. 
The Library's own computer-based circulation control 
system continues to perform well at a reasonable 
cost. For the moment the strategy to approach 
computer-based cataloging remains use of the OCLC 
data base to create a machine readable file of items 
in the collection, edited according to AACR2 and 
local conventions. These can be produced on tape 
by FEDLINK at the appropriate time. 

Much of the Library's work is aided by the Librarian's 
active role in professional organizations. She 
continued as a FEDLINK delegate to the OCLC 
Users Council and served on the Council's Executive 
Committee as Council Secretary. She completed the 
first of two years as a Director on the Executive 
Board of the District of Columbia Library Association, 
an organization of some 1 ,000 Washington area 
librarians. 



Nevertheless, there is reason to be optimistic for the 
immediate future. Both the Librarian and the 
Information Officer have plans to meet these 
challenges. Dr. Prewitt's appointment to an endowed 
position at Ohio University will allow her to pursue 
her work in the coming year without some of the 
current limits on funds, personnel, and equipment 
that exist in the Division. 



Plans 

Plans for the coming year will follow the general 
lines of work in FY82. The current Federal emphasis 
on control of publications and audiovisual products 
will offer a significant restraint for providing 
information about DCRT activities to the NIH 
community and to the many groups outside NIH that 
have expressed interest in DCRT work. The 
continuing trend toward limited budgets in the face 
of persistent inflation will require efforts to cut costs 
while maintaining a high level of service in the 
Library and a high quality of products from the 
Information Office. 



37 



Publications and Presentations 

Computers at NIH: Tools for the Advancement of Medicine. NIH Publication 
No. 82-1039, reprinted 1982, 20 pp. 

Data l^anagemenl Branch. NIH Publication No. 81-1927, Fall 1981, 16 pp. 

DCRT Library. January 1982, 24 pp. 

Division of Computer Research and Technology Fiscal Year 1981 Annual 
Report, Volume 1. October 1981, 44 pp. 

Division of Computer Research and Technology Fiscal Year 1981 Annual 
Report, Volume 2. October 1981, 100 pp. 

Dwyer, A.J., Glaubiger, D., Ecker, J.G., Doppman, J.L., Prewitt, J. M.S., and 
Plunkett, J.: The Radiograptiic Followup of Patients with Ewing Sarco- 
ma: A Demonstration of a General Method. Radiology (in press). 

Dwyer, A.J., Prewitt, J. M.S., Ecker, J.G., and Plunkett, J.: The Use of the 
Hazard Rate to Allay the Peril of Inappropriate Followup: An Optimiz- 
ation Approach to Patient Management. Journal of IVIedical Decision 
Making (in press). 

Herron, R., Dwyer, S.J., and Prewitt, J. M.S.: Computer Graphics for Medical 
Imaging. NGGA Tutorial T-11 Summary, 1981. 

Kroop, D.O., and Prewitt, J. M.S.: Computer Security in Medicine. MED- 
COMP '82, Philadelphia, PA, September 23-26, 1982 (in press). 

Laboratory of Applied Studies. NIH Publication No. 82-1928, April 1982, 16 
PP- 

Miller, P.O.: Text-to-tape Copy Preparation. In Torrence, S.R. (Ed.): Pro- 
ceedings of the 1981 Annual Conference of the National Association of 
Government Communicators. 1981, pp. 262-265. 

Prewitt, J. M.S.: How to Asses Image and Signal Processing Technology in 
Health Care. Feierliche Erotfnung des Neubaus des Instituts fur Medi- 
zinische Informatik und Systemforschung (MEDIS) der Gesellschaft fur 
Strahlen- und Umweltforschung (GSF), Neuherberg, Germany, June 14, 
1982. 

Prewitt, J.M.S.: Medical Computer Science: Whither or Wither? MEDCOMP, 
IEEE Computer Society, Silver Spring, MD, 1982. 

Prewitt, J. M.S.: New Developments in Image Processing and Analysis: Par- 
allel Computer for Medical Application. SIAM National Visiting Lecture, 
University of Florida, Gainesville, FL, October 14, 1981. 

Prewitt, J. M.S.: Pattern Recognition in Medicine. World Congress on Medi- 
cal Physics and Biomedical Engineering, Hamburg, Germany, Septem- 
ber 7-9, 1982. 

Prewitt, J.M.S.: Segmentation of Cell Images: Art or Science? Cytometry 
2:122, 1981. 

Prewitt, J. M.S.: The Diagnostic Performance of Leukocyte Counters. Cyto- 
metry 2: 122, 1981. 

Prewitt, J. M.S., Clyman, J., Lander, W.B., Lehman, J.S., Ranade, S.M., and 
Wu, S.C: Cytologic and Nuclear Shape: Fourier Characterization. Inter- 
national Conference on High Resolution Ceil Image Analysis. North 
Hollywood, CA, January 24-26, 1982. 

Prewitt, J. M.S., and Lander, W.B.: Cytologic and Nuclear Shape: Fourier 
Characterizations. Analytical and Quantitative Cytology 4: 156, 1982. 

Prewitt, J.M.S., Plantholt, M., Simpson, M., Edberg, T., and Sanfeliu, A.: The 
Graph-Theoretic Characterization of Tissue Textures. Cytometry 2: 1 22, 
1981. 

Prewitt, J.M.S., Ranade, S.M., Wu, S.C, and Lehman, J.S.: Cytologic and 
Nuclear Shape: Fourier Characterizations. Cytometry 2: 1 22, 1 981 . 

Ranft, U., Fu, K-S, and Prewitt, J. M.S.: Segmentation of Transverse Section 
Pictures of Muscle Tissue. World Congress on Medical Physics and 
Biomedical Engineering. Hamburg, Germany, September 7-9, 1982 (in 
press). 

Schuette, W.H., MacCollum, M.A., Smith, C.E., Prewitt, J.M.S., and Shack- 
ney, S.E.: An Iterative Method for the Decomposition of DNA Histo- 
grams. Annual Meeting of Cell Kinetics Society, March 1982. 

Schuette, W.H., Shackney, S.E., Smith C.E., and Prewitt, J. M.S.: An Iterative 
Method for the Decomposition of Gaussian Distortions from DNA Histo- 
grams. MEDCOMP, Philadelphia, PA, September 23-26, 1982 (in 
press). 



38 



DCRT 



Division of Computer Research and Technology 
National Institutes of Health 
Bethesda, Maryland 20205 



DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



FISCAL 

YEAR 

1982 



ANNUAL 
REPORT 



VOLUME 2 



I 



DOXRIW 

LALajajJaj 



Foreword 



The Division of Computer Research and Technology DCRT programs focus on three primary activities: 
has primary responsibility for incorporating the power conducting research, developing computer systems, 
of modern computers into the biomedical programs and providing computer facilities. 

and administrative procedures of NIH. DCRT serves jhe fiscal year 1982 annual report describes our 
as a scientific and technological resource for other work in two volumes: 

parts of PHS, and for other Federal organizations Volume 1 gives an overview of the work of 

with biomedical and statistical computing needs. each group, highlighting the year's 

accomplishments; 

Volume 2 gives details about the projects and 
activities of each group. 



Contents 



Computer Systems Laboratory 

Clinical Research, Patient Care, Epidemiology 1 

Laboratory Investigation 3 

Program Management and Administration 4 

Biomedical Communications and Conference 

Support 5 

Computer Research and Technique Development 5 

Research Projects 

Computer Support for Flow Microfluorimetry/Cell 

Sorters (FMF) 5 

Cardiac Scintillation Probe 7 

Computerized Radiation Therapy 8 

Medical Intensive Care Unit Patient Monitoring 

Computer System 9 

Automated Management of Critically III 

Patients 11 

Positron Emission Tomography (PET) Scan Image 

Analysis in Aging Studies 11 

Computer Assisted Tomography (CAT) Scan 

Image Analysis in Aging Studies 13 

Computer Analysis of Autoradiographic Images 

of Recombinant DNA Colonies 13 

Cataract Grading via Computerized Slit-Lamp 

Image Analysis 14 

Robust Boundary Detection of Necturus Gall 

Bladder Cells 15 

Rehabilitation Medicine Computer System 16 

Positron Emission Tomography (PET) 

Facility 18 

Computer Assisted Hematology Morphology 

Data Handling System 19 

Automated Pulmonary Physiology Testing 19 

Assessment of Tongue Motion During Speech 

Using Ultrasonic Imaging Techniques 21 

Anesthesia Computer System 22 

Medical Information Technology Project 23 

Molecular Graphics and Sequence Analysis 24 

Computer Analysis of Gel Electrophoresis 25 

Morphometric Analysis of Normal and Neoplastic 

Tissue Cultures 26 

Virus Structure As Determined by Image 

Processing of Electron Micrographs 27 

Image Processing of Electron Micrographs 28 

Potentiometric Titration Controller 29 

Metabolic Energy Measurements 30 

Electron Energy Measurements 30 

Electron Microanalysis Facility 31 



Molecular Interactions Laboratory Data 

System 33 

Californium-252 Plasma Desorption Mass 

Spectrometer Data System 34 

Distributed Laboratory Data Acquisition and 

Control System 34 

Image Processing Facility 36 

Analytic Models of Computer System 

Performance 37 



Laboratory of Applied Studies 

Clinical Research and Patient Care 41 

Laboratory Investigation 42 

Computer Research and Development 42 

Research Projects 

Statistical Research in Clinical Pathology 43 

Mathematical Models of Binding Equilibria 44 

Mathematical Modeling of Substrate Transport in 

Physiological Environments 44 

Analysis of Coupled Transport and Biochemical 

Kinetics 46 

Nonlinear Equations 48 

Numerical Approximation Techniques for the 

Solution of Reaction-Diffusion Systems in 

Biology 50 

Monitoring of the CNS in Critically III 

Patients 51 

Computer-based Studies in Pulmonary 

Pathophysiology and Respiratory 

Disease 52 

Investigation of Hybrid Computing for the 

Construction of Simulation Models and for 

the Analysis of Physiologic Signals 53 

Computer Systems for Nuclear Medicine 53 

Computer-Aided Analysis of 

Electrocardiograms 54 

Computer-based Studies in Ultrasonography 55 

Computer Based Analysis and Image 

Processing in Electron Microscopy and X-ray 

and Electron-Loss Spectroscopy 56 



Physical Sciences Laboratory Data Management Branch 

Summary of Activities 59 Clinical Research, Patient Care, Epidemiology 83 

Research Projects Laboratory Investigation 85 

Consulting Services 60 Program Management and Administration 86 

Studies in Mathematics and Statistics 61 Biomedical Communications Applications 87 

Theory of Biochemical Separation Computer Research and Technique Development 88 

Techniques 61 

Correlation Function Spectroscopy/Laser Light 

Scattering 62 Computer Center Branch 

Cell Motility and Chemotaxis 63 Summary of Projects 91 

Two-dimensional Fourier Transform Nuclear Research Projects 

Magnetic Resonance Spectroscopy 63 Computer Representation of Virus and other 

Precision in the Measurement of NMR Macromolecular Assemblies 93 

Parameters 64 

Theory and Measurement of Intermolecular 

Forces 64 Office of the Director 

Analysis of Intracellular pH by 31 P Nuclear c, x a *• x- 

Magnetic Resonance Spectroscopy 65 Summary of Activities 95 

Quantitative Analysis of Cell Electronmicroscopy research Projects 

and Plasma Membranes 65 Electronic Typesetting Methods 96 

Diffusion of Molecules on Cell Surfaces and 

Light Scattering from Fluids 66 

Control of Actin Assembly In Nonmuscle 

Cells....67 
Computerized Typesetting of Scientific 

Papers 68 

Laboratory of Statistical and Mathematical 
Methodology 

Clinical Research, Patient Care, Epidemiology 72 

Laboratory Investigation 72 

Program Management and Administration....73 
Biomedical Communications Applications.....73 
Computer Research and Technique Development 

Automated Data Processing of Medical 
Language 73 

Cluster Analysis 74 

Research Topics in Computer Science 75 

Discrete Mathematics and Applications 76 

Mutivariate Statistical Analysis 77 

Linear Methods in Statistics 78 

Non-numerical Programming Techniques and 
Applications 78 

Topics in Geometry and Analysis 80 



Computer Systems 
Laboratory 



Alan M. Demmerle, Chief 



Clinical Research, Patient Care, 
Epidemiology 

Computer Support for Flow Microfluorimetry 
(FMF) /Cell Sorters (NCI, NHLBI), This project 
provides support for the acquisition, display, and 
analysis of data from four Becton-Dickinson FACS- 
11 and one Coulter MDADS FMF in NCI and NHLBI. 
All five systems currently use Digital Equipment 
Corporation PDP-11 computers with RT-11 single 
user operating system. CSL is developing an RSX- 
11M multi-user system to replace RT-11 in some 
high volume applications. This system will feature an 
LSI-11 microcomputer (satellite) that will interact 
independently with the FMF operator during 
parameter entry and will acquire data and send it to 
a PDP-1 1 computer (host) for storage through an 
interprocessor link. 

Cardiac Scintillation Probe (CC, NHLBI). This 
nonimaging ECG-gated scintillation probe, when 
used in conjunction with left ventricular (LV) 
catheterization, permits simultaneous quantification 
of the variation of LV volume and pressure. The 
system can continuously derive parameters such as 
LV compliance, ejection fraction, filling and ejection 
rates, and various temporal relationships. The probe 
continues to be used to study the effects of 
nephidipine and verapermil on patients with 
asymmetric septal hypertrophy. In addition, a new 
protocol studying drug effects on patients with 
coronary artery disease was initiated. The probe is 
also being used to monitor the left ventricle 
performance of patients in the Medical Intensive 
Care Unit. 

Nuclear Medicine Computer Systems (CC). CSL 
has continued consultation and support for the 
imaging systems in the Nuclear Medicine 
Department, CC, to assess their changing needs and 
anticipated growth in the new Ambulatory Care 
Research Facility (ACRF). 



In the ACRF the three existing systems were placed 
in individual scan rooms and a central system was 
purchased to provide a central viewing station for 
processed studies, data storage, and increased 
program development capability. A distributed 
system was implemented to provide communication 
between the systems. Investigations using high 
efficiency camera systems and their potential 
applications were also initiated. 

Computerized Radiation Therapy (NCI). CSL 
developed a computer system, now in clinical 
operation in the Radiation Oncology Branch, NCI, to 
use the detailed contour and density information 
available from computer assisted tomography to 
improve radiation treatment planning. This system for 
external beam treatment planning is based on a 
generalized 3-D dose field model that covers photon, 
electron, and neutron beams. 

The computer program and most of its clinical 
implementation has been completed for the photon 
and electron fields available from the local 6 MV and 
12 MV linear accelerators. The current capabilities 
include interactive simulation of most irradiation 
techniques devices. The system enables the display 
of dose distributions computed in several transverse 
contours and overlaid on corresponding CT scans. 

Medical Intensive Care Unit Patient Monitoring 
Computer System (CC). The dynamic events 
occurring within the Clinical Center's Medical 
Intensive Care Unit are monitored by a unique 
multiple-computer system. Capabilities of the system 
include data acquisition and analysis, medical 
recordkeeping, tabular and graphical data display, 
and feedback control as required in support of 
patient care and research protocols. 

The facility contains a state-of-the-art catheterization 
laboratory with flexible computerized physiologic 
monitoring features, and a high resolution x-ray 
system with digital subtraction angiography 



capability. Of primary interest is the study of the 
etiology and therapy of septic shock. 

Automated Management of Critically III Patients 

(CC). This new research project is concerned with a 
systems approach to the management of critically ill 
patients in a clinical setting. The ultimate goal is the 
utilization of computer-based instrumentation to aid 
in the differential diagnosis of disease states, and 
the implementation of therapeutic modalities through 
automated technology. 

A state variable approach is utilized in the 
mathematical modeling of pertinent pharmacokinetic 
and physiologic processes. Several alternative 
methods for closed-loop automated medical 
interventions are being investigated. 

The Biomedical Image Analysis Project (NHLBI, 
NEI, NCI, NIADDK, NIDR). This project is oriented 
toward the development of general-purpose 
algorithms and techniques for image input (including 
digitization), image enhancement (including contrast 
enhancement), feature extraction (including edge 
detection, contour extraction, contour following, 
contour coordinate compression, and shape and 
texture analysis), three-dimensional representation, 
image reconstruction (including Fourier filtering, 
combining images, symmetrization), and other 
techniques of image processing and image 
reconstruction. The capability is used in work with a 
number of NIH researchers. 

Automated ECG Processing (CC). The Clinical 
Center's Heart Station was automated last year with 
a computer system that provides for the online 
acquisition, analysis, storage, and retrieval of 
diagnostic electrocardiograms. The newest versions 
of the vendor's turnkey software and diagnostic 
criteria packages were installed and the system was 
placed in routine clinical operation. The ECG 
analysis package will be modified by CSL as 
necessary to customize the ECG analysis process in 
order to satisfy NIH requirements. 

Rehabilitation Medicine Computer System (CC). 
This project involves the development of computer 
techniques in collaboration with the Department of 
Rehabilitation Medicine of the NIH Clinical Center. 
CSL has recommended computer techniques to 
automatically acquire anatomical and performance 
information, perform calculations, and display the 
results to the medical staff. The automated 
techniques include the measurement of body forces 
(hand and ground reaction forces), electromyograms 
(electrical activity of the muscle), and body 
kinematics (the position and angles of the limbs and 
and joints in space and time). The medical staff will 
enter additional data into a data base with computer 



generated forms displayed on a terminal screen, and 
perform inquiries and generate reports using the 
accumulated data. A competitive procurement is 
underway and system installation is targeted for the 
early spring of 1 983. 

Positron Emission Tomography (PET) Facility 

(CC). Various NIH Institutes use the PET Facility of 
the Nuclear Medicine Department. To meet 
increased demand for PET scans and analysis, CSL 
ordered and installed peripherals on the PET 
computer system to handle the increased data flow. 
To increase the image data analysis capability, CSL 
implemented an offline minicomputer system and 
modified existing image analysis software. In 
collaboration with Institute scientists, this system was 
programmed to compute local cerebral metabolic 
activities with radioactive deoxyglucose utilization. 

Computer Assisted Hematology Data Handling 
System (CC). In February 1 982, CSL, in 
collaboration with the Clinical Pathology Department, 
CC, completed the design and installation of a 
microcomputer-based system that allows entry of 
manual white cell differential, red cell morphology, 
and platelet estimates. It replaces the old method of 
manual offline counting and transcription to mark 
sense cards. A menu-driven CRT display at four user 
stations is used with a variety of dynamic formats. 
Data can be compared with automated cell counting 
results, which are displayed at the top of the screen. 
Results are then transferred online to the Clinical 
Pathology Laboratory Computer System via a direct 
serial communications line. 

Automated Pulmonary Physiology Testing 

(NHLBI). Fully-automated lung static compliance and 
inspiratory muscle strength procedures are now 
routinely performed in the Pulmonary Branch's 
pulmonary physiology/exercise laboratory. Under the 
control of a MINC 11 /03 computer system, data is 
acquired and analyzed in realtime, with graphical and 
textual reports produced at the completion of each 
procedure. 

Steady-state treadmill exercise testing has been 
partially automated. Although data is manually 
entered, analysis and report generation are fully 
computerized. Work is in progress to enable 
automatic realtime acquisition of exercise data with 
breath-by-breath analysis and thus makes the 
procedure entirely noninvasive. Once the steady 
state procedure is successfully implemented, a 
nonsteady state protocol is planned. 

Pulmonary Branch Support (NHLBI). This project 
assists the Pulmonary Branch in its computer and 
data processing needs. CSL continued to help 



maintain the computer portion of the two Collins 
automated pulmonary function analyzers. The 
BRIGHT software package, operating on the 
DECsystem-10, was identified as one means for 
managing pulmonary patient data. CSL provided 
introductory training and helped organize a general 
scheme to enable investigators to develop and 
maintain individualized data storage and analysis 
capabilities using BRIGHT. 

Assessment of Tongue Motion During Speech 
Using Ultrasonic Imaging Techniques (GC). A 
collaborative effort with the Departments of 
Rehabilitation Medicine and Diagnostic Radiology, 
CG, GSL, and the Physical Sciences Laboratory, 
DGRT, initiated to develop both a system and an 
analytical technique for realtime ultrasonic imaging of 
the tongue. Using an existing realtime diagnostic 
scanner, several normal subjects were scanned 
during the utterance of specific phonemes. The 
resultant images were digitized and analyzed for 
reliability and repeatability. fVlathematical techniques 
are under investigation for describing the patterns of 
tongue motion obtained. New instrumentation is 
being evaluated to implement this technique in the 
new AGRF Speech Lab scheduled for operation in 
FY83. 

Anesthesia Computer System (GG). This is a 
collaborative effort with the Anesthesiology Service, 
GG, to evaluate improved instrumentation techniques 
and to identify and investigate ways that automation 
can benefit anesthesia. Project emphasis is on 
adjunctive monitoring and automated recordkeeping 
in the operating room. Efforts this year centered on 
the development of a project plan to guide future 
work. 

Medical Information Technology Project. This 
project is concerned with the development of better 
ways to automate the essential physician 
contribution to the health care record. This year, in 
collaboration with a practicing dermatologist, we are 
field testing an ambulatory patient care treatment 
system. It is designed to help the physician generate 
patient information and treatment schedules, 
pharmacy prescriptions, medical and surgical reports, 
laboratory test orders, and referral letters to other 
doctors. Interaction with the system consists of high- 
speed friendly menu selections with many default 
fields preselected. Because most of the clinical 
software is table driven, it can be adapted to other 
clinical care and research environments. 

Laboratory Investigation 

Molecular Graphics and Sequence Analysis 

(NIADDK. NGI, NIDR). The sequence of some 



regular proteins, together with other structural 
information such as data from x-ray diffraction, fiber 
diffraction, electron microscopy, and spectroscopic 
analysis can be used to evaluate models of the 
protein structure. Four projects have been using both 
modeling techniques developed at NIH. We have 
recently published a new interpretation of the x-ray 
diffraction data for collagen fibrils. Analysis of 
cyanogen bromide fragments of keratin filaments are 
being studied to understand their structure and to 
compare keratin with other filamentous proteins. 
Analysis of myosin and streptococcal M proteins is 
continuing as sequences become available. 

Potentiometric Titration Controller (NHLBl). The 
Potentiometric Titration Gontroller previously 
developed by GSL and LGB, NHLBl, has been 
greatly enhanced by replacing the original 
spectrophotometer with a rapid scan 
spectrophotometer (RSS) that is capable of 
capturing and storing multiple complete optical 
spectra. As before, the solution potential is 
established by microcomputer-controlled electric 
currents, and the amounts of electron transport 
carriers are determined from spectral data. With the 
new system, it is possible to acquire scans in 
milliseconds instead of the 20 to 30 seconds needed 
by the earlier spectrophotometer. The system is now 
being used to study the electron transport chain in 
mammalian mitochondria. 

Metabolic Energy Measurements (NHLBl). A 
microcomputer-based instrument has been designed 
to study the energy transduction phenomena of 
respiring membranes. Electrodes interfaced to the 
microcomputer via an A/D converter allow 
membrane potential and protonmotive force to be 
calculated. The computer also monitors a pH 
electrode and an oxygen-measuring electrode. The 
oxygen uptake rate of the cellular material is 
calculated as is the proton extrusion rate and the 
proton-to-oxygen ratio. Users can observe all 
quantified parameters on a multipen plotter as they 
alter the medium of the experiment. 

Electron Microanalysis Facility (DRS). GSL is 
collaborating with BEIB, DRS, to develop an 
automated electron microanalysis facility consisting 
of two electron microscopes interfaced to a PDP-11/ 
60 computer system. The facility will be used for 
research into the elemental composition of biological 
specimens, and for the development of new 
techniques in electron microscopy. GSL is designing 
and implementing the computer system, which will 
acquire and display the spectra and images resulting 
from Electron Energy Loss (EEL) and energy 
dispersive x-ray spectrometry (EDS). This year, 



software for acquiring EEL and EDS spectra and 
EEL, EDS, and electron current signal images was 
completed. Basic software for displaying this data 
was also completed. A large data base of empirical 
x-ray emission and absorption data was compiled 
and validated, and retrieval software implemented. 
Software for the analytical spectrometer, and for 
'housekeeping' data acquisition and calibration was 
improved. A user interface for most data acquisition 
and display functions was designed and 
implemented. 

Electron Microanalysis Computer System 

(NINCDS). Late in FY81, the Laboratory of 
Neuropathology and Neuroanatomical Sciences, 
NINCDS, requested assistance in setting up a 
minicomputer system to be connected to a JEOL 
JEM-100CX electron microscope equipped with a 
Kevex 7000 Analytical Spectrometer, an Energy 
Dispersive x-ray detector, and an Electron Energy 
Loss Spectrometer. CSL recommended the purchase 
of a system compatible with the BEIB Electron 
Microanalysis Facility so that software developed for 
that project could be used without modification. A 
PDP-11/34 was delivered. CSL installed the 
operating system along with the Kevex 7000 and x- 
ray emission data base software developed for the 
BEIB facility. 

Microelisa Data Logger (NHLBI). A microprocessor- 
based instrument has been developed to record data 
from a Dynatech Microelisa Autoreader, a 
commercially available spectrophotometer. This 
special purpose instrument is another variation on 
CSL's digital cassette data recorder. The instrument 
receives ASCII data from the Microelisa via an 
RS232 interface port, edits and formats the data for 
recording, and writes blocks of data onto a digital 
cassette tape. The instrument will either be 
transparent to communications between terminal and 
modem or, when directed by the operator, will read 
tape records and transmit the data over a 
communications line via the modem. Using a STD 
microprocessor bus has resulted in a more compact 
instrument requiring less power and less unused 
circuit board 'real estate' than our previous designs. 

Molecular Interactions Laboratory Data System 

(NHLBI). This microcomputer (PDP-03) data system 
supervises the acquisition and processing of 
information from an analytical ultracentrifuge and a 
circular dichroic spectropolarimeter used in MDB, 
NHLBI, to investigate the interactions between 
human lipoprotein subunits. Preprocessed data are 
transferred to the DECsystem-10 for further analysis 
under MLAB using predefined procedures invoked by 
a few simple commands. The results of a study of 
system performance with very dilute solutions were 



presented at the 1 982 FASEB meeting in New 
Orleans. Plans are now underway for the 
modification of the ultracentrifuge interface to 
provide greater noise immunity and for the addition 
of a digital plotter to the system. 

Californium-252 Plasma Desorption Mass 
Spectrometer Data System (NHLBI). After delays 
caused by hardware design problems, software 
upgrades, and facilities renovations, both the 
spectrometer and a data system modeled after one 
in use at Texas A&M University have been installed 
and are now functional. This instrument provides NIH 
the capabilities of mass analysis for compounds 
difficult or impossible to analyze by other mass 
spectrometric means. It also extends the range of 
mass analysis to compounds with molecular weights 
in excess of 5000. 

DLDACS Project (NIADDK). An integrated 
laboratory data acquisition and processing system 
has been developed for LCP and LMB, NIADDK. The 
system is configured with thirteen satellites coupled 
through a local network to a host processor. At each 
satellite, a dedicated microcomputer system 
performs data acquisition from and control over an 
instrument/experiment. Although acquired data files 
may be stored locally, they are normally transferred 
via the network to a host storage medium. The local 
network allows the host storage medium to appear 
as a virtual storage device to the satellites. The hub 
of the network, the concentrator, utilizes DMA 
hardware on all communications links and performs 
a file store and forward function. Processing 
software provided at the host allows LDACS data 
files to be added, subtracted, averaged, smoothed, 
baseline corrected, integrated, differentiated, 
multiplied by a constant, or added to a constant. The 
results may be displayed graphically on a Tektronix 
terminal, typed at a terminal, printed on the line 
printer, plotted on an X-Y plotter, or transmitted to 
the NIH DECsystem-10 for additional processing. 

Program Management And 
Administration 

Small Animal Section Data Base Management 
System (DRS). In FY82, CSL revised and expanded 
the functional specifications developed in FY81 to 
generate a Request for Proposals for a small animal 
data management system for the Small Animal 
Section (SAS), VRB, DRS. We issued a Sources 
Sought to test the market for currently operating 
animal data management systems. Sixteen vendors 
responded, of which only three had the necessary 
experience to be evaluated as qualified. The RFP will 
be issued in the final quarter of FY82. Responses 



will be evaluated by early FY83. We anticipate that a 
contract will be signed in the second quarter of 
FY83. 

Library Automation (DRS). A CSL study performed 
in FY79 recommended that the NIH Library purchase 
a commercially available library system. Because 
none of the available systems could completely 
satisfy the Library's requirements, the study 
emphasized system flexibility in order to allow CSL 
programmers to make modifications and additions to 
the system. Due to administrative delays the 
procurement process did not start until late FY81. 

During FY82 , CSL has been heavily involved in a 
series of complex contract negotiations occasioned 
by an early admission from the contractor that work 
could not be performed In the promised time frame. 
The period of the contract was eventually extended 
for several months in return for substantial cost and 
technical concessions. In addition, CSL is 
supervising site preparation and providing 
programming support for conversion and editing of 
part of the data base not covered by the contract. 

Biomedical Communications And 
Conference Support 

Computers in Cardiology Conference. CSL 

continued its support of the annual International 
Conference on Computers in Cardiology. The 
Conference provides a forum for direct interaction 
and exchange between physicians, computer 
scientists, and engineers who are involved in various 
aspects of clinical computer systems in the field of 
cardiology. 

Computer Research And 
Technique Development 

Image Processing Facility. This project is intended 
to provide a utility to display and analyze digital 
images. The system will consist of a powerful 32-bit 
computer with a mixture of medium and high 
resolution video displays. Also, the system will 
include a microdensitometer to allow precise 
digitization of images. The computer and peripherals 
have been purchased and procurement of the 
display subsystem is in progress. Site preparation is 
also underway and completed system installation is 
forecast for early 1983. 

Analytic Models of Computer System 
Performance. This new project involves the 
development of analytic models that can be used to 
evaluate the performance of computer systems. 
During the past year, tools for modeling and 
analyzing computer systems using the graph 



theoretic model called Timed Place-Transition (P-T) 
Nets were developed. A method was developed for 
evaluating computer system performance with Timed 
P-T Net models. This method was used to model 
and analyze the bus arbitration techniques that occur 
in digital systems. Important upper bounds on the 
average time a device waits for the bus and the 
average time the shared bus is not used were 
derived. In addition, a state variable P-T Net model 
of the interconnection of two or more 
microprocessors was developed. This model 
provides a framework for determining the avoidance 
of deadlock and the maintenance of throughput in 
multiple microprocessor systems. 

Research Projects 

Computer Support for Flow Microfluorimetry/Cell 
Sorters (FMF) 

This project provides PDP-11 computer support at 
various levels for four Becton-Dickinson FACS II and 
one Coulter MDADS flow cytometry/electronic cell 
sorting instruments. Data acquisition is via an NIH 
designed interface to the computer. Data display and 
analysis for high sample throughput is the principal 
system feature. Software currently running under the 
RT-11 operating system is being converted to 
function under the RSX-11M operating system in 
order to allow more sophisticated recordkeeping and 
more effective support of current and anticipated 
workloads. New hardware and software capabilities 
are being added during the conversion effort. Upon 
completion the RSX System will be installed on the 
Immunology Branch, NCI computer. 

Objectives and Methods: Since FY75 CSL has 
provided engineering, system integration, and 
software support necessary to meet the data 
acquisiton, data display, and analysis needs of 
several investigators using flow cytometry/electronic 
cell sorting instruments at NIH. Software 
development and testing is done on a Digital 
Equipment Corporation (DEC) PDP-11/40 computer 
system owned by CSL. This allows investigators to 
have full use of their systems while new software is 
being developed. 

All systems are currently using the RT-11 single-user 
operating system. The systems allow data collection 
of up to four parameters on individual cells. Typically 
these are light scatter, two frequencies of 
fluorescence, and cell volume. The data can be 
collected in single parameter or correlated dual 
parameter modes. Data analysis and display 
programs allow the experimenter to produce various 
statistics and hard copy displays from the acquired 



data. The displays include three-dimensional 
pictures, contour maps, and vertical slice sections. 

Progress in FY82: In FY82, CSL continued to 
support FACS-ll/PDP-11/34 systems for I, NCI; 
LP,NCI; EA,H; EEB, NCI; and a Coulter MDADS/ 
PDP-11 /34 system for MO.NCI. 

The major effort in FY82 was continuing the 
conversion of RT-1 1 programs to run under the RSX- 
1 1 M multi-user operating system as well as adding 
functionality to these programs. The RSX-1 1 M 
system is initially being developed for one system in 
order to provide more effective support of current 
and anticipated workloads and more sophisticated 
data acquisition and recordkeeping functions. This 
system will be available to other NIH FMF sites as 
needs require after its implementation in the 
Immunology Branch, NCI. 

It was decided in FY81 to replace the DEC VT-1 1 
graphics display device with a Tektronix 4025 and to 
support this terminal on the RSX system. A new 
graphics software package was needed to drive the 
4025 initially for the RSX-11M system. In conjunction 
with the T4025, a contract was negotiated with 
Electronic Data Systems, Inc., to assist CSL 
personnel in developing software packages for 
displaying graphs on any terminal that is capable of 
performing Tektronix 4010 style graphics. 

All four of the major FMF data display and analysis 
programs were rewritten to run under RSX-1 1 M, 
using the new graphics software package and the 
Tektronix 4025 terminal that was delivered in the first 
quarter of FY82. Several minor improvements to the 
RT-1 1 display and analysis programs were made in 
FY82 in order to accommodate immediate needs of 
our supported users. 

The new data acquisition system environment 
consists of a PDP-1 1 host computer running RSX- 
11M and up to eight LSI-11 based satellites, each 
running RT-1 1 , connected to the host via an 
interprocessor link. 

Data acquired by a satellite is usually sent over the 
link and stored at the host site (Remote Storage 
Mode). However, in the event of a link failure, data is 
stored at the satellite site (Local Storage Mode). The 
satellite link software is common to all satellites, but 
distinct from the host link software. Together, the 
host and satellite link software provides file transfer 
capability. 

The development of the link software was completed 
in FY82 and the hardware for a complete satellite 
system has been acquired. The acquisition portion of 
the LSI-1 1 acquisition system has been written and 
is being tested. 



An important feature of the satellite system is the 
ability to create a 'laboratory notebook' as a 
permanent hard copy rather than continuing this as a 
manual task as in the existing PDP-1 1 /34 system. 
This 'notebook' concept is an integral part of the 
software that provides interaction with the operator 
via a DEC VT-1 00 terminal. Special features of the 
VT-1 00 are used to provide the operator with an 
easy-to-use single screen menu. Errors are reported 
in detail on the terminal screen. This interactive 
software was completed in FY82. 

During the third quarter of FY82, the DEC PDP-11/ 
40 computer system used for development of 
software for CSL-supported FMF systems was 
relocated to Building 12A from Building 10. CSL has 
also responded to many external requests and has 
provided copies of the interface hardware 
schematics, software, and documentation to FMF 
sites in the U.S., Europe, and Australia. 

Proposed Course: In the forthcoming year, CSL 
plans to complete the first RSX-1 1 M based FMF 
system and LSI-1 1 based data acquisition system 
and put them into operation at the I, NCI facility. If 
resources permit, the RT-1 1 software will be 
rewritten to use the T4025 graphics terminal as a 
replacement for the VT-1 1. CSL will also continue to 
support existing RT-11 based FMF sites at NIH. 



f^.'^'l^.'^X'^rT\lTA'^^T 


"'™'-'So!|\j8fJ=™]|^'^ 


ZOl CT00050-03 CSL 


""nctober"!, 1981 to September 30, 1982 


Computer Support"Vo" no"Microfluorimetry/Cel 1 Sorter (FHF) 


PI: R. J. Romanoff Computer Specialist CSL, nCRT 

OTHERS: R. Fico Electronics Engineer CSL, nCRT 
E. S. Loiederman Computer Aid CSL, DCRT 
S. 0. Sharrow Chemist I, NCI 
A. R. Schultz Chief. Processor Design 

Section CSL, nCRT 


I, NCI: LP, NCI: EEB, NCI: HO, NCI: EA, NHLBI 


L»fl/BBANCH 








CHECK APPHOPRIAie eox{£s) 

Dt«l) "INORS a (.J) I'HERVIEWS 


EU«H*H. OF WOHK (SOO -or,.» or l.„ - und.rlio, ..,»ord,) 

This project provides PDP-11 computer support at various levels for four 
Becton-Dlckinson FACS II and one Coulter MDADS flow cytometry/electronic 
cell sorting instrument. Data acquisition is via an NIH- designed 
interface to the computer. Data display and analysis for high sample 
throughput is the principle system feature. Software currently running under 
the RT-11 operating system is being converted to function under the RSX-UM 
operating system in order to allow more sophisticated recordkeeping and 
more effective support of current "and anticipated workloads. 



Cardiac Scintillation Probe 

Background and Objectives: The development of the 
cardiac scintillation probe is a continuation of CSL's 
collaboration with the Nuclear Medicine Department, 
CC, and the Cardiology Branch, NHLBI. Originally 
this collaboration resulted in the development of a 
noninvasive cardiac imaging technique known as 
ECG-gated scintigraphic angiography. In addition to 
visualizing global LV cardiac function, the images 
produced by this technique can be processed to 
produce a time-activity curve that represents 
changes in ventricular volume over a cardiac cycle. 
This time-activity curve (LV volume curve) can be 
used to calculate various parameters of cardiac 
function such as ejection fraction, peak ejection rate, 
peak filling rate, and their temporal relationships. The 
imaging technique has been used with great success 
to characterize various cardiac related diseases. 
However, if the images are not required, then this 
time-activity curve could be generated by a much 
smaller and simpler system using a single small Nal 
detector and microcomputer system. In 1977 CSL 
began the development of a cardiac scintillation 
probe system, which could produce this time activity 
curve. The system is easily transportable and allows 
continuous monitoring of cardiac function at the 



'"i^)Xi,w 


net INfOMtTIM EICNMCt: 

IS. HOT ui. (M. .p.c;i 


- 


"■^'irfilki"""''"' 








IblUIMaM. HLIUMM nUiftUI 




PUICO COVERU 






October 


1. 1981 to September 31 


. 1982 


Cardiac 


CI (M Eh.r.ct.r. or U..) 

Scintillation Probe 












nRtOHNa EHSlCtO ON IHt MIOJE 


T 


PI: 


H. G. Ostrow 


Electronics Engineer CSL. DCRT 


OIHEBS 


S. I. Allen 


Medical Besearch Analyst CSL. OCRI 




S. Bacharach 


Physicist NH. CC 
Physicist NH. CC 




R. Bonow 


Cardiologist CB, NHLBI 




D. Rosen 


Cardiologist CB, NHLBI 


COOPtHlIINC t 


IIS (if .-,) 




Nuclea 


Medicine, CC, Cdrdldo 


gy Branch, KHLBl 








iKfS™"'" 


r Syitemi i ahnr.if,nrj 




Processor Design Section 
DCRT. NIH. Bethesda MO 20205 


:](.) Hiii*N 


.0 I l.O 

Mt m(€S) 

uBJtcis a (b) m 


1 


D(.0 "i^w 


U Ui) 1NU«»I(«S 




SUWt*R<r or hOR* (200 .er4. or It.. . undtrL 


n. b.p,«r<l<) 


CSL cont 


nued the dewelopment o 


Its Cardiac Sclnttllatlon Probe System 


begun In 


1977. This nonimaging 


ECG-gated scintillation probe, when used in 


conjunct 


on with left ventricul 


r (LV) catherizatlon. permits simultaneous 


quanttf 1 


atlon of the variation 


of LV volume and pressure. By simultaneously 


rneasurln 


LV volume and LV pres 


ure. parameters such as LV compliance can be 


continue 




ion to such measurements as ejection fraction. 


filling 


nd ejection rates, and 


temporal relationships. This year the probe 


continued to be used to study t 


e effects of nephldlpine and verapermil on 


patients 


with asymmetric septal 


hypertrophy. In addition, a new protocol 


Studying 


the drug's effects on patients with coronary artery disease was 


tnltUte 


. The probe is also tw 


\nq used to monitor the left ventricle 


per forma 


ce of patients in the 


edical Intensive Care Unit. The pressure-volum^ 


relation 


hips produced by the probe system allowed the effects of drugs to be 


quant ita 






permit t 


e system to be easily 


sed on a routine basis by Clinical Center 


the proh 


and In quantifying the llmtfatlon nf IhP tPrhninne. 1 



bedside or other location in the Clinical Center 
outside the Nuclear Medicine Department. 

Methods: The system consists of a 3-inch diameter 
Nal scintillation probe, probe electronics, 
microcomputer system, and display. The system is 
programmed to acquire scintillation data from the 
probe, process the data, and plot and display various 
parameters of left ventricular (LV) function. This 
nonimaging, ECG-gated probe, when used in 
conjunction with ventricular catheterization, permits 
simultaneous quantification of the variation of LV 
volume and LV pressure. Parameters such as LV 
compliance can be continuously monitored. In the 
catheterization laboratory, pressure-volume 
measurements are used to study the effects of drugs 
on patients with various heart diseases. 

Progress in FY82: This year the probe continued to 
be used in the catheterization laboratory to study the 
effects of nephidipine and verapermil on patients 
with asymmetric septal hypertrophy (ASH). The 
pressure volume data from these studies were 
transferred to the PDP-10 system and software was 
written to analyze the data to determine additional 
parameters to assess the effects of these drugs. The 
verapermil studies have been concluded. The 
pressure-volume data has provided information to 
help understand the positive effects of the drug in 
this particular disease. 

A new protocol studying drug effects on patients with 
coronary artery disease was initiated. As with the 
ASH study the data is being analyzed to determine 
the parameters that best characterized the effects of 
the drug. 

Results of the drug intervention studies indicated 
changes in ventricular parameters previously not 
used. Nine studies were designed and performed in 
order to quantitate the limitations of techniques in 
measuring these parameters. 

The Medical Intensive Care Unit, CC, has started to 
perform ECG-gated Scintiangiographic studies. 
Because using the Pho/Gamma camera on a routine 
basis at the bedside is difficult, software was written 
on the HP system to use the probe in this 
application. 

Significance to Biomedical Research: Nuclear 
Medicine techniques provide a relatively noninvasive 
procedure to access left ventricular function. The 
cardiac scintillation probe permits this capability to 
be used for clinical research studies at the bedside 
and in the catheterization laboratory. The pressure 
volume relationship produced by the probe system 
allows the effects of drugs to be quantitated in a 
manner not before possible. 



Proposed Course: Development activities in 
response to new applications are expected to 
continue. CSL will investigate making the probe and 
camera systems compatible. Making the systems 
compatible to the extent possible will reduce the 
resources required to support the probe system and 
allow new capabilities developed for the camera 
systems to be implemented quickly on the probe 
system. 

Computerized Radiation Therapy 

CSL has developed a computer system, now in 
clinical operation in the Radiation Oncology Branch, 
NCI, to use the detailed contour and density 
information available from ultrasound or computer 
assisted tomography to improve radiation treatment 
planning. This system for external beam treatment 
planning is based on a generalized 3-D dose field 
model that covers photon, electron, and neutron 
beams. 

Both the computer program and most of its clinical 
implementation have been completed for the photon 
and electron fields available from the local 6 mv and 
12 mv linear accelerators. The current capabilities 
include interactive simulation of most irradiation 
techniques, including the effects of most beam 
modifying devices. The system enables the display 
of dose distributions computed in several transverse 
contours and overlaid on corresponding CT or 
ultrasound scans. 

Background and Objectives: To develop and 
implement a generalized system for computer- 
assisted radiation treatment simulation. 

Methods Employed: The dose field model developed 
by Jan van de Geijn was implemented in RSX-1 1 M 
FORTRAN and experimentally tested to cover 
irregularly shaped beams as well as irregularly- 
shaped shielding blocks. A facility has been 
developed that enables the computation and display 
of dose distributions in planes perpendicular to the 
beam axes. 

Progress in FY82: The capabilities of the graphics 
input system, the use of CT and ultrasound images 
in addition to mechanically-obtained patient contours, 
and the color display system have been further 
expanded. Eight user terminals and two DeAnza 10 
5000 color display systems have been added to 
allow the PDP 1 1 /70 computer system to be used 
for treatment planning and system development. 

I^ajor Findings: The system, although continuing to 
be expanded, is in routine use for clinical treatment 
planning, in comparison to other existing systems, it 
offers high speed computation and display of 



complete dose distributions in multiple slices 
superimposed on CT or ultrasound images, including 
effects of wedge filters, irregular shielding blocks, 
and diaphragm rotation. Several modes of display 
are available. 

Proposed Course: Future plans include: 
implementation of the beam's eye view option for 
regular and irregular electron fields, extension of the 
current software package to include point source 
(seed) calculation, and extension of the capabilities 
to compute and display dose distributions in sagittal, 
coronal, and beam's eye view sections of the patient 
on an interactive basis. 

Pubiications: 

van de Geijn, J., Chien, l-Chu, Cheng, C. P., and Fredrickson, H.A.: A 
Unified 3-D Beam Model tor External Beam Dose Distributors. Proceed- 
ings of the VIII International Conference on Computers in Radiottierapy, 
Tokyo, Japan, 1 980. 



pSijEcrNfiMaefi'fo^MorS^fthis =""S5" 




PROJECT HU«BE« 








ZOl CT00052-03 CSL 




""october^l. 1981 to September 30, 1982 | 


TITLE OF PFfOJECI (80 ch.r.clrrs or U=s) 




Computerized Radiation Therapy 




pSoF£blSHr"HloNNEL'™«£J^H*^riE'p^^^^^ *'*" ^"^^^ "^ PAIN^IPAL INVEST ICATOHS ANO ALL OTHER 




PI: H. Fredrickson Computer Systems Analyst CSL, DCRT 




OTHERS: J. Van de Geijn Head, Radiation Physics 




Automation Section ROB, OCT, 


NCI 


D. Syed Head, Systems Design Section CSL. DCRT 




E. Glatstein Chief, Radiation Oncology 




Branch ROB, OCT, 


NCI 


B. Fraass Staff Fellow ROB. OCT. 


NCI 


L. Freeman Computer Programmer CSL, DCRT 




COOPEHATIIIC UNIT! {If any) 




Radiation Oncology Branch, NCI 




Computer Systems Laboratory | 






Systems Design Section 




DCRT, NIH, Bethesda, HD 20205 


^^ 1 PROFESSIONAL, ^^^ jOrHER, 


CHECK APPROPRIATE aOx(ES) 




D {.) HUMAN SUBJECTS Q (b) HO-AH TISSUES Q {.) NEITHER 




DM) MINORS a (.2) iriTERV.E-S 








CSL has developed a computer system, now in clinical operation in 




the Radiation Oncology Branch, NCI, to use the detailed contour 




and density information available from ultrasound or computer 




assisted tomography to improve radiation treatment planning. This 




system for external beam treatment planning is based on a 




generalized 3-D dose field model that covers photon, electron, and 




Both the computer program and most of its clinical implementation 




have been completed for the photon and electron fields available 




from the local 6 mu and 12 mv linear accelerators. The current 




techniques, including the effects of most beam modifying devices. 




The system enable?, the display of dose distributions computed in 




several transverse contours and overlaid on corresponding CAT or 













Medical Intensive Care Unit Patient Monitoring 
Computer System 

The dynamic events occurring within the Clinical 
Center's Medical Intensive Care Unit are monitored 
by a unique multiple-computer system. Capabilities of 
the system include online data acquisition and 
analysis, medical recordkeeping, tabular and 
graphical data displays, and feedback control, as 
required in support of patient care and research 
protocols. Elements include a minicomputer-based 
Patient Data Management Subsystem, a Software 
Development Subsystem, and a Medical Mass 
Spectrometer Subsystem. 

The facility also contains a state-of-the-art 
catheterization laboratory that includes a flexible 
computerized Vascular Research Subsystem, with 
physiologic waveform processing features, and a 
high-resolution x-ray system with digital subtraction 
angiography capability. 

Of primary interest is the utilization of the Medical 
Intensive Care Unit's computer systems in the study 
of the etiology and therapy of septic shock. 

Background and Objectives: The Medical Intensive 
Care Unit (MICU), which is administered by the 
Department of Critical Care Medicine in the Clinical 
Center, receives critically ill patients from clinical 
programs of NIH. The MICU comprises a five-bed 
ward area, a pair of isolation beds, and a vascular 
research laboratory. The research goals of this unit 
include the development of techniques for 
automated patient monitoring and noninvasive 
measurements of the cardiovascular and respiratory 
systems. In addition, catheterization studies are 
performed as necessary to obtain data that are 
available only through invasive methodology. 

Working with Clinical Center staff, CSL contributed to 
the engineering design of the intensive care unit. 
CSL also undertook the specification, procurement, 
and installation of the bedside patient monitoring 
equipment and the four computer systems: 

1 . a Patient Data Management System used for 
automatically monitoring patient variables, manually 
entering patient data, retrieving information online, 
and keeping medical records; 

2. a Vascular Research Subsystem used for 
acquiring and processing cardiovascular pressure 
waveforms, measuring cardiac output, displaying 
measured results online, and generating a cardiac 
catheterization report; 

3. a Software Development Subsystem used for 
developing software for the above described 
systems; and 



4. a Medical Mass Spectrometer Subsystem used 
for monitoring both the patient ainway gases and the 
gases delivered by the patient's respirator at all 
seven MICU beds. 

The first three systems were purchased from the 
Hewlett-Packard Corporation and all use identical 
minicomputers. The Chemetron Corporation 
manufactures the microcomputer-based mass 
spectrometer system. This year, additional 
subsystems have been incorporated within the MICU 
complex. 

Major Findings: The automation of the MICU has 
aided the medical staff by managing the large 
amount of data needed for the care of the critically ill 
patient, performing desired calculations, and allowing 
measurements that would not othenwise be possible. 

Progress in FY82: Modifications to the Vascular 
Research Subsystem were completed to allow the 
data generated to be collected by the Patient Data 
Management Subsystem's computer, in addition to 
the Vascular Research Subsystem's computer. A 
mobile noncomputerized Vascular Research 
Subsystem was added to the MICU to allow bedside 
monitoring with research quality instrumentation. This 
mobile subsystem also interfaces to the Patient Data 
Management Subsystem's computer. 

In order to provide pulmonary function monitoring for 
the critically ill population within the MICU, a Collins 
Pulmonary Function Testing Subsystem was 
integrated into the MICU. This microprocessor-based 
subsystem contains the spirometer and gas 
analyzers necessary for the calculation of pulmonary 
function parameters, such as vital capacity and lung 
volumes, and for the generation of flow-volume 
loops. 

A Hewlett-Packard Respiratory Research Subsystem 
was obtained to provide for the computation of 
pulmonary mechanics parameters, such as work of 
breathing and lung compliance and resistance. This 
subsystem also utilizes airway pressure and flow 
transducers to develop pressure-volume loops, in 
addition to flow-volume loops. 

A Microprocessor-Controlled Arrhythmia Monitoring 
Subsystem was added to the Nurses' Station in the 
MICU to provide for the central monitoring of cardiac 
arrhythmias. This display station provides an 
electrocardiogram memory and provides 
simultaneous hard copy records of the 
electrocardiogram and arterial pressure waveforms, 
as aids in the detection of transient arrhythmias. 



In order to supplement the electrocardiographic and 
cardiac catheterization data descriptors of cardiac 
function, a Hewlett-Packard Microprocessor 
Controlled Ultrasound Imaging Subsystem was 
installed in the MICU. This subsystem provides 
multiformat displays of cardiac structure and allows 
the visualization of intracardiac abnormalities. In 
addition, this Ultrasound Imaging Subsystem 
interfaces to the Software Development Subsystem 
to allow sophisticated image processing with NIH- 
developed programs. 

A Cardiac Probe, which was developed jointly by 
CSL staff and the Clinical Center's Nuclear Medicine 
Department, is being adapted for use in the MICU. 
This device provides left ventricular volume data by 
counting gamma ray induced scintillations, after the 
administration of injectable radioisotopes. Software is 
being completed to allow data from the Cardiac 
Probe to be collected, analyzed, and displayed 
utilizing the Software Development Subsystem's 
computer. 

A Philips X-ray Imaging System was installed within 
the MICU Catheterization Laboratory. CSL staff 
assisted in the specification and supervision of 
catheterization laboratory alterations necessary for 
system installation. Shielding was installed to protect 
the MICU computer room from exposure to 
electromagnetic interference from the x-ray system's 
generator and power supplies. The x-ray Imaging 
System includes a Computerized Digital Vascular 
Imaging Subsystem that provides the capability for 
digital subtraction angiography. A video switching 
device added to the Vascular Research Subsystem 
allows video displays produced by the Vascular 
Research Subsystem, the Patient Data Management 
Subsystem, and the Digital Vascular Imaging 
Subsystem to share a single Philips Video Monitor. 

Significance to Biomedical Researcti: This project is 
focused on the application of multifaceted diagnostic 
modalities to clinical research and the care of the 
critically ill patient. Any new developments made on 
this project will benefit many users of automated 
systems, as well as patient care and clinical 
research within the MICU at NIH. 

Proposed Course: Future efforts will center on 
hardware and software modifications necessary to 
enhance the system's ability to support patient care 
and reserch protocols. Possible modifications to the 
primary Patient Data Management Subsystem 
include the addition of urine output measurement 
scales and the computerization of fluid infusion 
therapy utilizing existing microprocessor-controlled 
infusion pumps. 



PM"ct"iS!lBEB'[D"piOT"«"hlSI -"«ll" 


, heal;h'and"h"man''L?vi 


ES 


PHOJECt «n..E« 






«T 


ZOl CT00054-03 CSL 


October 1. 19fil to Seotember 30. 1982 


Medical Intensive Care Unit Patient Monitoring 


Computer System 


UAHIS, UeORATORY AMD INSTlTUie AFFILIATIONS, AND TITLES OF PRIH 


IP«L INVJSrl««IORS .«0 ILL DIHM 


PI: K. M. Kempner Electronic 


s Engineer CSL, OCRT 


OTHERS: J. E. Parrillo. H.D. Chief, Cri 

L. W. Freeman Computer P 

S. L. Huntley Supv. Crit 

Technician 


tical Care 

CCHO, CC 
s Engineer CSL. OCRT 
rogrammer CSL, OCRT 
cal Care 
s CCMn, CC 


Critical Care Medicine Department. Clinical Center 


LAB/BHAriCH 

Computer Systems Laboratory 


Systems Design Section 


OCRT, NIH, Bethesda, MD 20205 


1.0 1 1.0 1 

CHECK APPflOPBIATE BOX(ES) 


a (c) m„c. 


□(.1) MINORS D(«2) INTERVIEWS 




SUMMARY OF WORK (!00 »ord. or 1«= - underlio. kcy-ord = ) 




The dynamic events occurring within the Clinica 
Intensive Care Unit are monitored by a unique mu 
system. Capabilities of the system include onl 
and analysis, medical recordkeeping, tabular an 
displays, and Teedback control, as required in 
care and research protocols. Elements include 
Patient Data Management Subsystem, a Software D 
and a Medical Mass Spectrometer Subsystem, 


Center's Medical 
tiple-computer 
ne data acquisition 
d graphical data 
support of patient 

evelopment Subsystem, 


The facility also contains a state-of-the-art c 
laboratory that includes a flexible computerize 
Subsystem, with physiologic waveform processing 
high-resolution x-ray system with digital subtr 
capability. 


d Vascular Research 

features, and a 
action angiography 



^-^----^' 


INIRABUfiAL'*REsLHCH PflOJECT 


ZOl CT00099-01 CSL 




TITLE OF PROJECT (8D characters or less) 




Automated Management of Critically 111 Patients 




'pR*SF"ssl3HAL"^S0NNEl'EHG«£J^N*^Ti£'p^^^^^ *'"' ^'^""^ "^ '^"'""'' 


INVESTIGATORS AND ALL OTHER 


PI: K. M. Kempner Electrc 


nics Engr. CSL, OCRT 


OTHERS: J. E. Parrillo M.n. Chief, 
N. fleClaris, Sc.fl. Profes 


Critical Care 

Medicine CCMO, CC 

or. Electrical 

Engr. Oept. Univ. of MO 


COOPERATING yillTS (.1 4"y) 




Critical Care Medicine Department, CC 
Electrical Engineering Department, Univ. of MD 




LAB/BRANCH 


























CM HUMAN SUBJECTS Q (b) HUMAN TISSUES 


n {=) NEITHER 


D (-1) MINORS aU2) INTERVIEWS 




SUMMARY OF WORK (200 wgrda or Ui= - underline koy-ordi) 




This research project is concerned with a system 
management of critically ill patients in a clinic 


approach to Che 
al setting. The 


ultimate goal is the utilization of computer-base 
to aid in the differential diagnosis of disease 
implementation of therapeutic modalities through 
technology. 


d instrumentation 
tates and the 


ft state variable approach is utilized in the matti 
of pertinent pharmacokinetic and physiologic proc 
clinical data and realtime monitored values are 
validation. Several alternative methods for clos 
medical interventions are being investigated. 


ematical modeling 
esses. Empirical 
tilized in model 



10 



Automated Management of Critically III Patients 

This research project is concerned with a systems 
approach to the management of critically ill patients 
in a clinical setting. The ultimate goal is the utilization 
of computer-based instrumentation to aid in the 
differential diagnosis of disease states and the 
implementation of therapeutic modalities through 
automated technology. 

A state variable approach is utilized in the 
mathematical modeling of pertinent pharmacokinetic 
and physiologic processes. Empirical clinical data 
and realtime monitored values are utilized in model 
validation. Several alternative methods for closed- 
loop automated medical interventions are being 
investigated. 

Background and Objectives: Noninvasive diagnostic 
and therapeutic techniques generally involve the 
application of sophisticated electronic technology 
and mathematical modeling techniques to the 
detection of pathophysiologic states. Particularly 
interesting and important problems involve 
cardiovascular disorders that give rise to low output 
syndrome. 

There is no singular cause for this syndrome, and 
therefore effective therapy requires the differential 
diagnosis of numerous contributory disturbances in 
cardiovascular homeostasis. Effective therapy 
principally involves the administration of one or more 
fluids and/or drugs in a critical care unit 
environment. 

Methods Employed: In order to accomplish the goal 
of developing systems capable of assisting in the 
medical management of a critically ill patient on a 
closed-loop basis, it will be necessary to develop 
validated models. Calculated physiologic parameters 
will be compared to measured physiologic data as 
the patient's response to the selected therapy 
progresses. 

A mathematical formulation of the relevant 
subsystems will be developed for a patient in a 
critical care unit setting. This includes the modeling 
of three principal subsystems: Pharmacokinetics, 
Drug/Receptor Interactions, and Cardiovascular 
Dynamics. 

Progress In FY82: An extensive literature search and 
a formal analysis were performed on the three major 
relevant subsystems. The literature search indicated 
that these three areas have previously been treated 
as distinct and unconnected problems. There has 
been little effort to combine them in a manner 
suitable for addressing the problem proposed. 



The mathematical formulations necessary to 
describe these systems are being finalized. Particular 
care is being paid to the selection of variables so 
that the subsystem models will couple in a 
physiologically sound as well as a computationally 
efficient manner. 

Significance to Biomedical Research: The use of 
automated systems in the implementation of 
therapeutic protocols within a critical care unit adds 
a new treatment modality and will have a major 
effect on protocol design. It will afford improvements 
in protocol design for patient care, clinical drug trials, 
and the study of the etiology and therapy of specific 
disease entities. In addition, the automation of 
therapeutic interventions, as proposed, will 
significantly expand the clinical and research data 
bases. 

Proposed Course: Following the development of the 
mathematical formulation of the three major 
subsystems, these subsystems will be implemented 
in software on the DCRT Central Computer Facility. 
The software will be utilized in a simulation mode to 
analyze actual patient data, and to generate 
recommendations for therapy along with predicted 
physiologic data values. 

Existing critical care protocols will be investigated to 
identify those components in which automated 
therapeutic modalities can easily be accommodated, 
within the framework of this research effort. An 
important aspect to be evaluated is the risk to the 
patient versus the realizable benefits. 

Selected protocols will be implemented utilizing the 
closed-loop techniques developed in this project, 
with the objective of carrying out controlled clinical 
trials and quantitatively evaluating their effectiveness. 



Positron Emission Tomography (PET) Scan 
Image Analysis in Aging Studies 

Procedures for transporting Positron Emission 
Tomography (PET) Scan and Computer Assisted 
Tomography (CAT) Scan images from the NIH 
Clinical Center to the DCRT Image Processing 
Facility have been established. An interactive 
computer procedure for delineating anatomical areas 
of interest on a CAT scan and computing metabolic 
activity from the corresponding area on the related 
PET scan has been developed. Improved methods 
for establishing external coordinates to align 
corresponding PET and CAT scans continue to be 
explored. 



11 



Background and Objectives: Positron Emission 
Tomography (PET) scanning performed in the 
Nuclear Medicine Department of the NIH Clinical 
Center provides a spatially-sequenced series of 
images of regional cerebral glucose metabolism in 
man. The Laboratory of Neurosciences of the 
National Institute on Aging wishes to incorporate 
PET scanning technology in the study of diseases 
associated with aging. The initial goal of this project 
is to delineate brain substructures represented in 
spatially sequenced Computer Assisted Tomography 
(CAT) scan images and to determine metabolic 
activity in these substructures from corresponding 
spatially sequenced PET scan images. 

Methods Employed: Computer Assisted Tomography 
(CAT) Scan and Positron Emission Tomography 
(PET) Scan images are transported from the NIH 
Clinical Center to the DCRT Image Processing 
Facility via magnetic tape. An analyst trained in 
neuroanatomy aligns associated CAT and PET Scan 
images by means of reference markers seen on both 
images. These markers are provided by a head- 
holding device worn by the subject during CAT and 
PET scanning. The analyst then 'draws' contours 
around regions of interest on the CAT Scan. The 
computer determines corresponding regions on the 
PET Scan and computes metabolic activity in that 
region of the brain. 

Progress in FY82: Radiation-to-glucose conversion 
equations were finalized and further investigation of 
a suitable head-holder to provide reference markers 
external to the brain continued. Image processing 
software is close to completion. A project entitled 
'Computer Assisted Tomography (CAT) Scan Image 
Analysis in Aging Studies' closely related to this work 
was started and is reported in this Annual Report. 

Significance to Biomedical Research: It is anticipated 
that this work will provide a basis for evaluating the 
utility of PET scanning in studying diseases 
associated with aging. Successful implementation of 
an external coordinate system should provide for 
accurate anatomical region designation via higher 
resolution CAT scan images to measure 
physiological processes from corresponding lower 
resolution PET scan images. 



PfiOJWT^HuHBER'loSUfSsMh?" ^"c!!!" 


,l™sfi!lfZi°i 


ZOl CT00083-02 CSL 




Positron Emission Tomography (PET) Scan Image Analysis in 


PI: J. M. neLeo Computer Systems Analyst CSL, DCRT 
OTHERS: S. I. Rapoport Chief LN. NIA 


Laboratory of Neurosciences (LN). NIA; Nuclear Medicine 
Department (NM). CC; Oiagnostic Radiology (DR), CC 


LAB/BR«NCH 


Systems Desiqn Section 


nCRT. NIH. flethesda. Mn 20205 


.1 1 .1 1 


CHECK APPROPRIATE BOX(eS) 

D[*l) M.NORS Q(.3) IHICRVIEWS 


Procedures for transporting Positron Emission Tomography (PET) 
Scan and Computer Assisted Tomography (CAT) Scan images from the NIH 
Clinical Center to the DCRT Image Processing Facility have been 
established. An interactive computer procedure for delineating 
anatomical areas of interest on a CAT scan and computing metabolic 

been developed. Improved methods for establishing external 

explored. 



Zni CT00086-01 CSL 



I September 30, 19i 



iisted Tomography (CAT) Scan Image Ana' 



S. I. Hapoport Chief 



alyst CSL, nCRT 
I LN, NIA 



DCRT. NIH. Bethesda. MP 20205 



D{'?) INI 



interactive image analysis computer procedure to measure 
rious parameters from Computer Assisted Tomography (CAT) scans 

the human brain has been designed and implemented on the DCRT 
age Processing Facility. 



12 



Computer Assisted Tomography (CAT) Scan 
Image Analysis in Aging Studies 

An interactive image analysis computer procedure to 
measure various parameters from Computer Assisted 
Tomography (CAT) scans of the human brain has 
been designed and implemented on the DCRT 
Image Processing Facility. 

Background and Objectives: This new project has 
been initiated to study changes in the human brain 
structure during normal aging and during brain 
disease processes associated with aging by means 
of measurements made from Computer Assisted 
Tomography (CAT) scans of human brains. 

Methods Employed: CAT scans are transported to 
the DCRT Image Processing Facility via magnetic 
tape. Through interactive analysis of the CAT scan 
images, an investigator is able to obtain a wide 
variety of descriptive measurements such as sizes 
and attenuation values of brain substructures and 
percent composition of white matter, grey matter, 
and cerebral spinal fluid. 

Progress in FY82: Software was developed to 
determine the following measurements from CAT 
Scans of the human brain: 

1) Bicaudate Index 

2) Ventricle/Brain Area Ratio 

3) Width of Ventricles 

4) Percentage composition of CSF, white matter 
and gray matter 

5) Cortical measurements 

6) Sylvian Fissure 

7) Interhemispheric Fissure 

8) General Length and Area measurements 
Using this software, measurements were made on 
scans of several normal subjects. 

Significance to Biomedical Research: This 
quantification methodology will greatly augment 
visual interpretation of brain CAT scans. It may 
provide a deeper understanding of brain structure 
changes during normal aging and disease processes. 
It is also possible that this work will produce new 
diagnostic tools. 

Proposed Course: Further refinement of the software 
is planned. A study to determine morphological brain 
changes related to aging will be conducted. 



Computer Analysis of Autoradiographic Images 
of Recombinant DNA Colonies 

A computerized methodology for analyzing 
autoradiographic spot images associated with 
recombinant DNA bacterial colonies has been 
developed in collaboration with scientists in NCI. 
This system represents a unique refinement in a 
method to directly identify cloned DNA sequences 
complementary to messenger RNA that are 
developmentally or hormonally induced. 

Spot density measurements are computed from 
digitized images produced via microdensitometry. 
These measurements are corrected for variability in 
exposure and local background, calibrated to 
hybridization standards, and normalized for 
comparison purposes. The system provides a variety 
of graphical and tabular output that effectively 
summarizes experimental results and identifies 
significant induced hybridization events. 

Background and Objectives: NCI scientists have 
been refining techniques to identify cloned DNA 
sequences complementary to messenger RNA that 
are developmentally or hormonally regulated. This 
refinement has led to a methodology that produces 
autoradiographic spot images representative of the 
amount of hybridization. The objective of this project 
is to provide an automated procedure for a 
quantitative analysis of these images. 

Methods Employed: Cloned bacteria are grown on 
agar in microliter wells, transferred to filter paper, 
and hybridized to end-labeled mRNA or cDNA 
probes. Autoradiographs of the filters are digitized 
and the density of each spot relative to background 
is established by means of CSL-developed image 
processing software operational on the DCRT Evans 
and Sutherland PDP-11/70 computer system. 
Compensation for variations in background, film 
exposure conditions, and hybridization are included 
in the methodology. A variety of graphical output 
including scatter diagrams, histograms, and listings is 
provided. 

Progress in FY82: Software development was 
completed and several experiments were conducted. 
Two papers are in preparation. The first paper 
discusses both the application of this system to 
analyzing the in vivo response of rat liver to 
glucocorticoids, as well as the application to other 
biological systems. The second paper will describe 
the computerized image analysis methodology and 
procedure in detail. 



13 



Significance to Biomedical Researclr. The 
methodology developed allows quantitative 
hybridization studies on a large number of 
sequences. Earlier qualitative assessment of 
autoradiographic spot images is now superseded 
with automated procedures yielding more accurate, 
more reproducible data. Computer graphic 
presentation of results greatly facilitates identification 
of significant experimental events. 

Proposed Course: Additional experiments are 
planned. It is anticipated that the methodology 
developed here will be discovered to have 
applicability to other areas of biomedical research, 
as demonstrated by inquiries from other NIH 
biomedical research investigators. 



jhTihsohiah science inforuation exchange 


U.S. OEPARIHENT OF 












™^'S.!Kk'8F"»"" 


ZOl CT00084-02 CSL 








PERIOD CDVERtD 


October 1. 1981 to September 30. 1982 


TITLE OF PROJECT (80 ch.r.ctcrs or l.ss) 


Computer Analysis of Autoradiographic Images of Recombinant DNA 


Colonies 


TOF^SslZL^Ersora'ENjwE^OriiE'pJrECT *"" "^'"^ "^ """^'"'" """""""^ *'" *'"'■ """ 


PI: J. H, DeLeo Computer Systems Analyst CSL, DCRT 


OTHERS: Floyd Taub Research Associate LR, C 


Brad Thompson Section Chief LB, C 


Laboratory of Biology (LB), C 








nCRT. NIH. Bethesda. HD 20205 

TOTAL «*N»E«S: 1 PROFESSIONAL. OtHEfl, 


t).:i 1 n.ii 1 

CHECK APPROPRIATE aOK{ES) 


a(.l) MINORS n{"?) INTERVIEWS 


SUHMART OF uORK (200 yonSs or less - underline key-ord,) 


A computerized methodology for analyzing autoradiographic spot images 


associated with recombinant DNA bacterial colonies has been developed 


in collaboration with scientists in NCI. This system represents a 


unique refinement in a method to directly identify cloned sequences 


complementary to messenger RNA that are developmentally or hormonally 


induced. 


Spot density measurements are computed from digitized images produced 


via microdensitometry. Tliese measurements are corrected for variability 


in exposure and local background, calibrated to hybridization standards. 


and normalized for comparison purposes. The system provides a variety 


of graphical and tabular output that effectively summarizes experimental 


results and identifies significant induced hybridization events. 



Cataract Grading via Computerized Slit-Lamp 
Image Analysis 

A new interactive image analysis procedure for 
analyzing and comparing slit-lamp camera images of 
human eye lenses has been developed and 
implemented on the DCRT Image Processing 
System. This procedure features television camera 
image digitization, interest area delineation by means 
of a flexible oval template, data calibration and 
standardization, and computation of a variety of 
statistical measurements for descriptive and 
comparative purposes. Also, an experiment has been 
designed to analyze the contribution to measurement 
variance due to photographic procedure, film and 
film processing, digitization, and analyst judgment 
factors. 

Bacl<ground and Objectives: A major problem for 
cataract researchers has been the lack of an 
objective, reproducible, in-vivo cataract classification 
scheme. Subjective classification methods are 
currently depended upon. With the tremendous 
variability in the morphology of cataracts, it is difficult 
to rely on such methodology either in survey works 
or in longitudinal studies. Development of an 
objective cataract grading scheme is seen as a high 
priority item among cataract researchers. 

i\4etiiods Employed: Photographs of human eye 
lenses obtained from Topcon and Zeiss slit-lamp 
cameras are digitized via microdensitometry or 
television signal quantitation. Resulting digitalized 
images are transported via magnetic tape to the 
DCRT Image Processing System for interactive 
analysis. The image analyst first defines and 
positions an oval window template over the area of 
interest using dials and then points to a 5-band 
calibration strip (an integral part of the image) using 
a graph-pen. A histogram distribution of the pixel 
values within the template is formed, scaled, 
calibrated, and processed to generate various 
statistical measurements used to describe and 
compare images. Colored isodensitometric 
representations of the eye lens may be displayed. 
Time-spaced photographs of an individual patient 
may be processed to produce an historical summary 
of lens opacity changes over time. This is viewed to 
be useful as an epidemiological tool for advancing 
our understanding of the cataract disease process, 
as well as in evaluating the effectiveness of 
therapeutic protocols such as the administration of 
anticataract drugs. 



14 



sissiriM'fss^ir:?:^;?: i'^".t 


""'"■' Sis'i"'"'"" 


ZOl CT00085-02 CSl 


October 1. 1981 to <ientember 30. 19R? 


Cataract Grading via Computerized Slit-Lamp 
Image Analysis 


M«ES, L»80R*I0KT MID mSTITUIE tfFILUTIOHS, UtO TITLES Of PRINCIPAL INV£STIG»IOBS AUD »U OTH£H 

PI: J. H. OeLeo Computer Systems Analyst CSL. OCRT 
OTHERS: R. 0. Sperduto M.D. BE. NEI 


CMPmilM WITS (i1 .r,,) 

Office of Biometry S Epidemiology (BE), NEI 
Division of Ophthalmology. Harvard Medical School 




Svstems Design Section 






CHt« APPflOPHUIE B0.(t5) 

D{.) «U*»N SUBJECTS Q(t) t-g-M TISSUES U(.) NEITHER 

n(.l) -INORS D(.0 INUBVK-S 


A new Interactive image analysis procedure for analyzing 
and comparing slit-lamp camera images of human eye lenses 
has been developed and implemented on the DCRT Image 
Processing System. This procedure features television 
camera image digitization, interest area delineation by 
means of a flexible oval template, data calibration and 
standardization, and computation of a variety of statistical 
measurements for descriptive and comparative purposes. Also, 

variance due to photographic procedure, film and film processing, 
digitization, and analyst judgment factors has been designed. 



S'8S=.i 



201 CTOOOa7-01 CSL 



Oct ober 1 . 1961 to Sept ember 30, 1J82 

.i or "PflOJiCI (eO'.h.r.cl.r. er U.,) 

Robust Boundary Detection of Necturus Gall Bladder Cells 



PI: J. M. OeLeo 

OTHER: K. Spring 



Computer Systems Analyst CSL, OCRT 
Research Physiologist LKEH, NHL8I 



MINC will (II .. 
LKEN. KHL6I 



Computer Systems Laboratory 



Systems Design Section 

DCRT, NIH. Bethesda. HD 20205 



lange hist( 



from video 



Ma/or Findings in FY82: The new interactive innage 
analysis system, which produces descriptive 
measurements and colored isodensitometric images 
from slit-lamp camera images of human eye lenses, 
is fully operational and easy to use. Measurement 
variance due to image analyst techniques seems to 
be insignificant. Calibration and standardization 
procedures should considerably reduce variance due 
to film, film processing, and digitization factors. The 
largest potential source of variance is in the quality 
control exercised by the photographer. This factor 
requires further study. 

Significance to Biomedical Research: Development 
of an objective cataract grading scheme is seen as a 
high priority item among cataract researchers. 

Proposed Course: A vigorous analysis of variance is 
planned after adequate study and improvement of 
image capture methodology. Use of supplemental 
views, such as frontal and retroillumination views, is 
being considered. 



Robust Boundary Detection of Necturus Gall 
Bladder Cells 

A robust boundary detection algorithm for automated 
planimetry of Necturus gall bladder cells has been 
designed to enhance an existing methodology that 
computes cell volume change histories from video 
images of cells visualized in a light microscope. 

Bacl<ground and Objectives: Epithelial cells of 
Necturus gall bladder regulate their volume after a 
change in osmolality of their bathing solution. The 
Laboratory of Kidney and Electrolyte Metabolism, 
NHLBI, has developed a computerized methodology 
for time-tracking cell volume changes through 
interactive planimetry of video images of cells 
visualized in a light microscope. The Computer 
Systems Laboratory has been requested to develop 
a specialized robust cell boundary detection 
algorithm to enhance overall throughput processing 
efficiency. 



15 



Methods Employed: The specialized robust cell 
boundary detection algorithm conceived operates as 
follows: 

1. The investigator points to the center of the cell. 

2. Opacity values are collected along 72 rays 
emanating from the center. The rays are of fixed 
length and 5 degrees apart. 

3. Each ray of opacity values is processed as 
follows: 

• Scale opacity values (0 to 255). 

• Accept a single minimum as a tentative edge 
point. 

• Confirm edge point by scaled gradient and 
localized texture parameters if necessary. 

4. Check neighboring pairs of tentative edge 
points, rejecting pairs that fail the test. 

5. Attempt to specify missing edge points by step 
3 applied over a narrower segment of the ray as 
determined by radii of a certified near neighbor edge 
point and a priori data specifying expected edge 
point radii lower and upper range values as a 
function of the angular distance to the nearest ray 
having a certified edge point. 

6. Sequentially connect the edge points. 

7. Compute the enclosed area. 

8. Repeat steps 1 to 7 over all cell slices to obtain 
the required volume. 

Significance to Biomedical Research: Application of 
quantitative light microscopic techniques to study cell 
volume changes due to fluid and ion transport in 
living epithelial tissues has already proven to be a 
powerful and effective research tool. An accurate, 
efficient, robust cell boundary detector algorithm 
would greatly improve upon the utility efficiency and 
throughput speed of this methodology. 

Proposed Course: It is planned to implement, test, 
and refine the robust cell boundary detector 
algorithm described above on the DCRT Image 
Processing Facility. Upon successful development, 
consideration will be given to alternative approaches 
for implementing the algorithm in production mode. 



Rehabilitation lUledicine Computer System 

This project involves the development of computer 
techniques in collaboration with the Department of 
Rehabilitation Medicine of the NIH Clinical Center. 
CSL has recommended computer techniques that 
can be used to automatically acquire anatomical and 
physiological information from patients, perform the 
required calculations on the data obtained, and 
display the necessary results to the medical staff. 
The automated techniques include the measurement 
of body forces (hand and ground reaction forces), 
electromyograms (electrical activity of the muscles), 
and body kinematics (the position and angles of the 
limbs and joints in space and time). An Automated 
Biomechanics Laboratory System that provides 
these measurements will be purchased in late FY82. 
The computer part of the system will allow the 
medical staff to enter patient and staff data into a 
data base with computer-generated forms displayed 
on a terminal screen, and to perform inquiries and 
generate reports using the accumulated data. In 
FY83, the physical space will be designed to 
accommodate this system in a new area of the 
Clinical Center. 






ZOl CT00081-02 CSL 



nrtnhpr 1 1 IB ! Hi ^ppt«.h»r in IW? 



ine Department Computer System 



OTHERS: H. 0. Jarret 
W. Schneiderw 
N, L. Gerber 



,, Biomechanical Engineer 
Physical Therapy ServiCi 
Rehabilitation Medicine 



CSL, DCRT 

RH, CC 

RM, CC 

R«. CC 



itation Medicine Department, Clinical Center 



computer Systems Laboratory 



Systems Design Section 



DCRT. niH. B ethesda. H P 20205 



This 

medic 
NIK r 


proje 




llaborat 


to au 


tomat 


cally acquir 


perto 


rm th 




quired ca 


(elec 
angi 


trica 
s of 


a 


tivity of 
limbs and 


F»82 


The 


Sy 


outer pa 



11 alio 



16 



Background and Objectives: The Department of 
Rehabilitation Medicine provides physiatric 
evaluation and treatment, physical therapy, 
occupational therapy, and speech therapy for NIH 
Clinical Center patients referred by Institute 
physicians. In addition, it develops various indices to 
evaluate these services. This department supports 
the efforts of, and collaborates with. Institute 
physicians engaged in research relevant to physical 
rehabilitation medicine. It also initiates both clinical 
and basic research independent of Institutes in the 
rehabilitation of mentally and physically handicapped 
individuals. 

In support of these goals, CSL is developing a 
computer system. Initially, the department will use 
the system for the following three projects: 

1 . The Automated Biomechanics Laboratory: a 
laboratory that will be used to automatically measure 
the position of the limb segments in space, the 
forces in the lower limbs, and the electromyographic 
signals from the muscles in the limbs; 

2. The Hand Dynamometer Instrument: a device 
that will be used to measure the magnitude and 
direction of the forces in the hand and to develop 
clinical tests to diagnose the mechanical and 
functional status of the hand, arm, and shoulder; 

3. The Physical Therapy Quality Assurance 
System: a data base system that will be used to 
assess medical staff effectiveness in providing the 
types of patient care needed, determine staff 
workload and scheduling, and identify areas for 
clinical research for the Physical Therapy Service. 
Progress in FY82: During the past year, CSL 
determined the instrumentation and computer 
requirements for the Department of Rehabilitation 
Medicine. A considerable amount of specialized 
instrumentation is needed to perform the required 
automated measurements. This includes: five motion 
cameras with light sources that are used to acquire 
the spatial coordinates of anatomical points on the 
patient's body with reflective markers, force plates 
that are used to measure patient ground reaction 
force, and hard wired or telemetry electromyogram 
acquisition equipment that is used to measure 
patient muscle activity. CSL prepared a Request for 
Proposals for an Automated Biomechanics 
Laboratory System that will be used to obtain the 
necessary transducers, instrumentation, and 
computer hardware and software. 

The Physical Therapy Quality Assurance Data Base 
System was initially implemented on a small 
computer system. It will be transferred to the larger 
computer that is purchased with the biomechanics 
laboratory system. The Biomedical Engineering and 
Instrumentation Branch of NIH's Division of 



Research Services continued development of the 
hand dynamometer instrument. 

Also, during the past year, a collaboration was 
initiated with the Gait Analysis Laboratory, 
Department of Orthopedic Surgery, Children's 
Hospital Medical Center and Harvard Medical 
School. In the future, computer programs, patient 
data, and engineering and medical expertise will be 
exchanged with this group. 

Significance to Biomedical Research!: The computer 
system will be used with arthritic, orthopedic, and 
neurological patients and with amputees in order to 
evaluate drug therapy, orthotic and prosthetic 
devices, and medical interventions. It will also be 
used as a teaching tool to help these patients learn 
to function with their disability in an efficient manner. 
Many hospitals in the United States are presently 
establishing automated biomechanics and gait 
analysis laboratories. Therefore, any new 
developments made on this project will benefit users 
of these automated systems, as well as patient care 
and clinical research within the Department of 
Rehabilitation Medicine at NIH. 

Proposed Course: The Department of Rehabilitation 
Medicine expects to purchase an Automated 
Biomechanics Laboratory System during the coming 
year. As the Department is moving to a new location 
in the Clinical Center, the new area will be designed 
to accommodate the cameras, force plates, EMG 
equipment, computer hardware, and patient 
measurement area. 

Also, during the coming year, a study will be done to 
determine the type of force plate that will best meet 
the measurement requirements of the Department. 
Comparisons will be made between small modular 
versus large force plates and piezoelectric versus 
strain gauge force plate transducers. In addition, 
methods for accurately determining the velocity and 
acceleration of anatomical points from acquired 
motion data will be investigated including the 
required camera resolution and frame rate and digital 
differentiation techniques. 



17 



Positron Emission Tomography (PET) Facility 

The PET facility of the Nuclear Medicine Department 
is used to collect and analyze images of the human 
brain for diagnosis and scientific research. The 
facility includes a PET Scanner that receives data 
consisting of gamma emissions from patients and a 
minicomputer system that operates the scanner, 
reconstructs the data into cross-sectional slices, and 
performs other analysis. The facility also includes an 
offline minicomputer system having an image array 
processor and color display that is used to 
interactively perform numerous image enhancement 
and analysis functions. Various NIH Institutes use 
this facility to research the aging process, 
schizophrenia, epilepsy, and other brain functions 
and disorders. 

Background and Objectives: In late FY81, the 
Nuclear Medicine Department requested assistance 
in improving their PET computer facility. At that time, 
the PET facility was receiving increased usage by 
various Institutes and had recently lost some of its 
technical staff. Our goal was to improve the existing 
system's hardware and software, to establish 
guidelines for collecting and storing patient data, and 
to provide an image analysis system that could be 
readily operational. 

Progress in FY82: To improve scan data flow and 
acquisition, we purchased and installed larger disks, 
a tape drive, and a floating point unit. To improve 
image data analysis, we purchased the offline 
minicomputer system and installed and modified 
image analysis software provided by NIMH. We also 
programmed the analysis system to compute local 
cerebral metabolic activity with radioactive 
deoxyglucose utilization. 

Proposed Course: NIH will be purchasing new PET 
scanners and there will be increasing demand for 
image analysis and for the storage and retrieval of 
large quantities of data. We will analyze these future 
demands and recommend appropriate solutions. 



'Sc!l£CrNlXlB£RlS"HOl">e"il!!»Si=T' 


„t.u;-.;™E"L?.,cis 






raBUtjHEJLIJ^.EB.ICE^ 


ZOl CTOOlOn-01 CSL 


PEHioo coy^ERm^^ ^_ ^qgj jg September 30. 1'982 




Positron Emission Tomography (PET) Facility 




P«OF£SS10NAL PEflSONHEL ENCACED OH THE PROJECT 


PI: A. J. Pashayan Computer Specialist CSL, nCRT 


OTHERS: W. L. Risso Electronics Engineer CSL, nCHT 


R. M. Kessler. M.n. Hearl. Positron Emission 


Tomography Section MM, CC 


Nuclear Medicine, Clinical Center 


t.*B/BB*NCH 


Computer Systems Laboratory 




Systems flesign Section 




OCRT, NIH, Rethesda, HP 20205 


^^^^|PflOFE3S10ri*L. ^^^ |l)Ih£R, 


CHECK APPROMIUICiMES) 


J(0 KUMW. SuejECIS a(b) HUb.«ni3SUES a(clHEIIH£B 


D(.1)«IN0H5 a 1.2) INTEHVIE-S 




The PET facility of the Nuclear Medicine Oepartment is used to 


collect and analyze images of the human brain for diagnosis and 


scientific research. The facility includes a PET Scanner which 


receives data consisting of gamma ei^issions from patients and a 


minicomputer system that operates the scanner, reconstructs the 


data into cross-sectional slices, and performs other analysis. 


The facility also includes an offline minicomputer system 


having an image array processor and color display that is used to 


interactively perform numerous image enhancement and analysis 


functions. Various NIH Institutes use this facility to research 


the aging process, schizophrenia, epilepsy, and other hrain 


functions and disorders. 



JSiLE^nS!BfS'!So"KorS«*Ii?;f^'?:!i" 


HEAL'TH-ANo'S'sEfviCES 


PROJECT NUMBER 








ZOl CTn0096-0 


CSL 








Computer Assisted Hematology Morphology Data Handling System 








PflOFESSIONAL PEHSONNEL ENCAGED ON THE PflOJECI 




PI: n. C. Songco Electronics Engineer CSL 


ncRT 


OTHERS: L. Wang Electronics Engineer CSL 


t)CRT 


J. A. nonlon. M.O., Ph.fl. Staff Physician CP 


CC 


A. Faust Prograrmer CP 


CC 


£. W. Lundy Technologist CP 


CC 


6. L. Wages Chief Technologist CP 


CC 


COOPEHAdNC UNIIS (i( ..,) 




CPD, CC 




LAB/BRANCH 














CHECK APPROPRIATE B0x(ES1 




DU) HUyAN SUBJECTS D (b) HUMAN I ISSUES a(c)NE.TH£R 




Dt.l) MINORS Q(,2) INTERVIEWS 




SUMMARY OF UORK (20O .ord= or le=i - unitrV.n, k.,.ortO 




Cell morphology evaluation is a major component of the workl 


ad 


of the Hematology Service of the Clinical Pathology Departme 


t. 


CC. Manual white cell differential counting is the predomin 


nt 


method of analyzing white cell morphology. Previously, 




technologists performed manual differentials and recorded th 




data on mark sense cards for later entry into the Clinical 




Pathology Laboratory Computer (CPLC). This approach led to 




of technologist time. The goal of this project was to desig 


^ 


and develop an alternative method of handling the cell 




morphology evaluation. 





18 '"•■■ 



Computer Assisted Hematology Morphology 
Data Handling System 

Cell morphology evaluation is a major component of 
the workload of the Hematology Service of the 
Clinical Pathology Department, CC. Manual white cell 
differential counting is the predominant method of 
analyzing white cell morphology. Previously, 
technologists performed manual differentials and 
recorded the data on mark sense cards for later 
entry into the Clinical Pathology Laboratory 
Computer (CPLC). This approach led to transcription 
errors, turnaround time lags, and inefficient use of 
technologist time. The goal of this project was to 
design and develop an alternative method of 
handling the cell morphology evaluation. 

Background and Objectives: CSL, in collaboration 
with the Clinical Pathology Department, CC, has 
developed a computer-assisted hematology 
morphology data handling system with the following 
characteristics: 1) allows direct entry of manual 
differentials, red cell morphology, and platelet 
estimates; 2) displays Coulter automated cell 
counting data for comparison; and 3) links in realtime 
to the CPLC. 

Progress FY82: System development and installation 
was completed this year. 

The technologists were involved in the human 
factors design of the system including specifying cell 
types, screen and keyboard layout, procedure 
definition, and 'help' and system messages. The 
acceptance and use of the system was both 
immediate and enthusiastic. Based on an average 
daily workload of 230 analyses, the system has 
decreased the manpower requirements by 50%, 
changed turnaround time for release of certified 
results from hours to minutes, and reduced 
transcription time and errors. 

A Digital Equipment Corporation (DEC) LSI-11 
microcomputer with dual floppy disk drives was used, 
with four VT100 CRT terminals as user stations. 
Extensive use of DEC forms software was made to 
allow flexibility in redefining screens and 
nomenclature. The system is linked to the CPLC via 
a direct 1200-baud serial line. The user responds to 
menus at each stage of the procedure. Help 
messages are available at any time providing online 
instruction and verification of procedures. 

Proposed Course: Because the present system is 
already used to capacity, it will be necessary to add 
or replace hardware and software in order to add 
additional tasks. Plans are now undenway to increase 
the capability of the present system by using 



cartridge disk drives and modifying the system 
software or upgrading from RT-1 1 to the more 
powerful RSX11-M operating system. If this is done, 
local data storage would be available for a 
morphology quality control program. Composite lists 
of platelet counts versus platelet estimates could be 
generated. Additional user stations could be added if 
the workload increased. This expansion is, however, 
predicated on the availability of financial and 
manpower resources. 

Publications: 

Donlon, J. A.. Wang, L., Lundy, E., Wages. B., Faust. A.. Songco. DC; A 
Computer Assisted Hematology Morphology Data Handling System. 
Symposium on Computer Applications in Medical Care, 1983 (in press). 



Automated Pulmonary Physiology Testing 

Procedures such as exercise testing, pulmonary 
compliance, and muscle strength have been found 
successful for evaluating pulmonary function. By 
exercising a patient on a treadmill and gradually 
increasing the workload (i.e., speed and incline), the 
physician can better assess cardiopulmonary 
disease, which in its early stages generally does not 
manifest itself except under physical exertion. In 
order to help the physician perform these procedures 
more effectively, a microcomputer system has been 
developed to enable automated realtime collection, 
analysis, and display of pulmonary compliance and 
inspiratory muscle strength data. Steady-state 
treadmill exercise testing has been only partially 
automated. Although data is manually entered, data 
analysis and report generation are fully 
computerized. Work is in progress to enable 
automatic realtime acquisition of exercise data with 
breath-by-breath analysis. The breath-by-breath 
technique allows determination of the anaerobic 
threshold noninvasively, without the need for arterial 
catheterization. The anaerobic threshold is used as a 
measure of an individual's 'fitness.' Patient data is 
stored on a local disk data base for future reference. 

Background and Objectives: Physicians monitor 
pulmonary parameters during exercise to better 
assess pulmonary function and to diagnose 
pulmonary dysfunction that only manifests itself 
under physical exertion. Procedures such as 
pulmonary compliance and inspiratory muscle 
strength also give insight into respiratory function. 



19 



Until last year, pulmonary treadmill exercise testing 
was performed manually at NIH. Data were written 
down and later entered into a programmable 
calculator for determination of results. Additional 
summary statistics and a final report were prepared 
by hand. Inspiratory muscle strength and pulmonary 
compliance measurements, done in the same lab, 
likewise were performed manually. 

In order to speed both exam and data analysis time, 
and to improve accuracy, these procedures were 
automated with a microcomputer system. 

Methods Employed: The microcomputer system is a 
DEC MINC-11/03 (Modular Instrument Computer) 
containing an LSI-1 1 microprocessor, 32K words of 
memory, auxiliary disk storage, and analog-to-digital 
and digital-to-analog conversion capability. There is 
also a video graphics display, a keyboard console, a 
hard copy unit for printing the video display, and a 
line printer. 

In determining pulmonary compliance, 
transpulmonary pressure (the difference between 
alveolar pressure, i.e., mouth pressure with mouth 
shutter closed, and esophageal pressure, as 
measured by a balloon transducer swallowed by the 
patient) and lung volume (measured with a wedge 
spirometer) are determined by the computer as the 
physician repeatedly closes a mouth shutter 
throughout a patient's inhalation or exhalation. A 
graphical plot of the data and an exponential least 
squares curve fit of the data is then produced to aid 
in evaluating the 'stretchability' of the patient's lungs. 

During the steady state treadmill procedure, the 
computer monitors expired volume and flow via a 
Tissot spirometer and pneumotach, respectively, as 
the patient is subjected to stepped increases in 
exercise, each time starting from a resting state. 
Expired oxygen, carbon dioxide, and nitrogen 
concentrations are monitored via a Perkin-Elmer 
mass spectrometer gas analyzer. Traditionally, in 
order to determine the patient's anaerobic threshold 
(i.e., point where the body begins to rely heavily on 
anaerobic metabolism and produce lactic acid), the 
patient is catheterized in order to obtain arterial 
blood samples at each steady state level. Acid/base 
and gas concentrations are determined offline by a 
blood gas analyzer from a sample of the patient's 
arterial blood, and entered at the keyboard. 
Pulmonary volumes, flows, and oxygen 
consumption-a measure of how hard the patient 
actually works to perform a given level of exercise- 
are then calculated. 



When one has the capability to automatically monitor 
air flow and expiratory gas concentrations in 
realtime, the anaerobic threshold can be determined 
noninvasively without the need for blood p02 
obtained from arterial catheterization. Anaerobic 
threshold is determined from measures of exhaled 
02, C02, and respiratory quotient. Breath-by-breath 
analysis also allows the performance of nonsteady 
state exercise testing, where the patient is subjected 
to continuously increasing levels of exercise. This 
methodology provides a more dynamic picture of the 
patient's cardiopulmonary performance. 

Progress in FY82: The MINC computer system is 
now used routinely to perform the static pulmonary 
compliance and inspiratory muscle strength 
procedures. The graphics capability of the MINC VT- 
105 console terminal was found to fall short of the 
lab's needs due to the one 'Y' per 'X' plotting 
limitation. As a result, a Retrographics VT-100 
terminal was procured, providing Tektronix 
compatible bit-mapped raster graphics capability. 
Using a CSL graphics package, all pulmonary 
graphics routines were rewritten for Retrographics 
compatibility. 

Throughout the year, several programs were written 
to aid pulmonary personnel in the analysis and 
organization of data obtained from other areas of the 
pulmonary clinical and research service. 

Work continued on fully automating the treadmill 
exercise system. However, a change in clinical 
priorities re-directed efforts towards the development 
of a breath-by-breath steady state exercise system 
rather than simply automating the technique currently 
performed. Completion of this goal is planned for the 
end of FY82. 

Proposed Course: Once the breath-by-breath steady 
state exercise procedure is completed, the system 
will be adapted to perform a nonsteady state 
procedure as well. Although there are now no 
immediate plans to add additional procedures to the 
system, having developed a general purpose tool for 
pulmonary data collection, new procedures or 
modified techniques can be easily incorporated into 
the existing protocol. For example, the computer 
may enable closed loop control of treadmill speed. 



20 



By monitoring heart rate and dynamically varying 
treadmill speed in response to heart rate changes, it 
should be possible to apply a more constant 
workload to the patient, thus leading to more stable 
results. In addition to the potential for performing 
new physiology procedures, additional mathematical 
analyses can be applied to the data in order to gain 
further insight into a patient's pulmonary function. 

Publications: 

Nadel, L.D.: Automated Pulmonary Analysis by an Online Microcomputer, In 
Nair, S. (Ed.): Computers in Crillcal Care ar)d Pulmonary Medicine (in 
press). 

Nadel. LD.: Breath-to-breath Pulmonary Exercise Testing Using an Online 
Microcomputer. First IEEE Computer Society International Conference 
on Medical Computer Science. Computational Medicine (MED- 
COMP'82), Philadelphia. September 25, 1982, 

Keogh, B., Gadek, J., Price, D,. Nadel. L.. and Crystal, R.: Remarkable 
Similarities in Exercise Gas Exchange Parameters in Markedly Dispa- 
rate Diseases: Comparisons Between Idiopathic Pulmonary Fibrosis 
and 1 -Antitrypsin Deficiency. American Review ol Respiratory Disease 
125: 157, April 1982. 



Assessment of Tongue Motion During Speech 
Using Ultrasonic Imaging Techniques 

This project is directed at developing a system 
capable of ultrasonically imaging the tongue in 
realtime. In addition to obtaining and integrating the 
necessary hardware, mathematical techniques must 
be identified and developed to analyze and describe 
images of continual tongue motion. Thus far, the 
tongues of several normal subjects have been 
imaged using a realtime ultrasonic body scanner in 
conjunction with a videotape recorder. Anatomical 
features of interest, as the subject vocalizes specific 
phonemes, are extracted using a graphic tablet 
followed by preliminary computer analysis. Efforts 
are presently being devoted to identifying and 
developing mathematical techniques for analyzing 
and describing the patterns of continuous tongue 
motion. Once we complete an evaluation of the 
ultrasonic scanners commercially available, we plan 
to purchase and build where necessary the hardware 
required to implement this technique in our new 
speech laboratory, scheduled for operation in FY83. 

Background and Objectives: Speech investigators 
have traditionally used radiographic methods for 
studying tongue motion during speech and 
swallowing. These techniques, however, are 
somewhat cumbersome and are not suitable for 
general screening, diagnosis, or therapy due to the 



™'''""""^-™ ■■•'"' ■'"•'" 


""''"^'un^S'i'""'"" 


zni cTonoss-oa csl 


(Ictober 1, 19fil to September 30. 1982 


Automated Pulmonary Physiology Testing 


PI: L. n. Nadel. Ph.n. Staff Fellow POS. rSL, DCRT 

OTHERS: B, A. Keogh, M.n. Expert PB, IR NHLRI 
P. S. Plexico Chief, Project nevelopment 

Section CSL. nCBT 


COOPEfiMINC UNirS (.1 .r,J 

Pulmonary Branch, NHLBI 


Computer Systems Laboratory 


Project nevelopment Section 


OCRT. NIH. Bethesda. HD 2020S 


lOIAL WNUWS. ^ |«.0ftSSION*L= jOtHER. 


CHECK *p™opruie eo«(ts] 
QUO NINONS a(,;) iNiERvms 


SUMH^RY gf WORK (JOt. -ord. or 1 od.rll.. k.^ord*) 

Procedures such as exercise testing, pulmonary compliance, and 
muscle strength have been found successful for evaluating pulmonary 

the workload {i.e., speed and incline), the physician can better assess 
cardiopulmonary disease, which in its early stages generally does not 
manifest itself except under physical exertion. In order to help the 
physician perform these procedures more effectively, a microcomputer system 
has been developed to enable automated realtime collection, analysis and 
display of pulmona'-y compliance and inspiratory muscle strength data. 






ZOl CT0O094-O1 CSL 



nrtnhpr l,^iq«l *" ^Jeptpmher 30. 



Tongue Motion During Speech 



PI: 



L. n. Nadel, Ph.D. 



OTHERS: R. C. Sonies 
T. H. Shawkei 
H.L. Stone, i 



Chief 
Oept. 



' Diagnostic Radiology. CC 
Consultant, Speech Pathology 
Oept. of Rehabilitation Med., 
Chief. Physical Sciences Laboratory 
PSL. OCRT 



Department of Rehabilitai 
Radiologiy, CC; PSL. OCRT 



, CC; Oept. of Diagnostii 



Project Development Section 
DCRT. NIH, Bethesda. MP 20205 



veloping a systej 



r. the tongues i 



nual tongue 
'i dent ape 



mi nary compi 
ifying and ■ 



extracted using a graphic tablet followed I 
ling mathematical techniques for analyzing . 



21 



harmful effects of x-rays. It is hoped that the 
development of an ultrasonic system for visualizing 
lingual function and performance will prove valuable 
for the general screening and diagnosis of speech 
pathology. Used as a tool for biofeedback, such a 
system might be of assistance in helping a patient to 
correct speech difficulties. As one learns more of the 
details of tongue motion in swallowing and speech, 
this imaging system may be of potential value in the 
fields of neurology and dentistry. 

Methods Employed: Using an existing realtime 
ultrasonic diagnostic body scanner (ATL), several 
normal subjects were scanned during the utterance 
of specific phonemes. The resultant images were 
photographed and digitized using a graphic tablet 
interfaced to a DECsystem-10 computer. The data 
was statistically analyzed for reliability and 
repeatability. Although mathematical techniques for 
describing the patterns of tongue motion obtained 
are presently under investigation, some interesting 
observations already have been made. New 
instrumentation is presently being evaluated to 
implement this technique without the need for an 
intermediate photograph. In order to observe one's 
natural tongue motion, a critical task will involve 
developing a means to mount the ultrasonic probe 
so as not to interfere with or influence the patient's 
speech. Additionally, variable positioning or 
nonconstant pressure on the transducer will affect 
the final image. 

Progress in FY82: In order to demonstrate the 
feasibility of the above-mentioned technique, the oral 
cavities of several normal subjects were scanned 
using a realtime ultrasonic body scanner, by placing 
the ultrasonic transducer two centimeters behind the 
mental symphysis of the patient's mandible. Each 
subject was asked to repeat specific phonemes 
while mouth images were continuously recorded on 
a videotape recorder. Using the recorder's freeze- 
frame capability in conjunction with a multi-image x- 
ray formatter, the desired images were copied onto 
negative film. The negatives were developed into 
positive prints. By placing the resultant prints on a 
graphic tablet connected by telephone to a 
DECsystem-10 computer, the polar coordinates of 
anatomical areas of interest, namely, the tongue 
surface and the genioglossus muscle, could be 
readily determined. The digitization technique was 
repeated by several observers in order to statistically 
determine the accuracy and reliability of both the 
ultrasonic imaging and digitization techniques. 
Preliminary observations were made regarding 
tongue motion for vocalizing particular phonemes. 

Since the Rehabilitation Medicine Department wishes 
to outfit its new speech lab with an ultrasonic 



imaging and analysis capability, various equipment 
and methods are being evaluated for integration into 
a tongue analysis system. 

Proposed Course: The required system hardware will 
be ordered early in FY83. The development of 
methodology and software for analysis of tongue 
images will continue. A means for placing an 
ultrasonic transducer with constant position and 
pressure two centimeters behind the mental 
symphysis of the patient's mandible will be further 
investigated and developed. Once all the system 
hardware is obtained, the various components will be 
interfaced and the necessary control and processing 
software will be written. Normal volunteers will then 
be scanned in order to obtain baseline values of 
tongue motion. Shortly thereafter, patients with 
speech problems will be observed as well. In the 
future, we also plan to use this system to study 
swallowing defects. 

Anesthesia Computer System 

This project involves evaluating improved 
instrumentation techniques and identifying and 
investigating ways that automation can benefit 
anesthesia. Project emphasis is on adjunctive 
monitoring and automated recordkeeping in the 
operating room. 

Background and Objectives: While computers and 
automation have been used in intensive care 
settings for some time, little previous work has been 
reported on their application in operating rooms. Two 
areas of potential benefit with an anesthesia 
computer have been identified. 

1. Adjunctive monitoring, i.e., using the computer 
for monitoring and display of patient parameters. The 
main goals are a unified, easy-to-read display; limit 
detection and trend analysis of the parameters; and 
archiving of the measurements for later use in 
research or anesthesia mishap analysis. 

2. Automated recordkeeping, in which the 
computer would not only record the results of 
monitoring, but also would make provision for a 
record of drug administration, for free text notes by 
the anesthesiologist, and for producing a printed 
record suitable for inclusion in the patient's record. 
An advantage of such a system is that it will allow 
the anesthesiologist to devote more time to the 
patient by simplifying the tasks of observing and 
recording measurements. 

Progress in FY82: Most effort this year has been 
expended on developing a project plan for future 
work. 

Proposed Course: The project plan will be 
completed, needed resources will be sought, and 



22 



US'lSc" 



^.r 



ZOl CT00093-01 CSL 



nrtnhPr 1 1981 to Seoter 



Anesthesia Computei 



Anesthesia 5eri 



Computer Systems Laboratory 



Project Be velop ment Sectii 



DCRT. WIH. Bethesda. HP 20205 



This project involves evaluating improved instrume 
techniques and identifying and investigating ways 
can benefit anesthesia. Project emphasis is on ad 



r.V.tlTlUVni'.^'T\lT.lS'' 


^"^'-'SoirtcE'i"'""" 201 CT0006S-03 CSL 


October 1. 1981 to September 30. 1982 


Medical Information Technology Project 


PI: S.I. Allen Medical Research Analyst CSL. ncUT 

OTHERS: 0. C. Songco Electronics Engineer CSL. nCRT 
C. S. Brown Consulting Dermatologist 
P. S. Plexico Chief. Project Development Sec. CSL, DCRT 
A. W. Pratt Director DCRT 


COOPtRMlNCUN.IS (i. .n,) 








m.L«M»RS, ^^ pc*ESS,(«.L,^^ |OTKD>. 


CHECK «PPROPRr*IE BOX(£S) 

□ (.) HWM. SUBJECTS (0 HU-M 11 SSUtS a (0 NtlI«B 

a(.0 "INOfiS n(.2) IN7£RVI«S 


This project involves the application of microprocessor technology 
and improved man-machine interface methods to permit physicians and 
their associates to more directly cormunicate with computer record 
systems. A pilot study involving medical transactions input directly 
by practicing physicians is underway. The goal is to develop better 
ways to automate the essential physician contribution to the health 
care record that is used in both research and patient care. 



efforts to accomplish the project plan will proceed to 
the extent that resources become available. The 
most likely scenario is the development of a 
'demonstration' operating room to test the 
hypotheses that adjunctive monitoring and 
automated recordkeeping are advantageous. 

Medical Information Technology Project 

This project involves the application of 
microprocessor technology and improved man- 
machine interface methods to permit physicians and 
their associates to more directly communicate with 
computer record systems. A pilot study involving 
medical transactions input directly by practicing 
physicians is underway. The goal is to develop better 
ways to automate the essential physician 
contribution to the health care record that is used in 
both research and patient care. 

Background and Objectives: The use of computers in 
medical and hospital practice is increasing as the 
cost of systems is decreasing due to technological 
innovation. However, few physicians are comfortable 
with current machine interfaces. With this in mind, 
we are investigating devices and methods that 
provide a more capable, attractive interface while 
maintaining an acceptable level of flexibility and 
efficiency. The ultimate aim is to increase physician 
productivity in patient diagnosis and treatment and to 
increase patient understanding of disease processes 
and management plans. 

Methods: Much of the clinical software is table- 
driven to allow the physician to add and modify the 
data bases. This approach also provides a 
convenient means of adapting the programs to other 
clinical care and research environments. Both the 
clinical data base and processing software are being 
developed and tested on the CSL time-shared 
computer system. Finished programs, ready for use 
in patient care, are then transfered to a compatible 
microcomputer system situated in the physician's 
office. 

Progress in FY82: In collaboration with a practicing 
dermatologist, we are field testing an ambulatory 
patient care transaction system. This system allows 
the physician to input, store, retrieve, and 
disseminate patient data needed by various 
members of the health care team as well as by the 
patient. The immediate data processing focus 
includes machine generation of patient information 
and treatment schedules, pharmacy prescriptions, 
medical and surgical procedure reports, laboratory 
test orders, and referral letters to other doctors. 

Disease-specific and problem-specific protocols are 
used to lead the user through a restricted tree- 
structured hierarchy cf relevant diagnoses, 



23 



treatments, drugs, tests, and procedures. Where 
appropriate, protocols are modified by such factors 
as patient age, sex, weight, disease stage, and 
therapeutic response specified by the physician. 
When all workups and treatments are indicated, the 
computer then produces hardcopy treatment plans 
for the patient, record summaries for the doctor, 
prescriptions for the pharmacist, and test requests 
for specified laboratories. 

Proposed Course: Selected physician-operated 
modules will be tested to support critical diagnostic 
and therapeutic functions in ambulatory care. 
Programming logic to support isolated patient 
encounters also will be expanded to handle followup 
visits. The conventional CRT and keyboard terminal 
employed now will be replaced with faster I/O 
devices that are tailored to this medical application. 
For example, we plan to use graphic input to 
facilitate the capture of anatomic disease 
descriptions and keyboard substitutes to speed 
menu item selection. 

Molecular Graphics and Sequence Analysis 

The sequence of some regular proteins, together 
with other structural information such as data from x- 
ray diffraction, fiber diffraction, electron microscopy, 
and spectroscopic analysis can be used to evaluate 
models of the protein structure. Four such analyses 
are studies of collagen (with NIDR), keratin (with 
NIADDK and NCI), myosin (with Brookhaven), and 
streptococcal M proteins (with Rockefeller). 

The crystallographic structure of type I collagen 
fibrils had previously contained some controversial 
assumptions. A new model has been proposed this 
year that seems to resolve these difficulties and 
encompasses all of the experimental evidence into 
one structurally simple model. 

As the sequence of keratin cyanogen bromide 
fragments becomes available, an anaylsis of the 
sequence is proceeding by studying the periodicities 
in the sequence, and by predicting conformational 
properties of the specific amino acids in local regions 
of the chain. It is anticipated that the experimental 
results may be able to clearly specify if any of the 
proposed two or three models are correct. 

Analysis of myosin and streptococcal M proteins is 
continuing as sequences become available. This 
project is presently capable of easily evaluating new 
sequences for periodicities or regularities. 

Background and Objectives: While it would be 
somewhat idealistic to attempt to predict the 
structure of a globular or irregular protein, it is 
currently possible to convincingly model and predict 



the structure of regular (helical) proteins. With the 
current knowledge of the structure of the collagen 
helix, synthetic protein analogues of collagen, 
tropomyosin, and other regular proteins, one can 
extend this technology to new proteins as the 
sequence is experimentally determined if there are 
known points of similarity. 

Significance to Biomedical Researcii: Many proteins 
do not form three-dimensional crystalline solids 
whose structures can be analyzed by classical x-ray 
diffraction. However, if these proteins are regular, 
comparison and analogy with related proteins can be 
used to model the unknown structures in order to 
understand the structure and functioning of the 
proteins. In addition, one can use computer models 
to analyze two or more possible candidates and 
determine the most likely protein structure. 

Progress in FY82: A new model of collagen has 
been proposed that reconciles previously diverse 
data from a variety of experimental sources. A new 
analysis has begun that will use the sequence of 
keratin filaments to compare the structure to 
proposed models and to other proteins whose 
structure has been well characterized. 

Methods Empioyed: Standard Fourier methods have 
been used to analyze the sequences and to cross- 
correlate sequences. These sequence regularities 
are usually correlated with structural features, such 
as the collagen triple helix, the alpha helix, or the 
tropomyosin double-stranded alpha helix. In addition, 
software was written to model the collagen helix and 
double-stranded alpha helices on the Evans and 
Sutherland Picture System. This unique hardware 
allows three-dimensional analysis of proposed 
structures, both using traditional wire models, and by 
using CPK 'ball' models in three dimensions, where 
the size of the ball is related to the size of the 
individual amino acid, and the color of the ball is 
related to the function of the amino acid. 

Proposed Course: As new sequences of regular 
(helical) proteins become available, it is relatively 
easy to model these sequences and describe their 
structures both graphically and quantitatively. 

Publications: 

Trus, B. L., and Elzinga, M.: Computer Modeling of A 17,000 Dalton Frag- 
ment of tvlyosin. In Balaban, M., Sussman, J. L., Traub, W., and Yonath, 
A. (Eds.): Structural Aspects of Recognition and Assembly in Biological 
Macromolecules, Rehovot, Israel (in press). 

Piez, K. A., and Trus, B. L.: A new model for packing of type-l collagen 
molecules in the native fibril. Bioscience Reports 1:801-810, 1981. 



24 



OMIIN SCIENCE INFOM 
T HIM6CR (Oe 101 <■*• 



ZOl CT00090-01 CSL 



October 1, 19fll i 



PI: B. L, Trus 

OTHERS: A, C. Stevei 
P. M. Steini 
R, L. Jerni. 



Research Chemist 



LPB, NIAOOK 
OB, NCI 
Chemtst LTB.nCBO 



Brookhaven Nat 
The Rockefelle 
The RocVefelle 



LPB, NIAnilK; OB. NCI; LTB, NCI; Oology nept.. BNL; Microbiology. 
Rockfeller University 



_ Systems Pes! gn S ectlq i 



nCRT. NIH . Bethsda, HP 20205 



The sequi 
diffract 



NCI), myosin (w 



r proteins, toi 

iscopy, and spei 

th NIDR). keral 
en), and strepl 



'JI(c?'iSelR't!"iiJi'™"h!; !S!3" 






™''s,^is;!f""';" 






ZOl CT00080-02 CSL 


October 1. 1981 to Sept 


mber 30. 1982 






Computer Analysis of Ge 


Electrophoresis 










PI; B.L. Trus 


Research Chemist CSL. OCRT 


OTHERS: R. Goldman 


Staff Fello» LBP. NIAOOK 


V. NUocJem 


Staff Fellow CE. NIAOOK 


J. E. Ran 


Oirector NIAOOK 


R. felstcd 


Research Chemist BCRP*. NCI 




* Baltimore Cancer Research Program 


CI>Ofl««tl»G »»IIS (If ..,1 




LBP. NIAOOK; CE.NIAODK; 


BCRP.NCI 


LM/BRtNCH 1 

Computer Systems Laboratory | 


















tOtm ll*NTE«RS> jPflOfESSim*Li lOIHERl 1 


CHeCK *RPROPfll»It BlUlES) 




] (.) »Ul» SutJECIS a ( 




D(.0">.O.S □ (.,) ■.■ER.Ii.E 






.rli., l.,..ra,) 


This project was designed 




and accurately quantitate 


ne- and two-dimensional gels. Quantitative 


comparisons of two gels is 


semi-automatic, and one project has 


used methods developed her 


to separate the results of double-labeled 


radiography of protein gel 


using color negative film and appropriate 




because 3H and spillover of 14C 


are recorded in the hlue s 


nsitive layer of the film while HC alone 


is recorded in the green o 


red sensitive layer. This method was 


used to analyze the effect 


of growth rate and medium composition 


on the relative levels of 


ndividual proteins in a pathogenic 


strain of Escherichia coll 





Computer Analysis of Gel Electrophoresis 

This project was designed to allow NIH scientists to 
easily and accurately quantitate one- and two- 
dinnensional gels. Quantitative comparisons of two 
gels is semi-automatic, and one project has used 
methods developed here to separate the results of 
double-labeled radiography of protein gels using 
color negative film and appropriate filters. This is 
possible because 3H and spillover of 14C are 
recorded in the blue sensitive layer of the film while 
14C alone is recorded in the green or red sensitive 
layer. This method was used to analyze the effect of 
growth rate and medium composition on the relative 
levels of individual proteins in a pathogenic strain of 
Escherichia coli. 

Background and Objectives: The primary objective of 
this project has been to develop experimental 
techniques and computer software to easily and 
automatically quantitate two-dimensional gels. In 
addition, analysis of one-dimensional gels is equally 
accurate and feasible. Initially only Coomassie blue 
stained gels were analyzed, but currently 
autoradiographs are equally amenable to processing. 

Significance to Biomedical Researcli: Use of gel 
electrophoresis and autoradiographs is 
commonplace in chemical, biochemical, and 
biomedical research. However, the quantitation of 
these gels is difficult. We have developed systems 
that accurately and easily provide this quantitation to 
the scientist. A number of laboraties outside of NIH 
have requested our software for private use. 

Progress in FY82: This project has produced many 
useful results to a number of scientists at NIH. As 
new gels require analysis, further fine tuning of the 
methods will continue to improve the product. In 
addition, we have used the methods to analyze color 
negative film (rephotographed through appropriate 
color filters) so as to analyze the growth rates and 
medium composition on the relative levels of 
individual proteins in a pathogenic strain of 
Escherichia coli. These results are being submitted 
for publication. 



25 



Methods Employed: Gels were rephotographed onto 
Ektapan 4162 black and white film. Color films were 
photographed through appropriate color filters. The 
black and white negative was scanned on the 
Perkin-Elmer microdensitometer and stored on tape 
for later processing. A computer program CINT was 
used to analyze the two-dimensional gels, and 
another program OVERLP was used to correlate two 
gels when necessary or desired. PIC was used in the 
one-dimensional analyses. 

Proposed Course: Computer software is being 
expanded to provide for better matching of two gels. 
All software is essentially machine independent so 
as to be transferred to the newly-acquired image 
processing laboratory. Additional options are being 
added to the software so as to provide additional 
flexibility to the research scientist. 

Publication: 

Nikodem, V. M., Trus, B. L, and Rail, J. E.: Two-dimensional gel analysis of 
rat liver nuclear proteins after thyroidectomy and thyroid hormone treat- 
ment. Proceedings of the National Academy of Science 78:441 1 -441 5, 
1981. 



lylTHSDNlAH SCIENCE IHFORBATICW EXCHANGE 


HFAI 


H-ANoTuirSE^iCES 


PROJECT NUNBER 








ZOl CT0n091-01 CSL 




INTRA 


mUL RESEARCH PBOJECT 






198? 






Horphometric Analysis of Normal 


and Neoplastic Tissue Cultures 






PfiCFESSIOMAL PEHSONNEL ENCAGEO ON THE PROJECT 




PI: R. L. Trus flesearc 


h Chemist CSL. flCRT 


OTHEflS: K. K. Sanford Chief, 


In Vitro Carcinogenesis Section 




LCMB, NCI 


G. Jones Microbi 


ologist LCMB, fJCl 


mm 


H 


CCOPE«AT,^C ONHM^. .".) :^-.<J| 






"^i^' 


L*B/BR*riCH 


uraraar 


Systems Design Section 1 


INSTIIUIE AJJO LOCATION 








CHECK APPROPRIATE B0X{ES) 




3(.) HUNAN SUBJECTS Q (">) HW 


^TISSUES D(c) NEITHER 


D(.l) NINORS Q(,2) INTERVIEWS 




SWUIAHY OF «0R»! {im .oM^ or U„ - u.d.rli. 


l..,-onjO 


This project was designed to s 


tudy the morphonietric differences 


between normal and tumorigenic 


fibroblastic cell lines. Initially, 


human, rat, and mouse cell lin 


es were selected for analysis. The 


cells were photographed from 


iving cultures without staining or 


fixing. The types of criterii 


n being used by the computer to aid 




shape, and chromatin texture and clumping. 



Morphometric Analysis of Normal and Neoplastic 
Tissue Cultures 

This project was designed to study the morphometric 
differences between normal and tumorigenic 
fibroblastic cell lines. Initially, human, rat, and mouse 
cell lines were selected for analysis. The cells were 
photographed from living cultures without staining or 
fixing. The types of criterion being used by the 
computer to aid in differentiating between normal 
and tumorigenic cells include nucleus and nucleolus 
size and shape, and chromatin texture and clumping. 

Background and Objectives: This project, which was 
begun this year, uses standard techniques of image 
processing as applied to these low contrast 
unstained specimens as well as techniques 
developed at NIH. We hope to demonstrate that it is 
possible and practical to differentiate between 
normal and tumorigenic cells in a nondestructive 
manner. We are using many of the same criterion 
used by the pathologist in differentiating stained and 
fixed sections. 

Significance to Biomedical Research: We hope to 
demonstrate that this nondestructive method can be 
used with confidence to determine if a culture is 
normal. This method would be important for studies 
of carcinogenesis in cultures. 

Progress in FY82: Software was developed to 
perform a pilot study on three types of cultures. 
Preliminary results suggest that it may be possible to 
determine statistical differences between normal and 
abnormal cells. 

Methods Employed: Cell cultures were photographed 
through a light microscope onto 35 mm black and 
white film. The film was digitized by a Perkin-Elmer 
1010G microdensitometer with a 50 microns squared 
aperture. Images were viewed on a video frame 
buffer, and interactively processed. Results are 
stored in log files for each sample, and files are 
pooled for each type of culture yielding better 
statistics. The mouse and rat cultures underwent 
spontaneous neoplastic transformation, while the 
human fibroblast line was exposed to chemical 
carcinogens to generate the tumorigenic line. 

Proposed Course: After the analysis of the three pilot 
studies, we expect to continue analysis of additional 
cell lines, and are considering nonlethal staining 
techniques. 



26 



Virus Structure As Determined by Image 
Processing of Electron Micrographs 

A new virus structure, that of bacteriophage T7, has 
been determined by image processing of electron 
micrographs. We analyzed T7 poiycapsid tubes 
because these structures are more amenable to 
image processing. Optical diffraction revealed that 
the polycapsids were based on cylindrical foldings of 
a hexagonal lattice with a spacing of 12.6 nm, which 
is similar to the lattice constant for other complex 
icosahedral phage capsids defined to date. However, 
the details of the T7 capsomer differ from the other 
results. 

Background and Objectives: Viruses are significantly 
smaller than bacteria, and as a result are not seen in 
a light microscope. Information about their structure 
usually comes from electron microscopy, which is 
limited by resolution, low contrast, and noise. If 
staining is used, then the resolution is limited by the 
size of the stain, and often has noise as a result of 
uneven staining. However, because virus structures 
are generally periodic, they are a perfect candidate 
for image processing. 

Virus shells are composed of one or a few proteins 
that form simple repetitive geometric forms. The 
forms or containers can be, for example, cylinders, 
icosahedra, or spheres. There are classes of 
structures, and knowledge of the fine structure of 
one coat protein can be used to understand the 
structures of other similar viruses in the class. It is 
our primary objective to add to the pool of 
information, and to be able to use this information to 
increase our understanding about how virus structure 
relates to function and activity. This project was 
described under project number Z01 CT00082-01 
CSLin FY81. 

Progress in FY82: The results of a study of a virus 
previously determined by us, beet necrotic yellow 
vein virus, were published. In addition, the results of 
the structural determination of the T7 virus were 
presented at two meetings, and are being submitted 
for publication. These results are especially 
significant because another virus (polyoma), which 
has significant similarities to T7, was recently 
reported to have significantly differing geometry. 

Significance to Biomedical Researcti: This project 
should be considered as basic research whose aim 
is to increase our understanding of the structure and 
functions of viruses in general, as well as subclasses 
of viruses similar to those studied to date. 

Mettiods Employed: The micrographs were taken 
with a Philips EM400T microscope, and the best 
negatives were preselected by optical diffraction. 



Negatives were digitized on a Perkin-Elmer 1010G 
microdensitometer and analyzed by means of the 
PIC computer system. Results were photowritten on 
the Perkin-Elmer microdensitometer. Typical 
processing of the images consisted of Fourier 
filtering of up to 50 unit cells and symmetrization of 
the results as needed. 

Proposed Course: We anticipate evaluating other 
viruses for suitability for examination with these 
methods, and continuing with this project to 
determine the structure of various classes of viruses. 

Publications: 

Steven, A, C, Trus, B. L-. Putz. C. and Wurtz, M.; The Molecular Organiza- 
tion of Beet Necrotic Yellow Vein Virus. Virology 113:428-438, 1981. 

Steven, A. C, Serwer, P., and Trus, B. L.: Molecular Packing in Bacteno- 
phage T7 Capsid Visualized at 2.5 nm Resolution in Computer Proc- 
essed Electron Micrographs. Eighth Biennial Conference on Bacteno- 
phage Assembly, Fall Creek Falls Park, Tennessee, May 9-14, 1982. 

Trus, B. L., Sender, P., and Steven. A. C: Capsid Fine Structure of Bacterio- 
phage T7 Determined by Image Processing of Electron Micrographs. 
Tenth International Congress on Electron Microscopy, Hamburg, Ger- 
many, August 17-24, 1982. 






ZOl CI0009Z-01 CSL 



October 1. IQfi l to September 30. 198? 



I As netermined by Image Processing or Electron 



Research Chemist 
Visiting Scientis 



LPB, NIADDK; The I 



I Science Center 



Computer Syste ms Laboratory 



Systems Design Section 



DCRT, NIH, Rethesda. 



s structure, th 
essing of elect 


t of bacteriophage T7 
on micrographs. We an 


has been determined 
alyzed T7 poiycapsid 


ffraction revca 
1 foldings of a 
imilar to the 1 
1 phage capsids 
the 17 capsome 


ed that the poiycapsid 
hexagonal lattice with 

defined to date. Howe 
differ from the other 


to image processing 
s were based on 

a spacing of 12.6 ntt 
er complex 
.er, the details of 

results. 



27 



Image Processing of Electron Micrographs 

This project was designed to facilitate structure 
determination from electron microscopy by providing 
suitable software, hardware, and scientific expertise 
to allow other scientists, primarily at NIH, to use 
image processing and computer reconstruction to 
determine or understand a specimen's structure. 

Two new applications that began this year are 
analysis and identification of small particles by 
electron beam excited x-ray microanalysis as applied 
to aqueous suspensions of vertebrate retinal rod 
cells and the analysis of the microtrabecular lattice 
and the cytoskeleton to determine volume, surface 
area, and the diffusion of molecules. 

A study of densitometer techniques was completed 
and published. Studies continued from FY81 include 
analysis of keratin, membrane structure, and muscle 
structure. 

Background and Objectives: The objective of this 
project is to develop a general-purpose software 
package for the analysis of electron micrographs. In 
addition, the computer analysis requires optimal 
utilization of the available hardware and the 
availability of a research scientist capable of 
providing logistical support. Techniques and software 
developed in this project have been used 
independently of this project both at NIH and at 
laboratories outside NIH. 

Significance to Biomedicai Researcii: Computer 
analysis of electron micrographs is still a relatively 
recent addition to the tools available to scientists for 
structural analysis. Few laboratories have the 
combined software and hardware capability to 
perform the image processing and image 
reconstruction available at NIH. These techniques 
are especially powerful when applied to two- 
dimensional crystalline structures. In addition, we can 
correlate and align similar particles that are not 
crystalline, and correct for a number of artifacts and 
experimental problems. 



Progress in FY82: This project has had some growth 
in software, but primarily has grown in the utilization 
of programs and the PIC system. It is feasible for an 
NIH scientist to bring in a problem and obtain 
preliminary results in a relatively short period of time. 
Then a decision is made to expand the preliminary 
study into a project, or to use the results that were 
obtained. 

One study, in collaboration with NIADDK, used the 
computer to analyze digital information to analyze 
small particles by electron beam excited x-ray 
microanalysis for particles in aqueous suspension. 
This novel approach was applied to the isolated 
outer segments of vertebrate retinal rod cells and 
was used to study the distribution of K, Os, P, and 
45Ca in unstable objects. 

Another study, in collaboration with the Physical 
Sciences Laboratory, DCRT, examined the 
microtrabecular lattice and the cytoskeleton. Images 
were digitized and analyzed for the fraction of 
interlinked slender strands versus the amount of 
open spaces. 

Proposed Course: This project will continue software 
development as needed and will be converted to use 
the new image processing facility as it becomes 
available. In addition, new biological structures that 
become available for analysis will be examined. 

Publications: 

McGee, P. A., Trus, B. L., and Steven, A. C: Techniques to Evaluate the 
performance of Scanning Microdensitometers in the Digitization of 
Electron Micrographs. Micron (in press). 

Trus, B. L, and Steven, A. C: Computer Processing of Electron Micro- 
graphs of Periodic Biological Specimens. Washington Society of Elec- 
tron Microscopy Annual Picture Meeting, Uniformed Services, University 
of the Health Sciences, Bethesda, MD, May 6, 1982. 

Gershon, N. D., Porter, K. R., and Trus, B. L.: The Microtrabecular Lattice 
and the Cytoskeleton. Their Volume, Surface Area and the Diffusion of 
Molecules Through It. Aharon Katzir-Katchalsky Symposium on Biologi- 
cal Structue and Coupled Flovws, Rehovot, Israel, June 6-11, 1982. 

Hagins, W. A., Foster, M. C, George, J. G., and Trus, B. L: Combined X-ray 
Microanalysis and Radioautography of Diffusible Elements in Aqueous 
Suspensions of Cells and Cell Fragments. Proceedings of Microbeam 
Analysis Society, August 1982. 

Trus, B., and Steven, A.: Digital Image Processing of Electron Micrographs- 
The PIC System. Journal of Ultramicroscopy Q: 383-386, 1981. 



28 



s!;"™iiii.M 


foS wf'SJrfhu ii^T' 


HUL'.H'.No'S'Lfv.C.S 


fflojici Nu.e«. 








201 CT0008e-02 CSL 


October 1. 1981 to Septanber 30, 1982 


TITLE Of PflOJ£ 


CT (80 chir*c(*r» or I.»] 


Image 


recessing of Electron Hicrographs 


™»KstS"" 


trsd'ENjwiJ'oJ'Tie'SSi"'; *"" '"'" " """"*'■ '"""'"""* *"" *^^ """ 


PI: 


B. L. Trus Research Chemist CSL, DCRT 


OTHERS : 


A. C. Steven Visiting Scientist LPB. NIADDK 




H. A. Hagins Chief. Section on Membrane Biophysics 




LCP, NIAODK 




H. C. Foster IPA Appomtee/Guest Worker LCP. NIADDK 




J. G. George Laboratory Technician LCP. NIADOf. 




N. D. Gershon Senior Staff Fellow dsl. DCRT 




K. Porter University of Colorado and Fogarty 




Scholar 


eO0P.«*IIHt UN 


IS tM .',) 


LPB, Hit 

L<|/S>.«C. 


DDK; LCP. NIADDK; PSL. DCRT; University of Colorado 




Computer 


Systons Laboratory 


Systems 


Design Section 






DCRT. NIH. Bethesda, HD 2020S 


■2 1 .2 1 


J (.) MWAh S 


It BO<((S) 


Dl.M-.NOBS 


D{.2) iNimit-s 


This project was designed to facilitate structure determination 1 


from el 


ectron microscopy by providing suitable software, hardware. 




entific expertise to allow other scientists, primarily at NIH 




image processing and computer reconstruction to determine or 


underst 


and a specimen's structure. 


Two new 


applications that began this year are the analysis and identi- 


ficatio 


n of small particles by electron beam excited x-ray microanalysis 


as appl 


led to aqueous suspensions of vertebrate retinal rod cells and 




ysis of the microtrabecular lattice and the cytoskeleton to 




ne volume, surface area, and the diffusion of molecules. 



ZOl CTO0095-01 CSL 



QcLQb er - 1 . lisi tn ;;pnt.emt?pr 3(1 . iqfl? 



-Project DevelQpmeat SgclJ.flg_ 



L M*hu?Js! "^"' '^^'^'"j'^'V "" ''"^"'' 



The original configuration was usee 
bacteria system. This year a new sj 
rapid scan spectrophotometer (RSS) 
mammalian eel 1s. 



Potentiometric Titration Controller 

A Potentiometric Titration Controller was developed 
to study the electron transport chain at the cell level. 
The details of this system have been reported in 
previous years. The original configuration was used 
to study the E. coli bacteria system. This year a new 
system incorporating a rapid scan 
spectrophotometer (RSS) is being used to study 
mammalian cells. 

Background and Objectives: The exact nature of the 
respiratory chain in the mitochondria of mammalian 
cells is still not known. The carriers are various iron- 
containing molecules that have different affinities 
(and corresponding voltages) by which they hold 
electrons. The amount of energy liberated by 
passage of an electron from one carrier to another 
can be determined by the difference in redox 
potentials of these carriers. The redox potential is 
the voltage at which the carrier is equally reduced 
and oxidized and it can be determined by a Nernst 
relationship. 

The original Potentiometric Titration Controller could 
fix solution potential using electric currents under 
program control. The amounts of oxidized and 
reduced transport carriers were determined by 
calculations based on voltages obtained from the 
spectrophotometer corresponding to optical 
transmittance and wavelength. A wavelength scan 
motor was driven under computer control to change 
the wavelength and obtain complete spectra. 

Progress in FY82: The Potentiometric Titration 
Controller has been greatly enhanced with the 
addition of a Rapid Scan spectrophotometer (RSS) 
in place of the original device. As before, the 
solution potential is fixed by using electric currents 
and the amounts of oxidized and reduced electron 
transport carriers are determined using spectral data. 
Now, however, complete optical spectra can be 
taken in milliseconds instead of the 20 to 30 
seconds required by the earlier system. Instead of 
controlling a wavelength drive motor to acquire 
spectra, entire scans are initiated under 
microcomputer control. Complete scan data is stored 
In an internal memory buffer of the RSS. Either 
single scans or up to 64 rapid successive scans can 
be taken and stored in the buffer. Data is then 
transferred to the controller via a 9600 baud serial 
link and recorded on disk. As before, selected data 
is then transferred via modem to the DECsystem-10 
for further analysis and graphics. 



29 



Proposed Course: The earlier Potentiometric Titration 
Controller was used to study the electron transport 
chain of a bacterial system, E. coli. A similar 
approach using the new RSS will now be used to 
extend these studies to the chain in mammalian 
mitochondria. In addition, the new system will enable 
us to study other important kinetic features of 
respiratory systems not possible with the original 
system. 

Metabolic Energy Measurements 

A microcomputer-based instrument has been 
developed to study cellular energy transduction 
phenomena. Specially designed electrodes have 
been constructed and interfaced to the 
microcomputer to calculate membrane potential and 
protonmotive force. Derived paramaters are recorded 
in realtime on a multipen plotter via D/A converters. 
The user can observe all parameters as he perturbs 
alterations to the medium in which the respiring and 
energy transducing systems are suspended. 

Background and Objectives: Cellular energy is 
derived from the oxidation of substrates. Electrons 
removed from these substrates are passed through a 
chain of respiratory carriers-eventually, oxygen. 
During the process, energy is stored in the form of a 
gradient of protons across the cell or mitochondrial 
membrane. The difference in the number of protons 
inside the cell and outside the cell results in an 
electrochemical potential or protonmotive force. This 
electrochemical potential has two components, an 
electrical potential across the membrane, and a 
chemical potential characterized by a difference in 
pH. These quantities are very difficult to measure 
and current techniques are cumbersome and time- 
consuming. 

Progress in FY82: In FY80 and FY81 we developed 
a microcomputer-based system that is capable of 
realtime monitoring of these components. In FY82 
specific electrodes have been designed and 
constructed that respond to changes in electrical 
potential and to changes in pH. The voltage signals 
from these electrodes are amplified and sent to the 
microcomputer via an analog-to-digital converter. 
The computer stores the signals after digitally 
filtering out noise and correcting for the nonlinearity 
of the electrode transfer functions. It then computes 
the electrical potential and the difference in pH from 
the measured electrode voltages, and from these 
determines protonmotive force. 



The microcomputer also monitors signals from a pH 
electrode and an oxygen-measuring electrode. The 
computer program corrects for relaxation time delays 
in electrode responses and filters out noise. The 
oxygen uptake rate of the cellular material is 
calculated as are the protron extrusion rate and the 
proton-to-oxygen ratio. All derived parameters are 
calculated in realtime and are output via D/A 
converters to a multipen plotter. In this manner the 
user can observe all quantified paramaters in parallel 
as he alters the medium in which the respiring and 
energy-transducing system is suspended. 

Proposed Course: Hardware and software 
development are now complete and all programs 
reside in ROM and therefore do not require 
downloading as before. The emphasis next year will 
be on acquiring and analyzing data. New hardware 
and software will be developed as the need arises. 



SSi'i 


?"1S= 


''^"■'"'"•-"^" 


""™'-'SoSfS'™|'" 


ZOl CT00O62-03 CSL 


October 1, 1981 to September 30, 19S2 1 










Net 


bolic Energy Measurements 


™f 


\^,Z°« 


*PMloi*Ha'wG«e"oH'TH£'pfiOJECT *"" "^'■" ^ "'""="'*'■ '""ESTICATOflS AMD ALL OTHES 




PI: 


D. C. Songco Electronics Engineer CSL, DCRT 




OTHERS: R. V. Hendler Chief, BES LCB, IR, NHLBI | 






0. H. Setty Visiting Associate LCB. IR, NHLBI 






R. Shrager Mathematician LAS, DCRT 






W. Friauf Chief, EEES BEIR. DRS 




LCB, 


UNH5 (if .r,,) 

IR. NHLBI 










Comp 


ter Systems Laboratory 


:.t(;i 


Proj 


ct Development Section 


INSI 


prRT 




"' ""'.e \—. r 1 


D{. 


"""" 


RIAIE B0«(£!;) 


ni- 


1 »\m 


S □(,!).«,ERViEWS 






microcomputer-based instriwient has been developed to study 
ellular energy transduction phenomena. Specially designed 
lectrodes have been constructed and interfaced to the 
icrocomputer to calculate membrane potential and 




protonmotive force. Derived paramaters are recorded in realtime I 






n a multipen plotter via D/A converters. The user can observe 






11 parameters as he perturbs alterations to the medium in 






hich the respiring and energy transducing systems are suspended. 



30 



Electron Microanalysis Facility 

CSL is collaborating with BEIB, DRS, in developing 
an automated electron microanalysis facility 
consisting of two electron microscopes interfaced to 
a computer system. The facility will be used for 
research into the elemental composition of biological 
specimens, and for the development of new 
techniques in electron microscopy. CSL is designing 
and implementing the computer system, which will 
acquire and display the spectra and images 
produced by Electron Energy Loss Spectrometry, 
Energy-Dispersive x-ray Spectrometry, and 
Wavelength Dispersive x-ray Spectrometry. (See 
also: Z01 RS1 0058-04 and Z01 RS1 0059-04.) 

Background and Objectives: The Computer Systems 
Laboratory is designing and implementing a 
computer system as part of the BEIB Electron Beam 
Imaging and Microspectroscopy Facility. The facility 
consists of two electron microscopes and will be 
used for research into the elemental composition of 
biological specimens and for developing new 
techniques in electron microscopy. 

One of the electron microscopes is a Hitachi H-700H 
200 keV Scanning Transmission Electron 
Microscope (STEM) equipped with: 

• a lithium-drifted silicon (Si(Li)) detector 
connected to a Kevex 7000 Analytical 
Spectrometer for performing Energy-Dispersive 
x-ray Spectrometry (EDS) 

• an electron spectrometer for performing 
Electron Energy Loss Spectrometry (EELS) 

• detectors for bright and dark field electron 
current signals. 

The other electron microscope is a Cameca 50 keV 
Electron Microprobe equipped with: 

• a Si(Li) detector for performing EDS 

• three Wavelength Dispersive x-ray (WDS) 
spectrometers 

• detectors for bright and dark field electron 
current signals. 

A PDP-11/60 computer system is being interfaced to 
both microscopes to perform the following functions: 

• control electron beam position, stage position, 
and the various detectors 

• acquire spectral and image data from all 
detectors 

• process and display the spectral and image data 

• monitor and display a wide variety of 
'housekeeping' parameters, including: lens 
currents, lens temperatures, beam current, beam 



energy, pump temperatures, coolant flow, 
vacuum pressures, water leak detectors, power 
supply voltages, room temperature, and room 
humidity. 
Progress in FY82: CSL's software efforts this year 
have been concentrated on these aspects of data 
acquisition from the STEM: 

• improvement of EEL spectral data acquisition 
and display 

• implementation of EEL and EDS image 
acquisition and display 

• implementation of fast electron current signal 
acquisition and display 

• retrieval of empirical x-ray spectral data 

• improvement of the Kevex 7000 and 
housekeeping data acquisition software 

• development of a user interface. 

EELS data acquisition and control of the STEM 
beam position is done by a satellite processor 
connected to the PDP-11/60 by a high-speed link. 
Software has been written that allows the STEM 
operator to define areas of a specimen as targets for 
data acquisition and to collect EDS, EELS, and 
electron current signal data from the target areas. 

Three data acquisition modes have been 
implemented: 

• SPECTRUM mode produces a single x-ray and/ 
or EEL spectrum from the target area. 

• IMAGE mode produces any combinatin of EEL, 
EDS, or electron current signal images from the 
target area. 

• FAST IMAGE mode produces only current signal 
images, but at high speed. 

Software has been developed this fiscal year for 
display and scaling of EEL spectra using the Kevex 
7000. Also, EEL, EDS, and electron current signal 
images can be displayed on the DeAnza ID5400 
Display System, where zoom, scroll, and contrast/ 
brightness enhancement can be performed. 

Software to retrieve empirical x-ray spectral data 
produced by electron energy transitions within 
ionized atoms was completed. This software allows 
an operator or another programmer to specify an 
element and the transitions or absorption lines of 
interest using either Siegbahn or Shell-Pair notation. 
It then looks up the associated energies and relative 
peak intensities. Conversely, an energy range may 
be specified, in which case the transitions and 
absorption lines within that range are retrieved. The 
x-ray data base was assembled by BEIB from 
multiple sources, and contains over 3000 entries. 
The contractor validated this data by fitting each x- 
ray line series to a model based on Moseley's Law 



31 



and flagging for inspection any entry showing a large 
deviation from the theoretical value. The resulting 
data base is probably the most complete and 
accurate available in machine-readable form, and 
has been sent to over ten extramural requestors. 

EDS data acquisition is done by the Kevex 7000, 
which is connected directly to the computer. 
Software has been developed to allow programs on 
the 1 1 /60 to control the KEVEX 7000 and to save or 
restore spectra to or from disk files. During this fiscal 
year this software was improved to make it much 
more efficient and easier to use. It was also installed 
on a PDP-1 1 /34 computer system connected to a 
JEOL JEM-100CX electron microscope for the 
Laboratory of Neuropathology and Neuroanatomical 
Sciences, NINCDS. 

Housekeeping parameters are acquired by the 
computer by means of an Analogies AN5400 data 
acquisition subsystem. Software has been developed 
to acquire, monitor, and display the STEM beam 
energy, magnification, lens currents, pump 
temperatures, and coolant flow. Improved calibration 
parameter maintenance utilities also have been 
developed. 

A menu selection scheme simplifies the operation of 
the data acquisition and display software. The menu 
selection software is completely table-driven so that 
it is easy to add new functions as they become 
available. Menu selection terminates with the display 
of a form for entering or modifying the parameters 
for the chosen function. Each form records the 
parameters last used and restores them as defaults 
the next time it is selected. This greatly simplifies 
operation, because most forms have dozens of 
parameters, few of which are changed with each 
use. Currently, forms for data acquisition and display 
functions, housekeeping parameter display, and 
specimen target definition can be activated through 
menu selection. 

The microprobe was connected directly to the 1 1 /60 
with a serial interface, and BEIB has developed 
software for acquiring WDS data. 



Proposed Course: Next year, we expect to: 

• implement primitive image processing functions, 

• enhance the data acquisition software, and 

• automate reading of the STEM beam current. 
Publications: 

Fiori, C. E., Gorlen, K. E., and Gibson, C: Comments on the Computeriza- 
tion of an Analytical Electron Microscope. Proceedings of the Thirty- 
Ninth Annual (Meeting of the Electron Microscopy Society of America. 
Baton Rouge, Claitor's Publishing Division, 1981, pp. 246-249. 

Fiori, C. E., Myklebust, R. L, and Gorlen, K. E.: Sequential Simplex: A 
Procedure for Resolving Spectral Interference in Energy Dispersive X- 
ray Spectrometry. Energy Dispersive X-ray Spectrometry. Gaithersburg, 
IvID., National Bureau of Standards Special Publication 604, 1979, pp. 
233-272. 

Fiori, C. E., Swyt, C. P., and Gorlen, K. E.: Application of the Top-Hat Digital 
Filter to a Nonlinear Spectral Unraveling Procedure in Energy-Disper- 
sive X-ray fvlicroanalysis. fi/licrobeam Analysis. San Francisco, San 
Francisco Press, Inc., 1981, pp. 320-324. 



teSHiS.iS' 


S^'no;"™"!?; .S!!!!" 


ntCn Sm'\tnm: 




r«.j£c> 


"""'" 












ZOl CTOonci 


03 CSL 


nctob 


r 1, 1981 to September 30, 1982 


















Elect 


on Microanalysis Facility 








"^i%,\'ST\\ 


;io;r£L'"«";;s'(»''TO'™s ""° """ " """" 


I ,».K,IO., 


«s .«0 < 




PI; 


K. E. Gorlen Electronics Engine 


er 


CSL 


OCRT 


OTHEBS: 


C. E. Fiori Physical Scientist 

L. K. Barden Electronics Engine 

J. S. Del Priore Mathematician 

P. S. Plexico Chief, Project flev 

C. C. Gibson 

H. S. Eden 

J. R. Ellis 

G. R. Hook 

R. n. Leapman 

C. R. Swyt 


clopment 


BEIB 
CSL 
CSL 
CSL 
BEIR 
BEIB 
BEIB 
BEIB 
BEIB 
BEIB 


DRS 
OCRT 
nCRT 
DCRT 
DRS 
ORS 
DRS 
ORS 
DRS 
ORS 


C0OP£H*riNC ml 


s (i< .»,) 








nRS; NKLBl; NIADDK; NIMH; NINCnS 








L«6/aHAriCH 
Compute 


Systems Laboratory 








Project 


nevelopnent Section 








OCRT. N 


H. Bethesda. m 20J05 








3 


1 3.0 1 

! ■o.(ES) 

JStIS D(t) »li««« IISSUES 


D (■) «Eii 


ER 




a(..).i».«s 


n (.!) I«IS«»IEWS 








CSL 


Is collaborating »1th DRS/8EIB In develo 
tron microanalysis facility consisting o 
oscopes Interfaced to a computer system 


ping an a 

f two ele 

The fac 


t^r 


111 be 


Elec 
Spec 


Imens, and for the development of new te 
oscopy. CSL Is designing and Implementi 
1 win acquire and display the spectra a 
:ron Energy Loss Spectrometry, Energy-Oi 
trometry, and Wavelength Dispersive K-ra 


ng the CO 
nd Images 

y Spectre 


produc 

-ray 

etry. 


system, 
ed by 


See 


also: ZOl RS1005e-04 
Zni RSin059-l)4 









32 



Molecular Interactions Laboratory Data System 

This microcomputer (PDP 1 1 /03) data system 
supervises the acquisition and processing of 
information from an analytical ultracentrifuge and a 
circular dichroic spectropolarimeter used in MDB, 
NHLBI, to investigate the interactions between 
human lipoprotein subunits. 

Objectives and Methods: As a supplement to the 
ultracentrifuge data system, a microcomputer-based 
data acquisition and analysis system was developed 
for use with a Gary Model 61 CD spectropolarimeter. 
The system consists of a simple, flexible CD 
spectropolarimeter/microcomputer interface and an 
interactive data processing program system by which 
CD spectra may be acquired, averaged, subtracted, 
converted to mean residue ellipticities, printed, and 
stored for future use. Stored data may also be 
transferred conveniently to a large computer facility 
for semi-automatic conformation analysis. The 
system overcomes some of the difficulties 
encountered in attempting the visual interpretation of 
noisy CD spectral recordings and in providing 
additional data manipulation capabilities not easily 
realizable with manual methods. 

The CD spectropolarimeter is interfaced with the 
microcomputer system via a special highly noise- 
immune interface scheme based on the conversion 
of the signals from the spectropolarimeter to pulse 
trains. These pulse trains are then transmitted via 
current loops to a CSL-designed timer/counter 
interface board in the microcomputer. The operating 
software consists of two programs that interact with 
the user through a standardized combination of 
menus and dynamically alterable displays. One 
program provides data acquisition, processing, 
output, and storage functions, while the other 
program provides all but acquisition and includes the 
capability to edit the operating parameters of a scan 
file stored on the diskette. 

Progress in FY82: Considerable effort was spent this 
fiscal year in testing the system and in developing 
tools on the DECsystem-10 for the graphic display of 
spectra. The results of a study of system 
performance with very dilute solutions were 
presented at the 1982 FASEB meeting in New 
Orleans. The CD data system has proven to be 
efficient and easy to use; over 900 scans on several 
hundred samples have been processed in the past 
year. The time required to process a CD scan into 
mean residue ellipticities has been reduced from 
several hours to less than 30 seconds. In addition. 



an arbitrary number of scans may be averaged, 
difference spectra may be generated quickly, and 
data at any stage of reduction may be stored for 
future retrieval and manipulation. With this system 
usable CD measurements have been obtained at 
protein concentrations below 1 microgram per 
milliliter. 

Proposed Course: Plans for the future include 
modification of the ultracentrifuge interface to 
provide greater noise immunity and the addition of 
support for a digital plotter that has recently been 
added to the system. 

Publication: 

Tate, R., Schultz, A., and Osborne, J.: Computer-Assisted Analysis of apo- 
lipoprotein subunit interactions. Federation Proceedings 41: 874, 1982 



?sij«nSeM'ioS'«o"«:*tih: \'^.T 


Mui;'.H**So'*;uiri"v.c.. 




Kl»e» 




'""^■Soyta^""-'" 


701 






IBIRUMU. lEtUHH nO.KCI 








Molecular Interactions Laboratory Data System 


NueS, LtBOnilORY HDD IMSTIIUIE tfFILUIICNt, MO ItlLiS tf miMClrU. IHVtIIICA 




raOFESSIONAL PCRSOMNa ENCACEO ON 1H( PftOJECT 




PI: R. L. Tate. Ph.D. Computer Specialist 


CSL, IKRT 


OTHERS: 0. C. Osborne, Ph.O. Research Chemist 


MPB. NHLBI 


A. R. Schulti. Jr. Head, Processor Oeslan 




Section 


CSL. DCRT 


COOPtRMINO UNIIS (.« ..,) 




HOB. NHLfll 




LAB/ennNCH 








Processor Design Section 


^~ PCPT. NIH. Bethesda. TO 20205 


,.,*. «-.,»«..^ |«.Qrj.ON.t. |O.H£fl. 


3[.) HU-«N5u8*m a(0««it«.nssu£S a Ic) wi 


N(« 


Uf.l] HINORS a {.2} IfURVK-G 




SUWI*RT Of WORK {JOO .««, »r 1 a.,1.., s.,.e^O 




This microcomputer (POP U/03) data system supervises th 


acquisition 


and processing of information from an analytical ultrace 


trifuge and a 


circular dichroic spectropolarimeter used in TOB. NHLBI. 




the interactions hetween human lipoprotein subunits. Cur 


ent capabilities 


include acquisition, display, and preprocessing of data 


rom the 


ultracentrifuge and transfer of preprocessed data files 




OECsystem-in for further analysis under MLflB using prede 


ined procedures 


invoked hy a few simple cotmands. Software support for tl 


e spectro- 


polarimeter includes the ability to add. subtract, and av 




spectra and to transfer files to the POP-IO for further a 


nalysis. The 


results of a study of system performance with very dilute 


solutions were 


presented at the 198? FASfR meeting in New Orleans. Plans 




underway for modification of the ultracentrifuge ioterfac 


e to 


|^0V|ide qp||ter noise immunity and for the addition of a 


digital plotter 



33 



Californium-252 Plasma Desorption Mass 
Spectrometer Data System 

Background and Objectives: The Californium-252 
time-of-fliglit (TOF) mass spectrometer employs 
nuclear fission fragments to ionize samples that 
frequently have proven intractable to other methods 
of analysis. In this instrument, fission fragments 
generated by the radioactive decay of a thin film of 
252-Cf impact on a thin layer of sample deposited on 
a conductive plastic film, producing a localized 
plasma. The sample molecules produced within this 
plasma are extracted by an electric field and briefly 
accelerated down an evacuated tube toward a 
microchannel plate ion detector. The elapsed times 
between the ionization event and the arrival of the 
ions produced are measured with an ultraprecise 
clock capable of measuring time intervals of 
hundreds of milliseconds with a resolution of 800 
picoseconds. The elapsed time measurements are 
then sent to a computer where they are sorted, 
tallied, converted to mass units, and displayed. The 
extended range of the timing clock coupled with the 
unique characteristics of the ionization process make 
this mass spectrometer ideally suited to the 
investigation of the high molecular weight 
compounds typical of biological materials. The TOF 
mass spectrometer, which is not commercially 
available, was developed at Texas A & M University 
by Dr. Ronald Macfariane under a NHLBI contract. 
The data system was specified to be compatible with 
interface hardware and software available from Dr. 
Macfariane. The need for realtime sorting of a large 
volume of input data puts unusual and stringent 
demands on the data system that controls the 
spectrometer and acquires and processes its data 
output. Realtime performance and the ability to 
access very large data arrays in main memory are 
key considerations. 

Progress in FY82: After delays necessitated by 
hardware design problems, software upgrades, and 
facilities renovations, both the spectrometer and a 
data system modeled after one in use at Texas A & 
M University have been installed in the laboratory of 
Dr. Henry Fates, LC, NHLBI, and will soon be 
functional. This instrument will provide NIH the 
capabilities of mass analysis for compounds that 
have proven difficult or impossible to analyze by 
other mass spectrometric means, also extending the 
range of mass analysis to compounds with molecular 
weights in excess of 5000. 



Proposed Course: Expansion of the computer 
system to include a line printer and additional 
memory is currently planned. Modifications to the 
mass spectrometer are also in progress to enhance 
its safety and reliability. Plans are being developed 
to implement computer control of sample changing 
as well. 

Distributed Laboratory Data Acquisition and 
Control System 

An integrated laboratory data acquisition and 
processing system has been developed for LCP and 
LMB, NIADDK. The system is configured with 
satellites coupled through a local network to a host 
processor. At each satellite a dedicated 
microcomputer system performs data acquisition 
from and control over an instrument/experiment. 
Although acquired data files may be stored locally, 
they are normally transferred via the network to a 
host storage medium. The local network allows the 
host storage medium to appear as a 'virtual' storage 
device to the satellites. 

Bac/<ground and Objectives: A system of 
microcomputers capable of independently controlling 
and acquiring data from an instrument/experiment 
was proposed in December 1976 as the best system 
architecture of upgrading laboratory data processing. 
A prototype laboratory data acquisition and control 
system (LDACS) computer and essential elements of 
the communication system were developed. 

Satellites perform the realtime data acquisition and 
instrument control functions. Their configuration 
includes a Digital Equipment Corporation (DEC) LSI- 
11 microcomputer, a 28K word memory, low density 
random access storage, graphics terminal, and all 
the necessary I/O hardware to interface the 
instrument/experiment. Software developed by CSL 
for each satellite, running under DEC'S RT-1 1 
operating system, provides the user with a 'turn-key' 
system. Presently, the system is configured with 
eight satellites, supporting eleven instruments, 
connected (via the communications computer) to a 
DEC PDP1 1 /70 host processor. Instruments 
connected to the network include: 
spectrophotometer, CARY 14, CARY 118, CARY 
219, two Perkin-Elmer 580B's, a 
microspectrophotometer (designed by NIADDK); 
spectropolarimeters, CARY 60, Jasco J500A; a 
Varian Electron spin resonance spectrometer; I. S. 
Co. Model 1440 liquid chromatograph; and a 
stimulus response retina experiment. 



34 



siSEiriSB^s'r-sm;?: v^.t 


,«2iHiii;'iifl 


ZOl CT00058-03 CSL 


October I. 1<581 to September 30. 193? 


Cdlifornium-252 Plasma nesorption Mass Spectrometer Data System 


NUItS, LABOHHOBT *ND INSTIIUT£ AfFILUtlWS, AHO (lILtS Of PftlNCIML l»V££TIC*TOBB W.0 *LL OIHtfi 

PI: R. L. Tate, Ph.n. Computer Specialist CSL, DCRT 
OTHfR: H. M. Fales. Ph.D. Chief LC. NHLRI 


COOI-tfiAIIHC UN.IS (.t ..,,) 

LC. NKLBI 




Processor Design Section 


DCRT. NIH. Bethesda. MD 2020B 


L .« A., ^^ pOFtSS.p, |OI»£H. 


CHtCK APPflOPSI«I£ eO»(ES) 

QUI) «.N0«; n(.J) INICBVIt-S 






T*ie Californium-2S2 plasma desorption mass spectrometer puts unusual and 
stringent demands on the data system that controls the spectrometer and 
acquires and processes its data output. Realtime performance and the 
ability to access very large data arrays in main memory are key 
considerations, ftfter delays necessitated by hardware design problems, 
software upgrades, and facilities renovations, both the spectrometer and 
a data system design modeled after one in use at Texas ARM University 
have been installed and are now functional. This instrument now provides 
NIH the capabilities of mass analysis for compounds that have proven 
difficult or impossible to analyze by other mass spectrometric means. 
It also extends the range of mass analysis to compounds with molecular 
weights in excess of 5000. 



Sil^iriSJBtg'tS^M'rii?: T.T 


"'"'JJ^S'sf'"'"''"" 


ZOl CT00n56-03 CSL 




Distributed Laboratory Data Acquisition and Control System 


««tS, L*BM*IO«r AM INStlTUIt AfFILUTIONS, UU> THUS Of PfllNClPM. INV£SIIC*TO«S AND ALL OIHtR 

PI: J. I. Powell Electronics Engineer CSL. nCRT 

OTHERS: W. H. Jennings Physicist LCP, NIADDK 
E. R. O'Bryan Electronics Engineer CSL. nCRT 
A. R. Schultz, Jr. Chief. Processor Design 

Section CSL. nCRI 


LCP and LHB. NIAODK 


LAe/BAANCH 

Computer Systems Laboratory 


Processor Design Section 


DCRT. NIH, Bethesda. HD 20205 


3.0 [3.0 1 ' 


CM£M APPROWtlAIE BO.(tS) 

J (.) HUIIAM SOflJECIS a M HMW tlSSUtS Q t=) NEITHER 


SUIWAR) W kOB« (JOO .on), or lt» - -r-d.rlin, t,,.ord.) 

An integrated laboratory data acquisition and processing system has 
been developed for LCP and LHB, NIAODK. The system is configured 
with satellites coupled through a local networli to a host processor. 
At each satellite a dedicated microcomputer system performs data 
acquisition , rom and control over an instrument/experiment. 
Although acquired data files may be stored locally, they are 
normally transferred via the network to a host storage mediuifl. The 
local network allows the host storage medium to appear as a 
'virtual' storage device to the satellites. 



The local network includes a software nnodule, 
installed as a handler under the RT-11 operating 
system, at each satellite. Each satellite is connected 
via a hardwired serial link to a front end 
communications computer. The communications 
computer performs a file store and forward function. 
Received files are placed on a first-in-first-out queue. 
Files are transferred from the queue to the host via a 
parallel DMA link. The communications task running 
on the host maps the files to the appropriate 
directory area based on the identity of the satellite 
that originated the transfer and the extension of the 
file being transferred. 

The host processor, a DEC PDP 11/70, is configured 
with: 640K words of memory, a high speed printer/ 
plotter, an X-Y plotter, a 9-track magnetic tape drive, 
dual floppy disk drives, and two large capacity disk 
drives. DEC's multiuser, multitasking operating 
system, RSX-1 1 M, is used to service the processing 
needs of the users. User access to the host is 
provided by hardwired links between terminals and 
host timesharing ports. Processing software provided 
at the host allows LDACS data files to be: added, 
subtracted, averaged, smoothed, baseline corrected, 
integrated, differentiated, multiplied by a constant, 
and added to a constant. The results may be 
displayed graphically on a Tektronix terminal, typed 
at a terminal, printed on the line printer, plotted on 
an X-Y plotter or transmitted to the DECsystem-10 
for additional processing. 

Progress in FY82: One new satellite, supporting a 
second Perkin-Elmer 580B spectrophotometer, was 
added to the system in FY82. LDACS software 
supporting the acquisition of ESR data versus time, 
and software supporting acquisition of MSP data 
versus wavelength (wavelength controlled by the 
LDACS) was added this fiscal year. 

General processing programs were developed and 
installed for the PDP-11/70 allowing manipulation of 
LDACS data files and printing/plotting of the results. 
Special purpose programs were provided for 
processing of spectrophotometer melting run data 
and processing of data files from an MSP 
experiment. 

The last vestige of the 1 1 -year-old centralized data 
acquisition and processing system, the Honeywell- 
516 computer, was removed from the configuration. 

User's manuals for the CARY 219, PE 580B's, and 
the stimulus response retina experiment were written 
under contract. 



35 



Documentation of minor modifications and additions 
to the software libraries developed for LDACS were 
completed under contract. Retrofitting of LDACS 
units was initiated with programs utilizing the 
software libraries and incorporating (where required) 
the multipurpose counter/timer module, and the 
temperature control unit (both developed by CSL). 

Proposed Course: Support for the system will 
continue. The retrofitting of LDACS units with up-to- 
date software will be completed. Documentation of 
the system will be given a high priority, with the 
objective of completing an LDACS User's Guide for 
each LDACS and completing all systems 
documentation. 

The original scope of the project (upgrading the H- 
516 centralized system) is approaching completion. 
However, it is anticipated that some level of long- 
term support will continue. New satellites or 
instruments may be added to the system, and 
existing LDACS acquisition programs may be 
modified to enhance data acquisition or to 
incorporate new instrument/experiment setups. 
Occasionally, special-purpose programs are required 
to process a set of experimental data. 

Publication: 

Powell, J. I., Fico, R., Jennings, W. H., O'Bryan, E. R., Schultz, Jr., A. R.: A 
Local Network for Distributed Laboratory Microcomputers. Tutorial-Mi- 
crocomputer Network, 1981, pp. 263-268. 




, Engineer CSL, DCRT 

CSL. DCRT 
CSL, nCRT 



DCRT, SIH, Bethesda. HP 20205 



-rays, micrographs i 
e system is complett 



ility to display and analyze 
of a powerful 32-bit 
gh resolution displays. 

other images. The computer 
construction of the physica 
Procurement of the first of 
expected early in FY83. 



Image Processing Facility 

This project is intended to provide a utility to display 
and analyze digital images. The system will consist 
of a powerful 32-bit computer with a mixture of 
medium- and high-resolution displays. Also, the 
system will include a microdensitometer to allow 
precise digitization of x-rays, micrographs, and other 
images. The computer and peripherals have been 
purchased, and construction of the physical space to 
house the system is complete. Procurement of the 
first of the displays is underway, with delivery 
expected early in FY83. 

Bac/<ground arid Objectives: This project arose in 
response to a critically overcrowded situation that 
exists on the present DCRT Evans and Sutherland 
Graphics computer. As image processing 
applications at NIH have increased, the limited 
resources of that graphics system have been 
saturated. During FY80, CSL, in collaboration with 
present and potential users designed a new general- 
purpose computer facility to aid the acquisition, 
display, and analysis of images such as electron 



^ii^^nil2B^«^'fss^irs?rIi?^^^^^T^ 


J*^';^"^SS='"'''"' 


ZOl CT00fl97 


ni CSL 


October 1, 1981 to September 30, 1982 


Analytic Models of Computer System Performance 


PI: R. L. Kartino Electronics Engineer CSL 

OTHER: R. W. Newcomb Professor, Electrical 

Engineering Oepartment Univ 


DCRT 
of HD 


COOPEHArinC UNITG {H .ny) 

University of Maryland 


Computer Systems Laboratory 


Systems Oesign Section 


DCRT. NIH. Bethesda, HD 20205 


TDTAL-mUaS:^^^ | PflOFESEl^AL: |orHE«. 


CHECK *PPBOPHIAIE BOX(ES) 


This project involves the deyelopment of analytic models that 
used to evaluate the performance of computer systems. During 
past year, tools for modeling and analyzing computer systems u 
the graph theoretic rnodel called Timed Place-Transition (P-T) 
were developed. This included the introduction of four specia 
nodes, the determination of new methods for finding net invari 

structure. Using these results, a method was developed for 
evaluating computer system performance with Timed P-T Net mode 
This method was used to model and analyze the bus arbitration 
techniques that occur in digital systems. In addition, a stat 
variable P-T Net model of the interconnection of two or more 
microprocessors was developed. This model provides a framewor 
determining the avoidance of deadlock and the maintenance of 
throughput in multiple microprocessor systems. In FY83, this 
with Timed P-T Nets will be continued. 


an be 
ing 

for 



36 



micrographs, CAT scans, and radiographs. This 
facility will be available for use by the NIH 
community. 

Progress in FY82: The system will be based on a 32- 
bit, one megabyte computer, with a smaller 16-bit 
processor to handle image acquisition. A multidisplay 
raster scan frame buffer will provide several users 
concurrent access to the central processor. Images 
will be digitized when necessary through a 
microdensitometer or a vidicon camera. Hard copy 
will be provided by a camera system. 

All currently budgeted equipment and software have 
been ordered; some already has been received and 
the rest is scheduled for delivery in the next few 
months. The physical space has been completed 
with all power and cooling. 

Significance to Biomedical Researclr. Study of 
images obtained in the biomedical laboratory is 
proving more and more fruitful as technology is able 
to supply the proper tools at a reasonable cost. 
Biomedical scientists are employing image analysis 
for a wide variety of research goals, and the use of 
such techniques is expected to grow very rapidly in 
the near future. 

Proposed Course: As the equipment is delivered, it 
will be integrated into a system. The development of 
systems and applications software, as well as the 
transfer of existing image packages from the Evans 
and Sutherland system, will proceed. 

Analytic Models of Computer System 
Performance 

This project involves the development of analytic 
models that can be used to evaluate the 
performance of computer systems. During the past 
year, tools for modeling and analyzing computer 
systems using the graph theoretic model called 
Timed Place-Transition (P-T) Nets were developed. 
This included the introduction of four special nodes, 
the determination of new methods for finding net 
invariants, and the derivation of new relationships 
among net variables from net structure. Using these 
results, a method was developed for evaluating 
computer system performance with Timed P-T Net 
models. This method was used to model and 
analyze the bus arbitration techniques that occur in 
digital systems. In addition, a state variable P-T Net 
model of the interconnection of two or more 
microprocessors was developed. This model 
provides a framework for determining the avoidance 
of deadlock and the maintenance of throughput in 
multiple microprocessor systems. In FY83, this work 
with Timed P-T Nets will be continued. 



Background and Objectives: There are two major 
approaches to evaluating the performance of a 
computer system: simulation and analytic modeling. 
Simulation models have been a popular form of 
modeling for years but can be difficult and costly to 
construct, validate, and run. Recent advances in 
analytic modeling techniques, which can be used to 
model many aspects of a computer system, have 
provided new tools for evaluating computer system 
performance. 

There are two major types of analytic modeling 
techniques: graph theoretic and queueing theory 
models. A number of graph theoretic models have 
been found to be useful for the analytic modeling of 
computer systems. These include such graph 
models as Place-Transition (Petri) Nets, Parallel 
Program Schemata, Computation Graphs, and 
Marked Graphs. Queueing theory models have also 
been found to be useful for the modeling of 
computer systems because they can capture 
important features of actual systems, and algorithms 
that solve the equations of these models are 
available as queueing network evaluation packages. 

These analytic models provide useful tools when 
designing computer systems and deciding among 
alternative hardware or software configurations. In 
particular, with the integrated circuits that are 
currently available, it is technically and economically 
feasible to build systems consisting of many central 
processing units. Many processor and memory 
configurations also are possible now that memories 
can be placed in close proximity to the processors. 
Methods are needed for designing systems now 
possible with this new technology. Various structures 
must be considered and analytic methods for 
evaluating alternatives must be developed. 

Methods Employed: Timed Place-Transition (P-T) 
Nets are the modeling technique used to develop 
tools for evaluating computer system performance. 
The advantages of modeling with these nets are 
that: large and complex systems can be represented 
in a manner that is easy to understand due to the 
graphical and precise nature of these nets; the 
behavior of the modeled system can be analyzed 
using developed results of the P-T Net theory; and a 
system can be synthesized hierarchically with the 
ability to use different levels of abstraction and 
refinement. In addition, the usefulness of P-T Nets 
as models results from their ability to represent both 
concurrency and conflict in a system. Concurrency 
occurs when more than one event is taking place in 
a system at one time and conflict occurs when a 
decision must be made among alternatives. In order 
to evaluate the performance of computer systems 



37 



including sucii things as waiting times and 
ttiroughputs, the time parameter is added to the P-T 
Net model. 

Progress in FY82: During the past year, tools for 
modeling and analyzing computer systems using 
Timed P-T Nets were developed. This included the 
introduction of four special nodes, the determination 
of new methods for finding net invariants, and the 
derivation of new relationships among net variables 
from net structure. Using these results, a method 
was developed for evaluating computer system 
performance with Timed P-T Net Models. This 
method was used to model and analyze the bus 
arbitration techniques that occur in digital systems. 
From these models, important upper bounds on the 
average time a device waits for the bus and the 
average time the shared bus is not used were 
derived. This work demonstrates that the results that 
can be obtained concerning the performance of a 
computer system using Timed P-T Net models is 



dependent on the structure of the net that accurately 
portrays the system. 

A state variable P-T Net model of the 
interconnection of two or more microprocessors with 
input and output devices was developed. This model 
provides a useful framework for modeling and 
analyzing multiple microprocessor systems. For 
example, it can be used to determine the avoidance 
of deadlock and the maintenance of throughput in 
such systems. 

Proposed Course: The work on using Timed P-T Net 
models for evaluating computer system performance 
will be continued, including the derivation of more 
relationships among net variables based on the 
structure of these nets and the development of a 
Timed P-T Net model where time is associated with 
both the places and transitions of the net. Also, this 
graph theoretic model will be compared with 
queueing theory models of computer system 
performance. 



38. 



Laboratory of 
Applied Studies 



Eugene K. Harris, Chief 



Clinical Research and Patient 
Care 

Computer-aided analysis of electrocardiograms. 

J. Bailey, M. Horton (LAS); cardiologists and 
biomedical engineers in the U.S.A. and abroad. The 
purpose of this project is to evaluate the utility of 
leading connputer programs for ECG interpretation, 
and to search for optimal computer-based methods 
of extracting medically significant ECG patterns. A 
study of the components of variance in ECG 
parameters has been completed, using data from the 
Framingham Heart Project; a manuscript is being 
prepared. 

Computer systems for nuclear medicine. J. 

Bailey, M. Douglas, R. Burgess (LAS); H. Ostrow 
(CSL); M. Green, et al. (CC, Nuclear Medicine). This 
project involves development and aplication of 
computer systems to such diagnostic imaging 
activities as ECG-gated radionuclide ventriculography 
and dynamic scintigraphic studies of other organs 
(e.g., kidneys, lungs). A study of 79 radionuclide 
ventriculograms has revealed average signal-to-noise 
characteristics, optimum filtering, and optimum 
segmentation for detection of regional abnormalities; 
a series of manuscripts is being prepared. A study of 
segmental artery sterosis in canines using functional 
maps of renal scintigraphic data has been published. 

Computer-based studies of pulmonary 
pathophysiology and respiratory disease. J. 

Bailey, R. Burgess, M. Horton, E. Pottala (LAS); R. 
Crystal, A. Nienhuis (NHLBI); A. Jones (CC, Nuclear 
Medicine). These studies attempt to achieve better 
understanding of pulmonary pathophysiology through 
use of computer-based models of pulmonary gas 
exchange and respiratory mechanics, comparing 
predicted values with real patient data. A 
minicomputer-based system for analyzing gas 
exchange during exercise was purchased and 
installed. Development of programs to analyze mass 
spectrometer and flowmeter data and to control the 
treadmill and bicycle is nearly complete. 



Statistical research in clinical pathology. E. Harris, 
M. Horton, A. Albert (LAS); G. Shakarji, F. VanSant 
(DMB); clinical chemists and others in the U.S.A., 
Europe, and Japan. This research involves 
application of statistical theory to clinical laboratory 
tests, including serial studies of blood chemistries in 
health and disease.Multivariate subject-specific 
reference regions were shown by computer 
simulation studies and applications to real data to be 
substantially more specific against false positive 
results than corresponding univariate reference 
ranges. A collaborative study to explore relative 
sensitivities in various diseases has begun. A 
statistical method for deriving reference differences 
as criteria for evaluating the significance of observed 
changes has been extended and applied to a 
comprehensive data base of serial data from healthy 
subjects. A method for sequential assessment of risk 
in acute disease has been developed by combining 
discriminant function with analysis of response 
curves. The method has tested successfully in 
application to patients under intensive care following 
myocardial infarction. 

Computer-based studies in ultrasonography. M. 

Douglas, J. Bailey, E. Pottala (LAS); B. Maron 
(NHLBI). Ultrasonography allows noninvasive 
visualization of many organs without the hazard of 
ionizing radiation. This project involves development 
of minicomputer systems for image enhancement, 
pattern recognition, and three-dimensional 
reconstruction from ultrasound data sources, 
principally wide-angle phased array 
echocardiography. A principal difficulty in 
echocardiogram studies is separating gross cardiac 
motion from regional wall movement in a quantitative 
way. To resolve this problem, we have begun to 
study the use of reliable fiducial points, such as the 
papillary muscles or the intersection of mitral cusps. 
Further progress in this project awaits upgrading of 
the DeAnza system by the vendor. 



41 



Laboratory Investigation 

Mathematical modeling of biological processes. 

J. Fletcher (LAS); R. Schubert (Louisiana Tech. 
University). Scientists are developing and applying 
mathematical models in studies of substrate 
transport in the microcirculation, in diffusion 
processes in physiology, and in macromolecule- 
ligand binding equilibria. A new unified model has 
been developed for the microcirculation during a 
perfused organ experiment. Parametric studies of 
this model's properties are underway. A final 
summary report on mathematical models for 
equilibrium binding experiments was published and 
distributed. A new investigation of capillary 
hematocrit and oxyhemoglobin unloading effects in 
capillaries was begun. 

Mechanisms of active transport/biochemical 
kinetics. B. Bunow (LAS); A. Kaplan (NCI); D. 
Mikulecky (Medical College of Virginia); J. Kernevez 
(University of Tech., Compiegne, France). 
Experimental and mathematical studies of the energy 
mechanisms for active transport and of multistate 
biochemical kinetics in cells and membranes 
continue. Network modeling methods were applied 
through collaboration with NIH scientists to problems 
including cellular metabolism, neural networks, nerve 
conduction, and tissue oxygenation. A realistic 
mathematical model for isoelectric focusing has 
been derived and programmed for computer solution. 

Hybrid computing to analyze physiologic signals 
and construct simulation models. E. Pottala, B. 
Bunow (LAS); T. Colburn (NIMH); various NIH and 
FDA scientists. This project uses the LAS 
minicomputer system (MAC-16) for hardware 
simulation of physiologic functions and for analysis 
of analog signals (myogram, EEG, etc.). Two 
network simulation languages were implemented on 
the IBM System 370 and the VAX system (NIMH). 
This will facilitate model building and make 
simulation models more accessible for investigators 
on campus. 

Image processing in electron-loss spectroscopy. 

M. Douglas, (LAS); J. Costa (NIMH). This project 
involves the development and implementation of 
mathematical models and image enhancement 
techniques to analyze computer-acquired information 
from electron-loss and x-ray spectra indicating the 
location of extremely small quantities of important 
chemical elements and active protein molecules 
within cells. Activities in this area have been delayed 
while the vendor completes upgrading of the DeAnza 
system as originally planned. 



Computer Research and 
Development 

Mathematical and computational methods for 
nonlinear equations. R. Shrager, R. Hendler 
(NHLBI); A. Schechter (NIADDK). Work continues in 
the study of methods of fitting nonlinear models and 
mathematical methods of spectral analysis. An 
algorithm was developed for rapid solution of one 
nonlinear equation in one unknown. The method 
does not require derivatives and guarantees the 
answer to full machine precision. The algorithm has 
been installed in MLAB. A program to simulate the 
oxygen saturation of hemoglobin from the primary 
regulators pH, PC02, and 2,3-DPG is now under 
development. Also under development is a program 
to optimize the wavelength selection in the design of 
hemoglobin saturation analyzers. 

Numerical methods for the solution of 
mathematical models describing reaction- 
diffusion and other processes in biological 
systems. M. Bieterman, J. Fletcher, B. Bunow (LAS); 
I. Babuska (University of Maryland). Ongoing study, 
development, and implementation of efficient, flexible 
numerical methods for the solution of nonlinear 
ordinary and partial differential equations is involved 
in modeling dynamic physiological processes. 
Theoretical work was completed on the adaptive 
finite element method for solving systems of reaction 
diffusion equations. Program packages have been 
implemented on the IBM System 370 and are 
currently undergoing testing. Preliminary applications 
have been made to problems in nerve conduction, 
facilitated diffusion in tissues, developmental biology, 
and ecosystems. 



42 



Research Projects 

statistical Research in Clinical Pathology 

Univariate and multivariate time series models and 
discriminant techniques are being applied to various 
data bases consisting of short series of 
measurements of serum biochemistries in healthy 
subjects and patients with myocardial infarction. The 
purpose is to gain practical experience in the use of 
these statistical predictive techniques to detect 
changes and trends within individuals, taking into 
account biological variation and measurement error. 
The time scale of these series varies from daily to 
weekly, 6-month, and 12-month intervals between 
observations. Parallel computer-based simulation 
studies are also underway, particularly to estimate 
the relative sensitivities and specificities of 
multivariate and univariate forecasting methods. 
Mathematical investigations into the properties of a 
new stochastic model of linear change are 
continuing. 



m!j"tl"iSlsl«'i"?«?™'li!IISJ.°"!" 


lanUMMAL aEUARGH nOJCCT 


ZOt CT00007-14 US 


ram amm oj,,^,, , _ j,b 


to September 50, 1 y«^ 




statistical Res 


earch in Clinical Pathology 




i. ShakarJI 
M.R. Horton 
H. Slln 
a.2. l(lllia.a 


fleaearch Pelloa (Beljiiim) 
Sapv. Systems Analyst 
Compoter Systems Anslyat 
Clinical Pathology 
Inatltote for Health 
San Francisco, CA 
PL Hedical Service Sept. 
Osaka, Japan 


LAS WRT 

3JIB te«T 

LAS KRT 


None 


Laboratory of Applied ^tudiea 


SECTION 


1;CKT. HU. Betheada. NO 


2020i 




torn »»i„i.«i, inor 






CHia Amiamm M.lei) 
2 M •ma mjteis 

D(.i) noons u(.!) i»I!»vit«s 


a M BB><« Iiisois u l.l «iii 


.. 


SOauO or ,««« 1!00 ..«. ,r 1... 
techniquea are being app 


Z 


trli». k.j-ord,) 


onaiatlne of abort 
althy sub.)ects and 


r!m Mlchelistri.e'tn'h^ 




e*?e?hi 


(^uea to 


biolDtiical variation and 


'ii 


se statistical predi^tiv 


^'re.e"rer'ror!''Thru« 


iiiri: 


vitlsa and 


series varie. from doily 
between observations. P 
also underway, partioula 
specificities of .oltiva 
«atne«atloal investijatio 


to weekly, 6-Donth. and \2- 

na into the propertiea of a 
e continuing. 



Objectives: To investigate applications of statistical 
theory, particularly the use of variance components, 
discriminant analysis, and the theory of discrete and 
continuous time series, to the interpretation of serial 
clinical laboratory measurements in healthy subjects 
and patients with acute and chronic disease. 

Progress during FY82: The study of the relative 
specificities of univariate and multivariate reference 
regions, applied to subject-specific serial 
observations, has been completed and published. 
Results from computer simulations were confirmed 
by analysis of semiannual series of clinical chemistry 
measurements in 700 longterm participants in health 
maintenance programs in Tokyo and Osaka, Japan. 
Multivariate reference regions were shown to be 
much more specific than corresponding univariate 
regions, thus capable of avoiding many false alarms. 
A statistical comparison of regional data bases of 
clinical chemistries in healthy individuals has begun. 
This work is part of a general investigation into the 
transferability of clinical data among population 
groups and geographic areas. 

New research was undertaken on the development 
and application of a statistical method for estimating 
'reference changes,' that is, critical differences 
between successive measurements of a biochemical 
constituent in an individual. At the present time, such 
criteria are left to the judgement of the individual 
physician and are known to vary widely even among 
residents in the same hospital. The proposed 
method, allowing for serial correlation, has been 
tested on serial observations of calcium and alkaline 
phosphatase in healthy subjects. A paper describing 
the method and its uses is in preparation. 

The main research project under A. Albert's Fogarty 
International Research fellowship has been 
concerned with the development and testing of 
models to assess changing risk probabilities in 
patients with acute diseases. Several methods have 
been proposed and successfully applied to existing 
data bases of patients with acute myocardial 
infarction, and children treated in intensive care 
units. Besides this project. Dr. Albert has also 
undertaken studies on ways to improve laboratory 
data interpretation and made available to the NIH 
research community a program for multiple group 
logistic discrimination. 

Proposed Course: A study of the relative sensitivity 
of multivariate reference regions will be initiated in 
collaboration with Dr. Yasaka. This will entail 
incorporating into the data bank followup information 
on the clinical status of individuals examined. 



43 



Further development and applicaton of the statistical 
methodology proposed for calculation of reference 
changes will be undertaken for selected patient 
groups in collaboration with members of the Clinical 
Pathology Department at the University of Virginia 
Medical School. Continued joint efforts in this area 
and in the general application of stochastic time 
series to patient data will be initiated with Dr. Albert, 
Fogarty Research Fellow in this Laboratory, upon his 
return in 1982 to the University of Liege. 

The statistical comparison of regional data bases in 
Japan will continue. 

Publications: 

Albert, A.: Atypicality indices as reference values for laboratory data. Amer. 
J. Clin. Pathol. 76: 421-425, 1981. 

Albert, A.: Discriminant analysis based on multivariate response curves: an 
approach to dynamic prognosis. Statistics in Medicine (in press). 

Albert, A.: On the use and computation of likelihood ratios in Clinical 
Chemistry. Clin Cliem. 5: 1113-1119, 1982. 

Albert, A., Chapelle, J. P., Heusghem, C, Kulbertus, H.E., and Harris, E.K.: 
Evaluation of risk using serial laboratory data in acute myocardial in- 
farction. In Heusghem, C, Albert, A., and Benson, E.S. (Eds.): 
Advanced Interpolation of Clinical Laboratory Data. New York, H/larcel 
Dekker (in press). 

Harris, E.K.: Further applications of time series analysis to short series of 
biochemical measurements. In Grasbeck, R., and Alstrom, T. (Eds.): 
Reference Values in Laboratory Medicine. Chichester, U.K., John Wiley 
&Sons, 1981, pp. 167-176. 

Harris, E.K.: Regression, least squares, and correlation. In Seligson, D., 
M.D. (Ed.): Handbook of Clinical Chemistry. CRC Press (in press). 

Harris, E.K.: Use of statistical models to detect subject specific changes. In 
Yasaka, T. (Ed.): Proceedings of the International Conference on Auto- 
mated Multiphasic Health Testing & Sen/ices. Amsterdam, Excerpta 
Medica, 1981, pp. 35-44. 

Harris, E.K., Yasaka, T., Horton, M.R., and Shakarji, G.: Comparing Multivar- 
iate and Univariate subject-specific reference regions for blood con- 
stituents in healthy persons. Clinical Chemistry, 28: 422-426, 1982. 



Mathematical Models of Binding Equilibria 

The objective of this project is the study of 
mathematical models of ligand-receptor or ligand- 
macromolecule binding studies at equilibrium. The 
models are examined for mathematical as well as for 
conceptual validity and are studied to determine their 
suitability for fitting to experimentally obtained 
laboratory data. The appropriateness of various 
model fitting criteria are studied and general 
guidelines and computational algorithms are 
designed for computer-aided interactive model fitting. 

Progress in FY82: Numerous requests for copies of 
exportable computer algorithms were honored and a 
number of B/l/D consultations were provided. A 
summary report including collected results from 
fifteen years of research in this area was completed. 
It is now available for general distribution to the 
biomedical community. 

Proposed Course: Applications of existing and new 
methodology to data analysis will continue to be 
made as they are requested by collaborating 
laboratories. Computer programs, reprints, and 
reports continue to be provided to requesting 
consultees. Analytical development of new models 
and continued research in fitting methodology in this 
area will emphasize validation of experimental 
techniques, multi-receptor models, and 
conformational changes in macromolecules due to 
binding of ions. 

Publications and Abstracts: 

Fletcher, J E.: The Analysis of Equilbrium Binding Data by the Fitting of 
Models. July 1982. 



Mathematical Modeling of Substrate Transport in 
Physiological Environments 

Mathematical models of microcirculatory structure 
and function are developed from conceptual models 
into systems of coupled ordinary and/or partial 
differential equations. Methods of solution of these 
nonclassical formulations are developed and tested 
and satisfactory cost effective methods are used to 
explore the properties of these models. The results 
are interpreted in terms of microcirculatory 
physiology and are published in the scientific 
literature. 



44 



^iS'ss^^TiiXi^r 


,«i!iiiH?iiyt'fl 


::; 


CT NUMBER 

CT00005-1 2 LAS 


"''(i"toS"r'^ , 1961 to Septenber 30. 1962 






Kathenatical Models of Binding Equilibria 




NAMES. L»BOfi»TOflY AND IN5IIIUTE AfriLIATIWS, ANO TITLES Of MIINCIPaL INVEST 


GATORS ANO ALL OTHER 


PI: J.B. Pleteher Chief, Applied Nftlheraa'.ica S 


ection LA3 DCRT 


OTHEHS: H. Shrager Hathenatician, «i:j 


US DCHT 


P. Abramaon Visiting Scientist 
(U.W. Univeralty) 


NCI 


J. Dunn PhyalcBl Chemist 


NCI 


Laboratory of Carclnogenes 




„0„ 


LAfl/ BRANCH 




Laboratory of Applied Studies 




Applied Kathenatics Section 






DCKT, NIH, Betheada, HD 20205 




IOI.yAN,lARS, 1 PROFESSIONAL. lOtHCR. 


CHtCK APPROPRIATt eO«fES) 




J I.) HO-AN SUBJECTS a(fc) MUHAN tISSUtS Q (c) 


EITHER 


a(.l) MINORS a (.2) INUflVlt« 




SJMMAfll or wOfiK (200 .orUt or U»i - undcrlU* li.,-6nl») 




The Objective of this pr 


ject ia the study of nathem 


tical models of 


ll^and-receptor or lirtand-ai 

exponaientttlly obtained lab 
model fitting criteria are 


ratory data. The appropria 
tudied and general guidelin 


conceptual 

for fitting to 


computational algorlthma ar 
nodal fittioK. 


dealgned for computer-aide 











"-^■"'-"■^' 


""^"""^'So?K""^'''"' 


ZOl CT Of 014-04 US 


uctober I, l*i1 to September 


3U, 19a2 




Natheiaattcal Hodellne of 3u 
cinvironnenta 


atrate Transport in Physiological 




PI: J.B. flet'oher 

«. Bieteraan 
B. Oulinis 


Chief, Applied Mathematics LAS OCHT 

Aaaoc. Prof. Louisiana 
Dept. of Biomedical Tech Univ. 
Encineering 

Professor of Physiology Charlottes 
Univ. of Virginia 


.uu 


„on„ 


Laboratory of Applied Studlea 




DCRT. HIH. Betheada, HD 202 


05 


— 


TOTAL MANVEAHSi IMDFESSI 




C.«« APPROmiATE 80«{tS) 

J(.) HUMAN SUBJECTS 0( 

U(.ll MINORS a (.2) INIENVIEM 


) HUMAN TISSUES G («) NEITHER 




developed froia conceptual lao 
partial aifferential equatio 


microclrculator structure and function are 




dels into ayatems of oou 




tory physiolofc-y and 
la in an effort to 


The results are interpreted 
are published in the sclent 
One objective of thia i 
ti3»ue levol phenomena by n 
determine relatlonahipa bet* 
to physiologic challengea. 


e developed and tested e 
o explore the properties 



One objective of this project is to study whole organ 
and organ tissue level phenomena by means of 
mathematical models in an effort to determine 
relationships between variables that govern the 
organ response to physiologic challenges. 

Progress in FY82: The details of a new, intricate, 
mathematically correct, solution to the generalized 
Krogh cylinder model were completed. A second 
effort was begun to investigate the kinetics of red 
cell and free hemoglobin offloading during capillary 
transit. This work is still in the exploratory stages. A 
manuscript describing the first effort has been 
accepted for publication. Two preliminary reports on 
the findings from these studies were presented at 
international meetings. 

It is anticipated that the research course of this 
project will have the following stages. 

1. Reexamine the Krogh cylinder models in the 
literature and their adequacy for the representation 
of perfused organ microcirculation. 

2. Develop exact mathematical solutions for 
extended Krogh models that exhibit tissue axial 
diffusion and capillary axial diffusion for the steady 
state constant metabolic rate experiments with 
perfused organs. 

3. Develop or modify numerical algorithms that will 
compute substrate levels for nonconstant metabolic 
rates and other nonlinear effects. 

4. Develop algorithms for the direct comparison of 
distributed substrate level computations with 
experimentally obtained microelectrode 
measurements. 

5. Identify those critical ranges of parameters that 
control organ response to physiologic challenge. 
Publications and Abstracts: 

Fletcher. J.E.; Diffusional Transport Coupling in an Ideal Capillary Tissue 
structure. SI AM 30th Anniversary Meeting. Stanford University. 1982. 

Fletcher, J.E,. and Schubert, R.W,: On the Computation of Substrate Levels 
in Perfused Tissues. Mathemalical Biosciences (in press). 

Fletcher, J.E., and Schubert. R,W.: The Theoretical Prediction of Substrate 
Levels and Their Histograms in Cell Free Perfused Tissues. Proceed- 
ings of the International Meeting of the OTT Society. Detroit, Michigan, 
August 1981, Plenum Press (in press). 

Schubert, R.W., Fletcher, J.E., and Reneau, D.D.: A Simplified Model for 
Predicting Myocardial P02 Histograms. The First Southern Biomedical 
Engineering Conference. LSU Medical Center. Shreveporl, LA, 1982. 



45 



Analysis of Coupled Transport and Biochemical 
Kinetics 

This project investigates six fundamental problems in 
biology: (1) the role of dynamic patterns in 
embryology and evolution, (2) the kinetics of 
enzymes located in cell membranes, (3) the kinetics 
of enzymes derived from malignant and normal ceils 
in culture, (4) mathematical modeling of isoelectric 
focusing studies, (5) thermodynamics of bioenergetic 
mechanisms in mitochondria, and (6) development of 
a new paradigm for biological modeling based upon 
topological representation and use of network 
modeling languages. Simulation on digital computers, 
particularly with network modeling languages, 
numerical solution of differential equations, and 
nonlinear regression analysis are the main tools in 
these investigations. While these problems are 
diverse in their biological background, they ail share 
in a common basis of mathematical and physical 
content through the role played by conservation laws 
and the mathematical methods involved in their 
resolution. 

• Dynamic Patterns 

Progress in FY82: This project is now essentially 
complete. Two manuscripts are still in press. 

• The Kinetics of Enzymes in Membranes 
Background and Objectives: Studies of the 
mechanism of membrane transport and energy 
transduction by membranes are generally less 
conclusive than studies of the mechanisms of 
enzymes in solution. This uncertainty arises because 
it is difficult both to manipulate the environment of 
the interior of a biological membrane and to measure 
responses there. The objective of this project is to 
determine the extent to which the actual organization 
of membrane-associated processes can be correctly 
inferred from the application of models to the kinds 
of experimental measurements currently made. 

Significance for Biomedical Researchi: Studies of 
membrane-associated enzymes, such as those of 
mitochondria, for example, are made by measuring 
external concentration changes, from which one 
attempts to infer biochemical organization. This 
process is quite unreliable, as witnessed by the 
continuing controversy over the biochemical 
organization of bioenergetics. By our work, we wish 
to demonstrate that this unreliability is intrinsic, that it 
results from the nature of the methods used to study 
such systems, and that it is not to be remedied by 
performing yet another experiment of the kinds 
currently popular, no matter how ingenious. 



Progress in FY82: An experimental system involving 
an enzyme immobilized in a polymer film had been 
previously prepared and studied. A manuscript 
describing the behavior of this system has been 
completed and will soon be submitted. This project is 
now complete. 

• Kinetics of Lactate Dehydrogenase (LDH) from 
Normal and Malignant Hepatocytes 

Progress in FY82: The major collaborator in this 
project has been inactive this year because of 
complications associated with a move to a new 
laboratory in Frederick. 

Future Course: During FY83 the computer system 
will be installed in the new lab. Hopefully the 
experimental program can be reactivated. 

• Mathematical Modeling of Isoelectric Focusing 
Bacl<ground and Objectives: Isoelectric focusing is a 
very widely employed experimental tool in 
biochemistry. Its purpose is to separate and identify 
the protein components of mixtures, e.g., the 
extracted contents of cells and tissues. Despite the 
effectiveness of currently used procedures, the 
application of isoelectric focusing proceeds on an ad 
hoc basis, and it is quite likely that these current 
procedures are less than optimal. The goal of this 
research is to place isoelectric focusing on a firm 
footing of physical chemistry through the derivation 
and solution of predictive equations to describe the 
process and its dependence upon experimentally 
controllable parameters. 

Significance to Biomedical Research: Separation of 
proteins has an important role in the preparation of 
biologicals such as enzymes and vaccines for 
medical research and treatment. Improvement in the 
techniques of separation will lead to reduced 
expense for production of these materials and 
increased supplies and variety. Identificatin of 
proteins plays an essential role in diagnostic clinical 
chemistry as well as basic biological research. 
Improvement in identification procedures by reducing 
the sample size or time to resolution will permit more 
extensive studies on precisely defined preparations. 

Progress in FY82: The physical-chemical principles 
involved in isoelectric focusing have now been 
identified. The mathematical equations turn out to be 
highly coupled, nonlinear partial differential equations 



46 



with very unusual boundary conditions. A FORTRAN 
program providing for numerical solution of these 
equations has been developed but not completely 
debugged. Unfortunately, the computer time to run 
this program was too great to permit its completion 
with the scheduling algorithm in place on the DCRT 
computer during FY82. 

Future Course: The scheduling algorithm on the 
DCRT computer has been modified to permit 
efficient running of the FORTRAN program. The 
program will be completed to the point of verification 
of its functional correctness. Run time for production 
using the program will still be too excessive to permit 
extensive studies. A Cray-1 computer in France is to 
be used for preliminary production runs. 

• Thermodynamics of Bioenergetic Systems 
Background and Objectives: The mechanism by 
which the generally reduced components of nutrients 
are oxidized in mitochondria is still quite obscure, 
although most of the components of this pathway 
have been identified. The membrane association of 
the components of the pathway makes it difficult to 
proceed in the usual biochemical manner of 
molecular dissection and reconstitution. Most 
experimental studies are made on systems that are 
quite structurally complex. Nevertheless, interest 
focuses on the usual biochemical question: What is 
the sequence of molecular forms involved in the 
bioenergetic pathway? The role of ubiquinone in this 
pathway is the particular object of our interest in this 
project. 

Significance to Biomedical Research: An 
understanding of the mechanism of the central 
energy-yielding process of living organisms is clearly 
essential. Thermodynamic analyis has shown that 
the accepted explanation for oxidant-induced 
reduction of cytochrome b in the presence of 
antimycin cannot be correct. Thermodynamic 
analysis combined with new experimental techniques 
developed by one collaborator (R. Hendler), and 
analyzed with methods developed by another 
collaborator (R. Shrager), promises to provide a 
basis for a correct explanation of this phenomenon. 

Progress in FY82: During this year, the problem 
was identified, and the thermodynamic inconsistency 
of the current models for oxidant-induced reduction 
of cytochrome b was demonstrated. This work has 
been presented in seminar form, and a manuscript is 
in preparation. 

Future Course: During FY83, we will attempt to 
provide an explanation that is consistent with 
thermodynamics, and to design and perform 
experiments to test its correctness. These 
experiments will be based on the recently 



demonstrated ability to characterize small spectral 
shifts in cytochromes undergoing conformational 
change. 

• Network Modeling in Biology 
Background and Objectives: Mathematical modeling 
in biology is especially difficult because of the need 
to be familiar with both the biological basis of 
problems and the mathematical tools required for 
their solution. Network modeling, supplemented with 
effective modeling languages, largely obviates the 
need for extensive mathematical sophistication, and 
makes the process of modeling accessible to 
biologists lacking such skills. Topological modeling is 
particularly appropriate to biological problems 
because the objects of study generally satisfy 
conservation laws. In biological systems, the 
processes of flow, accumulation, and transformation 
are fundamental; these are likewise the basic 
operations in network modeling. 

Significance for Biomedical Research: The choice of 
a model for a biological process strongly conditions 
the design of experiments to confirm and extend it. 
By making the analysis of models sufficiently simple, 
we intend to permit an investigator the freedom to 
consider many models. From comparisons among 
the models using simulation, it should be possible to 
develop incisive experiments that permit scientifically 
valid, rather than arbitrary, selection among the 
models. The network languages nicely complement 
the MLAB system in permitting users to model 
phenomena that are too complex to be conveniently 
described in MLAB. 

Progress in FY82: Demonstration studies on network 
modeling have been conducted in several areas: (1) 
oxygen distribution and uptake in capillary beds, (2) 
nerve impulse transmission in an axon model, (3) 
electrical excitation of synaptic membranes, (4) 
electrotonic excitation of dendritic trees, and (5) 
metabolic networks in brain tumors. These studies 
have shown that functional and correct models can 
be constructed quickly and easily. By using available 
network modeling languages, the process of model 
building is focused on structures readily understood 
by biological investigators, rather than on verification 
of mathematical aspects. Several of these studies 
will prove to be directly useful in subsequent 
biological investigations. The network language, 
SPICE2, has been installed on a dedicated 
minicomputer in NINCDS. A course in SPICE2 
programming for neurophysiologists was delivered to 
representatives from several labs in this institute. 
Network models are now being regularly employed in 
their research. The demonstration studies described 
above will be described in manuscripts in 



47 



preparaiion, and the monograph on topological 
modeling is well underway. 

Future Course: During FY83, collaboration with 
groups using network modeling will continue. The 
DCRT course in network modeling will be repeated. 
The monograph should be completed this year. A 
major deficiency of available network languages is in 
the area of data fitting. A major effort will be made to 
facilitate the interconversion of models expressed in 
network terms into forms compatible with languages 
such as MLAB on which data fitting is convenient. 
DELIGHT, an extension of SPICE with facilities for 
optimization, device libraries, and more effective 
graphics, has been ordered and will be installed in 
FY83. 

Publications: 

Bunow, B.: All things flow and change-some thoughts on the role of 
reaction and transport in biology. J. Wash. Acad. Sci. (in press). 

Bunow, B.: Turing and the physico-chemical basis of biological patterns. In 
Prewitt, J. (Ed.): IEEE Turing Memorial. 1982 (in (in press). 

Bunow, B., and Mikulecky, D.G.: On the feasibility of using flux meas- 
urements to distinguish among active transport models. Polish Winter 
School of Membrane Transport, (in press). 

Bunow, B., and Mikulecky, D.C.: Where does metabolic energy couple into 
the active transport process? J. Theor. Biol, (in press). 



PROJEcft'uSeER'^D^MOI^s^thii apacJ) 


ZiisiiffifcT 


Z01 CTO0035-06 LAS 


October 1, 19S1 to September W, 1983 


Analysis of Coupled Transport and Biochemical Kinetics 




INVESTIGATORS AND ALL OTHER 


0tLr°; TLr^ev,. ?™'?a3sor 


LflS/DCKT 


A. Kaplan Hesearch 


DCCP^NCl' ''""''^ 


D. Mikulecky Professor 


Medical College 


•---- »■-"-"" 


LAS. DCRT 


„o„. 


LAB/BHANCH 


Applied Mathematics Section 








D (=) NEITHE« 


a{«l) M.HOHS D(«2) INTEBVIEWS 






tal problems in biology: 

;hermodynamics 
d (6) development of a 
n topological representation 


£U the role of dynamic patterns in embryology 

modeling of isoelectric focusing studies, (5) 
of bioenergetic raechanisraa in mitochondria, an 


coiaputers, particularly with network raodelinG languages numerical 

are the main tools in these inveatigations- While these problems are 
diverse in their biological background, they all share in a common 
basis of mathematical and physical content through the role played by 



Nonlinear Equations 

Methods are developed for solving nonlinear 
equations frequently encountered at NIH, usually in 
the context of constrained nonlinear least squares or 
in the solution to nonlinear differential equations. 
Related problems, such as asymptotic error analysis, 
and the efficient treatment of sparse systems, are 
also considered. 

Progress in FY82: 

• MLAB Projects 

Gary Knott (DCRT). The root-finder has been refined 
so that it performs more efficiently than the best 
previously published methods. A paper has been 
submitted to l\/latliematics of Computation describing 
the procedure and the test results. A separate 
integration routine, independent of the differential 
equation solver, has been installed, so that integral 
expressions may now appear in differential 
equations. Some problems with the curve-fitter, 
reported by MLAB users, have been traced to poor 
scaling of the problem (e.g., unfortunate 
combinations of large and small effects of changing 
this or that parameter). A rescaling procedure is 
being designed to overcome the difficulty. 

• Equilibrium Studies of Magnesium Phosphate 
Lev Jacobson (DCRT). Improved range of 
experimental data has enabled us to produce a 
simple yet adequate model of Mg-H-HP04 binding 
with reliable equilibrium constants. A manuscript 
detailing this work has been submitted for 
publication. Our contribution described how to 
propose alternate physical models, express them in 
reliably computable form, and suggest experiments 
for model validation. 

• Hemoglobin Projects 

Robert Berger (NHLBI); Robert Winslow (CDC 
Atlanta); Luige Rossi-Bernardi, Giulio Dossi, Michele 
Samaja, and Massimo Luzzana (Ospedale San 
Raffaele, Milan, Italy). 

1 . Two possibly conflicting criteria govern the 
design of a hemoglobin analyzer (design being the 
selection of wavelengths of light to best detect 
various binding states of Hb). The first, the product 
of variances of estimated fractions of the various 
states, is designed to reduce variance from random 
errors in the data. The second criterion, norm of the 
sensitivity matrix, is designed to minimize the effect 
of instrumental drift (aging, misalignment, etc.), which 
is very important because it cannot be 'averaged out' 
by repeated measurements. Procedures are now 
being tested that minimize either criterion or a 
weighted sum of both. 



48 



2. Simulation of 02 Saturation of Hb involves four 
governing quantities (02 pressure, C02 pressure, pH, 
and 2,3-DPG concentration). An efficient model 
expression is required incorporating these quantities 
with a manageable number of parameters so that 
various types of Hb can be modeled. We are in the 
midst of designing both the validating experiments 
and the model. 

• Cytochromes in Mitochondria 

Richard W. Hendler (NHLBI); O.H. Setty (Visiting 
Associate); K.V. Subba Reddy (Visiting Fellow); Barry 
Bunow (AMS, DCRT). 

1. An article on analysis of titration data by Single 
Value Decomposition techniques has just been 
published in Analytical Biochemistry. 

2. A paper on this topic is being presented to a 
mathematical conference at the SIAM 30th 
Anniversary Meeting. 

3. Work on the midpoint potential of E. coli 
cytochromes is now being extended to beef heart 
mitochondria and the other mammalian cells. By 
using a new rapid-scan spectrophotometer, we can 
now do kinetic studies as well as more reliable 
equilibrium studies. 

4. A suspension of cells is given a pulse of 02. 
The uptake of 02 and the production of H is 
monitored by computer, which must smooth the data 
and compensate for 02 and H electrode delay times, 
in addition to several other monitoring functions. The 
program is running and producing good results. 

5. In collaboration with B. Bunow, we are 
attempting to explain the oxidant-induced reduction 
of one of the b-cytochromes by developing a 
thermodynamically consistent model that also fits the 
observations. 

• Phytic Acid Titration 

William Evans (Agr. Res., New Orleans). An article 
on the titration of twelve H binding sites of phytic 
acid has appeared in J. Am. Oil Chem. Soc. 

• SVD Kinetic Studies 

Ray Tate (DCRT); Jim Osborne (NHLBI); Randy 
Kincaid (NHLBI). SVD is useful in resolving the 
spectra of individually titrating species in a mixture 
because the mathematical form of the titration is 
known. Likewise, the form of kinetic relaxation may 
also be known (e.g., sum of exponentials). 
Therefore, complete spectra gathered as a function 
of time may also be analysed by SVD. An initial set 
of data involving the association of Calmodulin and 
Ca using fluoresence and CD spectra is being 
prepared. 



Proposed Course: The curve-fitting scaling 
procedures need more tests. Work continues on 
every aspect of the cytochrome problem. Some 
actual designs of Hb analyzers should be produced 
this year. The Hb-02 saturation model should be 
ready for use. 

Publications: 

Evans. W J . McCourlney, E.J.. and Shrager, R.I.: Titration Studies ot Phytic 
Acid. J. Am. Oil Chemists' Soc. 59: 189-191, 1982 

Shrager, R.I., and Hendler, R.W,: Titration of Individual Components in a 
Mixture with Resolution of Difference Spectra, pK's, and Resox Transi- 
tions. Anal. Chem. 544: 1147-1152, 1982. 



"""'■" '"""™'"'-^^' 


.«z;^iii«iil 


MOJiei Niun 
ZOt CTOOOKM)^ US 




nonlinear Bquatlona 


MUll, LASORATOin UO INJIITUIt tfFlirnTieHI, MD TITtlt V nlHCIPM. IHVtSTICaTOItt UQ UL OIHER 

ncrettiOMU. tuittrntx. £iMtw» oh the ntojcci 

PI: R.I. ShraBor Mathomatlclan LAS DCRT 

OTHERS: G,D. Knott Coapiter Sp.cl.Uat L3M DCRT 

J.E. Fletcher Rooenrch Nathonsticlnn US DCRT 


LB, HHLBI 


Laboratory of Applied Studies 


Applied Hatheiutics Section 


DCBT, NIH, Bethesda. HD 20205 


I.ML .«.««,. |««ESS.«.t^ |oiH«. 


CHtCi AW.0M1.TE BO.{ES) 

Ul.) Hm.«K SUBJECTS a (6) HUllW IllSUtS a(«)M'IH£« 

Q(..) IXNORS D(.I) tNIERVIWS 


SUNHlflir Of «0«« (200 .»rt» or l.» - wvd.rl,". l«^ord.) 

Methods are developed for solving nonlinear equattona frequently 


leoaC aijuarea or in the aolu 


tion to nonlinear differential equations. 1 


Related problems, such as asymptotic error analysie, and the efficient 



49 



Numerica! ApproximatiOuj Techniques for the 
Solution of Reaction-Diffusion Systems in 
Biology 

A novel numerical method of lines, used to 
approximately solve partial differential equations 
governing models of reaction-diffusion systems in 
biology, has been developed and analyzed. The 
somewhat general program FEMOL 1 , which 
implements the computational procedure, includes 
various user-oriented adaptive features. These 
features include automatic space and time mesh 
selections, which are made by the computer and are 
appropriate for the solution of any particular problem. 

Progress in FY82: Various goals were achieved this 
past year in the investigation of numerical 
approximation techniques for the solution of 
reaction-diffusion systems in biology. The most 
significant of these fall into three categories: 

• Software Development 
A somewhat general and easy-to-use method of 
lines program, FEM0L1, was developed, 
implemented, and tested. FEM0L1 is presently 
applicable to systems of two nonlinearly coupled 
time-dependent reaction-diffusion equations in one 
space dimension. This program was designed 
primarily as a tool for use by laboratory researchers 
and other NIH scientists in parametric studies. Many 



^si;^^rre^g'fss^rs:ru?^%",T^ 


>""T'Hii"lS, 


-"--"- "= 




Numerical Approximation Techniques for the Solution of Reaction-Diffusion 
Systems in Biology 


OTHER: J-li. t'letcher Chief, AMS LAS DCRT 
B. BunoH Bioraathematicifln, AHS LAS DCRT 


„0„, 


L*B/BBAriCH 

Laboratory of Applied Studies 


Applied Mathematics Section 


DCRT, NIH, Bethesda, HD 20205 


TOTAL jmEAfiS= | PH0FESS10NA^L= jOIHEfi. 


°'-'™"'"™' 0... »»..». o.c.».™. 






A novel numerical method of lines used to aoo 




differential equations gove 


nine models of reaction- 




in biology, has been develo 
FEHOL 1 , which iraoleraenta t 


ed and analyzed. The somewhot general program 
e computational procedure, includes various 








of any particular prob 





Standard user-oriented 'black box' features available 
in much commercial software, such as automatic 
estimation and control of time discretization errors, 
are contained in the program. FEM0L1 also includes 
a posteriori estimates of space discretization errors 
and a novel adaptive procedure, in which space 
meshes are modified during a problem's solution by 
the computer in order to control the space 
discretization errors. 

• Application to Biological Models 

FEM0L1 was used in computational experiments to 
evaluate the effects of various numerical 
approximation parameters in the method of lines 
solution of equations governing models of the 
microcirculation, population ecology, and electrical 
impulse conduction in nerves. Such a study is 
important, in that the roles of biological parameters 
in models are often obscured by the effects of the 
specific approximation parameters used in 
computations. These experimental results form the 
basis of a future publication. 

• Development of Mathematical Theory 
Many new mathematical results were obtained 
concerning the effectiveness of the a posteriori 
estimators used in FEM0L1. This theory not only 
supports the computational procedures used, but 
also prescribes reliable and efficient means by which 
space discretization errors should be estimated and 
controlled in more general problems arising in 
practice and (as seen in the computational 
experiments) in cases where underlying 
mathematical assumptions break down. The 
contents of these results were included in two 
publications. 

Proposed Course: Future investigations of numerical 
approximation techniques used in the study of 
biomathematical models will comprise (1) the 
extension of the capabilities of the method of lines 
program FEM0L1 , and (2) the modification and 
implementation of other techniques and software 
that were developed previously in LAS and 
elsewhere for the solution of specific problems 
encountered by NIH researchers. 

Publications: 

Bieterman, M., and Babuska, I.: The Finite Element Method for Parabolic 

Equations. I. A Posteriori Error Estimation. Numerische Mathematik (in 

press). 
Bieterman, M., and Babuska, I.: The Finite Element Method for Parabolic 

Equations. II. A Posteriori Error Estimation and Adaptive Approach. 

Numerische Mathematik. (in press). 



50 



Monitoring of the CNS in Criticaly III Patients 

This new project is a joint effort between the 
Laboratory of Applied Studies and the Department of 
Critical Care Medicine to design, build, and 
implennent a highly clinically oriented, distributed- 
processing, microcomputer-based system for 
analysis and display of scalp-recorded neuroelectric 
signals. As a part of the total noninvasive monitoring 
effort, this tool will then be used to investigate the 
degree of dysfunction in neurologically impaired 
patients, correlate the indices developed with other 
measures of cerebral function, and evaluate the 
effectiveness of various therapeutic interventions. 

Background and Objectives: In the critically ill 
medical patient with multiple organ dysfunction, 
impaired brain function frequently coincides with 
deterioration of other major systems. However, the 
degree of damage and capacity for restoration of the 
brain does not necessarily parallel that of the rest of 
the body. In addition, assessment of the central 
nervous system is hampered by limitations imposed 
by procedures (e.g., endotracheal intubation) and/or 
drugs (e.g., Pavulon). 



sis^^riisseig'ts^iTS?:*!:?: jw 


"''^'"^■rojS'"-'''"' 


ZOl CT0009a-0I LAS 


October 1, 1981 to September 30. 1982 


Monitoring of the CNS in Critically III Patients 


noFEUioMu. PotsONHa £N«ueo on ihe project 

PI: R.C. Burgeaa Senior Staff I'eUow LAS DCRT 

C. Natanaon Senior Investisator CC CC 
N.K. Korton Computer Syatens Analyst UiS DCRT 
E.i(. Pottala Electronics Enfiirieer US WRT 
J.J. Bailey Chief, MA.S LAS DCRT 


Crlticftl Care Medicine Departnent. CC 


LiB/eNtNCK 

Laboratory of Applied Studies 


Hedical Applications Section 


'"^^jaf.'TPlh^'fe^^CheaOa. HD 20205 


IOI.L-WUWS. pOfESS.OH.L. OTHER. 


CHECK *PPRO«»l«IE BO>{ES] 

:(.) HIMU SUBJECTS P(b) HWU« I.UUtS G (ej WITHER 

a(»0 "IMORS n(*i) INIEBvrC-S 


tmUM Of NOftK (200 -erl. or 1... . „«lwli-. Uy.ordt) 

This neif project is a joint effort between the Laboratory of Applied 
studies and the Departoent of Critical Care Kedicine to design, b'jild. 
and inplewent a highly clinically oriented, diatributed-proceaalng, 
oicrocomputer-baaed system for analyaia and diaolav of acelp-recorded 


effort, this tool will then 
developed with other meaaur 


part of the total noni 
be used to investigate 
y impaired patienta. co 
3 of cerebral function. 


tBi[:Ci 




rapeutic interventions. 





The most frequently employed tool for evaluation of 
the central nervous system, the neurological exam, 
suffers from its discontinuous and subjective nature 
and is highly dependent upon the examiner's skill. 
Many investigators have attempted to apply various 
diagnostic and monitoring techniques to the problem 
of assessing the neurological status of the critical 
care patient. 

Past efforts have been hindered by equipment 
artifacts and have required a high degree of skill on 
the part of the technician and the interpreter. This 
project will overcome the great difficulty of 
undertaking detailed microvolt-level signal analysis 
through the innovative use of state-of-the-art 
technology and multidisciplinary approach. 

Progress during FY82: A detailed literature search 
and exhaustive commercial product evaluation has 
been carried out. Dr. Burgess attended the state-of- 
the-art Evoked Potential course at Duke University. 
System requirements have been determined, 
instrumentation design has been completed, and 
equipment purchases have been initiated. 

Significance: Apart from the technological task of 
demonstrating a system that will reliably and semi- 
automatically obtain and process data from MICU 
patients, we seek answers to the following 
questions: 

1 . Which electrophysiological parameters can be 
used to best follow the functional neurologic status 
of the patients? 

2. What is the optimal protocol for obtaining data 
in order to balance recording requirements and 
nursing care needs? 

3. How can the parameters be best combined into 
a meaningful profile and be best displayed to provide 
comprehensive yet easy-to-assimilate clinical 
information? 

4. How does the information offered by this 
system compare to other neurodiagnostic 
techniques? 

5. How does this system improve our overall care 
of the patient and our understanding of the 
pathophysiologic dynamics? 



51 



Proposed Course: The completed system will 
include: (1) a precision analog front-end with high 
noise immunity for detection, amplification, and 
filtering of the spontaneous and evoked EEG activity; 
(2) stimulators for delivery of visual, auditory, and 
somatosensory stimuli; (3) a central processor with 
intelligent peripherals for data acquisition, 
manipulation, calculation, and storage; and (4) a 
display capable of high resolution graphics and 
printout for presentation of current and past data, 
trends, and interpretive imaging. 

Extensive development of both hardware and 
software will be carried out during the coming year. 
Instrumentation to be built includes a CPU interfaced 
preamplifier/filter and a computer control for the 
stimulators. A hardware as well as a software 
interface for an array processor will have to be 
developed. Programs will enable simultaneous 
stimulation, data acquisition, storage, and display. 
Operating in the realtime environment and employing 
an array processor for rapid computation will require 
complex and flexible data handling routines. 

Publications: 

Burgess, R.C.; An instrument to add evoked potential capability to the 
standard electroencephalograph. EEG and Clin. Neurophysiol. 53: 33, 
1981. 



™"'""''^'""™"™'" 


imtb«ur*l"esearcm piioject 


Z01 CT0OO54-06 LAS 


Pen<OD COVERED 


TITLE Of PHOJEtl (BO ch.r.cl.rr or less) ' 

Computer-baaed Studies in Pulmonary Pathophyaiology 


R.G. Crystal Chief 

M.R. Horton Computer Syateraa Analyst 
E.W. Pottala Electronics Bngineer 
J.J. Bailey Chief, HAS 


CHB NHLBI 
LAS DCRT 


Clinical Hematology, Pulmonary Branch, NHLBI, Nuclear He 


icine Dept., CC 


Laboratory of Applied Studies 


Medical Applications Section 


DCRT, HIH, Bethesda. HD 2020!j 


lUIHL MnrtAHK. jPHOFESSIOHAL: lOIMEHi 


CHECK APPROPRIATE eOX(ES) 

3 («) HUMAN SUBJECTS D (b) HUMAH TISSUES g (c) NEITHER 

D{-1) MINORS a{.?) .NTERVIEWS 


Medicine Department, CC and the Clinical Hematology 
NKLBI — ia directed toward a deeper understanding of 


pulm 


^Imonary Branches, 


pathophysiology through the 


construction of computer 




d models of 








ugh comparisons of 


model predictions with real patient data. 



Computer-based Studies in Pulmonary 
Pathiophiysiology and Respiratory Disease 

This project-through a collaborative effort of LAS 
with the Nuclear Medicine Department, CC and the 
Clinical Hematology and Pulmonary Branches, 
NHLBI--is directed toward a deeper understanding of 
pulmonary pathophysiology through the construction 
of computer-based models of pulmonary gas 
exchange and respiratory mechanics and through 
comparisons of model predictions with real patient 
data. 

Progress in FY82: The system designed, specified, 
and purchased during FY81 for breath-by-breath 
analysis of pulmonary gas exchange has been 
assembled. Interfaces for the bicycle ergometer and 
treadmill controllers to enable online computer 
control and data acquisition were designed and built. 
Necessary support equipment, such as an expired 
gas cooler and an analog tape deck controller, were 
fabricated. LAS has been developing software to 
effectuate realtime exercise testing and has been 
performing calibration of the individual instruments 
as well as the integrated system. 

Proposed Course: Preliminary testing of the new 
system on volunteer subjects is just beginning and 
substantial additional data will be collected for both 
system validation and establishment of a baseline 
data base. Numerous studies of the gas transport 
system are possible. An experimental protocol to 
evaluate the use of Hydralazine in sickle cell disease 
in collaboration with the NHLBI Hematology Branch 
has been approved. Studies of cardiorespiratory 
ability in both patients and athletes are in the 
planning stages in collaboration with the NHLBI 
Pulmonary Branch. 

Publications: None. 



52 



Investigation of Hybrid Computing for the 
Construction of Simulation Models and for the 
Analysis of Physiologic Signals 

This project was undertaken to develop physiologic 
simulation models using hybrid computing and also 
to use hybrid computing techniques to analyze 
physiologic signals such as electrocardiogram, 
electroencephalogram, and electromyogram. 

Progress During FY82: The study of neural signals in 
the rat (hippocampus and sensory cortex) has been 
deferred by that investigator and only recently re- 
initiated. 

The Division of Cardio-Renal Drug Products, FDA, is 
investigating the early detection of cardiac toxicity 
resulting from drug therapy. Rat electrocardiograms 
are being used to determine the sensitivity of 
detection. The data has been redigitized and 
reformatted so that it can be analyzed by an 
automatic vector cardiographic program acquired by 
LAS several years ago. 

Two simulation languages, NET2 and SPICE2, have 
been implemented on NIH computer systems. Both 
languages are available on the IBM System 370, and 
SPICE2 is also available on the VAX computer 



; \T.'!f 



201 CT00004-1 ' 



, 1-4M1 t.n ?;tfptffinh..T- '' 



PI: E.W. Pottala 

OTHERS: J.J. Bailey 



evelop phyaiologK 



imputing technliue3 



belonging to the Technical Development Section, 
NIMH. The single retinal cone cell model has been 
used to verify the above systems; a manuscript is in 
preparation. 

Proposed Course: The MAC-16 system will have 
continued use for ECG processing from the 
Framingham Heart Study (see project report on 
electrocardiography). 

Analysis of FDA data on rat electrocardiograms will 
continue in FY83. 

Network simulation studies will continue with the 
development and implementation of additional 
physiologic models using SPICE2 and NET2. 

Publications and Abstracts: None. 

Computer Systems for Nuclear Medicine 

This project involves computer-based mathematical 
analysis, pattern recognition, and image processing 
in support of diagnostic activities in the Nuclear 
Medicine Department of the Clinical Center and in 
collaborating Institutes. Applications include 
computerized ECG-gated radionuclide 
angiocardiography and myocardial perfusion 
scintigraphy, renal dynamics, and pulmonary 
ventilation-perfusion relationships. 

Progress during FY82: 

• Renal Scintigraphy 

The results of preliminary work carried out in FY76- 
FY80 have been published in several journals. These 
studies demonstrated significant enhancement of 
routine renography using functional mapping 
techniques. After a technical development and 
software upgrading effort was accomplished in FY81, 
definitive evaluation of the technique was carried out 
in nine dogs. Contrast angiography, routine 
radionuclide renography, and functional renal 
mapping were performed before, one month after, 
and 12 months after unilateral renal artery ligation. 
Histopathological study of kidneys removed from the 
sacrificed dogs is now underway. A manuscript 
correlating the pathological findings with the 
functional maps is in preparation. 

• Cardiac Scintigraphy 

In collaboration with Nuclear Medicine and the 
Cardiology Branch, LAS has begun investigation of 
several parameters reflecting mobility of the heart 
wall including ejection fraction, regional emptying 
time, phase (of first Fourier harmonic), and maximum 
ejection rate. Programs have been written to 
compute these parameters globally or for any 
sectors of the heart image. The test data base 



53 



includes rest and exercise studies on 40 normal 
volunteers, 24 patients with coronary disease and 
known resting apical abnormalities (hypokinesis, 
akinesis, or dyskinesis), and 15 patients with 
cardiomyopathy. 

• Image Processing 
Signal-to-noise (S/N) ratio and Fourier harmonic 
content of global and regional time-activity curves 
(TACs) were thoroughly studied in this data base. 
These studies showed that the physiological signal is 
largely contained in four or fewer harmonics and that 
higher harmonics probably represent noise relating 
to the counting statistics of the TAG. They further 
showed that smaller regions have relatively higher 
noise and that regions smaller than 1/4 of the 
ventricular region-of-interest produce higher 
proportions of TACs that cannot be distinguished 
from background TACs in terms of S/N ratio or 
harmonic content. Hence, regional parameters 
extracted from such TACs are unreliable. Finally, 
these studies showed that global ejection fraction is 
a good classifier of normal and abnormal cases but 
regional ejection fraction does not appear to be 
additionally helpful in separating apical abnormalities 
(related to coronary artery disease) from diffuse 
abnormalities (related to cardiomyopathy). However, 
regional emptying time (time to the minimum of the 
TAC) does appear to separate apical from diffuse 
abnormalities. A series of manuscripts describing 
these studies is in preparation. 



SKITHSONIWI SCIENCE INFORMATION EXCHANGE 


■ ""^'^BLrH|TT*^^tE^'lc^'^^ 


:„: 


T NUMBER 

CT00005-I 1 LAS 


October 1. 1981 to September 50. 1982 | 


C»p„.. 


S.,te„ for »„=!«.. M.dicin, 




OTHEKS; 


S.L. Bacharach Physicist 

J.J. Bailey Chief, Hed. Appl. Sec. 
R.C. Burgeaa Senior Staff Fellow 
K.A. Douglas Comp. Syst. Analyst 
P.P. van Rijk Visiting Scientist 
H.V. Green Ch. Appl. Physics Sec. 
A.E. Jones Chief, Diagnostic Imaging 
H.C. Oatrow Engineer 


CB NHLBI 
LAS DCRT 
LAS DCRT 
LAS DCRT 


llZZ 


TS (if .",) 

Medicine Department, CC, NIH 
r Syatema Laboratory, DCRT, HIH 




LAB/ BRANCH 


□ry of Applied Studies 




Medical 


Applications Section 




DCHT. NIH. Betheada, HD 20205 


^.1 1 3.0 1 0.1 


:(.) HUMAN 
a(..) HINORS 


Alt BOK(ES) 

U8JECIS D(b) HUMAII TISSUES a(c)N 
D{.i) INIEHVIE-S 


,™. 


SUHMAHV OF UORK (200 .ordg or Icsi - un 

This project involves 
pattern recognition, and i 

and collaboratine Instttut 
KCC-Kated radionuclide an« 
aclotigraphy, renal dynani 


computer-tl'sed mathematical 


analyaia. 


nage processing in support diagnostic 

ocardiosraphv and myocardial perfusion 
a, and pulmonary vantilation-nerfusion 







Proposed Course: 

• Renal Scintigraphy 

Further evaluation and refinement of the functional 
mapping technique is planned utilizing alternative 
renal lesions in canine subjects. Limitations as well 
as advantages of functional mapping in 
pathophysiology of various etiologies will be 
explored. 

• Cardiac Scintigraphy 

A statistical analysis of the data base will be 
pursued, one possible outcome of which might be a 
discriminant function with various parameters to 
achieve optimal separation of normals from 
abnormals. Another interesting study will involve 
those patients with myopathy secondary to 
adriamycin therapy, using each patient before 
therapy as his own control. Other patients who have 
normal contractility at rest but abnormalities upon 
exercise form an additional interesting data base. 

• Image Processing 

When the expanded DeAnza system together with 
magnetic tape and disk drives are operational, it will 
facilitate study of paired myocardial (Thallium) and 
blood pool image sequences. Refined edge 
detection, assessment of wall motion abnormalities 
and perfusion, and more accurate determination of 
volumes are planned. 

A model to demonstrate the effect of known 
amounts of additive noise on the detectability of 
regional wall motion abnormalities is planned. 

Publications: None. 

Computer-Aided Analysis of Electrocardiograms 

These studies, continuing since 1 970, have been 
directed toward the evaluation of accuracy, clinical 
utility, and cost effectiveness of various computer 
systems for analysis of routine electrocardiograms 



54 



(ECG's). Further studies will involve new methods of 
feature extraction and design of criteria by computer 
techniques and their use in epidemiological studies. 

Progress during FY82: A set of ECGs was recorded 
on four different occasions on each of four staff 
members of the Framingham Heart Project. One 
ECG from each member was digitized at 1000 
samples/sec and the others at 250 samples/sec. 
From these data the analytic variation, within person 
variation, and group biological variation were 
estimated and compared with similar studies done by 
Simonson in 1949. A manuscript is in preparation. 

Proposed Course: Georgetown University Medical 
Center has acquired 12 simultaneous lead ECGs on 
a group of patients with documented cardiac 
disease. GUMC has proposed a collaborative project 
with LAS to evaluate three of the better-known 
programs using this data base. 

Meanwhile, LAS continues to study the 
epidemiologic significance of the routine ECG in 
collaboration with the investigators of the 
Framingham Heart Study. The ECG correlates of 
such heart diseases as coronary disease, mitral 
prolapse, and asymmetric septal hypertrophy in a 
free-living population are of particular interest. 

Publications and Abstracts: 

Bailey J.J.. Berson. AS., Jackson, L.K., Mllliken, J. A.. Stevens, J.M., Tolan, 
G.O., and Wolf, H.K.: Evaluation Mettiodologies for ECG diagnostic 
systems. In Bonner, R.E., and Pryor, T.A. (Eds): Computerized Inter- 
pretation of the ECG VI. New York, Engineering Foundation, 1981 (in 
press). 

Macfarlane, P.W., Ctien, C.Y.. and Bailey, J. J.: A comparison of point 
scoring tecfiniques for the diagnosis of LVH. In Macfarlane, P.W. (Ed.): 
New Frontiers in Electrocardiology. New York, Jofin Wiley & Sons, 
1981 (in press). 



Computer-based studies in ultrasonography 

This project involves collaboration between LAS and 
the Cardiology Branch, NHLBI. It is directed toward 
computer-based processing for image enhancement, 
pattern recognition, and three-dimensional 
reconstruction from ultrasound data. Wide-angle, 
phased array echocardiography is the principal 
source of data. 

Progress in FY81: Resolution of hardware and 
software defects in the upgraded DeAnza image 
processing system has forced deferral of further 
work on this project (see project on Computer Based 
Analysis and Image Processing in Electron 
Microscopy). 

Publications and Abstracts: None. 



SoJt6l"iiMw'[3"iOT'2?ll*(iit Jm"T' 


IITUMIUL NEtCUGH flOJECT 


201 CT0O0O2-12 LAS 


October 1, 1^1 to S«pt«fflber ?0, t9ti2. 


Coaputer- Aided Analysis of Bl«ctrooardl08»aui 


HMt£S, L«BOR*TOnY UO INSTITUTE iFflLUTIOHl, AM TIILtS Cf POIHCIPAL IKVC&TISATORS MO HA. OTHtH 

PI: J.J. Bslley Chief, HAS US DCRT 

QTHEBS: E.K. HorrlB ChlBf LAS DCfiT 
H.R. Horlon Conputer S/stems Analyst US DCRT 
D. Savage PrsBtnghBO Heart Study NHLBI 






Medical ApplicationB Section 






CHECK (PMOfftMIt BO«(tSj 

D (.)«»..« suijccls D(i.)«u.»iissut. DW«iiHS« 

Dl-O CilNOS D(.») KIMVUK 


~££lFi?;!rJj 


Yne'eS"'l970, have been directed toward the 




or snalysia of routine electrocardiORrama 


epidemiological studies. 


ODputer t«chnlque9 and their use In 



^oj£Cl"Ni«EH'fo"lOT'2!If*lhhI J"?J!!'^' 


IHTIUMU. aEUAMH mJECT 


ZOl 


CTOO045-O4 LAS 




Computer- baaed studies in ultrasonography 


KUttS, L»BORHORY m INSTITUTE tfflLUTIWS, Mt> IITltS Of PRINCIPM. INVfSTI 

J.J, Bailey 


LAS DCfiT 
CB SHLBI 


CDdPERMING UNITS {il .«,} 

Cardlolo^ Branch, NKLBI 


L*B/BS»NeH 

Uboratory of Applied Studies 




DCRT. HIH. Bethesda, KD 20205 


U.4 1 0.2 1 0.2 


C«l« .rPROPRUU BOl{tl) 

K.J Hu-w suBjeeis d (0 w"M nssuEs u (0 «'"'» 

a(.l) K.NOfiS DM) IhlERVU-S 


SUHKARl OF .^ORK (200 .on]» «r l<i> - undarlln* t,.,.ORl>) 

This project involves collaboration of LAS, with the 
Branch, NHLBI. It is directed toward computet^besed pro 
iBawe enhancetnent, pattern recognition, and three-dimens 
reconstruction fron ultrasound data. The principal sour 
Mide-angle, phased array echocardiography. 


Cardiology 
ceaainfi for 



55 



Computer Based Analysis and Image Processing 
in Electron Microscopy and X-ray and Electron- 
Loss Spectroscopy 

This project involves collaboration of LAS with 
several NIH Institutes. It is directed toward the 
development of computer-based mathematical and 
statistical analyses, pattern recognition, and image 
processing of data, principally x-rays and electron 
energy loss spectra, derived from biological 
specimens studied in an analytical electron 
microscope. 

Progress During FY82: The DeAnza image 
processing system has been upgraded from a 
maximum image size of 256 by 256 pixels to a 
maximum of 512 by 480 pixels. Testing of this 
upgraded system revealed many serious hardware 
defects. The vendor has spent most of FY82 
resolving these problems. The same vendor provided 
the hardware interface between the DeAnza system 
and the magnetic tape and disk drives. This interface 
has also been sent back to the vendor for 
correction. Consequently many projects depending 
upon this system have been delayed (see Ultrasound 
and Nuclear Medicine). 

Statistical analyses currently are being performed to 
determine the sample size necessary to detect 
reliably changes in density of dense bodies in 
digitally acquired electron micrographs of platelets. 

Proposed Course: The study of the basic physics 
and the formulation of appropriate mathematical/ 
statistical models needed to achieve the analytical 
capabilities will require extensive work with 
phantoms, i.e., specimens of known composition that 
are very thin, prepared by such means as vacuum 
evaporation. There will need to be extensive studies 
of the signal/noise ratio in phantoms and in 
biological specimens. Potential problems with 
contamination and with specimen destruction by the 
high energy beam also need to be studied. 
Sophisticated algorithms for element recognition and 
location, image enhancement, etc., need to be 
designed. 



LAS proposes to undertake some of these objectives 
in collaboration with participating wet laboratories. 
The DeAnza system has been upgraded from a 
maximum image size of 256 x 256 pixels to 512 x 
480 pixels. Images acquired at Brookhaven are 512 
X 512 pixels; when the expanded system with the 
new magnetic tape and disk drives is operational, it 
will allow more rapid processing of the images 
obviating the need for data compression or 
partitioning. 

Publications and Abstracts: None. 



5So;£ci"iSUe 


jij««!,j;f™«;i™ ™ 


'T »tS:'; mmS\i%Ki^ 


PBOJECT BB.BEB 






'""■'SoS'lf^™'" 


ZOl CT00042-04 LAS 


"ocL'bTr" 


1, 1981 to Septe 


■ber 30, 19S2 






1..=) 


Compute 


^^^''d X-rav^and 


nd Image Processing in Electron 
Klectron-Loas Spectroscopy 


IIAUES, LABOR 


TOBY AHO INSIIIUTE tf 


ILUIIOHS, AND TITLES Of PBIHCIP6L IHVESTIGAIOBS ABB ALL OTHEB 


OTHfflS: 


H.A. Douglas 
j!l. Costa 


Computer Systems Analyst LAS DCRT 




J.J. Bailey 




COOPERATINC 


»ITS (if .»,) 




Clinical Neuropharmacology, HIHH, Laboratory of Cheoistry, NIADDK. 1 


LAB/BR*NCH 
Laborat 




udies 




Applications Sec 


tion 


DCRI, IIIH, Bethcada, KO 2020? 


D(.) »mu 


1 1 2 

UtE BOI(ES) 


0(1.) BBMAH IISSIJES a(i)«ElIBEB 


(..)..«. 


S D (.^) l«tE.V.E.S 






OBK (200 ■.«. .r U:: 


collaboraUon^of LAS and several lUH 


™?Se,'a 


tical md'staUst 


ted toward the development o 


f computer-based 


nition, and image 


df?f5S 


ing or data, prin 
from biological 


lipally X-rays and electron energy loss spectra 


specimens studied in an analytical electron 









56 



Physical Sciences 
Laboratory 



George H. Weiss, Chief 



Summary of Activities 

Consulting Services. R. A. Brooks (SN, NINCDS); J. 
Shapiro (CC, DIR); A. Pikus (CC, OPD); J. D. Dillon 
(Walter Reed Medical Center); B. Sonies, M. Stone 
(CC, REHAB); L. Nadel (CSL, DCRT); G. Knott (LSM, 
DCRT); H. Edelhoch (CE, NIADDK). Members of PSL 
collaborate with and advise researchers at NIH in 
several areas of applied mathematics and applied 
physics. 

A study on the lifespans of German veterans who 
sustained head injuries in World War I has been 
completed. A comparison of data on these veterans 
and on uninjured veterans showed that the lifespans 
were little affected by injuries until approximately the 
age of 60, after which death rates in the injured 
group exceeded those in the control group. Further, 
the only reliable prognostic factor for the occurrence 
of early death was posttraumatic epilepsy, while 
other measures of severity of injury showed no 
predictive value. 

A theory that enables one to calculate the statistical 
errors in parameters estimated from positron 
emission tomography data has been developed. The 
validation of this theory by simulation will be 
undertaken and then application to the data of Dr. 
Rodney Brooks will be made. 

A joint effort with members of the Speech 
Rehabilitation Department of the Clinical Center and 
with Dr. Lawrence Nadel of the Computer Systems 
Lab has begun on the use and quantitation of 
ultrasonic imaging to the study of tongue 
configuration in speech. So far the ultrasonic images 
have been digitized, and various studies completed 
on the reliability of measurements processed in this 
way. 

Studies in Mathematics and Statistics. George H. 
Weiss (PSL). A meeting on random walks and their 
application to the physical and biological sciences 
was held at the National Bureau of Standards on 
June 28-July 1, with participants from around the 



world. PSL has worked on several applications of 
random walk theory to physical problems. One is the 
configuration of polymer chains near surfaces, and 
another is a problem derived from crystallography. A 
second project that has been completed relates to 
order statistics of diffusion processes that may or 
may not be spatially homogeneous. 

Several statistical tests for examining the 
relatedness of DNA sequences from different 
species have been developed. Heretofore, tests of 
relatedness were produced by simulation only; our 
tests, while not comprehensive, are nevertheless 
exact. 

Correlation Function Spectroscopy/Laser Light 
Scattering. Ralph J. Nossal (PSL). Studies have 
been completed on changes in mechanical 
properties occurring near the gel-sol transition in 
polymer gels. A theory is being developed to relate 
macroscopic measured parameters to the underlying 
microscopic structure of the gel. The techniques that 
have been developed in this project are currently 
being used to investigate properties of gels formed 
from glycoproteins and of clots formed from 
reconstituted human plasma. 

Two-dimensional Fourier Transform Nuclear 
Magnetic Resonance Spectroscopy. James A. 
Ferretti (PSL). Considerable progress has been 
made in applying two-dimensional NMR 
spectroscopy to the measurement of rate constants 
in enzyme reactions. In principle it has been shown 
that one can monitor all of the pathways in a 
complex exchanging system. 

Theory and Measurement of Intermolecular 
Forces. Adrian Parsegian (PSL). For the first time 
ever, measurements of the forces between 
molecules have been made using a combination of 
thermodynamics and crystallographic methods. In 
particular these measurements have been made for 
the repulsive forces between parallel DNA double 
helices. 



59 



Quantitative Analysis of Cell Electronmicroscopy 
and Plasma Membranes. Nahum Gershon (PSL). 
This project uses image processing techniques to 
interpret eiectronmicrograph pictures. Measurements 
have been made of the volume, surface area, and 
pore size of the cytoskeleton of cells. These 
measurements show that the cytoskeleton occupies 
a small volume of the cytoplasm contrary to what is 
usually assumed. A computer system is being 
assembled to study three-dimensional cell structure. 

Research Projects 

Consulting Services 

Consulting services are provided to NIH researchers 
on problems requiring knowledge of advanced 
techniques in applied mathematics, physics, and 
statistics. Projects have been completed relating to 
interpolation errors in computerized tomography, to 
the occurrence of audiologic defects in patients and 
relatives of patients having osteogenesis imperfecta, 
on the consequences of head injuries for life span, 
and on the occurrence and time course of 
posttraumatic epilepsy. 

A project has been started on the estimation of 
errors in parameters measured by positron emission 
tomography, and a theory has been developed to 
estimate expected errors from the data. Another 
project with the Clinical Center is on the use of 
ultrasonic measurements to characterize tongue 
position in speech. The biochemical characterization 
of coated and uncoated vesicles by a combination of 
ultracentrifugal techniques and light scattering is 
another project recently initiated. 

Publications: 

Rish, B. L, Caveness, W. L., Dillon, J. D., Kistler, J. P., Mohr, J. P., and 
Weiss, G. H.; Analysis of brain abscess following penetrating craniocer- 
ebral injuries in Vietnam. Neurosurgery B■.5'i5-5A^ , 1981. 

Shapiro, J. R., Pil<us, A., Weiss, G. H., and Rowe, D. W.: Hearing and 
middle ear function in osteogenesis imperfecta. J. Am. Med. Assoc. 
247:2120-2126, 1982. 

Weiss, G. H., Caveness, W. F., Einsiedel-Lechtape, H., and McNeel, M. L: 
Life expectancy and causes of death in a group of head injured veter- 
ans. Arch. Neurol, (in press). 

Weiss, G. H., Feeney, D. M., Caveness, W. F., Dillon, J. D., Kistler, J. P., 
Mohr, J. P., and Rish, B. L.: Prognostic factors for the occurrence of 
posttraumatic epilepsy. Arch. Neurol, (in press). 

Weiss, G. H., and Rice, J.: A combinatorial problem in pharmacology. J. 
Math. Biol, (in press). 

Weiss, G. H., Talbert, A., and Brooks, R. A.: The use of phantom views to 
reduce CT streaks due to insufficient sampling. Phys in Biol and Med. 
(in press). 



™'"" 


is.g'&f .!!':;:'! 


ON EACHANCE 


tHTR 


isiif 


PROJECT 


ZCl 


CT 00022-15 PSL 




n^rnh^r 1 1QR1 rn R«prPTnhPr Ifl . 19R? 


.o„„U.„sS„vU» 
















paofEss 


ONAL PERSONNEL ENGAC 


ED ON THE PROJECT 










PI: 


George H. W 


eiss. Chief, 


SL, DCRT 








Other 


Ja-es E. Ki 


efer, PSL. DC 


T 










Ralph Noss. 












COOP£fiA 


IHC UNITS (if ..,) 


R. A. Brooks. 


SN, NINCDS; 


J. Shapiro 


CC, DIR; A. Piku 




st;„° 


D; J. D. DiUon 


Walter Reed 


Rlfc"K'" 


"'isS: 5?r; 


; H. Edelhoch, CE 








DivU 


'„p"of°cl.'C»t« 


Research 5 le 


hnology 








TOTAL HANYE-HS: j PROFESSIONAL: |OT«£R: j 


n(.) 


PROmiATE hoXes) 


D (.) HO. 




D(0 < 


EITHER 




D(..) 


INORS D (■!) INTER 


>IE>S 












CF WORK (!00 -onJs c 


re'p'tovided't 


'»irresear 




blems re oiring k 


„„- 


ledge 


Of advanced tec 


hniqnes in ap 


lied mathem 


atics, phys 


cs, and statistic 




oJ pa 


aphy, to the oc 


teogeneSis^m 


erfec?a' on 


^Chrconsequ 


ences ofhlad'inj 


ries 


A°pro 


ect'ha"'b«n st 


arted°™"thr 


stimatioro 


f "err or fir 


sttraumatic epile 


7' 


bv po 


itron emission 


tomography, 3 


d a theory 


nas been dev 


eloped to estimat 




i'3 


use of oltraso 
ochemical chara 


i^-^ 


light scatt 


eEFifS 


thcr project tece 


atio, 
tly 



60 



studies in Mathematics and Statistics 

A comprehensive review article on random walks 
and their application in chemical physics and biology 
has been completed. Arrangements for an 
international meeting on random walks and their 
applications, sponsored by NIH and NBS were 
made, and the meeting was held from June 28 to 
July 1, 1982. Several statistical tests have been 
developed for measuring the relatedness of DNA 
sequences from different species. A study of first 
passage problems for diffusion processes with 
spatially inhomogeneous transition coefficients has 
been completed. 

Publications: 

Kiefer, J. E., and Weiss. G. H.: A comparison of two methods for accelerat- 
ing the convergence of Fourier series. Comp. and Math. 7:327-336, 
1981. 

Rubin, R. J., and Weiss, G. H.: Random walks on lattices: The problem of 
visits to a set of points revisited. J. Math. Phys. 23:250-253, 1982. 

Weiss, G. H.: Random walks and their applications. Amer. Sci. (in press). 

Weiss, G. H., and Rubin. R. J,: Random walks: Theory and selected appli- 
cations. Adv. Chem. Phys. (in press). 

Weiss, G. H.. and Shiesinger. M. F.: On the expected number of distinct 
points in a subset visited by an N-step random walk. J. Stat Phys. 
27:355-363. 1982. 



-'"'■'""" ■™-'"'" 




ZOl CT 00024-07 PSL 


October I. 1981 to September 30. 1982 


—"«—-—" 


Oth.,: /"".ieferJ^SL. KRT 


eooPHAIlNC »»MS (i' ■">) B. J. Rubin, Ph. D.. Senior 


Scientist, KBS; K. E. Shuler. 
erg, Ph.D.. Univ. of Clifornla- 


LAB/eR*NCH 

Physic.l Science. Ubor.tor, 




Division o£ Computer Research & Technology 


0.8 ' 1 "„""■' ■ 1°'""' 5 3 


CB!C« .ffRlBll.H I10.(£S) 

D(.| KUl... SUIJSCIB D(t)HB.U,IISS«lS 

D(.i)»i«o«s Dl.i) l».M.I£.S 


a (•) >si'«i» 




nd their application in chemical 
•cnts tor in international 
.ponsored by NIH and NBS were 



Theory of Biochemical Separation Techniques 

The techniques of applied mathematics and statistics 
are applied to the design and analysis of biochemical 
experiments. 

Several methods have been compared for estimating 
peak height and half-width of chromatographic 
peaks. These have led to a clear choice among 
methods currently being used. A similar investigation 
has been undertaken on Lorentzian peaks that arise 
in NMR measurements. A theory of kinetic tailing in 
chromatography has been developed using 
techniques devised for the study of hopping 
conduction in solids. 

Publications: 

Weiss, G. H.: Optimal parameters for the measurement of the half-width of 
a Gaussian peak. Sep. Sci. S Tech. (in press). 



5Si;r^fis.?£'tss\rs::*fi^s^^^:5'^ 


■I^isii!'!!!^ 


„ ZOl CT 00014-15 PSL 


October 1, 1981 to September 30, 1982 1 


Theory oE Biochemical Separation Techniques 




PI: George H. Weiss, Chief, PSL, DCRT 




COOPtBATIHG wm (it ««,] 

John Rice, Professor of Statistics, University of 


California-San Diego 


LAB/eniNCH 








CM£W *PPflOPBI*I{ 80«(£S) 


□(e) NEITNfH 




peak height snd half-width 
d for the study of hoppine 


graphy has been developed using techniques devise 



61 



Correlation Function Spectroscopy/Laser Light 
Scattering 

Methods have been developed that utilize dynamic 
light scattering techniques to probe the mechanical 
rigidity and internal viscosity of polymer gels. Recent 
emphasis has been on characterizing changes in 
mechanical properties that occur near the gel-sol 
transition, as crosslink density and polymer 
concentrations are varied. The influence of solvent 
viscosity on the dissipation of mechanical excitation 
is also being assessed, to facilitate the formulation 
of a mathematical model to explain the dependence 
of measured macroscopic parameters on 
microscopic gel structure. Polyacrylamide gels have 
been used as model networks in these 
investigations, in part because of the technological 
importance of such polymers in biochemical 
separation procedures. Biological materials that have 
been studied by these techniques recently have 
included gels formed from glycoproteins obtained 
from sputum and clots formed from reconstituted 
human plasma (cf. project Z01 CT 00017-10 PSL). 
Other activities mainly have involved obtaining 
information on particle size distributions needed by 
other investigators at NIH, including data on 
liposomes, 'coated vesicles' obtained from bovine 
brain and polymethane coating material scraped 
from implanted pacemaker electrodes. Collaborative 
studies to support the further development of a laser 
Doppler blood flowmeter also have been performed. 

Publications: 

Chen, S-H., Chu, B., and Nossal, R. (Eds.): Scattering Tectiniques Applied 
to Supramoiecuies and Nonequiiibrium Systems. NATO ASI Series B: 
73, New York, Plenum Press, 1981. 

Nossal, R.: Laser Light Scattering. Methods of Experimentai Ptiysics 20: 
299-336, 1982. 

Nossal, R.: Quasielastic Light Scattering from Polymer Gels. In Chen, S-H., 
Chu, B., and Nossal, R. (Eds.): Scattering Techniques Applied to Supra- 
moiecuies and Nonequiiibrium Systems. New York, Plenum Press, 
1981, pp. 301-320. 

Nossal, R.: Stochastic aspects of biological locomotion. J. Stat Phys. (in 
press). 

Nossal, R., and Jolly, M.: Shear waves and 'internal viscosity' in cylindrical 
gels. J. AppI Phys. (in press). 









62 



Cell Motility and Chemotaxis 

This project has been undertaken to study various 
aspects of cell locomotion, including the 
mathennatical basis of macroscopic assays for 
leukocyte chemotaxis. Procedures for measuring 
parameters of cell migration, e.g., computer-assisted 
tracking techniques, have been developed. 

Recent emphasis has been on perfecting laser 
quasielastic light scattering techniques for 
nonperturbative measurements of the mechanical 
properties of soft protein gels that, when examined 
by conventional rheometers, are structurally unstable 
(cf. project Z01 CT 00021-11 PSL). The objective is 
to examine cytoplasmic extracts from motile cells, in 
order to understand how the polymerization of 
cytoskeletal networks and the generation of 
contractile force therein actually takes place. In order 
to gain practical experience in acquiring such data, 
various measurements have been performed on 
polymer networks formed from fibrin, which in many 
respects is a similar but more readily available 
material. 



Two-dimensional Fourier Transform Nuclear 
Magnetic Resonance Spectroscopy 

Two-dimensional Fourier transform Nuclear Magnetic 
Resonance (NMR) spectroscopy has been applied to 
determine the structure of conjugates of glutathione 
and to study enzyme-catalyzed rates of exchange. 
Unidirectional rates have been determined for the 
phosphoglucose isomerase catalyzed isomerization 
and anomerization of glucose-6-phosphate and 
fructose-6-phosphate, for the adenyl kinase 
exchange of adenosine diphosphate to adenosine 
triphosphate with creatine to form adenosine 
diphosphate and creatine phosphate. From these 
studies mechanisms of the reactions were proposed 
and it was demonstrated that two-dimensional NMR 
spectroscopy is capable of monitoring 
simultaneously all of the pathways in a complex 
exchanging system. 

Publications: 

Jacobson, L., and Ferretti, J. A.: The determination of a phosphorus-phos- 
ptiorus nuclear Overhauser enhancement by two dimensional magne- 
tization exchange spectroscopy. J. Amer. Chem. Soc. (in press). 

Marks, T, J,, Pohl, L. R., Gillette, J. R., Hong, M., Highet, R. J., Ferretti, J. 
A., and Henson, J. A.: Stereoselective formation of bromobenzene 
glutathione conjugates. Chem. Bio. Interactions (in press). 



EHITHSOMUN SCIENCE INFORHlTIM EXCHANGE 
PfiOJECI NUMER (Oo HOT «• (hit •»<:•) 


iI.«iKil»™i!l 


201 CT 00017-10 PSL 


October 1. 1981 to Septeiiber 30. 1982 


IIILE I mOJEEI (M .».r..>.,. .r 1...) 
Ctll Motility md Che.ot.xi. 




" 




COOPE««II«C UNItS (II ..,) 

J. GlailnBr, Ph.D., LBC, NIADDK 


L»e/BR»NCH 

Physical Sciences Laboratory 


"""" 


Division of Cos^uter Research and Technology 


IQUt ...YE.R^S,^ |™<,ESSI».t. jO,»E», 


CHECK APmOPRUIE eox{ES) 

J (.) «U... SU.J.CIS a (.1 .C.» TISSUES 

Dl.D.I.ORS aW) iHIERvii.s 


DM 


EIEHE. 


Recent enphssis has been on perfecting laser qusa 


sspeets 
s.ys fo 

polymer 


r leukocyte chemotaxis. 
light scattering 

tr.ctile force therein 



siirir 


a.s'S!j'.K':r!Jf: sw 


HE."t'l5'.°«c'Sl".rs."lCES 






"•■" 










ZOl 


CT 00025-07 PSL 


Octobe 


r I, 1981 to Septeslier I. 198! 






ritu Of 


PROJECT (eO chincl.n or Iti.] 






Two-0 


mensional Fourier Transform Nuclear Magnetic 


Resonanc 


e Spectroscopy 


Tof"^.' 


'o.T™imrEL'£i"«ErorTit'"S"c; '"° ■""' " """"' 


1 I..ESTIC. 


TOR! .»« .tt OTHER 




Lev Jacobson, PSl' 

R. J. Highet! PhD.l Laboratory of Chemis 


;;.= 


logy, NHLBl 




LH/BRAN 

Physic 


"l Sciences Laboratory 






seen OH 


DiviBion of Computer Research and Technology 


I0T«l«N««5. pOfESSIONAU |(,IH«^ ^ 


CMtCK *P 


WOmiATE (O.(IS) 








«.H SURJECTS Q (t] HUIAII TISSUES 
IHORS a (.i) IHTEfldEwS 


U (■) HE 


T.IR 


suuHini 


C» -OflH (ISO .srds or l*>t - uHO.rll.<* l..,.onl.) 






IV 


mensional Fourier transform Nuclear Magnetic 


f"c"n"g^ 


e (NMR) spectros- 
tes of glutathione 


i 


study enzyme catalyzed rstes of exchsnge. 
een determined for the phosphoglucose isomer 

kinase exchsnge of adenosine diphosphate to 
ne to form adenosine diphosphate and creatin 
s mechanisms of the reactions were proposed 


!"ih§i 


e triphosphate with 











63 









Precision in the IVIeasurement of NIUIR 
Parameters 



^ 



A study of the precision in the estimate of peak 
positions in spectroscopy was carried out. This study 
was carried out by considering both instrumental 
noise and error due to a finite digitization rate. 
Different strategies for estimating peak position were 
compared. The comparison showed that it is always 
desirable to use some form of curve fitting the peak 
rather than using the position of the digital maximum 
as the peak position estimate. It was shown that it is 
very difficult to estimate changes in peak position 
that are less than ten percent of the line width. 

A method for the estimation of peak areas for 
nuclear Overhauser enhancement factors was 
developed. Although this study constituted a 
preliminary investigation, which did not directly 
consider all sources of error, it was found that the 
peak area can be determined with high precision in 
the presence of substantial phase error. This method 
(called the product method) involves multiplying the 
peak height with the width at one-half peak height to 
compute peak area. 

Publications: 

Weiss, G. H., Ferretti, J. A., Kiefer, J. A.: A study of precision in the 
measurement of chemical shifts. J. Mag. Res. 46:69-83, 1 982. 



P.O.SOT ««..£ 


'^£^^^ 


.H^iHii' 


:f 


^ 


ZOl CT 00026-07 PSL 


October 1 


1981 to September 30. 1982 






Theo., „a Me,s„e..„. of I«„„oU.„l„ Po„e, 


JiSfKsS 


P£flLrL'EN^'MEroriHE'™°eCT *"" ^'^""^ ^^ 


'"" 


«L iriVESTIt.IOBS m ALL 0,«SB 


PI; 


i\t:i-i-r.:TiT:^ii 


Ph. 


D., PSL, DCRT 




M. Prouty. Ph.D., HIADDK 

n! Fuller. 'ph.D.i'flrock University, 


c^:- 


ada 


LAB/BBAHCH 

Phvaicnl 










Division o£ Computer Research and Technology 


....L .,,,,^...1, P,.FES,,«,.^t, |„„B„ ^^ 


3 1.1 H»«.« 


IME BOXfES) 




D (=) "ElIBiB 


a 1.1) <l«0« 


D(.2) INTERVIEWS 






SUUIAfiT OF U 


HK (ZOO -ord, or 1«, - .oa.rline k,,.ordO 






decaying 

■ho« that 
ti.ely !, 
cloBely t 


Its magnitude depends strongly on 
m the predictions of polyelectrolyt 


"£ 


dependence on ionic strength 
ory. The results correspond 


SSj 


equivalent of phase diagrams for a 
rmination of thermodynamic paramctc 


ZTJ, 


u:rprote",°:it,;°thrc"„.e- 


HolecuUr 


graphics on the DCRT and NIADDK ays 
Inr contacts corresponding to mcaau 


™«'r 


are being uaed Co visualize 
ts of molecular aaaembly. 



Theory and Measurement of Intermolecular 
Forces 

During this year we have achieved the first direct 
measurement of a force between molecules. The 
repulsion between parallel DNA double helices is an 
exponentially decaying hydration force similar to that 
observed previously between bilayer membranes. Its 
magnitude depends strongly on the identity of ionic 
species bound to the DNA molecule. Its decay and 
lack of dependence on ionic strength show that at 
to 20 Angstroms separations intermolecular forces 
differ qualitatively from the predictions of 
polyelectrolyte theory. The results correspond 
closely to an intuitive macroscopic theory of water 
polarization and work of water removal from 
molecular surfaces. 

Creation of protein gels and crystals under osmotic 
stress has enabled us to create the equivalent of 
phase diagrams for assembling proteins with the 
consequent determination of thermodynamic 
parameters. 



64 



Molecular graphics on the DCRT and NIADDK 
systems are being used to visualize the molecular 
contacts corresponding to measurements of 
molecular assembly. In this way we hope to derive 
useful and accurate potentials for molecular contact. 

Publications: 

Lis, L. J., McAlister, M.. Fuller, N., Rand. R. P.. and Parsegian. V.A.: 

Interactions between neutral phospholipid bilayer membranes. Biophys. 

J. 37:657-666. 1982. 
Lis, L. J., McAlister, M., Fuller. N., Rand, R. P.. and Parsegian. V. A.: 

Measurement of the lateral compressibility of several phospholipid bi- 

layers. Biophys. J 37:667-672. 1982. 
Parsegian, V. A.: Protein-Lipid Interactions in Memtiranes. The Rockefeller 

University Press. 1982, 401 pp. 



Analysis of Intracellular pH by 31 P Nuclear 
Magnetic Resonance Spectroscopy 

Binding constants of various magnesium/ 
orthophosphate complexes were experimentally 
determined. Precision and accuracy of intracellular 
pH measurements based on orthophosphate 31 P 
NMR chemical shift were thoroughly analyzed in 
terms of intracellular magnesium ions availability and 
limited instrumental S/N ratio. It was demonstrated 
that the presence of free magnesium ions has only a 
marginal effect on pH-dependent Intracellular 
orthophosphate chemical shift and no effect on 
derived intracellular pH estimates. 






lalyais of Intracellular pH by P Nuclear Magnetic Resonance Spe< 



Research k Technology 



lium/orthophosphati 



Quantitative Analysis of Cell Electronmicroscopy 
and Plasma Membranes 

In this study, we have initiated a three-dimensional 
reconstruction of microtubule nucleatlon centers of 
the erythrophorus of Holocentrus as viewed by high 
voltage electron micrographs. 

So far, we have measured the functional volume, 
surface area, and pore size of the mictrotubecular 
lattice and the cytoskeleton of cells and found that 
they occupy only a relatively small volume of the 
cytoplasm, contrary to what was intuitively assumed 
before. These results were correlated with recent 
measurements of diffusion of molecules throughout 
the cytoplasm. 

In addition we have begun to construct a stereo 
image analysis system that Includes color graphics. 
This system will be used to study the three- 
dimensional structure of cells and their biological 
functions. 

We have developed quantitative methods to analyze 
electronmicrographs of biological systems. The 
methods include digitization of electron-micrographs, 
manually or automatically, and computational 



si;Eji''!s.iJ'f!".ii's:";f;'™.T' 


«,. ;•.;■."'"".'"■ "u>i "' 


UI .*.« 




,JZ!ifKiC'mL ^° 


CT 00041-04 PSL 


October 1, 1981 to SepceDl>er 30, 19S2 


IlILt OF PROJECT (80 chtrtcUri or l*»*} 




Quantitative Analysis of Cell Electronmcroacopy and Plasi 


na Henbranea 










PI: N. Gerahon. Ph.D.. Visiting Sci.nti.t. PSL. 


DCRT 


K. Porter, Ph.D., Fogarty Scholar, Fogarty I 


1 
1 




Physical Sciences Laboratory 


SECtroh 






Division of Computer Research and Technology 








0.8 1 0.7 1 0.1 




CKte< »fi«op«i.ii ii»(is) 




J(.) .u..^ susjttis 0(0 Miau iissuis Q !■) 


«n«£« 


D(><) >i>o>s Ol-'l iaiiivins 




Suaatfly V nOAK (200 .».«» er Itii - .u>d*rl<.< k.^ert.) 






nstruction of microtu- 


bule nucleatlon centers of the erythrophorus of Holoeentr 


us as vifwed by high 


voltsge electron micrographs. 




So far. we have •eaiured the functional voIu,ne. surface a 


r<-«. and pore sice 


of the mictrotubecular lattice and the cytoskeleton of ce 


lis and found that 


they occupy only a relatively small volume of the cytopla 


sm. contrary to what 






measurements of diffusion of molecules throughout the cyt 


oplasm. 


In addition ue have begun to construct a stereo-image ana 


lysis system that 






structure of cells and their biological functions. 





65 



analysis of their contents (e.g., proteins on 
membranes or cytoplasmic organelles and structural 
elements). 

All kinds of eukaryotic cells possess the capacity to 
control their form, their size and to regenerate lost 
parts. A general loss of these capabilities is 
characteristic of neoplastic cells. It seems, therefore, 
that there must exist in ceils a mechanism for form 
control, a structurally continuous system that fills the 
cytoplasm and derives functional properties from its 
organization around a single center or complexes of 
many centers. Those units of organization, in 
addition to giving the cell its form, account for the 
nonrandom disposition of better-known organelles 
such as endoplasmic reticulum, Golgi bodies, 
microtubules, and to a lesser extent, mitochondria. 
We are constructing an image analysis with color 
graphics that will enable us to study the organization 
of cells in three dimensions. We use electron 
micrographs produced by the high voltage electron 
microscope (a national resource sponsored by NIH) 
in Boulder, Colorado and other micrographs taken at 
NIH. This concept of the cytoplasm is best illustrated 
by the red chromatophore of the tropical fish, 
Holocentrus. We already have initiated a study of the 
three-dimensional recontraction of the microtubule 
nucleation centers in these cells. These centers are 
dispersed in patterns that repeat from cell to cell. 

In addition, using a stereo image analysis system 
and the PIC image processing system, we have 
measured the fractional volume of the various 
elements of the cytoplasm together with their 
surface area and the size of their pores. These 
results indicate that the microtubular lattice and the 
cytoskeleton occupy only a small fraction of the 
cytoplasm volume. These findings mean that these 
structures cannot physically obstruct the diffusion of 
molecules through the cytoplasm to a large extent 
but rather slow it by other means (e.g., chemical 
attraction). This result shows that the previous 
intuitive impression that these structures occupy a 
significant portion of the cytoplasm volume is not 
valid. 



Diffusion of lUlolecules on Cell Surfaces and Light 
Scattering from Fluids 

In this study we evaluated the effect of cell 
nonplanarity (e.g., due to the existence of microvilli 
and blebs) on the rate of diffusion of proteins and 
lipids in cell membranes using fluorescence 
photobleaching recovery. For simulated microvillous 
membranes, we found that the existence of 
curvature does not affect the measured diffusion 
coefficient by spot photobleaching recovery, contrary 
to naive intuition. On the other hand, diffusion along 
surfaces curved along one direction only depends 
strongly on the nonplanarity. It was found that the 
amount of nonplanarity needed to explain results of 
measured diffusion anisotropy in fibroblasts is far 
beyond what exists in nature. 

In the second part, we derived hydrodynamic 
equations and the light scattering spectrum from 
microelastic fluids. We studied two cases, fluid-like 
and solid-like viscoelastic fluids. 

Publications: 

Aizenbud, B., and Gershon, N. D.: Diffusion of molecules on biological 

membranes of nonplanar form - a theoretical study. Biophys. J. (in 

press). 
Aizenbud B., and Gershon, N. D.: Diffusion of Molecules on Microvillous 

Biological Membranes. In Perelson, A, C, DeLisi, C, and Wiegel, F. W. 

(Eds.): Cell Surface Phenomena. New York, Marcel Dekker (in press). 
Aizenbud, B., and Gershon, N. D.: Hydrodynamic equations and VH light 

scattering from viscoelastic (solid like) systems. II. Molecular approach. 

Physica A (in press). 
Aizenbud, B. M., and Gershon, N. D.: Hydrodynamic equations and VH light 

scattering from viscoelastic (solid-like and fluid-like) systems. Pheno- 

menological approach. Physica A 107:126-142, 1981. 



66 



miTH^WIW SCItHCt ltlFI»M*nON EXCHUGl 

•aojicT >oui> !o. <oi ... i>i. .».,;) 


«,.;•>« ."'S"i*„c 


s 


""'" 


"""• 






e. 


ZOl 


CT 00068-03 PSL 










Diffusion of Molecules on Cell Surfaces .nd Ugh 


Sca.terin 


g fro. Fluids 


r«i';st»r";i*«L'"Nj;;E"» 'i«';!i™ci ""° "'"° " "'" 


'"■ '"•"'" 


,o.s m »iL .t»w 


OEher: Boris Aizenbud, Ph.D., DeparCnent of 






, 






Physical Sciences Laboratory 








"""'"'"" p-'—'i'. p=- 0.. 




U (■) «s 


I.iR 


D(.i) "iim«s 0(.') i«it«.ii.s 






SUSUARr W WORK (JOO ..rdi or I... - ..d.rti-^. k.y.ortf.} 






In this stud, - evslus.ed the effect of cell no 


planarit. 


(e.g.. due to the 


anisotropy in fibroblasts is far beyond what exi 


Spot phot 


along surfaces curved 
ty. It was found 
f measured diffusion 


In the second part, we derived hydrodynaraie equa 


o°case". 


he light scattering 






ZOl CT 00040-04 I 



1, 1981 to Septeidier 20, 1982 



Control of Actin Assenbly in Nonmuscle Cells 



PI: Stephen L. 



of Computer Research and Technology 






Control of Actin Assembly in Nonmuscle Cells 

The protein actin is a major cytoskeletal component 
of all eukaryotic cells, serving both structural and 
motility-related functions. The G-actin monomer 
binds one ATP, which is hydrolyzed upon 
polymerization to polymeric F-actin. Studies in this 
laboratory are aimed at elucidating the detailed 
mechanism of polymerization, the role of ATP 
hydrolysis, and the cellular control mechanisms for 
actin assembly and disassembly. 

The actin polymer has two ends that, by virtue of 
ATP hydrolysis, can have different monomer/ 
polymer equilibrium constants. As a result, actin 
monomers may treadmill through the filaments with 
net monomer addition occurring at one end of the 
filament and net loss at the opposite end while the 
filament maintains constant length. Direct evidence 
for treadmilling has been obtained this year using 
actin covalently modified with a fluorescent probe 
(N-pyrenyliodo-acetamide) that has a 20-30-fold 
fluorescence enhancement when G-actin 
polymerizes. Trace amounts of this probe, when 
added to a G-actin/F-actin solution at steady state, 
are incorporated from the G-pool into F-actin with 
kinetics implying a treadmill mechanism. ATP 
hydrolysis rates, measured concurrently, indicate a 
high degree of efficiency for the treadmill, with as 
few as 1-2 ATP molecules hydrolyzed for every new 
actin protomer incorporated. The efficiency is a 
strong function of ionic conditions with no 
treadmilling occurring in the absence of free divalent 
cations. ATP hydrolysis is obligatory; no monomer 
incorporation occurs when G ADP actin is used. 
Cytochalasin, a drug that we have shown to cap the 
end of actin filaments, strongly inhibits monomer 
incorporation at steady state. 

We are currently isolating several proteins from 
motile nonmuscle cells to determine their effects on 
actin polymerization, steady state monomer-polymer 
exchange, and nucleotide hydrolysis with the goal of 
understanding the control of assembly and 
disassembly of the actin filament in these cells. 

Publications: 

Brenner. S. L., and Korn. E. D.: Stimulation of actin ATPase activity by 
cytoctialaslns provides evidence for a new species of monomeric actin. 
J. Biol. Chem. 256:8663-8670, 1981, 



67 



Computerized Typesetting of Scientific Papers 

The object of this project is to prepare computer disl< 
or magnetic tape versions of scientific papers 
intended for publication. This material can be sent 
directly--on the phone or tape--to publishers' 
copyediting/typesetting computer systems. 

Our current efforts emphasized facilitation of 
magnetic tape production and file preparation for 
transmission. This year we overcame formidable 
coding problems to write tapes, destined for outside 
use, on the IBM System 370 directly from WYLBUR 
files. We also initiated a series of telephone 
transmissions of WYLBUR files. Over 30 scientific 
articles have been transmitted this year either using 
one of the tape writing programs or by telephone. 

Such electronic conversion of texts has been shown 
to be cheaper, faster, and more accurate than the 
old way of retyping material by the publisher. 
Typesetting costs can be halved. Already one journal 
is offering a major discount in page charges to 
authors submitting 'compuscripts.' Others should 
follow. The ultimate savings to NIH are expected to 
be significant. 



™'"'""' 


'^--"-^' 


,,™isi 


i'™f 


"Z 


r'oLb- 


3PSL 


Ocwber"? 


1981 to September 30, 1982 








nitb or fno 


ECT {BO ch.racUri or le«) 








C„^.teri 


ed Typesetting of Scientific Pa 


ers 






'»°""'""- 


Pfsoma £«=.t£0 BN I»£ ™»;£ci 


DCHT 








H. McNeel. Computer Systems 








Ochers; 


H. Douglas, Computer Systems 


Analyst, LAS, DCBl 








M. Horton, Computer Systems f 


nalyst. LAS, DCRT 
















Rockefell 


r niversi y ress, cience re 


s. Biophysical Jou 


rnal, Wav 


cly Press 


LAB/BHWICH 










Physit.l 


ciences Laboratory 










Division of Computer Research and Technology 


D(.) B»«». 
D(.0 "i«m 


«I£ Bfl.(£S) 

BBJECIS D(») HOSJll IISS»£S 
D (.2) .«IE«VI£»S 


1 0.3 

a M "£ii 


"" 














of jcient 


tic papers intended tor publica 


3"mu'ma?eria 


gnetic ta 
1 can be 


e^version. 


directly- 


on the phone or tape — to publis 


ers' copyediting/t 


ypesettin 


computer 


file°Jrep 


t efforts emphasized facilitati 


n of magnetic tape 


rm'ISrie'' 


^d-:: 


directly 


«lBU™ilcs"'over "o sJienJiti 


programs or by tel 


elephone 
ephone. 
aper, fas 


t^d^th^" 


more accu 


ate than the old way of retypin 


material by the p 


oblisher. 


Typeset- 


ting cost 


can be halved. Already one jo 


mal is offering a 


major di 


COUht 


follow. 


he ultimate saving, to NIH are 


»pected to be sign 


ificant. 





68 



Laboratory of Statistical and 
l\/iathematicai i\/lethodology 



James E. Mosimann, Chief 



LSM activities can be divided into three areas: 
computation, consultation, and research. 

Computation 
A major part of LSM activity is the offering of 
statistical and mathematical systems/packages to 
the NIH user community. LSM accepts responsibility 
for evaluation of new program packages and their 
suitability for NIH. When LSM does support a 
system/package for the NIH community, it provides 
maintenance, documentation, instruction, and 
assistance for users to interpret the results. 

Statistical Systems/Pacl<ages Support. During this 
year, as in the past year, the Statistical Software 
Section of LSM maintained the following program 
packages and programs: 

• BMD, BMDP: Biomedical Computer Programs, 
UCLA. 

• SPSS, SCSS: Statistical Package for the Social 
Sciences, SPSS, Inc. 

• SAS, SAS/GRAPH, SAS/ETS: Statistical 
Analysis System, SAS Institute, Inc. 

• P-STAT: Statistical Package, P-STAT, Inc. 

• IMSL: International Mathematical and Statistical 
Libraries, IMSL, Inc. 

• MSTAT1: Collection of Mathematical and 
Statistical Programs, DCRT. 

During the year every system /package went through 
at least one major update. The SSS staff answered 
over 5,500 calls for assistance, and taught a total of 
twelve courses on these systems/packages; two 
each on the SPSS and BMDP packages and eight 
courses on the SAS system. 

The use of program packages continues to increase. 
The average accesses per month of all the statistical 
packages rose from around 33,000 during FY81 to 
over 45,000 in FY82. For the sixth year in a row, 
SAS experienced the largest increase of any of the 
packages. SAS averages over 37,000 accesses per 
month, up from 24,000 per month in FY81. The 
average number of accesses per month for SPSS 



decreased from 6,000 to 4,600. The average 
combined accesses of the BMDP and BMD 
packages was 2,500, about the same as in FY81. As 
an example of a package used for specialized 
purposes, PSTAT averaged 30 accesses per month, 
down from 60 average accesses per month in FY81. 
The main programs and subroutines in MSTAT1 
averaged 1 ,300 accesses per month, in contrast with 
650 in FY81. Accesses to the IMSL package cannot 
be counted, but it is estimated that usage increased 
during FY82. 

The DCRT Mathematical and Statistical Program 
Manual was updated in FY82. 

IVILAB Support and Incorporation of C-LAB into 
MLAB. The Biomathematics and Computer Science 
Section maintains the DECsystem-10 interpretive 
program MLAB, a package designed and 
implemented by BCS staff. During FY82, several 
hundred biomedical researchers at NIH used this 
package for modeling and graphical display tasks. 
MLAB is part of the NIH-funded Prophet system, the 
SUMEX-AIM system at Stanford University, and the 
NIH-EPA Chemical Information System. It has been 
distributed to various universities and research 
centers at their request. During FY82, BCS staff 
assisted in implementing overlay facilities on the 
DECsystem-10, and redesigned MLAB in a 
segmented form to use overlay. This reduces the 
load on the DECsystem-10, and user costs in many 
cases, because software for seldom-used operations 
is not loaded into computer memory except when it 
is needed. This made it possible to incorporate C- 
LAB, a previously independent package for 
clustering and multivariate data analysis, into the 
MLAB package. Other additions to MLAB were: new 
OMNIGRAPH character fonts for graphical displays 
of mathematical formulas or scientific text, and a 
new, more informative system of error messages. 
One advanced and three introductory courses were 
taught for MLAB. Two articles on MLAB techniques 
appeared in INTERFACE. The tenth edition of the 



71 



MLAB Reference Manual is being printed, and will be 
distributed in FY82. 

Support for the Unified Generator Pacltage. This 
package, developed by a BCS staff member, 
generates IBM System 370 assembly language 
programs. The compatibility of the package with new 
WYLBUR was maintained. As before, assistance was 
provided for users on request. 

Support for Other Software. BCS continues to 
maintain certain special-purpose software and to 
assist users upon request. The PROLOG package, 
obtained from the University of Edinburgh, is 
designed for analysis of non-numerical data by 
aggregation of procedural rules; it has been used in 
LSM linguistic research. A program developed by 
BCS for interactive construction of an index for a 
document file has been supported. Various LSM- 
created programs for analysis and reconstruction of 
biological shapes using the symmetric axis method 
have been supported. A procedure simplifying the 
generation of users' IBM System 370 data set 
listings was developed and made available. 

Consultation 
As in previous years there was considerable 
variation in the amount of time required for an LSM 
consultation. Some very brief consultations are 
successful, and are brief precisely because there is a 
known answer to the question posed. Other 
consultations involve extensive time and statistical/ 
mathematical/computer science research as well. 

LSM consultations in FY82 were of the following 
types: 

• Mathematical, statistical, and computer science 
advice with limited computer use (5%) 

• Mathematical or statistical advice with 
considerable computer use (55%) 

• Computational advice alone (40%). 

The large computer use in these figures results from 
the continued availability and use of general purpose 
statistical and mathematical packages like SAS and 
MLAB. The diverse nature of LSM consulting is 
indicated by the projects and activities listed below. 

Clinical Research, Patient Care, 
Epidemiology 

Cancer Survival Study. R. Wesley (NCI/DCT/BRB). 
Patient survival-time data was analyzed. LSM 
assisted in modeling and nonstandard curve-fitting 
for maximum likelihood estimates of survival 
distribution parameters. 

Obstetrical Care Study. P. Vietze (NICHD/MRDD). 
Long-term effects of two-day rooming-in after 
delivery were studied. Five categories of behavior 



were monitored by an observer of experimental an 
control mother-child pairs at five time points from 
birth to six months. LSM assisted with statistical 
advice on design and data analysis. 

Diagnostic Study of Systemic Lupus 
Erythematosus. T. Chused (NIAID/LMI). Eleven 
clinical variables were evaluated, for ability to 
diagnose and estimate severity, in 87 patients with 
SLE plus a control group. LSM assisted in linear 
regression and discriminant analysis, which 
determined that a variable obtained as a ratio of cell 
counts was a good discriminator between the normal 
group and the SLE patients. Three presence- 
absence variables were significant in predicting this 
cell count ratio. 

Laboratory Investigation 

Anti-Bovine Gamma Globulin Radioimmunoassay. 

Michael Miller (NIADDK/A&R). Tolerance to a 
thymic-dependent antigen was tested in autoimmune 
mice. LSM assisted in designing procedures using 
MLAB to prepare standard curves from control 
experiments. 

IVIedian Lethal Dose Analysis. R. Evarts (NCI/ 
DCCP). Maximum likelihood estimates of median 
lethal doses (LD 50's) of compounds administered to 
groups of mice were calculated. LSM assisted in 
application of probit analysis techniques. 

IVIultiple-Site Binding. J. Dunn (NCI/DCCP). In a 
series of chemical reactions, many F molecules bind 
stepwise to a G molecule, with distinct affinities. LSM 
assisted in modeling and analysis, using MLAB. 

Receptor Characterization. M. Bissonette 
(NIADDK/DD). A mathematical model for the 
characterization of VIP and secretin receptors in rat 
pancreatic acini in terms of receptor number and 
receptor affinity was studied. LSM assisted in 
simultaneously curve-fitting several nonlinear 
functional forms to respective data, using MLAB. 

Metal Ion Protein Binding. C. Chatterji (NIAID/LC). 
Optical absorbance experiments measured metal ion 
binding to a protein constituent of snake venom. 
LSM continued to assist in curve-fitting models to 
absorbance data. 

Analysis of Simultaneous Binding Reactions. L. 

Jacobson (NICHD/LCP). Simultaneous binding 
reactions are studied by obtaining NMR scanner 
absorbances at specific frequencies. LSM continued 
to assist in mathematical modeling of equilibrium 
constant estimates. 



72 



Zinc-Activated Enzyme IModei. P. B. Chock 
(NHLBI/IR/LB). A twelve-compartment model for 
zinc activation was studied. LSM assisted in 
developing combinatorial aspects of the model. 

DNA Sequence {Matching. P. Senapathy (NIADDK/ 
LEP). Various natural DNA sequences are studied in 
terms of sequence lengths between identical k- 
tuples. LSM provided software for generating gap 
statistics, and assisted in use of MLAB to compare 
data with negative binomial distribution models. 

Bile Secretion Modeling. E. Feytmans (U. of 
Namur, Belgium). Secretion of bile under stimulation 
by taurocholate was measured in a patient 
population. LSM assisted in modeling and curve- 
fitting problems arising from delayed laminar flow in 
a catheter. 

Duck Motion Study. W. Schleidt (U. of Maryland). 
Various aspects of duck motion were observed. LSM 
assisted in computer generation of graphical displays 
of idealized motions, using MLAB. 

Induced Stroke Experiments. P. Ting (NINCDS/ 
LNNS). Stroke was induced in one side at several 
sites in groups of dogs, and blood flow, pressure and 
blood gases were monitored. At several post-stroke 
sacrifice times the breakdown of the blood-brain 
barrier and related nerve damage were evaluated. 
LSM assisted with statistical advice on problems of 
experimental design and data analysis. 

Program Management and 
Administration 

Review and Verification of NRC Personnel Data. 

A. R. Frost, Jr. (NRC); C. Gellman (Technassociates, 
Inc.). Files of the NRC Automated Personnel System 
were reviewed, verified, and corrected. The Unified 
Generator Package was used to create software for 
updating the files and generating reports. LSM 
assisted in use of the package, use of generated 
programs, and in the design of related software. The 
resulting system was used on a production basis for 
over three months, with as many as six clerks 
working full-time to prepare input transactions, and 
performed completely satisfactorily. 

Investigator Career Profile Study. F. Harding 
(NHLBI/DBDR); C. Crafford (JWK International). 
Data on investigators associated with NIH grants are 
evaluated for the effects of past NIH-supported 
training programs and assessments of national 
needs and currently available researchers. LSM 
provided assistance in the use of the Unified 
Generator Package to create data base software, in 
the use of generated updating, reformatting, and 
reporting systems, and in the training of users. Work 



in the period covered by this report involved the 
addition to the data base of information on grantees 
in FY78, and preparation of reports analyzing this 
data. 

Biomedical Communications 
Applications 

Gastric Ulcer Data Base. C. Sniderman, S. 
Humphreys (NLM). Methods for natural language 
querying of gastric ulcer data are being developed. 
LSM assisted in design of a parser for English 
sentences and of related information retrieval 
routines. 

Computer Research and 
Technique Development 

Symmetric Axis Analysis. R. Webber, A. Davis 
(NIDR/CIBI). CIBI staff have developed a number of 
applications of the symmetric axis method for 
describing and analyzing biomedical images. LSM 
assisted in modifications of symmetric axis software 
and preparation of a PDP-1 1 export package for CIBI 
use. 

Automated Data Processing of Medical Language 

Research was continued on the compositional lexical 
semantics of medical terms derived from Greek and 
Latin. Medical compound words can be regarded as 
'conjunctions' of larger semantic classes, 
represented by their Greek-Latin components. The 
relation between compound words and their 
constituents may be characterized as hierarchial. 
Medical compound words often represent certain 
units of meaning that could be likewise expressed by 
phrases consisting of separated components of 
compound words in English or other foreign 
languages. Additional standardization through 
medical compound word processing will be beneficial 
to retrieval performance. 

A list was prepared of 40 semantically productive 
Greek and Latin terminal morphemes (-ITIS, - 
ECTOMY, etc.), the frequency of which was high. 
Medical compound words concurring with the 
selected terminal morphemes were analyzed 
according to semantic constituents and the 
compositional semantic patterns were established. 
The algorithm for semantic interpretation and 
paraphrasing of Greek-Latin terms into English was 
developed for -ITIS forms and surgical procedure 
forms (-ECTOMY, -STOMY, -TOMY, -PLASTY). The 
paraphrasing rules mentioned above will be used as 
a model for the development of additional 
paraphrasing rules for other medical components 



73 



derived from Greek-Latin. Tlie development of 
paraphrasing rules will increase substantially the 
interpretive power of the lexicon and make it 
possible to interpret synonymous phrases that are 
not contained in the dictionary but that occur in 
medical context. Morphological analysis for the 
identification of productive terminal morphemes as 
markers of parts of speech classes, and the set of 
morphosyntactic transformation rules by which 
canonical nominal forms are derived from adjectives 
and noun plurals, were tested on the MEDLAR 
corpus. It was necessary to add 37 transformation 
rules to the existing tree to cover the MEDLAR 
corpus. 

Work was continued on the preparation of syntactic 
and semantic rules for the Viral Hepatitis Data Base 
information system. 

Proposed Course: 

1. Combination of research studies in medical 
language at present level (morphology, syntax, 
semantics). 

2.Creation of lexicographic data base to be used 
for the merge of medical dictionaries and extraction 
of microglossaries. 

3. Continuation of collaboration in the encoding of 
surgical pathology data with the Laboratory of 
Pathology, NCI, to refine the medical dictionary and 
study the language of diagnoses. 
Publications: None. 



?!K"isi.y't!SMr:ri!?:.'3" 


liZsiiffifl 


ZOl CT 00001 11 LSM 


""""ocloSIr 1. 1961 through September 30. 1982 


Automated Data Processing of Medical Language 


PI: H. C. Pacak Computer Systems 

Other: G. Dunham Computer Programm 
S. Harper Computer Programm 


Inalyst LSB 


DCfiT 
DCRT 


None 


Laboratory of Statistical and Mathematical 


Methodology 


1 


Kedical Information Science Section 


'""'Ml*:'' WS^thesda. Maryland 20205 


lOTAL IIAI.VEAR3. 1 PSDFESSIONM.! lOTHER. 


■2{,) HW«. SUBJECTS D(.)HU««i..L5U£E 

D(*i} HrNons q (ji) ihiedviems 


D (=> «£ 


'"" 






lexical sera 




as "conjunctions" of large 

often represent certain un 
or other foreign languages 


lt= or neanl-g t 
eparated cotnpone 

ill Be beneflcla 


1 compound u 
3, represent 


ords can be regarded 
ed by their 

cal compound words 

nd words in English 
through medical 









Cluster Analysis 

The main objective of this project is the application 
of computer cluster analysis and related methods to 
NIH researcher problems. This year, nearest 
neighbor algorithms based on the latest published 
research and extensions to it were developed and 
tested, and algorithms for analyzing spacial point 
patterns were developed for testing patterns of 
retinal cones for regularity. 

An improved algorithm for the computation of the 
Delaunay triangulation of a set of points was derived, 
programmed, and published. The dual of the 
triangulation is the set of Voronoi regions for the 
points. They define neighboring points and the 
nearest neighbor regions around each point. 
Measurements on the regions, e.g., areas and 
angles, can be used to test for randomness of a set 
of points. 

The spacial pattern of blue cones, obtained from 
macaque retinas, appear to have some kind of 
regularity. Models of regular point patterns with 
different amounts of error were fit to the data in 
order to study the underlying nature of the regularity 
and to determine the amount of disorder. Three 
statistics based on nearest neighbor distances and 
angles between the pairs of points were used in 
testing models. Disturbed lattice models, using a 
regular lattice of points with random error at each 
point, could not be fit to the data. However, a model 
that considered each point (cone) as a soft ball with 
a minimum distance required between any pair of 
balls, fit the data very well. 

Proposed Course: Other areas of the retina and 
other point patterns will be studied with this model to 
estimate the degree of regularity of cones in different 
regions of the retina. 

Pubiications: 

Shapiro, M.: A note on Lee and Schacter's algorithm for Delaunay triangula- 
tion. International Journal of Computer and Information Sciences (in 
press). 

Yaar, I., Shapiro, M., and Pottala, E.: Spectral analysis of the EEG in 
hepatic encephalopathy treated with levodopa. Electroencephalography 
and Clinical Neurophysiology 52: 61 7-625, 1 981 . 



74 



^iyt'£ri*efS'fc!o"-« srii?: ^^^.T' 


i«i.Zi!rMiMj^"™Mci 


201 


a 00008-08 LSH 


October 1, 1961 through September 30. 1982 


Cluster Anal/sis 


PI: M. B. Shapiro Research Kathenatlolan LSH 

Other: F. de Monasterlo Head LVR 

S. Scheln Expert LVR 


«EI 


Laboratory of Vision Research 


Laboratory of Statistical and Kathenstlcal Kethodology 


Statistical Hethodology Section 


DCRT, NIH. Bethesda, Maryland 20205 


,.,..«.«..., ~|».„S,«.U -jO,K». 


CHICK iCPflOPBI*ie BCIJl[tS) 

3|.l «U1U»S,.XCIS a|.]»«.»l.S!UtS aUlMllHC. 

U(.l) "l«0«i at->l l«Ill.ll»S 


siaiiM. » .o» {110 ..«. ., i.„ . ,. 
Nearest neighbor algorithms 


Pasea on the latest puMlshed 


oped for 


AlKorlthms for analvilnR spacial point patterns were 
testing patterns or retinal cones for regularity. 





iyiTHSDNIM SCItNCE INfOHBHIION EICHAHCt 


"''"^'wi'i'™^"* 


z 


CT 00009-08 LSM 


October 1. 1981 through September 30, 1982 


TIIU OF MIOJECT (80 ch.r.cUr. or 1,.0 

Hesearch Topics In Computer Science 


P.I. CD. Knott Computer Specialist LSH 
Other : None 


DCRT 


None 


L.B/BRWCH 

Laboratory or Statistical and Mathematical Methodology 


Blomathematics and Computer Science Section 


DCRT. NIH. Bethesda, Maryland 20205 


TOm^NU.RS, pOf£^SS^«.l. |O.H£B. 


CHCCK AmOPRIMt BO«(ES) 

Jt.) Huii«N suejEcis aCO "WW iisiuis a(ON(n«efi 

n(.l) -INOHS 0{.2) INUHVIE^ 




zrz 


d. The 
computer 


Various storane and retrieval algorithms have been 


is useful because such algorithms are used In almost 
programs. Thus biomedical computation in particular 
Improved storage and retrieval methods. 



Research Topics in Computer Science 

The object of this project is to develop theoretical 
bases for new computer methods that will expand 
and improve the use of computing in biomedical 
computation. The methods used are the application 
of known algorithms and the development of new 
pertinent theorems involving combinatoric and other 
related mathematics. Research work in storage and 
retrieval algorithms and their efficiency has been the 
primary topic of concern. 

Various storage and retrieval algorithms have been 
studied. The development of flexible and efficient 
storage and retrieval algorithms is useful because 
such algorithms are used in almost all computer 
programs. Thus biomedical computation in particular 
can benefit from improved storage and retrieval 
methods. 

Currently, a study of hashing storage and retrieval 
methods is underway. This has resulted in the 
analysis of the performance of the hashing methods 
that resolve collisions using direct-chaining with 
coalescing lists. 

Concurrently, an exhaustive survey of storage and 
retrieval methods is underway. This includes the 
recently-introduced k-d tree method. Various 
improvements and refinements in both the 
algorithms and their analysis are being studied. 

Much effort has gone into studying the B-Tree data 
structure for large files and developing a deletion 
algorithm to efficiently remove items from B-Trees. 

Routines to store, retrieve, and delete items in a 
hash table, which employ direct-chaining with and 
without coalescing lists, have been prepared. The 
analysis of these algorithms has been recently 
completed and the results are to be published. 

Publications: 

Knott. G. D.: Fixed-bucket binary storage trees. J. of Algorithms (in press). 
Knott. G. D.: Graphics Facilities in MLAB. In Sproull. R. (Ed.): Computer 

Grapliics, in Chang. S. (Ed.): Har)dbook of Computer and Electrical 

Engineering (in press). 



75 



^-' "»"''>- "•'-■--'""'' 


,I»SiifffL, 


™,TI.oa.„ 


"""aftober 1, 1981 through September 30, 1982 | 




.„ 




liL;: S™e""'=^'"~" '"""' "■'""="'"" 


1 


None 


L*e/BR*NCH 

Laboratory of Statistical and Hathematical Methodology 


""'Biomathe.atic3 and Computer Science Section 


DCHT. NIH. Bethesda, Maryland 20205 


lOIAL WNTEARS: 1 PfiOFESSIOI.JL: lOIHEBi 


CHECK *ppRUPRi*TE eox{es) 


Inclusion relations between vector spaces and relat 

Preparation of scientific manuscripts by computer g 
printer-plotters on minicomputers was investigated. 


r_^ 


,„,.„.. „.„, 





Discrete Mathematics and Applications 

The project objective is to develop mathematical 
theory and computational techniques using discrete 
mathematics (algebra, combinatorics, and graph 
theory), and to apply such methods to problems of 
biomedical research and computer science. 

Studies of inclusion relations between theories of 
modules over a ring continued. A manuscript 
prepared in the previous fiscal year was revised and 
accepted for publication. New research was directed 
towards (1) interrelating different mathematical 
theories of modules, (2) determining whether the 
lattice equations satisfied for a theory of modules 
was sufficient to determine all of the theory, and (3) 
classifying the rings that lead to the same module 
theory. 

In computer science, the minicomputer software for 
computer generation of scientific manuscripts (text 
and figures) was completed and tested. Tests of the 
TEX system (developed at Stanford) for computer 
generation of scientific text were performed. 

Proposed Course: Study of module theory will 
continue in the areas indicated above. Computer 
software to generate scientific manuscripts will be 
augmented by creation of mainframe software for 
high-speed generation of page displays using the 
advanced capabilities of Tektronix 4114 graphical 
display terminals. 

Publications: 

Hutchinson, G.: A complete logic for n-permutable congruence lattices. 

Algebra Universalis 13: 206-224, 1981. 
Hutchinson, G.: Exact embedding functors between categories of modules. 

J. of Pure and Applied Algebra 25; 107-111, 1982. 



76 



Multivariate Statistical Analysis 

The objective of this project is the study of 
multivariate ratios or proportions. 

Study continued on multivariate statistical methods 
(size-shape methods) for analyzing ratios having a 
multivariate lognormal distribution. Studies also were 
continued on ratios that follow an Inverted Dirichlet 
distribution. A paper on special invariant discriminant 
analyses for size and shape variables (with J. N. 
Darroch) was written. The principal investigator 
presented a review of this work at Cornell University, 
Ithaca, N. Y. 

Publications: 

DeBlas, A. L., Ratnaparkhl. M. V., and Mosimann, J. E.: Estimation of the 
number of monoclonal fiybridomas in a cell fusion experiment. J. of 
Immunological Methods 45: 1 09-1 1 5, 1 981 . 

DeBlas, A, L,, Ralnaparkhi, M, V., and Mosimann, J. E.,: Estimation of the 
number of monoclonal hybndomas in a cell fusion experiment. In Vuna- 
kis, H. v., and Lagone, J. J. (Eds.): Immunochemical Techniques (a 
volume of Methods In Enzymology). Academic Press, New York, N. Y. 
(in press). 

Mosimann, J. E., and Malley, J. D.: The Independence of Size and Shape 
Before and After Scale Change. In Taillie, C, Patil, G. P., and Baldes- 
sari, B. (Eds.): Statistical Distributions in Scientific Work, Vol.4, Models, 
Structures and Characterizations. Dordrecht, Holland, D. Reidel Pub- 
lishing Co., 1981, pp. 137-145. 

Ratnaparkhi, M. V.: On splitting model and related characterization of some 
statistical distributions. In Taillie, C, Patil, G. P., and Baldessan, B. 
(Eds.): Statistical Distributions in Scientific Work, Vol. 4, Models, Struc- 
tures and Characterizations. Dordrecht, Holland, D. Reidel Publishing 
Co., 1981, pp. 357-363. 

Ratnaparkhi, M. V.: Some bivariate distributions of (X,Y) where the condi- 
tional distribution of Y. given X is either beta or unit-gamma. In Taillie, 
C, Patil, G. P., and Baldessari, B. (Eds.): Statistical Distributions in 
Scientific Work, Vol. 4, Models, Structures and Characterizations. Dor- 
drecht, Holland, D. Reidel Publishing Co., 1981, pp. 389-400. 

Roux, J. J, J., and Ratnaparkhi, M. V.: On matnx-vanate beta type I distribu- 
tion and related characterization of Wisharl distribution. In Taillie, C, 
Patil, G. P., and Baldessari, B, (Eds.): Statistical Distributions in Scientif- 
ic Work, Vol. 4, Models, Structures and Characterizations. Dordrecht, 
Holland, D. Reidel Publishing Co., 1981, pp. 375-378. 



m"^mUu"(S'm^\iT, JSfJT" 


""^^"i't'"'''"' 


ZOl cr 00013-08 LSH 


^""VEo^'or 1, 1981 ctirough Scptenbcr 30, 1982 


Multivariate Statistical Analysis 


PI: J.E. Moslraann Oil.r. LSK L3K KIIT 

H.V. Ratnaparkhi Associate ProTessor 

Wright state Unlverslt» 
Dayton, Ohio 


€00»C«<ll«t WHS (II ,.,| 

Hone 


Uboratory of SUtlatloal anil Hauieiaatlcal Xethodoloiy 


Office of the Chief 


OCRT. HIH. Bethesds, Harylani) 2020; 






0-3 1 0.3 1 




CHICK iPPBOdllMt B0JIU»1 

D(.] «».«su.«i> aWBUua'issoi. Q(>)l,l„«l. 

al.i).i«o«s D(.i) i«ir«.iiM 


SUtUIAR, Of tfODK {!00 .«r4i »r l.tt - und.rlt',. Uf^*) 

The objective of thia project Is tne atudy of aultlvarlate 
ratios or proportions. 



77 



■ 



s«s. 


S'fSS'.!!"!"!™ ."'".T' 


,iI2isiS!c. 


PHOJ 


201 Cr 00039-05 LSM 


""'Vto 


er 1. 198) through September 30, 1982 




TITLE Of Pfl 


Jjty (.0 .h.r.cl.r. .r I,.,| 




Lloea 


r Methods in statistics 




7ilu\Z 








""="""■ ="""• °" '"' '""'" 








^'•/•VMr 


atory of Statistical and Mathematical Methodology 




SfCIl* 


stical Methodology Section 




DCRT. NIH. Bethesda. Maryland 20205 


■ 0,.L^..«.£..S. j^P,OFESS,l».l: |O.KS«, 


D (.) K»«« 


RIAT£ eox(ES) 
SUiJECIS DO) MU.UMISSUES D (•) 


EIT-E. 


D(.l)«l«« 


S (.!) INIES.IEUS 












methods in statistics are applied to biomedical 


data 









Linear Methods in Statistics 

Linear methods in statistics, as applied to biomedical 
data analysis, continue to be studied. Additional 
results were obtained on statistical and algebraic 
independence of random variables. A study of linear 
methods in variance component estimation was 
undertaken and a paper submitted for review and 
publication. Further, a new procedure was developed 
to treat unbalanced multivariate analysis of variance, 
and was submitted for review and publication. The 
method allows well-specified, rigorous tests of 
ANOVA models in the context of unequal cell 
sample sizes in the design layout; this is the most 
commonly occurring kind of data in the multivariate 
setting. The method was applied to schistosoma 
mansoni resistance to re-infectivity and the results 
were submitted to the Journal of Tropical Medicine 
and Hygiene as well as to a biostatistics journal. 

Publications: 

Grimes, A. M., Mueller, H. G., and Malley, J. D.: Examination of binaural 
amplification in children. Ear and Hearing 2: 208-210, 1981. 

Malley, J. D.: Simultaneous confidence intervals for ratios of normal means. 
J. of The American Statistical Association 77: 170-176, 1982. 

Malley, J. D.: Statistical and algebraic independence. Annals of Statistics (in 
press). 



Non-numerical Programming Techniques and 
Applications 

Several applications of non-numerical programming 
techniques were pursued during the year. These 
projects involved use of the PROLOG and REDUCE 
computer languages, and the Unified Generator 
Package, and were mostly in the general area of 
computational linguistics. The two projects 
accounting for the most effort were an investigation 
into automated analysis of instructional text, and 
research on automatic interpretation of medical 
terminology in terms of the constituent morphemes 
of individual words. 



78 



Investigations into the development of a PROLOG 
program that can analyze material from an English- 
language textbook were continued during this 
reporting period. A program capable of analyzing 
paragraphs describing how to form legal BASIC 
expressions was extended into one capable of 
analyzing a discussion of the LET statement of the 
BASIC language. This extension involves the 
simultaneous treatment of syntactic and semantic 
considerations about the subject matter being 
'comprehended' by the program. As a result of 
analyzing the textual material, the PROLOG program 
assimilates enough knowledge not only to parse, but 
also to interpret a LET statement. A detailed report 
has been written about this research, emphasizing 
the interconnections between syntactic and semantic 
concepts. 

A second project in the area of computational 
linguistics involved the creation of a PROLOG 
program to analyze medical terms denoting surgical 
procedures in terms of their constituent morphemes 
(e.g., 'cysticolithectomy' is 'excision (-ectomy) of 
calculus (-lith-) from cystic duct (cystico-).' Six 
classes of surgical procedures were considered, an 
existing lexicon of morphemes was substantially 
enlarged, and the program/lexicon combination was 
tested on over 1 ,500 terms, taken mostly from 
medical dictionaries. Over 75 percent of the terms, 
including nearly all of those commonly used, can be 
interpreted automatically. The remainder presumably 
would have to be listed in a lexicon of whole words 
('full forms'). 

Other efforts during this reporting period included the 
formatting and editing of a data base of surgical 
pathology summary diagnoses, which were then 
indexed using the Unified Generator Package to 
obtain frequency data on the terms used in the 
diagnoses. These and similar frequencies will be 
used in a forthcoming LSM project on word 
frequency distributions. For another project, the 
symbolic algebraic manipulation language REDUCE 
was used to assist in factoring a sixth-degree 
polynomial of interest in the theory of iterated maps 
on the unit interval. 

Publications: None. 





K'-a-rKSiKsr* 


"'"^"-^'wE^'"''"' 


ZOI a 00017-0*1 LSH 


'"""oJlUbCT 


1. 1981 through Septenbvr 30, 1982 




wlcal Programing Techniques and Appllaatlons 


P.l. 


n WW ihtmoii un\.\iu»i, uo iiiui of mmcim imrESTieAieH uo ui. otwr 
H.G. Pacak Computer Systems Analyst LSM OCRT 


None 


ti (it .-,) 


Laboratory of SUtlstleal •!(] WtHeaatleal Hethodology | 


""Blo.ath 


ewatics and Computer Science Section 


"""'ocbtTn 


IH. Bethesda, Maryland 20205 


0.6 
J (.) «»<«. 1 


""""'0*6' P"' 




Several 
compute 


R (iOO .grO* or tf)» - und.rli'<* Mr.8,^») 

applications of non-numerical proitrannlnK technlauea wi>r«> ntir.iiicd 


the year. TTieae pr 
r languages, and th 


^SSJnirclneraSr ?IcJa'eriKrJ?e Sf! 


putatlonal UnKulstlcs. 


The two proj^ta 1 


s^of^^nsSucUoial^SK^.'^nd ^searcS^S'ai 


n into automated 




of medical ternlnology In terms or the constituent morphemea of Individual 



79 



Topics in Geometry and Analysis 

The project objective is to develop mathematical and 
computational techniques using geometry and 
mathematical analysis, and to apply such methods to 
problems of biomedical research and computer 
science. 

An algorithm for high accuracy identification and 
description of protein spots in two-dimensional 
electrophoretic gels was developed and coded. 
Testing of the program has begun. 

In order to study convex cones a parametrization of 
all (non-isomorphic) N-algebras was studied. To 
determine smoothness properties of this 
parametrization and for its intrinsic interest a related 
set of Lie groups endowed with a left-invariant 
Riemannian geometry was studied. 

At this point, the eye is the most accurate and 
effective detection device for protein spots in two- 
dimensional electrophoretic gels. The edges of the 
spots can generally be well fit by parabolic 
segments. An algorithm has been developed to 
model parabolic fitting by the eye utilizing a 
'parabolic spacial second derivative' and other 
analogues of cues used by the eye. This has been 
coded as a Pascal program on a VAX computer. 
(This work is in collaboration with LGCB, NIMH.) 



By developing alternate characterizations of some of 
the axioms of an N-algebra and studying N-algebra 
isomorphisms in these terms, a parameter space for 
the algebras can be constructed as the intersection 
of hyperplanes and a sphere in an appropriate vector 
space. A unique (with respect to N-algebra 
isomorphism class) parametrization space is then 
obtained as the quotient of the subset of the vector 
space by a tensor product of two lower dimensional 
orthogonal groups. By evaluating canonical 
Riemannian geometry of a cone of generalized 
positive definite symmetric matrices with respect to a 
particular field of bases, one uses the standard 
diffeomorphism from the triangular group of a T- 
algebra onto its related cone to endow the Lie group 
with a left-invariant Riemannian metric that is 
isomorphic to the geometry of the cone. This 
enables it to be seen that the parametrization is in 
fact polynomial. 

Publications: 

O'Connor, M. A.: Invariant metrics on cones. Proc. of the Conference on 
Invariant Metrics and Hoiomorptiic h/laps, Rome, Italy, Istituto di Alta 
Matematica F. Sever! di C.N.R. (in press). 



p!!i«?"(S.tS't!JMT'™Ji: =S=":5" 


«E."i;5'.rS"'sES.cts 


»0JECI <m> 


" 




,„,u«rj;ili ^atc 


201 a 


00079-02 LSH 


October 1, 1981 through September 30. 1982 | 


liTUS Of PSOJECT (•» ■h.r.i..r= .r I,,-.) 


1 


Topics in Geometry and Analysis 




"™F";stj;;j*"Koi!rEL'iJjKE"oirTiE'™i°j"CT "" '""* " ""'"'"■ """""""' 


m .IL OIHE« 


P.I. M.A. O'Connor staff Fellou LSH OC 




Other: C.R. Nerrll Senior Hesearch Scientist LGCB «I 




Uboratory of General and Comparative Biochemistry 


'■"/BtygHratory of Statistical and Mathematical Methodology 






Biomathematics and Computer Science Section 




DCRT. NIH, Bethesda. Maryland 20205 


,<„.l,j.^,E.,S, |™.F£SSJ»^E, 


CHECK APPHOPRIME 80K(ES) 




D |.) H».m SC.JECIS a W «U«» IISSUES D (c) »EIIME. 




D1.1)«I«0«S a (.2) |,„E«.I£«S 








«n algorithm for high accuracy Identification and descrlptio 




spots in two-dimensional electrophoretic ael s was developed 




In order to study convex cones a parametrization of all {nnn- 


isomorphic) 











80 



^ 



Data Management 
Branch 



J. Emmett Ward, Chief 



Clinical Research, Patient Care, 
Epidemiology 

Clinical Support Section. During this past fiscal 
year a number of projects were developed and 
successfully completed: 

• the development of a set of programs to 
translate and format MIS Raw Purged Data 

• the development of a system of programs for 
preprocessing MIS Purged Data 

• the design and development of a set of 
programs to build and update the integrated 
data base 

• the writing of programs to translate and 
preprocess clinical laboratory data for the 
integrated data base 

• the development of a procedure to automatically 
transmit chemistry and hematology data from 
the CIU data base to the PDP-10 System 

• the transfer of data sets from mountable disk 
packs to mass storage, and 

• the modification of a number of programs to 
support maintenance and retrieval of clinical 
data for the integrated data base. 

BRIGHT STAT-PACK. Brian Cole, Jeanne Grillo 
(DMB/SAS); David Rodbard, Peter Munson (NICHD/ 
BES); Jay Shapiro (CC). A computer system has 
been developed on the DECsystem-10 that enables 
Clinical Center investigators to analyze their own 
clinical data. Available thus far are t-test, basic 
statistics, weighted linear regression, chi-square, 
frequency distributions, and high-quality basic 
graphics. The system provides a convenient method 
of obtaining clinical data from the NIH Clinical 
Information Utility, and interfaces with both MLAB 
and SAS. 

Effects of Bromocriptine on Schizophrenic 
Patients. Diane Feskanich (DMB/SAS); Neal Cutler 
(NIMH/BP). Statistical analyses were performed to 
study the effects of Bromocriptine on physiological 
and psychological variables. Camera-ready graphs 



were produced, from which slides were made to 
illustrate a presentation by Dr. Cutler. 

Survival System. Diane Feskanich (DMB/SAS); 
Ardyce Asire (NCI). This life table analysis system 
was originally developed in the 1960's to support the 
End Results in Cancer studies of NCI. Maintenance 
and improvement of the system is now the primary 
goal. The system has been sent to tumor registries 
and hospitals both in the U.S. and elsewhere. During 
FY82 the system was expanded to handle more than 
two racial groups. 

Prevalence of Major Neurological Diseases: 
Nigeria. Diane Feskanich, Jeanne Grillo (DMB/SAS); 
Bruce Schoenberg (NINCDS/NS); Dr. Osuntokun 
(University of Ibadan). This WHO-sponsored study 
consists of four parts: (1) census and health screen, 
(2) evaluation of risk factors, (3) neurological exam 
results, and (4) followup. A pilot study was done for 
Part 1 to determine validity and usefulness of 
questions and goodness of the questionnaire. During 
FY82 a new pilot was begun, based on information 
from the FY81 pilot. New forms were developed, and 
edit and update programs were written. 

Alcohol and Memory. Mary Lee Dante (DMB/SAS); 
Elizabeth Parker (ADA). Test results from four types 
of tests measuring the performance of male college 
students under the effects of alcohol were 
computerized and scored, and a data base suitable 
for analysis was created. 

Cerebral Palsy/Neonates. Diane Feskanich (DMB/ 
SAS); Tatiana Kudrjavcev (NINCDS/NS). Ms. 
Feskanich received tapes of birth certificates and 
neonatal and fetal death certificates from the 
University of Rochester. During FY82 focus was on 
looking at low birthweight for age as a predictor of 
cerebral palsy in the newborn. Frequency 
distributions and counts were produced for this 
study. 

Nutrient Data Base. Diane Feskanich (DMB/SAS); 
Elaine Offutt (CC/NUTR). The USDA nutrient data 



83 



base was obtained and programs were written to 
calculate nutrient intakes from daily foods records for 
research studies. A comparison of the numerous 
nutrient data banks available for purchase is being 
conducted to determine the one best suited for use 
in an interactive setting by Clinical Center 
nutritionists. 

Combined Cardiology/Heart Surgery Data 
System. Larry Martin (DMB/ASPS); Roger Dailey 
(DMB/DBAS); C. Mcintosh, D. Rosing (NHLBI). This 
combined system provides a chronological record of 
the medical activity of NHLBI Cardiology and Heart 
Surgery Branch patients. In FY82 effort was directed 
toward meeting the routine and ad hoc reporting 
requirements and statistical needs of the NHLBI 
physicians and researchers. The system was 
expanded to include a graded exercise form and a 
physical examination form. A program and command 
processor were written to produce assorted 
bibliographies of pertinent papers and publications. 

NIADDK Study of the Incidence and Prevelance 
of Kidney and Urinary Tract Diseases in the 
Armed Forces. Darius Georg (DMB/ASPS); N. 
Cummings (NIAID). This study is being conducted to 
evaluate the research needs in this area and to 
correlate the research needs with the occurrence of 
morbidity and mortality of the disease. This system is 
currently in the raw data evaluation stage. 

Sleep Study System. Darius Georg, Peter Basa 
(DMB/ASPS); Christian Gillin (NIMH/BPE). This 
system is being developed to provide a 
computerized method for scoring sleep data. 
Analysis and design have been completed. Programs 
to edit, update, and report on the data have been 
written. Other information such as drug 
administration and behavioral and clinical ratings 
may be added to the system at a later date. 

Psychobiology Patient Information System. 

Dennis George, Steven Soroka (DMB/ASPS); Frank 
Putnam (NIMH/BP). The purpose of this project is to 
condense a large amount of data for a small number 
of patients into a format that is useful for research 
analysis. During the last year the data was converted 
into a fixed format and programs have been written 
to produce reports in both hard copy and graphic 
form. This data was extracted from the CC/MIS data 
base. 

Analysis of SLE Nephritis Patient Data. George 
Shakarji (DMB/OC); John Klippel (NIADDK). 
Collection of patient therapy data and 
implementation of our computer storage and retrieval 
system was begun. Complete chemistry and therapy 
data has been stored on over 1 00 SLE (Systemic 
Lupus Erythematosis) nephritis patients. Investigators 



have had ready access to the data base to get up- 
to-date information and analysis on trends in patient 
progress. 

Forecasting Trends of SLE Nephritis Patients. 

George Shakarji (DMB/OC); John Klippel (NIADDK). 
Trend and forecasting analysis systems are being 
implemented to detect and forecast relationships 
among three groups of SLE patients, namely: those 
who are on dialysis, those who have doubled their 
creatinine, and a control group. Results from DNA 
Binding, Serum Creatinine, C3, Serum Albumin, Qual 
protein, RBC/HPF, and Hemoglobin are being 
analyzed for correlational relationships among these 
data. It is hoped that analysis will indicate and 
predict certain long term outcomes based on the 
relationships of these data in the three groups. 

Subject Specific Reference Regions for Blood 
Chemistry Data. George Shakarji (DMB/OC); 
Eugene K. Harris (DCRT/LAS). This study is a part 
of the continuing studies on defining reference 
regions, both univariate and multivariate, as applied 
to subject specific variability in clinical chemistry 
results for blood. Programs were written and 
completed to examine both theoretical and empirical 
properties: first through computer simulation, then by 
application to serial clinical assays collected over a 
long period of time (seven to nine years). 

Analysis of Means and Variances of Chemistry 
Data in Normal Subjects. George Shakarji, David 
VanSant (DMB/OC); Eugene K. Harris (DCRT/LAS). 
A package was designed and generalized to analyze 
serial measurements of analytes for within and 
across subjects. This package computes tests of 
normality for within-subject data, computes serial 
correlations and moments for each individual's data, 
and then proceeds to perform analysis of variance 
and covariance over all the subjects. 

Dyslipidemia Computerized Recordkeeping 
System. George Roberts (DMB/SAS); Ernst 
Schaefer (NHLBI/DMB). This system keeps records 
on clinical laboratory data for normal and 
dyslipidemic subjects and provides for routine 
reporting, ad hoc queries, and preparation of 
selected subfiles for statistical analysis. During FY82 
work was begun on nutritional studies of these 
patients' diets. The analysis of the effects of 
Neomycin on cholesterol levels was also begun. 

Penicillin Study. Vivian Pelham, Charles Twigg 
(DMB/ASPS); Dorothy Sogn (DIR/NIAID). This 
system is being developed to collect data and 
provide reports from the clinical trials of skin testing 
with major and minor penicillin derivatives in 
hospitalized adults. Analysis and design were 



84 



completed during the year and the system is being 
implemented. 

Laboratory Investigation 

Smithsonian Tick Collection Query/Retrieval 
System. Diane Feskanich (DMB/SAS); Carleton 
Clifford, Jim Keirans (NIAID/RML). The Rocky 
Mountain Lab has catalogued their tick collection on 
tape at the Smithsonian Institution. DMB is supplying 
the ability to update and query this file from Montana 
using the DCRT central computer facilities. During 
FY82 Ms. Feskanich installed interactive programs 
for data entry on the DataPoint word processor in 
Montana, ensured that the software would interface 
with DCRT software, and trained RML personnel in 
use of the software. She has also been acting as 
intermediary in the data flow between RML and the 
Smithsonian. 

Monkey Management System. Diane Feskanich 
(DMB/SAS); Robert Williams (NICHD/PRB). A data 
base of the bibliographic and experimental history, 
plus the current medical status and experimental 
protocol for each monkey, was built. Programs were 
developed for data entry, editing, updating, and 
reporting. The system is being used to select 
appropriate individuals for specific experiments, and 
to prepare daily work assignments for caretakers and 
technicians. Future enhancements will include 
special reporting facilities. 

Graphics. George Roberts (DMB/SAS); David 
Rodbard (NICHD/BES); Doris Wallace (DRG/RAE). 
Mr. Roberts has provided assistance to NIH research 
scientists and administrators who are learning how to 
use the new graphics capabilities. During FY82 he 
also began the development of a front-end package 
that will be able to accept existing sequential data 
files, apply user requested transformations, and 
produce TELL-A-GRAF data files. The user will be 
able to generate his own TELL-A-GRAF data file by 
supplying the appropriate algebraic equations in 
Fortran notation. 

Molecular Modeling. Sigurd Knisley (DMB/SAS). 
During FY82 Mr. Knisley has been working on 
modifications to the shaded surface molecular 
display developed by Richard Feldmann and Tom 
Porter (DCRT). Currently available are a 
transparency option, which allows the viewer to see 
internal features or contact interfaces between 
molecules, and a variable illumination angle, which 
improves three-dimensionality. 

Seroepidemiology Data Processing System. Judy 
Mahaffey (DMB/ASPS); Paul Levine (NCI). The 
Clinical Studies Section, NCI Laboratory of Viral 
Carcinogenesis, is trying to find characteristics of 



serum samples that can be used to predict cancer. 
To this end, a computer system has been designed 
to manage all data necessary for efficient inventory 
control, test results feedback, and statistical analysis. 
The system is now operational and reports from the 
system are being sent to collaborating scientists in 
the U.S., Ghana, Greenland, and Singapore. During 
the past year a new contractor took over the running 
of this system. DMB provided assistance in setting 
them up to correctly run the system. 

Primate Colony Carcinogen Study. John Parks 
(DMB/ASPS); Susan M. Sieber-Fabro (NCI/DCT). 
The purpose of this project is to develop a system to 
maintain and search data generated from a colony of 
approximately 2,000 nonhuman primates. The 
system as originally requested became operational 
during the last year. Future enhancements will be 
added as requested by user. 

Canine Breeding Colony Data Processing 
System. Peter Basa (DMB/DBAS); Dennis George 
(DMB/ASPS); T. Wolfle (DRS/VRB/ACS). The goal 
of this project was to develop a system to assist the 
Veterinary Resources Branch, DRS, with its 
recordkeeping and work scheduling. The system is 
complete. DRS is now in the process of installing a 
word processor system (CADO) in Poolesville to 
handle all data entry, maintenance, etc. When this is 
complete, DMB will work on interfacing the two 
systems. 

Strain Specificities Reference System Steve 
Soroka (DMB/ASPS); David Sachs (NCI). A 
computer system is being developed for the Division 
of Cancer Biology and Diagnosis, NCI Immunology 
Branch, to assist in transplantation biology research. 
The system will be used to help locate existing 
cogeneric mouse strain products and/or to design 
mouse strain products having specific antigens that 
are used in experiments relative to the development 
of sera. The project has been temporarily suspended 
until new coding schemes and structures are 
developed and implemented by the sponsor. 

Finite Element Package. David VanSant (DMB/OC); 

Warren Pince (NIEHS). This is an easy-to-use finite 
element program for solving a large class of elliptic 
(steady state), parabolic (time dependent), and 
equivalent partial differential equation problems in 
general two-dimensional regions. This package has a 
preprocessor program that allows the user to supply 
the problem description in a greatly simplified form 
so that no knowledge of FORTRAN is required. 
Graphical output can also be produced. Scalar, 
vector, and stress fields can be displayed via the 
Calcomp plotter. 



85 



Program Management and 
Administration 

Administrative Data Base (ADB). Marvin Katz, 
Ron Wicks (DMB). This ongoing administrative 
project utilizes data base technology in support of 
NIH-wide materiel and financial management. As the 
Materiel Management System (MMS) entered its fifth 
year of development and operation, much time was 
spent in enhancing existing software. During FY82 
some 50 change control items successfully went into 
production. Several new developmental efforts were 
implemented: 

1 . Requirements analysis for development of a 
data base Financial Management System (FMS) has 
been completed. It is anticipated that development 
of the FMS using contractor personnel will 
commence in late FY82. 

2. Deployment of the delegated interface to MMS 
in the B/l/D Administrative offices continues. By the 
end of FY82 this effort will be essentially completed. 

3. The development of the stock inventory system 
is proceeding. This system will be phased into 
production during late FY82 and FY83. 

4. A vendor credit capability has been added to 
the accounts payable system. 

5. An ability to search the NIH vendor data base 
using alphabetic names directly at the ADB terminal 
has been added. 

6. The procurement system has been extended to 
include entry of DFM miscellaneous obligations and 
training orders. The online production of SF-147's 
has been added for reprints and for personal service 
contracts. 

7. The requirements for open market requisition 
processing are being reviewed. 

Full-Time Equivalency. Dennis George, Mike Letke 
(DMB/ASPS); George Roberts (DMB/SAS); John 
Hartinger (NCI/FMB). A system was built for 
monitoring the ceiling levels and full-time equivalency 
manyears for NIH. Data input, update, and query 
facilities are available, and report programs usable 
by all B/I/D's and by Central Budget have been 
provided. 

DRR Grants Subproject System. Vivian Pelham 
(DMB/ASPS); Jean Babb (DRR). The existing DRR 
Grants Subproject System that used CPS was 
evaluated. A proposal was made for the redesign of 
this system to make use of more current, 
supportable technology. The proposal was accepted, 
and the system was developed and turned over to 
the user during the past year. 

NIH Nutrition Grants Monitoring System. Judy 
Mahaffey (DMB/ASPS); Thomas VogI (OD). A 
system has been designed for the NIH Nutrition 



Coordinating Committee to assist them in monitoring 
and reporting data on biomedical and behavioral 
nutrition research at NIH and at other agencies 
within DHHS. The system is operational and Dr. 
Vogl's office is currently using it to answer inquiries 
from NIH directors' offices, the White House, fl 

Congress, and the public as they relate to dollar V 
amounts and percentages of grant money being 
spent in the area of nutrition. This is an ongoing 
project with the data base being created each fiscal 
year. 

Review and Evaluation Branch Grants 
Information System (GENIUS). Penny Brogan 
(DMB/ASPS); Harry Canter (NCI). The computerized 
Research Analysis and Evaluation Branch 
Information System, a highly specialized system, was 
designed and implemented for the Division of Cancer 
Grants, NCI. The system provides information on 
grants, contracts, intramural projects, and unfunded 
grants. The grants and contracts systems are 
'generalized' so they can provide information from 
any NIH Institute. In the future, a Training Grants 
system will be developed, and history file 
maintenance must be added to the intramural 
projects and unfunded grants systems. Additional 
programs must be written to pick up more contract 
information from the NCI-CMS system. 

NIH International Activities and Personnel 
Monitoring System. Penny Brogan (DMB/ASPS); 
Libby Low (FIC). A system provides the Fogarty 
International Center with the ability to maintain and 
query a data base with information on foreign 
scientists who are in the U. S. to perform health 
research. Most of these scientists are working at 
NIH. The system provides query capability as well as 
regularly scheduled preprogrammed reports. The 
existing system is being revised to include more 
accurate dating and editing capabilities, new reports, 
etc. Fourteen programs of the revised system are 
complete and operating. About sixteen programs still 
need to be written. 

Employee Health System and Accident Reporting 
System. Vivian Pelham , Steve Soroka (DMB/ 
ASPS); Julio Rivera, John Leach (ORS/S). A system 
is being developed to combine the employee health 
and accident reporting systems. Analysis and design 
has been completed and the system is currently 
being implemented. 

Committee on Academic Science and 
Engineering (CASE) Reports. Darius Georg (DMB/ 
ASPS); J. Bailey (OD/OPPE). This project involves a 
broad spectrum of data processing support required 
for the collection and reporting of DHHS obligations 
to institutes of higher education, research and 



86 



development centers, and nonprofit institutions. This 
is an ongoing project. 

MMS Query and Reports. Jane Blessley (DMB/ 
ASPS). This project is intended to provide an 
economical method for the selection and reporting of 
data from the NIH Administrative Data Base. Ms. 
Blessley provides recurring and ad hoc reports from 
the data base for all segments of the NIH 
community. During the past year she trained and 
turned over the responsibility for this project to 
another unit of DMB. 

System for Controlling and Monitoring 
Complaints of Discrimination at NIH. Darius Georg 
(DMB/ASPS); G. Yee, M. Williams (OD/DEO). This 
project establishes and maintains a file that provides 
statistical data, on a case-by-case basis, of formal 
and informal complaints of discrimination at NIH. In 
the past year Mr. Georg revised and simplified the 
retrieval process. 

System for Classifying NIH Research and 
Development Awards. Darius Georg (DMB/ASPS); 
William Rhode (OPPE/RA). The objective of this 
project is to test the feasibility of and then develop a 
computer system based on CRISP index terms for 
categorizing by basic research, applied research, and 
development and to show percentage distribution of 
dollars associated with each category. If the system 
proves feasible, the data will be used to prepare 
annual reports to the Office of Management and 
Budget (0MB) and the National Science Foundation 
(NSF). 

ARMS/TDCS Interface (TAPS). Dennis George 
(DMB/ASPS); B. Hughes (OPA/P); A. Amatucci (OA/ 
M). This project is intended to create an NIH 
Personnel System that is a composite of the current 
NIH personnel system (ARMS) and the DHHS 
Personnel System (TDCS). The system was 
completed and turned over to the Office of the 
Director/Systems and Action Branch during the year. 

Radiation Safety Control System. Charles Twigg 
(DMB/ASPS); R. Zoon (DRS/RSB). This system is 
designed to monitor the use and users of radioactive 
isotopes at NIH. When complete, this system will 
include five subsystems. They are: (1) Inventory and 
Bioassay, (2) Lab Survey and Airborne Release, (3) 
Waste Processed and Activity Balance, (4) Training, 
and (5) Film Badges. In the past year, online 
collection and update of Form NIH88 was tested and 
implemented. Development of the lab survey and 
airborne release subsystem was begun. All 
subsystems except the Waste Processed and Lab 
Survey have been completed. 

Electrical Safety Program System. Larry Martin, 
Steve Soroka (DMB/ASPS); Howard Metz (DRS/ 



BEIB). The chief of Scientific Equipment Services of 
the Biomedical Engineering and Instrumentation 
Branch has requested a system to help monitor 
maintenance of equipment at the Clinical Center. A 
system is being designed to computerize the results 
of routine electrical safety checks and of preventive 
maintenance performed on hospital equipment. The 
system will be used by DRS to schedule equipment 
checks, to provide reviews on instruments checked 
by contractors and by the Clinical Center, and to 
provide statistical information on repair histories of 
different types of equipment. The system was 
completed and turned over to the user during the 
year. 

Interferon Production Monitoring System. Dennis 
George (DMB/ASPS); Hilton Levy (NIAID/LVD). The 
purpose of this project is to develop a system to 
monitor the production and subsequent use of 
interferon on an experimental basis. Various 
production techniques and use protocols are to be 
monitored in both human and animal subjects. Initial 
analysis and design have been started. 

Information System of Extramural Scientists. 

Darius Georg (DMB/ASPS); William Rhode (OD/ 
OPPE). This project involves the creation of a data 
base drawn from various sources to perform analysis 
of patterns of involvement in NIH science review 
activities by extramural scientists. The data base has 
been created and reports are being run as 
requested. 

Medical Records Auditing System. Judy Mahaffey 
(DMB/ASPS); Gloria Burich (CC/MRD). The purpose 
of this system is to assist the Medical Records 
Department in the monitoring and reporting of the 
status of medical records from the time they enter 
the department until they leave. When the system is 
developed it should replace four manual systems 
now being used. The work-in-process portion of the 
system was completed during the past year. Work is 
now focusing on the 'audit' segment of the system. 

AIRS Personnel System. Steve Soroka (DMB/ 
ASPS); L. Lee Manuel (DCRT/OD). This project will 
involve a complete revision of the system due to the 
availability of the new TAPS file. Analysis, design, 
and implementation are currently in progress. 

Biomedical Communications 
Applications 

Selective Dissemination of Information. Sigurd 
Knisley (DMB/SAS). SAS has continued its support 
of the current awareness search for both Chemical 
Biological Activities (CBAC) and the Biosciences 



87 



Information System (BIOSIS). Retrospective 
searches are referred to the NIH Library staff. 

Editorial Data Base Management System. Brian 
Cole (DMB/SAS); Judith Prewitt (DCRT/OD). A 
system is being built using the interactive capabilities 
of the DECsystem-10 that will allow professional 
journal editors and conference chairmen to track 
information on paper submission and refereeing. 

Bibliographic Data Base. Sigurd Knisley (DMB/ 
SAS); Curtis Harris (NCI/DCCP). Bibliographic 
information and keywords drawn from Dr. Harris' 
reprints of scientific articles were entered into the 
computer files. A system of searching this 
information and printing it for direct inclusion into 
book and journal bibliographies was set up using the 
powerful new tools available in WYLBUR. 

Chinese Personalities and Institutions in 
Biomedicine. Judy Mahaffey (DMB/ASPS); Joseph 
Quinn, Joseph Lee (FIC). Due to a rapid increase in 
international exchanges in the field of biomedicine 
between the U.S. and the People's Republic of 
China, the Fogarty International Center has 
requested DMB services to design a system for the 
computerization of data on biomedical scientists and 
institutions in the PRC. The system will be used by 
the FIC officials in briefing NIH and non-NIH 
scientists interested in biomedical research in China. 

Computer Research and 
Technique Development 

SFOR (Structured FORTRAN) Compiler. Bob 

Magnuson (DMB/OC). The SFOR compiler, which 
generates block-structured IBM FORTRAN source 
code, was further enhanced to assist programmers' 
writing structured programs. There are six different 
kinds of blocks available to the FORTRAN 
programmer-CASENTRY, FOR, IF, LOOP, REPEAT, 
and WHILE. 

RMAG Products Support. Bob Magnuson (DMB/ 
OC). Necessary support is provided for RMAG, SLR, 
Logic Subroutines, Arithmetic Subroutine, SLANG, 
Voice Input, and SFOR. This ongoing support 
includes software maintenance, customer 
assistance, and the teaching of formal DCRT 
courses on the use of these products. In particular, a 
special effort had to be mounted to change over to 
the new WYLBUR format data sets. 

PDOC: Program Documentation System. Bob 

Magnuson (DMB/OC). PDOC is a tool used to 
document programs. It is a front end to the WYLBUR 
Document Formatter, allowing the users to employ 
all of the Document Formatter's powerful features, 
while adding several useful enhancements of its 



own. The PDOC system generates good-looking 
boxed comments that really stand out in explaining a 
program. PDOC permits symbolic referencing- 
forward or backward~of the generated code line 
numbers or of the generated document section 
numbers. There is a command for underlining. 
Various kinds of heads are created within the 
generated document, and are placed automatically 
within optional tables of contents. PDOC has an 
INCLUDE command, which permits inclusion of 
different files as parts of the PDOC source. A PDOC 
file consists of a source program to be documented 
to which you have added interspersed PDOC 
commands and WYLBUR Document Formatter 
commands. Hence, both the program and its 
documentation can be maintained within the same 
PDOC file. When running PDOC, the user can elect 
to have all or part of the contained code extracted 
and placed into an active file. There it can be 
compared, compiled, tested, saved, or run. 

CP Tools. Bob Magnuson (DMB/OC). CP Tools is 
an integrated set of WYLBUR command procedures. 
Stored as members of a partitioned data set, only 
one of these tools need be accessed by a single 
user-defined command. When that tool is executed, 
it selects the tool actually wanted, passing along any 
included keyword parameters (taken from the 
command line argument). The tools include an 
NIH7000 editor that permits TV editing' of the user's 
data sets, combined with all of WYLBUR editing, 
plus 'token editing' (for changing variables or 
keywords without affecting other parts with identical 
substrings). Another tool gives online help on the 
various tools. Other tools include JCL generators for 
running SLANG and SFOR, as well as for 
microfiching any number of hold jobs. There are 
tools for setting the NIH7000 tabs as well as its PF 
keys to whatever the user wants. One of the tools 
formats text into TYPE commands with NIH7000 
screen underlines, thus simplifying the task of 
creating screen help messages. 



I 



Computer Center 
Branch 



Joseph D. Naughton, Chief 



Summary of Projects 

New Software. 

TVEDIT, a powerful full-screen text editor, was made 
available on the DECsystem-10. TVEDIT provides an 
up-to-date, visible copy of the portion of the file 
being edited on the terminal screen. Nearly all of 
TVEDIT's commands are issued using labeled keys 
on the NIH7000 terminal keyboard, eliminating the 
need for memorizing commands. Editing changes 
are shown immediately, and the cursor may be 
moved to any location on the screen, making it 
possible for the user accurately to construct tables, 
diagrams, and flow charts. These features give 
TVEDIT remarkable versatility and ease of use. 

POSTER was developed to fill the need for an easy 
way to prepare slides, posters, and transparencies of 
textual material for use as visual aids at scientific 
meetings and symposia. In the past, titles, captions, 
and summaries had to be hand-lettered or typed, 
photographed, and then enlarged. POSTER 
eliminates these steps and prepares posters directly 
or produces high quality copies for slides and 
transparencies. A variety of formatting capabilitites, 
24 typefaces, numerous specialized symbols, and 
commands for underlining and generating sub- and 
super-scripts are all available in POSTER. 

The entire online data storage facility of the NIH 
Computer Utility was completely redesigned during 
the year. The installation of some 224 actuators IBM 
3380 Disk Drives, together with the use of the 3850 
Mass Storage System, provided a tremendous 
increase in capacity permitting the online storage of 
both larger and greater numbers of user data bases 
in a more reliable and cost effective manner. Three 
new data storage categories-Open, Controlled, and 
Dedicated-provide online data storage facilities for 
data sets ranging from a few bytes to more than 100 
million bytes online DASD. The Mass Storage 
System provides online storage for data bases 



exceeding this size. After extensive internal testing 
by the Computer Center, a plan to transfer all active 
user data sets to the new online storage structure 
was developed. Using a phased parallel approach 
with an elaborate back-up/recovery procedure some 
300,000 online data sets were transferred during the 
last six months of the year with no interruption in 
service to users. The transfer will be completed 
during the early part of FY83. When complete, the 
NIH Computer Utility will have more than 141 billion 
bytes of online DASD data storage capacity available 
to users. It is anticipated that this, together with 
availability of some 236 billion bytes of data storage 
space on the Mass Storage System, will be 
adequate to meet users' needs for several years into 
the future. 

Output Facilities. This past year saw the 
retirement of SPOUT, a facility that was originally 
designed to handle nonstandard output in an offline 
fashion. Improvements in the JES2 portion of the 
operating system and the use of high quality printing 
produced by the versatile 3800 laser printers made it 
feasible to allow the use of nonstandard forms 
online. Appropriate software changes were 
implemented gradually, and by July all standard 
forms as well as user-supplied forms became 
available online. Eliminating SPOUT reduced both 
processing overhead and manual handling of tapes, 
thereby increasing efficiency and accuracy. Two 
new CalComp 1055 Plotters replaced the old model 
1036 plotters that had been in use since July 1977. 
The new plotters operate at almost three times the 
speed of the old ones and have the ability to plot in 
four colors. Better quality output and faster 
turnaround was provided to graphics users, at no 
increase in cost. 



91 



Communications Linlt. MERCURY, the 
communications link between the IBM System 370 
and the DECsystem-10, is vital to obtaining the 
maximum effectiveness from both systems. Several 
hardware and software improvements were 
implemented this year to enable the MERCURY 
program to keep pace with new developments on 
the IBM System 370. The most significant 
enhancement allows the DECsystem-10 users to 
take advantage of the IBM 3850 Mass Storage 
System for infrequently-used data sets. As the year 
comes to a close, new hardware was acquired to 
improve both the speed and reliability of the 
MERCURY link. 

Documentation and Publications. Providing 
current documentation for all services was an 
important challenge in this year of transition. 
INTERFACE continued to be the users' most 
important source of up-to-the-minute information. 
Eight issues and an Annual Index were published 
during the fiscal year. The Computer Center Users 
Guide also reflected the increased pace of change. 
For the first time it was completely revised twice 
within one year and required four updates to keep 
users informed of the most current and complete 
system technical details at all times. A new contract 
was established to make printing of the Users Guide 
more timely and reliable. 

Other publications are oriented toward familiarizing 
users of the Computer Utility with the services, 
languages, and training available. Seven new 
publications were released this year and 12 others 
were either revised or updated. 

User Training. A variety of intensive courses 
designed to acquaint the user with the computer 
services and languages available were offered 
through the Computer Center Training Program. The 
Training Program accommodated 2,244 students in 
190 sessions of 64 different courses during the year. 
Special effort was made this year to designate 
courses at elementary, intermediate, and advanced 
levels, in order to meet the diverse needs of the user 
community. 

Many extra sessions of the introductory WYLBUR 
courses were offered to make the capabilities of 



1 



New WYLBUR available to users. Other courses 
covered: the many computer languages used at NIH; 
the software packages available for statistics, 
computation, and graphics; and the operating 
systems of the IBM System 370 and the 
DECsystem-10. 

Those unable to attend regular classroom courses 
were able to choose from 26 self-study courses, 
ranging from introductory surveys of computers and 
programming to advanced discussions of the IBM 
System 370 and DECsystem-10 operating systems. 

Customer Assistance and Systems IViaintenance. 

Customer assistance has always been an important 
part of the Computer Center's services, and even 
more so in FY82. The number of Programmer 
Trouble Reports (PTR's) researched and answered 
during the year rose to a record high of 5,148. This 
was partly due to the new, easier-to-use PTR 
command developed for WYLBUR. Although the 
Programmer Assistance and Liaison Unit continued 
to operate under restricted hours, it recorded more 
than 37,500 calls or visits from customers needing 
assistance. 

Installation of new hardware and software on the 
IBM System 370 required significant changes to the 
operating system. The number of SYSGENS 
installed during the year rose to 106. Nearly 11,000 
'fixes,' both preventive and corrective, were tested 
and applied to the system, and 1 new releases of 
current software packages were installed. 



92 



Research Projects 

In addition to the many activities, services, and 
facilities it provides NIH, the Computer Center also 
serves biomedical computing with its research work 
in molecular graphics. 

Computer Representation of Virus and other 
IMacromolecular Assemblies 

Over the past five years, a computer raster display 
has been used to represent the surface structure of 
macromolecules. Shaded spheres are the primitives 
of the representation. In globular protein and nucleic 
acid structures, each sphere represents one atom or 
at times one amino acid. In representing viruses, the 
sphere primitive has been used in some cases to 
represent one whole protein. In other cases, the 
spheres are used to represent the shape of a portion 
of a protein. 

Symmetry of the viruses plays an important part in 
making the computer model look realistic. Once the 
shape of a protomer is modeled, it is then iterated 
over the (icosahedral) symmetry of the assembly. A 
model of limulus hemocyanin has been constructed 
starting from image-enhanced electron microscope 
data. Twenty spheres are used to represent the 
kidney bean shape of each of the 48 70,000-dalton 
proteins in the assembly. Fifty spheres were used to 
represent the low resolution shape of the muscle 
actin obtained from electron diffraction. 

The generalization of surface shape representation 
obtained using spheres as primitives indicates that it 
will be possible to model subcellular organizations by 
computer. 

Publications: 

Berzofsky, J. A., Buckenmeyer. G.K.. Hicks, G.. Gurd, F.R.N., Feldmann, 
R.J., and MIna. J.: Topographic antigenic determinants, recognized by 
monoclonal antibodies to sperm whale myoglobin. J. Biol. Chem. 257: 
3189-3198, 1982. 

Feldmann, R.J.. Potter, M., and Glaudemans, C.P.J. : A hypothetical space- 
filling model of the V-regions of the galactan-binding myeloma Immun- 
oglobulin J539. Mol. Immun. 18: 683-698, 1981. 

Pawllta, M., fulushinski, E., Feldmann, R.J., and Potter, U:. A monoclonal 
antibody that defines an Idiotope with two subsites in galactan-binding 
myeloma proteins. J. Exp. Med. 1981: 1946-1956, 1981. 





ZOt CT00089-01 CCB 


October 1, 1981 through Septsabcr 30. 1982 


Computer Rcpreaentatlon of Virus and other ttooroMleoular Aaseobllea 


P.I. Hicnard J. Feldaann. CCB, DCItT. Coaputer Spealillst 


CMfEN^riNC W.tl (II ..,) 


Coaputer Center Branch 




DCRT. HJH. Bethesda. HD 20205 


t01*L -ANnUHD. iPRaFESSIONM.! lOIKOi 


CMtCK iPCftOMUIt BOll[ES] 


Over the past five yeara, a computer raster display haa been used to represent 

spnere represents one atom or at times one amino acid. In representing viruses, 1 
the sphere primitive has been used In some cases to represent one whole protein. 

protein. Syiwnetry of the viruses plays an important part In making the computer 

Iterated over the Clcosanedral) symmetry oi the assembly. A model oi limulus 
hemocyanin has been constructed starting from Image-enhanced electron microscope 
data. 7\<enty spheres are used to represent the kidney bean shape of each of the 
U8 70.000-dalton proteins In tne assembly. Klfty spheres were used to represent 
the low resolution shape of the muscle actln obtained from electron diffraction. 
The generalization of surface ahape representation obtained using spheres as 
primitives Indicates that It will be possible to model sub-cellular organizations 
by computer. 



Office of the 
Director 



Arnold W. Pratt, M.D., Director 



Summary of Activities 

Library Automation. E. Chu (OD); J. Mahaffey 
(DMB); J. Knight (CSL). In conjunction with other 
DCRT staff, the DCRT Librarian applies computer 
techniques to DCRT needs, advises other libraries, 
and maintains knowledge of work done outside NIH. 
In FY82 work on this project has been limited to a 
few modifications on the existing systems. 

DCRT Publication File. K. Griffin, P.O. Miller (OD); 
R. Baxter (DMB). This ongoing project was begun in 
1979 to create a file of citations for all papers 
published by DCRT authors. In FY82 additional work 
was done to correct errors in the file; a usable 
product should be available in the first quarter of the 
next fiscal year. 

Electronic Typesetting Methods. P.O. Miller (OD). 
This project is an offshoot of work begun in 1979 as 
a part of a joint PSL/LAS/OD effort. It involves 
creating a specially-coded magnetic tape of 
WYLBUR text that can be used directly by 
computerized typesetting equipment at GPO. The 
technique has been shared with NIH and other 
Federal public affairs communities, through the NIH 
Printing Committee and the National Association of 
Government Communicators. 

DCRT Communications Program. P.O. Miller, W.C. 
Mohler (OD). Previously called the DCRT Information 
Program, this is an ongoing project to develop 
improved and coordinated communication 
techniques to support DCRT activities. It has four 
parts: Analyzing Needs, Creating and Evaluating 
Products, Developing Resources, and Education. In 
FY82 work continued on developing and distributing 
products. As the OMB-imposed moratorium on 
publications and audiovisuals continues, new 
methods are being explored for communication 
between DCRT and the various groups that have 
needs for information about its work. 

Clinical Data Management and Analysis. W.C. 
Mohler (OD); B. Cole (DMB); D. Rodbard (NICHD); 



J.R. Shapiro (Clinical Center). In spite of the rapid 
growth in use of data management and statistical 
packages provided for NIH scientists on DCRT 
computers, there is a perceived need for facilities 
that would be easier to learn and use in NIH clinical 
research projects. In FY82 work continued using 
BRIGHT, a table-oriented data management/analysis 
package on the DECsystem-10, developing added 
data analysis and display programs. Two 
presentations were made to medical staff fellows 
and other scientists with an interest in clinical data 
processing. Discussion began to examine the utility 
of a seminar series on clinical computer applications. 

Multi-function Microprocessor Interface. A.W. 
Pratt (OD); D. Songco (CSL). This project, begun in 
FY80, seeks to adapt a variety of information 
acquisition techniques on a single microcomputer as 
a versatile data input/output interface for biomedical 
scientists and clinicians. Work in FY82 is discussed 
in detail in the Computer Systems Laboratory report. 

Medical Linguistics. A.W. Pratt (OD), et al. This is a 
long-term project to define a set of semantic and 
syntactic forms that can aid in the analysis and 
interpretation of written medical statements. 



95 



Research Projects 

Electronic Typesetting Methods. 

Using the WYLBUR system, text is first collected and 
stored for publication production. Then, a magnetic 
tape of the data is furnished to GPO for direct 
typesetting input. Typesetting costs have been 
reduced eighty percent. 

The technique has been made available to others in 
the public affairs community. 

Future plans include exploration of interfaces to link 
equipment over telecommunications lines without 
hand-carrying tape, along with participating in the 
development of Gen Code, a proposed ANSI 
standard for encoding text for machine processing. 



S'S?.J." 



Oc tober 1, 1981 through September 30, 1982 



Electronic Typesetting Methods 



Graphic Syatems Devel op ment Divtsioi 



PrlntlnB Office 



Dethes d a, .MP_ 2020 5 



Using the WYLBUR system, text Is fii 
production. Then, a magnetic tape i 
typesetting input. Typesetting cosl 



stored for publlcatioi 
lished to GPO for din 
:d eighty percent. 

: public affairs 



96 



DCRT 



Ovwon o( Computer Research and Technotogy 
Nabond ^isMutes ol HeaWi 
Bethaada. Maryland 20206 



Division of 
Computer 
Research and 
Technology 



Fiscal Year t9a3 
Annual Report 
Volume 1 



iiksi 



RtMENTOfWAem- 
AND HUMAN SERVICES 
Public Health Service 
National Institutes o! Health 










iim%* ^ tfm\wmmMmm\ { \^ ^ 



MATHEMATICSMEDICALAPPLICATIONSSTATISTK 







EEOPERSONALCOMPUfERSSYSTEMSPOLICYAND 



l>ii#ioinn off Fiscal Year 1983 u s department of health 

l^lwlOIVil \^M AND HUMAN SERVICES 

f^^%0W%m%nM^^^ Annual Report PuWic Heaim service 

W wlTI|JU &Vl National Institutes ol Health 

Research and ^°'^^^ ' 
Technology 



MATHEMATICALANALYSISBIOLOGICALPHENOMEI 
MISTRYTHEORETICALANALYSISCELLLOCOMOTIO 
SESLASERLIGHTSCATTERINGBIOCHEMISTRYCHE 
ROBABILITYDENSITYFUNCTIONMOLECULARFORC 
ELSCRYTALLOGRAPHICDATAKINETICSHYPRATIO 

COMPUTER-BASEDSTUDIESMATHEMATICALTHEC 

PUTINGSINGULARVALUEDECOMPOSITIONDATAF 

MATHEMATICSMEDICALAPPLICATIONSSTATISTK 

NRECOGNITIONIMAGEPROCESSINGBIOMEDICALI 

ESPARTIALDIFFERENTIALEQUATIGNSADAPTIVEF 

PROGRAMS/PACKAGESSTATISTICALSOFTWAREB 

MLABCURVE-FITTINGNUMERICALDERIVATIVESBK 

TATISTICALMETHODOLOGYMEDICALINFORMATIC 

ATORPACKAGEGRAPHICALDISPLAYSLINGUISTIC 

EDICALLANGUAGECLUSTERANALYSISDISCRETEI^ 

ENGINEERINGSPECIALIZEDCOMPUTERSYSTEMSS 

ECTDEVELOPMENTLABORATORYAUTOMATIONP/^ 

ERSMICROCOMPUTERSMICROPROCESSORSELEC 

NDDISPLAYMOLECULARGRAPHICSLABORATORYC 

NGINSTRUMENTINTERFACESPICTUREARCHIVING 

CLINICALSUPPORTCARDIOLOGY/HEARTSyRGER\ 

ITYNIHADMINISTRATIVED/n^/UBASEPpOCCPT 

GEMENTSYSTEMAPPLIEDSYSTEMSPROGRAMMIN 

PLICATIONSDATAMANAGEMENTSYSTEMSDATAB 

MMINGSYSTEMSPESIGNINFGRMATIGNPROCESSII 

MAINFRAMESINTELLIGENTTERMINALSfMJOGRAM^^ 

IPROGRAMMINGTIMESHARINGTEXTEDITm 

OCUMENTATIGNCENTRALCOMPUTERUTIU^ 

IDESYSTEMRESEARCHANDDEVELOPMENTOPER^ 

PMENTSERVICESFEE-F QRSERVICEUSERSERVICE 

SCIENTIFICANDTECHNICALCGMMUNICATIONADP 

ATIONADMINISTRATIVEFINANCIALMANAGEMENT 

UBLICATIONSCOMPUTERIZEDBIBLIGGRAPHYSEA 

EEOPERSONALCOM^^ 



Foreword 



ill 



The Division of Computer Research and Technology DCRT programs focus on three pnmary activities 

has primary responsibility for incorporating the power of conducting research, developing computer systems, 
nnodern computers into the biomedical programs and and providing computer facilities 

administrative procedures of NIH. DCRT serves as a jhe fiscal year 1983 annual report descnbes our work 

scientific and technological resource for other parts of in two volumes: 

PHS, and for other Federal organizations with Volume 1 gives an overview of the worV of each 

biomedical and statistical computing needs. group, highlighting the year s accomplishments; 

Volume 2 gives details atx>ut tf^ protects and 
activities of each group 



From the Director 



During 1 983 the Division of Computer Research and 
Technology was again broadly and actively involved in 
the conduct and management of science across all of 
NIH. The DCRT laboratories and branches are integral 
parts of the extensive biomedical computing activities 
within the core of NIH scientific excellence. Indeed, 
mathematics, statistics, engineenng, and other 
elements of computer science are now recognized 
throughout the world as essential elements in 
biomedical science. 

In 1963, when NIH first considered bringing these 
disciplines together in a new division, this was a novel 
idea. The concept has proved so successful that two 
decades later the task now facing us, DCRT and NIH, 
is to build upon the high level of technological success 
and move to one that can truly be called an Intellectual 
Era of biomedical computing. 

The Technological Era of computing is still a dominant 
force after three decades. Clever digital engineenng 
continues to pack more processing power, storage, and 
communication capacity into smaller, cheaper, and 
eventually more reliable boxes. The plethora of new 
hardware is matched by new offerings of software. 

However, better computer technology alone is not 
enough to create the intellectual linkage between the 
power of computing and the substance of modern 
science. The committee report in 1963 that led to 
formation of DCRT recognized this when it said at the 
outset; 

The Committee was deeply impressed by the power of 
this technology and the promise it holds for contributing 
a new level of insight into problems in the life sciences. 
It was equally impressed by the magnitude of the 
resources... especially intellectual that its large scale 
application demands. 



Results of great practical benefit m latjoratones. clinics, 
and offices across NIH demonstrate the importance of 
mathematicians, engineers, and computer experts as 
intellectual resources in accomplishing that linkage But 
the greatest benefits can occur only when an 
equivalent intellectual contribution emerges from the 
biomedical scientists in this collat>oration Then the 
intellectual union of computing and science can move 
forward to advance the underlying theory of biomedical 
science and to generate advances m the information 
sciences. Progress arises from the engagement of fwst 
class minds m a context where the focus on 
information processing is fully engaged m the 
environment of biomedical research 

The challenge to DCRT and NIH m 1983 is still the one 
presented in 1963, to amass and focus the best 
available intellectual resources on the important and 
pervasive opportunities for biomedical computing at 
NIH. 



Ui^^^d i\ 



Arnold W. Pratt. M.D. 
DCRT Director 



•J%c.^,Jf^ 



3 



Contents 



Physical Sciences Laboratory 1 

Laboratory of Applied Studies 5 

Laboratory of Statistical and 

Mathematical Methodology 11 

Computer Systems Laboratory 17 

Data Management Branch 23 

Computer Center Branch 27 

Office of the Director 3 3 



MATHEMATICALANALYSISBIOLOGICALPHENOMEI 
MISTRYTHEORETICALANALYSISCELLLOCOMOTIO 
SESLASERLIGHTSCATTERINGBIOCHEMISTRYCHE 
ROBABILITYDENSITYFUNCTIONMOLECULARFORC 
ELSCRYTALLOGRAPHICDATAKINETICSHYDRATIO 

t LIQlflHBBPROCESSINGBIOMEDICAl 



H "/.■VJ1!J.A'"^*^*''^'^«"CALDERIVATI\ 

I^AV;'iP*';^^yi'^X"^^<>*-^<3iYMEDICALINFORIviAit 



IS JB^yJiSSMIM^^'^^**"* /AUTOMATIONpil 

ES^yj£P?5SJ?.'*i^. ^OPROCESSORSELEC 

MRPJISkf.YJSStf.V^, ^PHICSLABORATORY 

c ■ 

S.^."!!! 'SI?/*^^""*'^ /STEMSPROGRAMMH 

PLICATION AGEMENTSYSTEMSDATAr 



k HARINGTEXTEDITINGABC 



*^'"*'f^P.%l5ytI?.V^'"''"'«NANCIAUWAMAC....v.:„ 
-'LL'SMfPPjyi?' IZEDBIBLIOGRAPHYSE/ 

t 3iiSffMftPi*M.tMTERSSYSTEMSP0UCYAMC 



Physical Sciences Laboratory 



ieorge H. Weiss, Chief 



function and Scope of Work 

;he Physical Sciences Laboratory works in areas of 
hysics, chemistry, and applied mathematics relevant to 
/■oblems in the biological sciences and medicine. Its 
rogram consists of two parts; internally-generated 
jsearch, and consultation and collaboration with other 
IIH scientists in its areas of expertise. 

SL's program originally involved theoretical analysis 
nly; it now includes a considerable experimental 
omponent as well. These expenments are in the areas 
f crystallography and light scattering and embody 
asults originally developed theoretically by members of 
ne laboratory. 

he PSL staff consists of five professionals and several 
isitors who work in the areas of: 

1. The structure and elucidation of forces determining 
the structure of membranes and other biological 
surfaces by a combination of crystallographic and 
thermodynamic methods. 

2. Light scattering from biologically interesting 
systems, and the determination of dynamic properties 
of these systems by suitable interpretation of such 
experiments. 

3. The use of image processing techniques for 
interpreting and reconstructing the three-dimensional 
structure of cells from multiple electron micrographs. 

4. The application of the techniques of applied 
mathematics to problems in the physical and 
chemical sciences. 

Because of the wide range of interests of members of 
the laboratory, a number of collaborative efforts, both 
with NIH scientists and with others, have been formed. 
Some of the data on intermolecular forces are being 
generated by scientists at Brock University, Canada. 
The electron micrograph images that are the 
experimental data for the image reconstruction project, 
are generated at the University of Colorado. 
Techniques for the analysis and optimization of kinetic 
experiments are being developed in response to 
problems in nuclear magnetic resonance and positron 
emission tomography that have ansen at NIH. 



FY83 Accomplishments 

This year has seen the methodology used to measure 
forces between DNA double helices extended to the 
study of the effects of ionic species bound to the 
molecular surface. Further results using the technique 
developed by PSL members allow measurement of the 
pressure needed to pack DNA in a viral head The 
experimental results suggest a variety of theoretical 
investigations on effects of hydration forces on bilayer 
deformation and on models of cell membrane fusion A 
significant finding in this project is that the commonly 
accepted double-layer theory used to analyze forces m 
the immediate vicinity of membranes is badly in error 
and must be replaced by a more realistic theory. 

Computational techniques have been developed for 
random walk models that have long been used in 
identifying crystallographic space groups from x-ray 
scattering data. Until now use of these models was 
restncted to a very small number of space groups, but 
the methods developed and still under development will 
allow routine use of statistical methods for the most 
commonly occurring space groups 

A study using light scattering techniques has been 
completed on the effects of subunit crosslmking on clot 
strength. These techniques are presently being applied 
to analyze the shear moduli of polymer gels, relating 
them to such system charactenstics as crosslink 
density and polymer concentration. This investigation, 
when completed, will shed light on how such variables 
affect the dissipation of mechanical excitations. 
A technique has been developed for the elimination of 
phase error in nuclear Overhauser effect 
measurements commonly used m nuclear magnetic 
resonance. Extensions of this technique will prove to 
be useful for the interpretation of kinetic data from two- 
dimensional Fourier transform NMR expenments 
A computer system has been assembled for three- 
dimensional reconstruction of cells from sets of 
electron micrographs 



!?J^JJnHEriCSMEOICALAPPI NSsriricT 

HnvrwriMnn 

MLABCURt 

tatisticalMe "wed" iiu?-?, 

ATORPACKAGE CALmlF i.^7.%ApT 



^?KASPRATORY 

yl IDBT AD^uiitf 1^1^ 



■ TY _.,iATIVEn^ ^ 

PUCATr^lJl^f?'^£?yfir MSPROGRAlASn 



JVI 



OCUMENtAtlONCEM iooun. iGABCi 

LPESYSTEIWRESEAPCh"^^^^^ 



SSSA?TS9'SPOLIC.VANp 



Physical Sciences Laboratory 



Beorge H. Weiss, Chief 



Punction and Scope of Work 

fhe Physical Sciences Laboratory works in areas of 

jhysics, chemistry, and applied mathematics relevant to 
|)roblems in the biological sciences and medicine. Its 
)rogram consists of two parts: internally-generated 
esearch, and consultation and collaboration with other 
>JIH scientists in its areas of expertise. 

'SL's program originally involved theoretical analysis 
jnly; it now includes a considerable expenmental 
component as well. These experiments are in the areas 
3f crystallography and light scattering and embody 
esults originally developed theoretically by members of 
:he laboratory. 

The PSL staff consists of five professionals and several 
>/isitors who work in the areas of: 

1 . The structure and elucidation of forces determining 
the structure of membranes and other biological 
surfaces by a combination of crystallographic and 
thermodynamic methods. 

2. Light scattenng from biologically interesting 
systems, and the determination of dynamic properties 
of these systems by suitable interpretation of such 
experiments. 

3. The use of image processing techniques for 
interpreting and reconstructing the three-dimensional 
structure of cells from multiple electron micrographs. 

4. The application of the techniques of applied 
mathematics to problems in the physical and 
chemical sciences. 

Because of the wide range of interests of members of 
the laboratory, a number of collaborative efforts, both 
with NIH scientists and with others, have been formed. 
Some of the data on intermolecular forces are being 
generated by scientists at Brock University, Canada. 
The electron micrograph images that are the 
experimental data for the image reconstruction project, 
are generated at the University of Colorado. 
Techniques for the analysis and optimization of kinetic 
experiments are being developed in response to 
problems in nuclear magnetic resonance and positron 
emission tomography that have ansen at NIH. 



FY83 Accomplishments 

This year has seen the methodology used to measure 
forces between DNA double helices extended to the 
study of the effects of ionic species bound to the 
molecular surface. Further results using the technique 
developed by PSL members allow measurement of the 
pressure needed to pack DNA in a viral head The 
experimental results suggest a variety of theoretical 
investigations on effects of hydration forces on bilayer 
deformation and on models of cell membrane fusion A 
significant finding in this project is that the commonly 
accepted double-layer theory used to analyze forces m 
the immediate vicinity of membranes is badly m error 
and must be replaced by a more realistic theory 

Computational techniques have been developed for 
random walk models that have long been used in 
identifying crystallographic space groups from x-ray 
scattering data. Until now use of these models was 
restricted to a very small number of space groups, but 
the methods developed and still under development will 
allow routine use of statistical methods for the most 
commonly occurring space groups. 

A study using light scattenng techniques has been 
completed on the effects of subunit crosslmking on clot 
strength. These techniques are presently being applied 
to analyze the shear moduli of polymer gels, relating 
them to such system charactenstics as crosslink 
density and polymer concentration. This investigation, 
when completed, will shed light on how such vanables 
affect the dissipation of mechanical excitations 
A technique has been developed for the elimination of 
phase error in nuclear Overhauser effect 
measurements commonly used in nuclear magnetic 
resonance. Extensions of this technique will prove to 
be useful for the interpretation of kinetic data from two- 
dimensional Founer transform NMR expenments. 
A computer system has been assembled for three- 
dimensional reconstruction of cells from sets of 
electron micrographs. 



Future Plans/Trends 

Experimental work will continue on the use of combined 
thermodynamic and crystallographic methods for the 
measurement of intermolecular forces, particularly 
focused on the control and characterization of phase 
transitions in membranes and helical molecules. In 
conjunction with this, a systematic theoretical 
examination of several models of membrane fusion will 
be undertaken, taking into account the forces that have 
been identified in PSL measurements. 

A molecular modeling study of the structure of sugars 
and polysaccharides will be begun. The structure of 
these molecules is not completely characterized 
because they are not crystallized readily. The study will 
make heavy use of the molecular graphics facility 
together with energy minimization and molecular 
simulations. 

Work, both experimental and theoretical, will continue 
on the physical properties of biological polymer gels. 
Mathematical theories are to be developed relating to 
gelatin kinetics, to the coalescence of fibrils to form 
fiber bundles, and to the mechanical dissipation 
properties of idealized fibrin and actin lattices. In 
addition, theoretical studies will be initiated on the 
factors affecting the percolation of macromolecules 
through polymer networks. 



In a collaborative project with NIADDK, LCP, PSL 
scientists have measured the forces between DNA 
molecules in aqueous solutions. This graph shows 
that the dominant contribution to the 
intermolecular force is the work of removal of 
water from the vicinity of the molecular surface. 
This observation is contrary to all previous 
expectations of polyelectrolyte interactions. 



An investigation into practical computational techniques 
for detailing properties of random walks used in 
crystallography will be continued. Considerable success 
has been achieved in these calculations so far for a 
small number of space-group related random walks. 
Future study will be devoted to enlarging the catalog of 
results available to crystallographers. 

A method for the three-dimensional reconstruction of 
cellular structure is being developed using electron 
micrographs. Software is being written for rotation of 
the resulting images. Further study will be devoted to 
estimation of the space available for diffusion in the 
cytoplasmic matrix. 

Dr. Adrian Parsegian has been installed as President of 
the Biophysical Society. Dr. James A. Ferretti has taken 
a position as section chief in NHLBI. 



i 



Divalent Cation Solutions 



8 r 



6 - 



OD A 


* Putrescine 10 mM 


° i 


□ MgCI, 25 mM 


o 


CaCI, 25 mM 



20 



25 



30 



35 



40 



Interaxial Spacing (l) 



Publications and Presentations 

Aizenbud. B. and Gershon. N Dillusion ol molecules on biological meni' 
branes ol nonpianar lorm-a theoretical study Biophys J 38 287293, 
1982 

Aizenbud. B. and Gershon. N D Dillusion ol molecules on microvillous bio- 
logical membranes In Perelson. A C . DeLisi. C . and Wiegel. F W (Eds ) 
Cell Surface Phenomena New York. Marcel Dekker (in press) 

Aizenbud. 6 . and Gershon. N D Hydrodynamic equations and VH light scat- 
tering Irom viscoelastic (solid-liKe) systems II Molecular approach PhysKa 
A 108A 583-588. 1981 

Balaban. R S . and Ferretti. J A Rales ol enzyme-catalyzed exchange deter- 
mined by two-dimensional NMR A study ol glucose 6-phosphate anomeri- 
zation and isomerizalion Proc Nail Acad So USA 80 1241-1245. 1983 

Gershon. N . Porter, K . and Trus. B Microtrabecular lattice and the cyloskele- 
Ion Their volume, surface area and the dillusion ol molecules through it 
Biological Structures and Coupled Flows Proc ol Aharon KalirKalchalsky 
Memorial Symposium Israel. 1982. pp 377-380 

Gladner, J A. and Nossal. R Edects ol crosslinking on the ngidity and 
proteolytic susceptibility of human fibrin clots. Thrombosis Res 30 273-278. 
1983 

Griffith. J P . Lee. B K . Murdock. A L . and Amelunxen. R E Molecular 
symmetry of glyceraldehyde-3-phosphale dehydrogenase from bacillus 
coagulans J Molecular Biology (m press) 

Gruen. D W R. Marcelja. S. and Parsegian, V A Water structure near the 
membrane surface In Perelson, A (Ed ) Membrane Surfaces New York, 
Marcel Dekker. Inc (in press) 

Jacobson. L The determination ol a phosphorus-phosphorus nuclear Over- 
hauser enhancement by two-dimensional magnetization exchange spec- 
troscopy J Megn Reson 49: 522. 1982 

Kiefer. J E . and Weiss. G H : The Pearson random walk AlP Proceedings (m 
press) 

Lee. B Calculation of volume llucluation for globular protein models Proc 
Nail Acad Sci USA 80 622-626, 1983 

Lee. B Partial molar volume from the hard-sphere mixture model J Phys 
Chem 87 112-118. 1983 

Loosley-Millman. M E . Rand. R P . and Parsegian. V A Effects of monova- 
lent ion binding and screening on measured electrostatic forces between 
charged phospholipid bilayers Biophys J 40 221-232. 1982 

Monks, T J , Pohl, L R . Gillette. J R . Hong. M . Highel, R J , Ferretti, J A . 
and Hinson. J A Stereoselective formation ol bromobenzene glutathione 
conjugates Chem Biol Inleraclions i^ 203-216, 1982 

Nossal, R Stochastic aspects ol biological locomotion J Slat Phys 30 391- 
400, 1983 

Nossal, R , and Jolly. M Shear waves and internal viscosity in cylindrical gels 
J AppI Phys 53 5518-5525. 1982 

Nossal. R . Weiss. G H , Nandi, P K , Lippoldt, R E , and Edelhoch, H Sizes 
and mass distribution ol clathnn coaled vesicles Irom bovine brain Arch 
Biochem Biophys (in press) 

Parsegian. V A Dimensions of the intermediate phase ol dipalmiloylphospha- 
tidylcholme Biophys J (in press) 



Parsegian. v A Membrane inleraclion and dafonnation Armalt N Y Acml 

Set (in press) 
Parsegian. v A , and Rand. R P Membrane ttaniport and Itt* activily o) miaWr 

near the membrane surface Ptryncal Mamodt n m» StuOy ot C^Um 

Biophysics (in p<ess) 
Parsegian. V A . Rand. R P . and GmgaN. LMtont lor tnt tludy ol 

membrane fusion from membrarw interacMxo m pho ip >x)H)id tytlemt 

Ctba Found Symp (in press) 
Rubin. R J . and Wetss, G H Limitir>g iNckneu ol an adiortMd pdymw 

chain J Crtem Phys 78 2039-2043. 1963 
Sterner, C A Litl. M , and Nossal. R Apptcabont ol dynamc hght tcatlanng 

to studies of mucin structure Proceedings ol 1983 Synvo$Ktm on N0m 

Techrvques in Biorheotogy (in pceM) 
Steiner. C A . Lift. M . and Nossal. R Effects ol c«lcum ion on the ttrucli«« ol 

canine tracheal mucin Biorheology (in press) 
Stone. M , Sonies, 8 C Shawker, T H , Weiss, G H , and NaiM. L Analyvs 

ot realtime ultrasound images of tongue contiguralion usirtg ■ gndK*g«x- 

ing system J Phonetics (in press) 
Weiss. G H Chromatographic kinetics artd the phanomerxwi ol IMkng Stp 

Sci S Tech 17 1609-1622. 1982 
Weiss. G H Random walks Encycl Stat Sa (in press) 
Weiss. G H Random walks and their appl«ations Am So 71 6S-7I. ^9t3 
Weiss, G H , Caveness, W F , Einsiedel-Lechlape, H , and McN««l. M L Uta 

expectancy and causes ol death m a group ol head in|ur«d veterans ol 

World War I Arch Neurol 39 741-743, 1982 
Weiss, G H , Feeney. D M . Caveness, W F . Dillon, D , Kistler. J P . Mohi. J 

P . and Rish. B L Prognostic factors for tt>e occurrertce ol posltrawnanc 

epilepsy Arch Neurol 40 7- 10, 1983 
Weiss. G H , Ferrelti, J A Kieler J E . and Jacobson. L A meffwd tor 

eliminating errors due to phase imperfection m NOE maasuremants J 

Mag Res (in press) 
Weiss. G H . and Kiefer. J E The Pearson random waNi with unequal slap 

sizes J Phys A16 489-495. 1983 
Weiss. G H , and Rice, J A A combinatonal problem in pharmacotogy J 

Math Biol 14 195-201, 1982 
Weiss. G H . and Rice. J Optimal parameters lor the measurement ol the hall- 

wtdlh of a Gaussian peak Sep Sa S Tech 17 tlOI-tflS. 1982 
Weiss. G M . and Rubin. R J (Eds ) Proceedings of ttw Symposium on 

Random Walks J Slat Phys 30 249-561. 1983 
Weiss. G H . and Rubin, R J Random walks tl>eory and seiacted appkca- 

tions Adv Chem Phys 52 363-505, 1983 
Weiss, G H , Shuler, K E , and Lindenberg, K Oder statistics lor hrst passaga 

times in diffusion processes J Stat Phys 31 255-278, 1983 
Weiss, G H , and Szabo, A First passage problems for a class ol rnatlar 

equations with separable kernels Physics (m press) 
Weiss, G l-l , Talbert, A , and Brooks, R A The use ol phantom vwws to 

reduce CT streaks due lo insufficient angular sampling Phys m Biol tnd 

Med 27 1151-1162. 1982 



:''^ ''— yii 



S'itT5IIS?fiL"ETICALANALYSISCELLLbc6M0f. 
SESLASERLIGHTSCATTERINGBIOCHEMISTRYCHB 
»«.,^o...,„^^..^,,„, ^^'ONMOLECULARFORi 

^ KINETICSHVDR^T!'" 

COMPUTER-BASEDSTUDIESMATHEMATICALTHEf 

£4X^',ffPJl't9>'ok^RVALUEDEC^OMPOSITION^^^^ 

M^Ji'iK'AI.'.'^SMEDICALAPPLICATIONSSTATISTIC 

^f£?i?il'^'JI<>''IMAGEPROCESSINGBIOMEC>ICALI 

ESPARTIALDIFFERENTIALEQUATIONSADAPTIVEf 



■ w ir-v» ■ s I 



: 3ISCRETEA 

ECTDEVELOPMENTt RATORYAUTOMATinMDS' 

ERSMJCROCOMPUT 

^ ^"^OLECULARG^^ 

GkiniUAi.^U ^ ^ 



^MENTOPER/ 

ECOPyR^'SKfPJJA'^y.TISJil^DB'Bi-IOGRAPHYSEi 
5^°of!l^S^tPil?l^'JTERSSYSTEiyiSPOLICYANl 



Laboratory of Applied Studies 



John E. Fletcher, Ph.D., Acting Chief 



Functions 

The Laboratory of Applied Studies (LAS) has three 
main purposes: 

1. in collaboration with biomedical scientists, to apply 
. mathematical theory and computing science to the 
[ development, testing, and improvement of 
I mathematical models of physiological processes-- 
I particularly reaction-diffusion kinetics, transport and 
I exchange of substrates in tissues, and the 
I descnption of metabolism within cells and organs; 
I 2. in collaboration with clinicians, to develop and 
apply mathematical or statistical theories and special- 
purpose computing procedures (analog or digital as 
required) to facilitate research projects aimed at 
improving diagnosis of disease, assessment of 
treatment, and understanding of basic physiological 
and pathophysiological processes; 
3. to engage in applied research in mathematics, 
statistics, and computer science as necessary to 
provide a sound theoretical basis for collaborative 
studies, and to insure that state-of-the-art 
mathematical and computational methods are 
available as research tools at NIH. 

Two sections carry out these primary LAS functions: 
Applied Mathematics Section -AMS"(John E. 
Fletcher, Ph.D., Chief). This staff of five includes 
specialists in applied mathematics, computer science, 
biomathematics, and biomedical engineering. 
Medical Applications Section -MAS -(James J. 
Bailey, M.D., Chief). This five-member staff includes 
physician-scientists, electronic engineers, and 
computer systems analysts. 

Dr. Harris, former Chief, LAS is a biostatistician with 
training in public health and the basic medical sciences. 
He retired in early FY83 and continues in LAS as a 
part-time guest worker. 



Scope of Work 



The Laboratory of Applied Studies works on projects in 
basic and clinical biomedical science. Largely, these 
involve collaboration with other groups at NIH, 
elsewhere in the U.S.A., and abroad. The collaborating 
investigators this year included: 



• clinicians in the Cardiology, Clinical Hematology, and 
Pulmonary Branches of NHLBI; in the Arlhntis and 
Rheumatism Branch of NIADDK; and in the Cntical 
Care Medicine and the Nuclear Medicine 
Departments of the Clinical Center 

• physiologists and biomedical engineers at the 
Louisiana Technical University and elsewhere in the 
U.S.A. and Europe studying the transport of 
substrates within the microcirculation and the 
autoregulation of tissue perfusion 

• biochemists and physicians at NIH, and at 
universities in the U.S.A. and in France working on 
models for receptors of drugs or other ligands. on the 
kinetics of enzymes in membranes, and on other 
problems in tissue metabolism 

• electrocardiologists and biomedical engineers in the 
U.S.A.. Canada, and Europe concerned with 
improved algonthms for computer-based 
interpretation of ECG's and evaluation of ECG 
interpretative programs 

• clinical chemists and pathologists at NIH (Clinical 
Pathology Department, Clinical Center) and 
elsewhere in the U.S.A., in Europe, and in Japan 
engaged in the collection and study of reference 
values in laboratory medicine 

Highlights of the Year's Activities 

Although FY83 was a transition year with retirement of 
some senior staff and departure of a visiting scientist, a 
number of collaborative projects made substantial 
progress. 

In collaboration with NHLBI, a minicomputer-based 
laboratory system for studying delivery of oxygen to 
tissues dunng exercise through breath-by-breath 
analysis of pulmonary gas exchange has undergone 
extensive development in FY82-83. The equipment 
interfaces, designed and fabncated by Dr. E Pottala m 
FY82, are now controlled by software interfaces linked 
to the main laboratory control programs 

Automatic operation of the entire system, including 
acquisition and analysis of ventilatory flow and gas 
concentrations, as well as control of the bicycle and 
treadmill, is now possible Before this report is issued. 



M. Morton expects to complete the systems 
programming that will allow automatic computation of 
noninvasive indices of patient functional status. 

Serial measurements obtained in this exercise testing 
laboratory provide an objective indication of severity of 
disease and efficacy of treatment in patients with lung 
and blood disorders. Dr. R. Burgess, in collaboration 
with the Clinical Hematology Branch, NHLBI, is carrying 
out a study of drug therapy in patients with sickle cell 
disease. 

In a joint project with the Nuclear Medicine Department, 
LAS has extensively analyzed the relationship of 
signal/noise ratios to harmonic content of regional 
time-activity curves in radionuclide ventriculography. 
Smaller regions were shown to have lower signal/noise 
ratios and a shift of harmonic content to higher 
frequencies, reflecting the effect of poorer counting 
statistics. 

Theoretical work by Dr. M. Bieterman on the adaptive 
finite element (FEM0L1) methods for the solution of 
reaction-diffusion equations was essentially completed 
in FY83, These routines are now available on the NIH 
central computer systems. In addition, the IMSL 
routines known as TWODEPEP, finite element 
programs of a general type, have been implemented on 
the IBM System 370. 

These programs are being work-tested on models of 
bioheat transfer in hyperthermia being studied by BEIB 
in collaboration with the Radiation Oncology Branch, 
CC, and on linear and nonlinear (Michaelis-Menten) 
metabolic models for tissue substrate diffusion and 
consumption. 

B. Bunow and E. Pottala have studied network 
modeling languages and have demonstrated that 
network models are useful for biological simulation. 
Implementation of network modeling software on 
dedicated scientific computer systems such as the VAX 
1 1/750-780 has made network modeling more 
available to investigators on the NIH campus and has 
established their utility on dedicated scientific 
computers. Some functional errors in the larger NET2 
system were discovered in FY83. 



The further use of this particular system on the IBM J 
System 370 computer will depend upon correction of ^ 
these errors by the package's developers. Presently, 
interested NIH scientists are being instructed in 
network methods, and exploratory applications are 
underway in collaboration with NIH researchers on 
problems of nerve conduction and of facilitated 
diffusion in tissues. 

A considerable advantage of these modeling systems is 
that a functional rather than a mathematical description 
of the biological process suffices as a requirement to 
initiate study of its stimulus-response characteristics. 

In FY82 a collaboration was initiated with FDA 
physiologists to use rodent ECG's as a means of 
testing for cardiotoxicity of drugs and fad diets (e.g., 
liquid protein). The frequency content of rodent ECG's 
has required redesign of instrumentation and 
development of wholly new ECG analysis software. 

The rodent heart rate of 400-600 beats per minute 
produces ECG's with a much higher frequency content 
than that found in human ECG's. These signal analysis 
problems were largely resolved in FY83 and currently 
the ECG analysis of 38 animals before and during 
various levels of drug treatment is proceeding. 

The LAS DeAnza image processing system has been 
upgraded from resolution of 256 x 256 pixels to 512 x 
480 pixels. This upgrade will permit greater resolution 
and the development of more sophisticated algorithms 
for edge detection, image enhancement, and image 
manipulation. 

During FY83 LAS staff members participated in various 
teaching and consulting, or advisory, activities. 

J. Fletcher continued to serve as Chairman of the 
Mathematics and Computer Science Departments of 
the Foundation for Advanced Education in the 
Sciences. He is currently serving on a Planning 
Committee for the Director, DCRT; on a DRR ad hoc 
committee to design a workscope for the NAS 
Modeling Workshops; and as the DCRT representative 
to the NIH Advisory Committee for Computer Usage. 

J. Bailey continued as a member of an NHLBI site- 
visiting team concerned with computer analysis of 



exercise ECG's. He also served as a consultant on 
Common Standards for Quantitative 
Electrocardiography, a program in medicine and public 
health, sponsored by the European Economic 
Community. 

Since his retirement from LAS, former Chief E. Harris 
has continued to be a consultant in applied statistics to 
the Food and Drug Administrations Division of Medical 
Devices and Diagnostic Products. Dr. Harris also serves 
as consultant statistician to the College of American 
Pathologists, to the International Federation of Clinical 
Chemistry (Expert Panel on the Theory of Reference 
Values), and is a member of the Board of Editors of 
Clinical Chemistry. 



Future Plans 

The minicomputer system for analyzing pulmonary gas 
exchange in exercise will be tested on healthy 
volunteer subjects and on selected patient groups 
Studies to evaluate cardiorespiratory abilities in patients 
and controls will be specified with protocols m 
cooperation with the Clinical Hematology and 
Pulmonary Branches of NHLBI. 

The analysis of the signal/noise charactenstics of 
various parameters of regional ventricular wall motion 
will continue lointly with the Nuclear Medicine 
Department, CC, and the Cardiology Branch, NHLBI, in 
an effort to refine noninvasive, differential diagnosis of 
coronary artery disease and cardiomyopathies. 



40 ao uo 




40 to 120 



40 10 110 




40 to 1M 




40 to 120 



Adaptive finite element methods are useful for 
solving reaction-diffusion equations. Here, the 
techniques solve a nerve impulse propagation 
problem. 



A new project in cooperation with the Department of 
Critical Care Medicine, CC, to investigate dysfunction in 
neurologically impaired patients will move forward to 
integration and implementation of microcomputer-based 
methods for analysis and display of evoked potentials 
This project was suspended in late FY83 because of 
procurement delays in hardware acquisition and by 
problems with the NHLBI equipment. 

A major effort will continue the development of a 
language that will facilitate the conversion of network 
models simulating biological processes into forms 
compatible with languages such as MLAB, which will 
permit access to powerful data-fitting algorithms. 

A scientific collaboration with the Yale University 
Computer Center is expected to result in new and 
improved numehcal software for both the solution of 
linear equations and for systems of differential 
equations. The possibility of the incorporation of these 
routines into the MLAB modeling package will be 
explored. 

The rodent ECG project is expected to extend to 
analyzing sequential changes in mice infected with pure 
strains of Trypanosoma Cruzi (Chagas' disease). This 
extension of the rodent ECG project will be in 
collaboration with investigators from the World Health 
Organization and the Laboratory of Parasitic Disease 
NIAID. 

Utilizing the newly upgraded DeAnza image processing 
system, a joint study with the Clinical 
Neuropharmacology Laboratory, NIMH, will continue to 
develop theory and methods for interpreting electron 
energy loss spectra in intracellular organelles, 
particularly in the examination of dense bodies in 
electron micrographs of blood platelets. 

Publications and Presentations 

Albert. A Discnmjnani analysis based on mullivanale response curves: a de- 
scriptive approach lo dynamic allocation Slalislics m Medicine 2 95-106 
1983 

Albert. A . Chapelle. JP. Heusghem. C . Kulbertus, HE. and Harns. E K 
Evalualion o( risk using serial laboratory data m acute myocardial mlarclion 
In Heusghem. C . Albert. A . and Benson. E S (Eds ) Advanced Inlerprela- 
lion ot Clinical Laboratory Data New York. Marcel Dekker, 1982. pp 117- 



Albert, A., and Rultimann, U : Prediction of an ordered categorical response 

variable from serial measurement. Biometrics (in press) 
Bacharach, S.L.. Green. M.V.. Vitale. D., Douglas. MA., White. G.. Bonow. R.O 
and Jones. A.E.: A minimum error method for temporal fourier filtering o( 
gated cardiac data information processing. Eighth IPMI Conference on 
Medical Imaging (in press). 
Bacharach, S.L., Green, M.V., Vilale, D„ White. G., Douglas, UA.. Bonow HO 
and Jones, A.E.: Optimum number of harmonics for filtenng cardiac volume 
curves. Journal ol Nuclear Medicine 24:5. 1983, 
Bailey, J.J.. Berson. AS., Jackson, U.K., Stevens. J.M., Tolan, G D and Wolf 
H.K.: Evaluation methodologies for ECG diagnostic systems. In Bonner' 
R.E., Pryor, T.A.. Laks. MM., and Cole. S.S. (Eds.): Computenzed Interpre- 
tation of the Electrocardiogram VI New York. Engineenng Foundation 
1981. pp. 53-62. 
Bieterman, M.: An adaptive method for reaction-diffusion equations in one 

dimension. SI AM National Meeting. Denver, Colorado June 1983 
Bieterman. M.: A Posterion error estimation and adaptive finite element gnds for 
parabolic equations. Army Research Office Workshop on Adaptive Meth- 
ods for Partial Differential Equations. College Park. Maryland February 
1983 (in press). 
Bieterman. M.: On using local solution information ,n a mesh modification 
strategy for time-dependent equations. NASA-ICASE Workshop on Grid 
Methods. Hampton, Virginia, September 1983. 
Bieterman. M,. and Babuska. I.: The finite element method for parabolic equa- 
tions, I. A posteriori error estimation. Numerische Mathemalik 40-339-371 
1982. 
Bieterman, M., and Babuska, I.: The finite element method for parabolic equa- 
tions, II. A posteriori error estimation and adaptive approach. Numensche 
Malhematik 40:373-406. 1982. 
Bunow, B.: All things flow and change. Proc Wash. Acad Sci 7243-60 1982 
Bunow. B.: Cellular Enzymology: The steady-state kinetics of compartmental- 
ized enzymes. Journal of Theoretical Biology 8A:6: 1-627. 1980 
Bunow. B.: Turing and the physico-chemical basts of biological patterns In 

Prewitt, J. (Ed): IEEE Turing Memorial (in press) 
Bunow. B.. and Mikulecky, DC: On the feasibility of using flux measurements 
to distinguish among active transport models. Polish Winter Sc/iool of 
Membrane Transport (in press), 
de Graaf. C.N.. Douglas. MA.. Findley. S.M.. van Rijk. P.P.. Bacharach S L 
Green. M.V.. and Bonow. R 0. Een algoritme voor het localiseren vari 
siructuren in scintigrafische beelden Nucleair Geneeskundig Bulletin 4:42- 

Douglas. M.A.. Bailey. J.J.. van Ri,k. P.P.. Bacharach. S.L . Bonow. R O . and 
Green. M.V : Analysis of regional function in radionuclide ventriculography 
Physiological signal, scintillation noise, and regional size Computers in 
Cardiology IEEE Computer Society. Silver Spnng. MD. 1983 pp 315-318 
Ferenci. P , Covell. D . Schafer. D.F , Waggoner. J.G.. Shrager. R . Berman M 
and Jones. A E Metabolism o( the inhibitory neurotransmitter-ammobutyric 
acid in a rabbit model of fulminant hepatic failure Hepatology (in press) 
Fletcher. J E . and Schubert, R W.: Capillary wall permeability effects ,n capil- 
lary-lissue structures Proceedings of 1 983 ISOTT conference HKi%\on LA 
1983 (in press) 
Fletcher. J E . and Schubert, R W Diffusional coupling m a hemoglob.n-free 
perfused capillary-tissue structure Proceedings of the 1982 ISOTT moot- 
ing Dortmond, Germany, July 1982 (in press) 
Fletcher. J E . and Schubert. R w On the computation of substrate levels m 

perfused tissues Mathematical Biosciences 62 7^-^06 1982 
Fletcher. J E . and Schubert. H w The theoretical prediction of substrate levels 
and their histograms in cell free perfused tissues Proceedings of the 
International Meeting of the Society of Oxygen Transport to Tissue -IV (in 
prossi 



Hams, E K : Addendum to recent paper on reference values (or change 
Clinical Chemistry (in press) 

Hams. E.K : Regression, least squares, and correlation In Seligson. D . M 
(Ed): Handbook ol Clinical Chemistry (in press) 

Harris, E K The effects of reductions in analytic variance on the early detec- 
tion of trends Proceedings ol the IV International Meeting on Clinical 
Laboratory Organization and Management Uppsala, Sweden, June 29July 
1, 1983 (in press) 

Harris, E K , and Yasaka, T.: On the calculation of a reference change' for 
comparing two consecutive measurements. Clinical Chemistry 29 25-30, 
1983 

Horlon, f^ H : Computing on a shoestring; Naive users, sen/ice organizations, 
and computers Chi '83 Conference on Human Factors in Computing Sys- 
tems, Boston, Mass , 1983 

Kernevez, J P , and Bunov* B Numerical exploration of bifurcating branches of 
solutions to reaction-diffusion equations describing the kinetics of immobi- 
lized enzymes In Absi, E , Giowinski, R , Lascaux, P , and Veysseyre, H 
(Eds): Numerical Methods lor Engineering Pans, Dunod, 1980, pp 65-79 

Kernevez, J P , Joiy, G , Cuban, M C , Bunov*, B , and Thomas, D Self organi- 
zation in enzyme systems Ina Novosibirsk Colloquium 1978 Novosibirsk, 
Nauka. 1982, pp 257-271 

IWacfarlane, P W , Chen, C Y , and Bailey, J J A comparison of point sconng 
techniques for the diagnosis of LVH In de Padua, F , and Macfarlane, 
PW,, (Eds); New Frontiers in Electrocardiology Wiley. 1981, pp 353-356 



LAS scientists and FDA physiologists are 
collaborating in a study of cardiotoxicity in rats, 
produced by drugs, food additives, or fad diets. 
This is a typical electrocardiogram taken from that 
study. The very high quality of this tracing is made 
possible by computer processing to extract a 
typical complex from the average of many cardiac 
cycles, thereby suppressing random muscle noise. 
This high quality is necessary in order to detect the 
earliest changes of cardiotoxicity. 



Schubert. R W . Fletcher, J E , and Reneau, An analytical nvxJ* for amal 
diffusion in the Krogh cylin<J«r Proc»»<tngs ol Iht 1983 ISOTT comm- 
ence Ruston, (JV, August 1983 (in press) 

Schubert, R W , Fletcher. J E , and Reneau, D A stmpMiMl mo<J»l lor p»»*c1- 
ing myocardial P02 histograms Prtjc»edmgs ol Itte Frsl Southtm BmnaO- 
ical Engineenng Conlerence LSUME, June 1982 (m prMt) 

Setty, O H . Hendler, R W , and Shrager, R I Swnoltan^out nmtatMtnm* ol 
PiyiF, delta pH, delta psi, an<j H/O ratios m tniaci E Coti Bnpltr* J (" 
press) 

Shrager, R I Analysis of optical spectra by SVO SIAM 1963 Nabooal MaMng. 
June 6, 1983 (in press) 

Shrager, R I Some piUalls m the use of denvatrve spectra PHtMchmm^y mi 
PhototHOlogy (in press) 

Shrager, R I SVO as a descnplioo ot chermcal uuation SIAM 30th ArvvvarMiy 
Meeting, Stanford Unrversity, California, July 19, 1982 

van Ri|k, P P , Bailey, J J . and de Graaf, C N GecompulenseenJe meltxxJeri 
voor de detectie van regionale ventnculaire contrscueabrxxinalrtertwi Nu- 
cleair Geneeskundig Bulletin 4 49-54, 1982 

Winslow, R M , Samaia. M . Winslow, N J , Rossi-Bwnard", L , and Shrager, R I 
Simulation of continuous blood 02' equUibrum curves over pTiysiologcal 
pH, DPG, and PC02 range Journal ol AppHed PItysiology 54(2) &24-S20, 
1983 




iWATMEMATICALAWALYSISBI OLOGICALPHEMQMg 
MISTRYTHEORETICALANALYSISCELLLOC^^ 
SESLASERLIGHTSCATTERINGBIOCHEMISTRYCHE 
ROBABILITYDENSITYFUNCTIONMOLECULAR^^ 

COMPUTER-BASEDSTUDIES MATHfeMATlCALTHgf 
PUTINGSINGULARVALUEDECOMPOSITION^^^ 
MATHEMATICSMEDICALAPPLICATIONSSTATISTil 
NRECOGNITIONIMAGEPROclsSINGBIOMEDI^^^^^^ 
L__.,,_ x UATIONSADAPTIVEf 



I^X'JiJi^^y^ETHODOLOG^^^ 



ersmicrocomputersmicroprocessorWelV^ 

Syf^Vi^^lgAIA^ANAGEMENTSYSTEMSDAflvB 
??"'-5f^fIg?l^°^^'g!J!' ?PO"WATIONPROCEs" 

?D^E"s?l¥iaS&''E¥RWA^HS^^^^ 
PMENTSERVICESFEE.FOR.SERVi rgfi^iSjgi?.™ 

SClENTIFIfcANDTECHNlCALCOMMU^lfcAtlGMA^ 
ATlONADMINISTRATIVEFINANCIAUMANAfiiM 

UBLICATIONSCOMPUTERZEO^BIBUOGRAP^ 

EEOPERSONALCOM^PUTERSSYSTEI^^^^^^ 

^i^ NFEORPPLA NMANAGERIAl Fimi 



Laboratory of Statistical and 
Mathematical Methodology 



James E. Mosimann, Chief 



Function and Scope of Work 

The Laboratory of Statistical and Mathematical 
y^ethodology (LSM) combines research in mathematical 
itatistics, mathematics, and computer and information 
"science with collaboration and service in these areas to 
^IH researchers and administrators. LSM staff interact 
with all NIH Institutes, with other Federal agencies 
outside HHS, and with biomedical researchers 
worldwide. 

In addition to the position of chief, the laboratory has 
jsixteen full-time professional positions distributed 
.among four sections; 

' The Statistical Software Section (SSS) provides 
I consultation to and collaboration with NIH 

researchers and administrators in all computational 
aspects of biomedical data analysis, including 
! selection and support of large systems/packages. 
Four specialists in scientific programming are led by 
a computer systems analyst whose specialty is 
statistics. 

The Statistical Methodology Section (SMS) works 
closely with the Statistical Software Section. Three 
professionals in mathematical statistics provide 
biostatistical consultation and do independent 
research. 

The Blomathematics and Computer Science 
Section (BCS), directed by a mathematician, 
performs independent research and provides 
consultation and collaboration in the specialties of its 
five computer and mathematical scientists. 
The IMedical Information Science Section (MIS) 
investigates and develops methods for application of 
information and computer science to medical 
language data processing. Two computer specialists 
work under the direction of a computer systems 
analyst who is an expert in computational linguistics. 

A major part of LSM activity is the offering of statistical 
and mathematical systems/packages to the NIH user 
community. LSM accepts responsibility for evaluation of 
new systems/packages and their suitability for NIH. 
When it offers a system/package to the NIH 
community, LSM makes three basic commitments: 



1 . Maintenance of the package, with adequate 
documentation, through NIH computer system 
changes, system/package updates, and corrections 

2. Rapid response to queries concerning user access 
to a system/package program, including job control 
language and program parameters. 

3 Assistance in interpretation of results 

As a result of LSM's policy of not only supporting the 
use of these systems/packages but also aiding m the 
interpretation of their output, the statisticians of the 
Statistical Methodology Section provide consultation 
over a wide range of scientific fields Some very brief 
consultations are very successful because there is a 
known answer to the question at hand Other 
consultations involve extensive time and statistical/ 
mathematical/computer science research as well 

Research projects in LSM vary widely, from studies of 
natural language processing for medical information 
systems and studies of efficient algonthms for 
information retrieval to studies in mathematics and 
statistical methodologies for biomedical applications. 

FY83 Accomplishments 

FY83 was LSM's ninth year as a separate entity withm 
DCRT. The volume of its computational and 
consultative sen/ices continued to expand; its research 
activities decreased slightly, with one project 
terminated 

Computation 

During this year, the Statistical Software Section of 
LSM maintained the following systems/packages and 
programs on the IBM System 370 of the DCRT 
Computer Center: 
. BMD and BMDP. Biomedical Computer Programs. 

UCLA. 
. SPSS. SPSS-X. and SCSS, Statistical Package for 

the Social Sciences. SPSS, Inc. 
. SAS, SAS/GRAPH. and SAS/ETS, Statistical 

Analysis System, SAS Institute. Inc. 

• P-STAT Statistical Package. P-STAT. Inc 

• IMSL. International Mathematical and Statistical 
Libranes. IMSL. Inc 



11 



I 



• MSTAT1, Collection of Mathematical and Statistical 
Programs, DCRT. 

In FY83 the SSS staff responded to over 7,500 queries 
concerning use of these packages. Also during this 
year, BMDP and IMSL went through a major update. 
NIH served as a test site for both SAS82 and SPSS-X. 
Both systems will become production systems during 
the next fiscal year. 



The Biomathematics and Computer Science Section 
maintains several systems/packages and specialized 
systems on the DECsystem-10 of the Computer Center. 
Foremost in use is the interpretive system MLAB, 
designed (by LSM scientists) for biomathematical 
modeling, for cluster analysis by C-LAB operators, and 
for computer graphics. The Unified Generator Package, 
written and maintained by a BCS staff member, runs on 
DCRT's IBM System 370. 



STATISTICAL ANALYSIS SYSTEM 



STf?flIN=B DRUO=B 




LEGEND: DOSE 



18 



■o— o— o 10 



20 



22 



24 



26 



50 



SAS(Statistical Analysis System), which can be 
used to draw graphs lil<e this one, is one of many 
software systems/packages supported by LSM. 



12 



LSM stresses the importance of teaching the effective 
use of systems/packages to the biomedical scientists 
and other users of DCRT. 

In FY83, LSM continued to expand teaching and 
documentation for supported systems/packages LStVl 
taught eight introductory courses for SAS, two for 
SPSS, and two for BIVIDP. In addition, two introductory 
courses and one advanced course were taught for 
MLAB, plus four introductory courses on computer 
graphics at NIH. The second edition of the Beginner's 
Guide to MLAB is being printed now. and will be 
distributed before the end of FY83. 

ECS staff augmented I^LAB in FY83 by adding several 
mathematical operators and by adding facilities to 
permit numencal derivatives to be used for curve-fitting 
of large models. Also, color graphics and scientific text 
display were enhanced. 

A separate program (GRAPH 1) for easy generation of 
graphs was developed by SMS staff and is now being 
used by the NIH community. With only a small amount 
of user preparation, publication-quality graphs can be 
generated. 

Consultation, Collaboration, and Research 

LSM consultation and research in FY83 was closely 
tied to the use of the computer. Most consultations (55 
percent) involved statistical advice combined with 
considerable computer use. Others (40 percent) 
involved computer use alone and a small fraction (5 
percent) involved mathematical or statistical advice with 
only limited computer use. The percentages are 
unchanged from FY82. 

In FY83, LSM research, collaborative, and consultative 
efforts were expressed in a number of studies. 
Statistical methodologies were developed for, or 
modified to suit, specific biomedical problems. 

A study reported in FY82 was the subject of a 
publication. Statistical and Algebraic Independence,' m 
the Annals of Statistics in March 1983. This study 
contributes to a knowledge of properties of the sample 
covariance matrix, which is the basis for statistical 
discriminant analysis. Studies of discriminant methods 
continued in collaboration with Dr. J. Darroch, Flinders 



University. South Australia, and Dr H Hoffman. DRS. 
including discriminant analyses of morphological 
vanation in inbred strains of laboratory mice witti 
reference to purity of breeding slocks Collaborative 
work continued with Dr P Turkeltaub (BB/DPB) on 
clinical symptoms and allergic reaction to pollen LSM 
participation in a study of Chagas disease (Dr F Neva. 
NIAID/LPD) was concluded when the edited tapes of 
data were prepared for the investigator 

Collaborative work in various studies of schistosomiasis 
(with Dr. A. Cheever, NIAID. LPD) continued One 
portion of this research precipitated the development of 
a new statistical methodology that gives an exact 
treatment for a multivariate analysis of variance with 
unbalanced data This analysis may be applied both to 
experiments with repealed measurements and to 
growth curve analyses A paper on these results, which 
includes multivanate as well as nonparametnc 
treatment of these designs, has been submitted to a 
statistical journal. 

A study of nonparametnc multiple comparisons was 
initiated in FY83. with particular attention being given to 
theoretical as well as to computer-simulated behavior 
of vanous procedures. The optimal selection of a 
sequence of items based on relative ranks with ties has 
been investigated, as well as an evaluation of tests for 
correlated proportions with incomplete data. 

A collaborative study of the spatial distribution of blue 
cones in the retina with Dr S. Schein (NEI/CB) and F. 
de Monasterio (NEI/LVR) was completed It was 
possible to eliminate possible models on the basis of 
the several statistical techniques developed Also, 
studies of size and shape' vanables were continued. 
These studies provide methods for studying random 
proportions or ratios of common occurrence in 
biomedical data. 

In computer science, work continued using PROLOG 
(the logical procedure language selected by the 
Japanese as the basis for their fifth-generation 
computer project), used here to formulate problems of 
medical linguistics. A program was developed to 
partially translate scientific text from an NIH input 
format to the input format used by the TeX manuscnpt 
13 



generation system at Stanford. Studies continued in 
data storage and retrieval and on mathematical 
questions concerning vector spaces. 

In FY83 research in medical linguistics was continued 
on compositional morphosemantic analysis of medical 
terms derived from Greek and Latin. A methodology 
was developed for automated morphosemantic 
segmentation and semantic interpretation (paraphrasing 
rules) of medical compound words derived from Greek 
and Latin that denote surgical procedures. This 
methodology can be used with terms in the Systematic 
Nomenclature of Pathology (SNOP). The establishment 
of morphosemantic distribution patterns of medical 
compound words and the subsequent determination of 
semantic relations among them are crucially important 
for automated semantic interpretation of such words. 

MIS also continued its collaboration with the Laboratory 
of Pathology, NCI and with the Clinical Support Section 
of the Data Management Branch, DCRT to maintain 
and improve the data base of Clinical Center surgical 
pathology reports. The automatic encoding system 
provided by MIS computed representations of the 
summary diagnoses of the surgical pathology report as 
written by the pathologist, in a language based on the 
SNOP vocabulary. Collaboration continued with Dr. 
Donald E. Henson, NCI, concerning changes in the 
SNOP dictionary. 

LSM collaborative research on computer analysis of 
two-dimensional gel electrophoresis was discontinued 
due to the departure of the principal investigator. 
Results and computer programs were made available 
to NIH collaborative researchers. 

LSM computer scientist Dr. Gary D. Knott received the 
Public Health Service Commedation Medal in June. The 
award was made for his continuing leadership and 
innovation in the development of MLAB, now used 
worldwide to advance science through biomathematical 
modeling and computer graphics. 

Future Plans/Trends 

No major shift in laboratory service or research is 
anticipated in the coming year. Current levels of 



statistical and mathematical systems/packages 
support, consultation, and user assistance will be 
maintained. Research projects will be continuations of 
those already initiated and reported here. 

Publications 

Campbell, G,: Asymptotic Properties of Several Nonparamelric Multivariate Dis- 
tribution Function Estimators Under Random Censonng. Survival Analysis 
In Crowley, J,, and Johnson. R, A. (Eds): Institute of Mathematical Statis- 
tics Lecture Notes: Monograph Senes. Haywood. Calilornia. 1982. pp. 243- 
256. 

Campbell, G.: Optimal Selection Based on Relative Ranks of a Sequence with 
Ties. Advances in Applied Probability (in press). 

Campbell, G., and Foldes, A.: Large-Sample Properties of Nonparamelric Bivar- 
late Estimators with Censored Data Colloquia Mathematica Societatis 
Janos Bolyai Vol. 32: Nonparamelric Statistical Inference. Budapest, Hun- 
gary, 1980, pp 103-121 

Cheever, A. W,. Minker, R. G.. and DuVall, R. N.: Schistosoma Japonicum m 
Rabbits: Differences in the Host-Parasite Relationship Over a Seven- Year 
Period. Am. J. Trap. Med Hyg. 31(3): 514-517, 1982. 

DeBlas. A. L., Ratnaparkhi, M. V . and f^osimann. J E.: Estimation of the 
number of monoclonal hybndomas m a cell-fusion expenmenl. In Vunakis. 
H. v., and Langone. J J (Eds ): Immunochemical Techniques Methods m 
Enzymology. Vol. 92 New York, Academic Press, 1983. pp. 36-39. 

Knott. G. D : Direct chaining with coalescing lists Journal o1 Algonthms (in 
press). 

Knott. G D : Fixed-bucket binary storage trees J. of Algonthms 3: 276-287. 
1982. 

Malley. J D.: Statistical and algebraic independence The Annals of Statistics 
11 (1): 341-345. 1983 

Mosimann. J E : Discussion of Professor Aitchison's paper (The Statistical 
Analysis of Compositional Data). Journal of the Royal Statistical Society B 
44 (2): 168-170. 1982. 

Mosimann, J. E.: Size and Shape Analysis In Johnson. N I . Kolz. S . and Read, 
C B (Eds): Encyclopedia of Statistical Sciences John Wiley and Sons. Inc. 
(m press). 

Norton, L M : Automated Analysis of Instnjctional Text Artificial Intelligence 20: 
307-344, 1983 

Norton, L M , and Pacak, M G Morphosemantic Analysis of Compound Word 
Forms Denoting Surgical Procedures Meltxxis of Information in Medicine 
22: 29-36, 1983 

O'Connor, M A : Invanant metrics on cones In Proceedings of the Conference 
on Invariant Metrics and Holomorphic Maps Rome, Italy. Islituto di Alta 
Matemalica F Seven di CNR Symposia Malhemalica. London and New 
York. Academic Press. 1982 

Shapiro, M A note on Lee and Schacter's algorithm lor Oelaunay lnar>gulation 
International Journal of Computer and Information Sciences 10 (6) 413- 
418. 1981 

Yaar, I., and Shapiro. MBA quantitative study of the Electroencephalographic 
Response to Levodopa treatment m Parkinsonian patients Oimcal Electro- 
encephalography 14 (2) 82-85. 1983 



14 



Uses per month of 

Statistical packages supported by LSM* 



75600 




JUN 75 JUN 76 JUN 77 JUN 78 JUN 79 JUN 80 JUN 81 JUN '82 JUN '93 



*Packag«s supported by th« Statlttlcal Software Section only. Ooee not Include 
packages supported by the Blomathematlcs and Computef Science* Section. 



15 



mathem^ 

ROBABILiTYDENSITYFUNCTIONMOI,lciaARFo"c 



^j'i?ii^\ 1 ni^ui^ i 4iU AL 1 tiL,l, 



i!?^JJt!EMATI,CSMEpiCAi.APPLIC^^ 

„., ,.. , ,. ■ _ ■ . . SA!>afiTiVFF 

MLABCURVE-FITTINGNUMERICALDEhiv iy^^au 
TATISTICALMETHODOLOGYMEDICAU^^ 

gg|ffigL\^^M%TE%?A%'gSg^^^^^^^^ 
NGINSTRUMENTINTERFACg SPICTUREARC^^^ 

~^nut^i.%j^ ..lii^AHTSURGERY 



''^■- r., » ! r « ;: t _,v ^^ 



n-YNIHADMIN StRATIVEDATABASEPDOCCPT^ 

ESIO">"|gf>r?fgaTm»fppn rr'ccii 

MAiNFHAmbsiNTEU 7 

oc"fiSE"i^7r%%'K*^*^^^^^^ 

«>E_SYAT_E_MR^SEARCHANDDEVEl.^^^ 



f;ty'^'^"*0''ALCoWF»t!TFI?'!:^VCTPWt<STTrM tr^v a lir- 

IU2££ 



Computer Systems Laboratory 



Alan M. Demmerle, Chief 



Function and Scope of Work 

The Computer Systems Laboratory--28 professionals 
representing the disciplines of engineenng, computer 
science, medicine, and chemistry--is the major source 
of expertise at NIH for minicomputer and 
microcomputer technology. 

CSL engineers and scientists, in collaboration with NIH 
intramural laboratory and clinical investigators, apply 
this technology in the areas of laboratory automation 
and patient care. Some projects are occasionally 
undertaken with NIH extramural program staff and with 
other Federal agencies. 

CSL's multidisciplinary approach aids both the 
recognition of problem areas that will benefit from 
automation and the interpretation of research needs in 
terms of computer methods. 

Computers may be used only in an adjunctive manner-- 
for example, as a more convenient means to acquire 
laboratory and clinical data--or they may be integral 
parts of an elaborate instrumentation system, such as a 
computer-controlled mass spectrometer. Advances in 
large-scale circuit integration (LSI)"the microprocessor 
revolution-have brought about the miniaturization of 
computer components and a dramatic decline in their 
prices and power requirements. CSL engineers are now 
able to use microprocessors to deal with problems that 
once defied solution because of cost, size, or 
manpower constraints. 

CSL projects range in size from consulting activities of 
a few days' or weeks' duration to large-scale efforts 
taking many manyears. Much CSL work involves the 
development of new methods or technology or is 
influenced strongly by the changing needs of research. 
Thus, it is often difficult to predict the long-term scope 
at the outset of a project. 

FY83 Highlights 

This year, CSL engineers and scientists worked on 28 
projects, representing collaboration with almost all of 
the NIH Institutes. Some of these projects require only 
limited resources, while others take many manyears. 
The latter deserve particular emphasis because of both 



their sheer magnitude and the importance of the patient 
care or research activity they support. 

One of these ma|or projects involves the automation of 
flow cytometers/electronic cell sorters (FC/ECS) 
These instruments are being used increasingly in 
biomedical research, particularly in fields of laboratory 
research, and in clinical studies in immunology, 
cytology, and oncology CSL support tor these 
instruments began in 1973, when NCI requested 
development of a computer-based data acquisition, 
processing, and display system for the prototype flow 
cytometers developed at Stanford and Los Alamos 
Scientific Labs. 

Over the years, a system has evolved that is based 
upon a Digital Equipment Corporation PDF 1 1/34 
computer using an RT-11 operating system System 
hardware includes a refresh CRT display, two disks, 
magnetic tape, 64K words of memory, incremental 
plotter, a link to the NIH Central Computer Facility, and 
an intertace to the FC/ECS. Programs were developed 
to display the data as two- or three-dimensional figures 

Two-dimensional contour maps at user-selected 
thresholds are also available. Integration of selected 
curve segments and statistics descnbing peaks are 
included upon request. The various data presentations 
are a powertui tool to assist researchers m data 
interpretations. Equally important is the increase m the 
number of samples processed. 

This system has been duplicated six times at NIH and 
the Naval f^edical Center, and, m addition, many copies 
of the system documentation have been requested and 
sent to research centers in the U.S.. Europe, and 
Australia. 

In order to accommodate a high volume worKload 
environment where sample throughput is important, a 
new FC/ECS Computer System was developed by CSL 
and installed at NCI dunng the summer of 1983 The 
new system uses a Digital Equipment Corporation PDP 
11/24 (host) computer running under an RSX-11M 
multiuser operating system. Tektronix 4025 graphics 
terminals, and Digital Equipment Corporation LSI/ 11 -23 
(satellite) microcomputers for independent acquisition 
of data from each of several FC/ECS instruments 

17 



The satellite computer sends acquired data through a 
high-speed direct memory access (DMA) link to the 
host computer where it is stored on magnetic tape or 
disk. The terminal at the satellite may be used for 
displaying acquired data. This is helpful in viewing 
collected data files immediately after they have been 
stored at the host. This terminal is also used during 
data acquisition for parameter entry, display of file 
recordkeeping, and error reporting. 

Simultaneously with data acquisition, one or more users 
may analyze data at the host using independent 
graphics terminals. Processed files may be queued to a 
plotter so that a terminal can be used for further 
analysis while plotting proceeds. The new system can 
function with a PDP-11/34 computer, however, the 
PDP-11/24 supports more memory, and therefore more 
users may analyze their data concurrently. CSL 
anticipates the purchase of Digital Equipment 
Corporation's new 11/70 chip set, when it becomes 
available, so as to upgrade from the PDP-11/24 and 
further improve system speed and performance. 

Another project requiring substantial investment of CSL 
manpower involves the development of a computer 
system for data acquisition, processing, and display, in 
support of the Electron Beam Imaging and 
Microspectroscopy Facility in DRS/BEIB. The Facility 
was developed by physicists, engineers, and computer 
scientists from BEIB and CSL as a research resource 
for NIH investigators. It contains an automated electron 
microscope system that can analyze and display a 
specimen's elemental chemical makeup as well as its 
morphological microstructure. 

Electron energy loss (EEL) images, generated by 
measuring the characteristic amount of energy lost by 
beam electrons that interact with the atoms of a 
specimen, are thought to be the first in the world of this 
type produced on a scanning transmission electron 
microscope (STEfvi). 

Another important development by project team 
members is a digital filtering technique to background- 
correct images produced by energy dispersive x-ray 
spectrometry (EDS). The presentation of this technique 
won the Corning Award for the best contributed 
scientific paper at the last Joint National Meeting of the 



Electron Microscopy Society of America and the 
Microbeam Analysis Society. 

During the past year, the first biological images were 
obtained showing calcium distributions in secretory 
ameloblasts. This study of the role of ameloblasts in 
tooth enamel formation, done in a collaboration 
between BEIB physicists and a NIDR visiting scientist, 
heralded the beginning of a shift from instrumentation 
and computer systems research and development to 
applying the system to biological research. 

Subsequently, collaborations have begun with scientists 
from several other Institutes. NINCDS investigators are 
obtaining aluminum and calcium maps in cells taken 
from the hippocampus of victims of Parkinsonian 
dementia. Nitrogen and oxygen distributions in 
chromaffin cells are of interest to NIADDK investigators 
in a study of the release of epinephrine, and the 
Clinical Center is examining the relevance of 
magnesium in cardiovascular disease by studying the 
distribution of magnesium in individual lymphocytes. 

To address some of the difficult problems presented by 
these and other studies, CSL presently is concentrating 
on improvements and refinements to the system data 
acquisition, imaging, and analysis capabilities. 

Development of a new Image Processing Facility for 
NIH-wide use is another of CSL's major projects. 
Designed to complement an existing Evans and 
Sutherland System, which has in the past supported 
both image analysis and molecular graphics 
applications, the new facility has become operational 
this year. The main components are a Digital 
Equipment Corporation VAX 1 1/70 computer and a 
DeAnza IP8500 Image Array Processor. Although only 
one user station-comprising a 512 x 512 resolution 
color monitor with joystick and digitizing tablet-is 
currently available, two more stations are being 
procured and should be operational soon. 

Software existing on the Evans and Sutherland System 
is being reprogrammed for the more powerful VAX 
machine, and new programs that utilize the extended 
capabilities of the DeAnza System are being 
developed. 



18 



The PIC software package, a mainstay of image 
processing users of the Evans and Sutherland System, 
is already operational, and is the backbone of scientific 
research on the new system. Current research centers 
in two fields. First, high resolution structural studies for 
several viruses (including vesicular stomatitis virus, 
bacteriophase T7, tobacco mosaic virus, and varicella 
zoster virus) are underway currently. Second, structural 
analyses of fibrous proteins such as keratin, vimentin, 
desmin, and actin are in progress. 

These ongoing projects in macromolecular structural 
determination have placed NIH at the forefront of 
research in the analysis of high resolution electron 
micrographs. 

As more and more images become available from an 
ever-increasing diversity of sources, the management 
of these images poses a considerable problem. This is 
particularly true in the NIH Clinical Center where it is 
desired that images obtained from a vanety of 
modalities (CT, PET, NMR and Ultrasound Scans, 
Digital Vascular Imaging, etc.) be stored, transmitted, 
cataloged, and displayed at viewing stations in 
dedicated viewing areas and physicians' offices. 

CSL has undertaken a study of the feasibility of 
implementing a picture archiving and communication 
system for the Clinical Center. Although the study has 
not involved a major CSL effort thus far, the 
implementation phases of such a network imply a 
massive investment of manpower over a protracted 
period of time. 

The study has involved an investigation of the latest 
advances in data storage and local network technology. 
Optical laser disks, which will eventually enable up to 
one year's worth of images to be stored in one 
'jukebox' system, are being considered, as are 
advanced local communications networks that permit 
the transfer of millions of bits per second. State-of-the- 
art image display and data management techniques will 
be needed and are being assessed. Funding and 
technological considerations may dictate phased 
implementation of such a system. Resource limitations 
may likewise require that substantial portions of the 
project be contracted out. The study is expected to 
provide the basis for such decision-making 



Future Plans/Trends 

FY84 can be expected to present an increased 
demand for computers in laboratory and patient care 
settings More complex research goals of biomedical 
research investigators point to a greater need for 
automation in the laboratory Technological 
developments in large-scale circuit integration continue 
to lead to lower costs and smaller sizes for computers 
The current popularity of personal' computers is 
resulting in greater awareness on the part of NIH 
scientists of the potential benefits of computers 

At the same time, CSL is faced with limited personnel 
and budgetary resources In response to the challenge 
imposed by this conflicting set of circumstances, CSL 
expects to maintain high quality engineering and 
laboratory computer support to NIH programs by 
continuing policies developed in the past for managing 
resource issues. CSL staff will be deployed on projects 
promising maximum impact to the NIH community- 
those that serve a significant number of scientists, 
affect the quality of patient care, or represent general- 
purpose developments 

Publications and Presentations 

Bonow, R . OSI'OW H Roi-ng D L^r.r^-. ':■■ A..i-. ^ '.'j.. _ -^. j^-'. 

S . Green M , »nd Ep»le.n. S V«f«p«n»( EWect* on L»«t ventneuw Pr». 

sore- Volume An«lysi« onth ■ Norwn«gmg SonMation Probe OcU««on (•» 

press) 
Cotnputer Systems LsborarcY June 1963 
Donlon, J . Wang L . Luntfy. E . Wages. B . Feust, A . and Songco. A 

Computer Assisted Hematology MorpKo»ogy Data Handing SyaMm Pro- 

(Medmgs ol lh« &tm Annus/ Srmpotmjm on Cam^^tr Afiptcmtont #» 

UeOcM Csr«. Sneralon Wasnwiglon Hotel. Wa»»wiglon. D C Ooofter 30- 

November 2. 1962. pp 270 273 
Foster. M . George. J . Trus. B . and Ha»n». W So<»um. P ot a aaxn. CWonna. 

Magneswm. and EichangeaWe 45 Cataum Ions in Rod Outer Segmarw 

by Combined x Ray Micfoana»ysrt and Ra<»oaulograprty ftaortj^ca' Joir- 

ntl*y (2) 341. 1963 
Gershon. N . Porter. K . and Trus. B The Cy»o**a»*c Maln» Its SirucM*. 

Volume Surface A/ea. and Space tor Oiftuaon Bupf>rfKl Jamal *y (2) 

S5. 1963 
Gershon N Porter K. and Trus. B The MKi o tf a becUar Lawoa and *m 

Cyioskeieton The« Volume, Stxfaca A/ea. and the OAuaon ol MoiaciMa 

Through It Joumi ot Cel Biotogy 95 (2) 406A, 1962 
Gershon N Porter. K . and rrus. B The lAcroMbwaMr Laf*» and *• 

Cytoskelelon The« Volume. S«*lace Area, and the Wtiaon at Uotaotfaa 

Throu(^ It Proc0t<tngs of ir» Aharon K»tt-KtKt\tmy tMnotml 9rnrpo- 

sum Bologcl Smxtum and Cotolta Flows Israel, mtamefconal Soarwa 

Services (m press) 



19 



Goldman, R., Trus. B., and Leh/e, L.: Quanlitalive Double-Label Radiography of 
Two-Dlmensional Protein Gels Using Color Negative Film and Complete 
Analysis European Journal of Biochemistry 131: 473-480. 1983 

Gorlen, K : A Computerized STEM lor Biological Research. Annual International 
Medical and Latioratory Instrumentation, Washington, DC. November 17. 
1982 

Gorlen. K . Barden. L.. Del Priore. J.. Kochar. A . Florl, C, Gibson. C. and 
Leapman, R : A Data Acquisition System lor an Analytical Electron Micro- 
scope. Proceedings of the FALL DECUS {in press). 

Green. M . Ostrow. H . Bacharach. S.. Allen, S., Bonow, R., and Johnston, G.: 
Realtime Scintillation Probe Measurement of Left Ventricular Function. 
Nuklear Medizine 20: 11 6- 1 23. 1 98 1 

Hook. G . Fion, C , Gorlen, K., Gibson, C, Garruto, R., Fukatsu, R., Uanagihara, 
R., and Gajdusek, D : Elemental Imaging of Brain Tissue Using a Computer 
Controlled Electron Beam X-ray Microanalyzer Conference on Aluminum 
Analysis m Biological Matenal, Charlottesville, VA, June 29-30, 1983. 

Leapman, R , Fion, C, Gorlen. K.. Gibson. C. and Swyt. D.: Combined Elemen- 
tal and Structural Imaging in a Computer Controlled Analytical Electron 
Microscope Journal of Ultramicroscopy (in press). 

Martino. R . and Gerber. L.: An Automated Biomechanics Laboratory Applied to 
Rehabilitation Proceedings of the Fifth Annual Conference of the IEEE 
Engineering in Medicine and Biology Society (in press). 

Nadel. L. Automated Pulmonary Analysis by an Online Microcomputer. Com- 
puters in Cnlical Care and Pulmonary Medicine 3: 103-1 13. 1983. 

Nikodem. v.. Huang. D.. Trus. B.. and Rail. J.: The Effects of Thyroid Hormone 
on In Vitro Phosphorylation. Acelylation. and Ribosylation of Rat Liver 
Nuclear Proteins. Hormone and Metabolic Research (in press). 

Sabnn. H . and Kertesz. A.: The Effect of Imposed Fixational Eye on Binocular 
Rivalry. Perception and Psychophysics (in press). 



Steinen. P., Rice. R.. Roop. D . Trus. B . and Steven, A.: Complete Amino Acid 
Sequence of a Mouse Epidermal Keratin Subunit: Implications tor the 
Structure of Intermediate Filaments, Nature 302: 794-800. 1983. 

Steven. A.. Hainfeld. J . Wall. J.. Trus. B.. and Steinert, P.: The Distribution of 
Mass in Heteropolymer Intermediate Filaments Assembled in Vitro: STEM 
Analysis of Vimentin/Desmin and Bavme Epidermal Keratin. Journal of 
Biological Chemistry (in press). 

Steven. A.. Serwer. P., Bisher. M.. and Trus. B.: Molecular Architecture of 
Bactenophase T7 Capsid. Journal of Virology 124: 109-120. 1983. 

Tate. R.: Microcomputer Systems in the Laboratory: An Introduction. Serono 
Symposium Series. Raven Press. 1983 (in press) 

Steven. A,, Serwer. P,, and Trus. B,: Molecular Packing in Bactenophase T7 
Determined by Image Processing of Electron Micrographs, Journal of Viro- 
logy 124: 109-120. 1983, 

Tate. R,: The Microcomputer Revolution: Enhanced Instrument Capabilities. 
Annual International Medical and Laboratory Instrumentation Symposium, 
Sheraton Washington Hotel, November 18, 1982. 

Trus, 8.: Companson of Diffraction Patterns from Stained and Unstained Helical 
Particles: Modeling Expenmenis with Actin-Like Filaments Mini-symposium 
on Image Processing in Electron Microscopy, University of Pennsylvania 
School of Medicine, Philadelphia, PA, March 24, 1983, 

Trus, B,: Image Processing of Electron Micrographs, Chesapeake Society lor 
Electron Microscopy, Uniform Services University of Health Sciences, Be- 
thesda, MD, October 20, 1982, 

Trus, B,: Particles by Heavy Metal Staining: Modeling Expenments with Actin- 
Like Filaments, Annual Picture Meeting of the Chesapeake Society for 
Electron Microscopy, University of Maryland Baltimore County, Baltimore, 
MD, May 5, 1983. 



20 



STEM COMPUTER SYSTEM 



Terminals 



/^ 



Link to 
DEC-10 



Tape 



V\ 



^ 




Jii. 



Satellite Computer 
LSI- 11/23 



Z 



Beam 
X - Y Control 



V. 



S^rr; 



Disk 



67 MB 67 MB 




Bright 
Dark Field 
Detectors 



EEL 

Spectrometer 

& Detector 



STEM 




:r 




Host 
Computer 
PDP 11 60 



A system for controlling an electron microscope's 
detectors and processing the data acquired from 
the detectors requires a considerable amount of 
computer hardware and software. The system 
developed at NIH for application in biology consists 
of a host computer and a satellite computer 
interfaced to a scanning transmission electron 
microscope (STEM). The STEM, which provides 
analytical signals, includes a magnetic sector 
electron energy loss spectrometer, an energy 
dispersive x-ray detector, and conventional bright 
and dark field detectors. 



21 



IVIA ■ ntti.g^A riUALAS^AL¥^£^l£BULUIjllWA&.^ncraumE 

MISTRYTHEORETICALANALYSISCELLLpCOMOTIC 

SESLASERLIGHTSCATTERINGBIQCHEMISTRYCHE 

ROBABILITYDENSITYFUNCTIONMOLECULARFOR< 

ELSCRYTALLOGRAPHICDATAKINETiCSHYPRATIC 

COMPUTER-BASEDSTUDIESMATHEMATICALTHEi 

PUTINGSINGULARVALUEDECOMPOSITIONDATAI 

MATHEMATICSMEDICALAPPLICATIONSSTATISTI 

NRECOGNITIONIIVIAGEPROCESSINGBIOMEDICAL 

ESPARTIALPiFFERENTIALEQUATIGNSADAPTIVEl 

PROGRAMS/PACKAGESSTATISTICALSOFTWAREE 

MLABCURVE-FiTTINGNUMERICALDERIVATIVESBI 

TATISTICALMETHODOLOGYMEDICALINFORMATK 

ATORPACKAGEGRAPHICALDi'^»> ^vsLINGUISTIC 

EDICALLANGUAGECLUSTER. ISDISCRETEjl 

ENGINEERINGSPECIALIZEDCOMPUTERSYSTEMSi 
ECTDEVELOPMENTLABORATORYAUTOMATIONP/ 
ERSMICROCOMPUTERSMICROPROCESSORSELEC 
NDDISPLAYMOLECULARGRAPHICSLABORATGRYI 

CLINICALSUPPORTCARDIOLOGY/HEARTSURGER\ 
ITYNIHADMINISTRATIVEDATABASEPDOCCPTOGL 
GEMENTSYSTEMAPPLIEDSYSTEMSPROGRAMMIN 
PLICATIONSDATAMANAGEMENTSYSTEMSDATAB 
MMINGSYSTEMSDESIGNINFORMATIONPROCESSII 



ii^riiMi¥ii^din I c^ft-iuicn i i cnsvizraAL.di'nuunMmi 
IPROGRAMMINGTIMESHARINGTEXTEDITINGABCI 
OCUMENTATIONCENTRALCOMPUTERUTILITYFED 
IDESYSTEMRESEARCHANDDEVELOPMENTOPERA 
PMENTSERVICESFEE-FOR-SERViCEUSERSERVICE 
SCIENTIFICANDTECHNI' ¥IMUNICA ADP 

ATIONADMINISTRATIVEi-inanCIALMANAueiiviENT 
UBLICATIONSCOMPUTERIZEDBIBLIOGRAPHYSEA 
EEOPERSONALCOMPUTERSSYSTEMSPOLICYAND 
NPCGORPiNATIONFEORPPLANMANAGERIALFUNI 



Data Management Branch 



J. Emmett Ward, Chief 



Functions and Scope of Work 

The Data Management Branch (DMB) provides advice 
and assistance to research investigators, program 
officials, and administrators throughout NIH in planning 
for and obtaining computer data processing services. In 
this role the branch is a central NIH resource for 
systems analysis, design, and programming. The 
Branch is also responsible for the development, 
maintenance, and processing of the NIH Administrative 
Data Base and the Clinical Center's Clinical Information 
Utility. There are currently 50 permanent full-time 
employees whose disciplines include computer science, 
mathematics, and statistics. 

DMB staff design and create computer-based data 
management systems that provide practical solutions to 
the unique mix of administrative, scientific, and 
management data processing problems encountered at 
NIH. Each new computer system user is provided 
comprehensive training in all system facilities and 
functions of the system provided by DMB. In addition 
DMB staff teach courses about programming tools; 
provide advice on data management techniques to NIH 
programmers; serve as consultants to the B/I/D's for 
obtaining and monitoring contracting services for 
computer systems development; and create and 
maintain general purpose, user-oriented programming 
tools to speed building and improve operation of 
applications systems. 

DMB comprises four sections. The Applied Systems 
Programming Section (ASPS) and the Scientific 
Applications Section (SAS) provide general computer 
systems analysis and programming services for all of 
the B/I/D's. The ASPS supports general data 
management, and the SAS handles those projects that 
require scientific data analysis. 

The Data Base Applications Section develops and 
maintains the central administrative data base for NIH 
materiel and financial management. The Clinical 
Support Section develops and maintains the Clinical 
Information Utility as a data base for research and 
patient care in the Clinical Center. 



FY83 Accomplishments 

The Clinical Information Utility is a long term effort that, 
when completed, will provide a unique archive of 
integrated data for use in patient care and research. 
Efforts to date have involved: 

1 . the development of software to acquire and to 
make available data from the NIH Clinical Center 
Medical Information System and the individual clinical 
service activities 

2. the integration of a number of these individual data 
bases, which allows random access to the integrated 
data, and 

3. the development of software that enables users to 
make online requests for information and to receive 
automatically-generated retrievals of weekly, monthly, 
quarterly, and specified-time-period reports. 

During FY83, the entire data base moved to Mass 
Storage; three access paths were developed to enable 
patient retrievals by way of the Registry File and the 
Inverted File and directly into the integrated data base; 
one conditional path was also added to allow access 
through the inverted file. Medications, vital signs, and 
Blood Bank data were added to the integrated data 
base, and automatic scheduling of weekly, monthly, 
and quarterly retrievals was implemented. 

The NIH Administrative Data Base is an ongoing 
developmental project that uses data base technology 
in support of NIH-wide materiel and financial 
management. Significant progress has been made 
during the past year in several areas. 

As of March 1983, DELPRO became fully operational 
throughout NIH. This effort places purchasing and 
receiving for delegated authorities in the B/I/D's, and it 
required the installation of 237 terminals and the 
training of more than 1 ,000 people. A new version of 
the central procurement system was implemented in 
May 1983. This new version is more efficient and has 
been designed to be highly portable and maintainable. 
Plans call for extending this software to accommodate 
the delegated functions. 

Because of the need to conform to the Office of 
Management and Budget initiative on cash 



23 



n 



management, a shift in pnorities occurred on the 
Administrative Data Base. This shift has caused delays 
in implementing the Stock Requisitioning and Central 
and Self Service Stores Inventory systems. 

Full cash management was achieved in the Accounts 
Payable System during FY83. These functions were 
phased in according to the requirements of the OfvlB 
mandate while improving support for Accounts Payable 
personnel. 

Stock requisitioning software was completed in March 
1983, and initial training and refinement of the training 
manual began in April. This system was made available 
to the B/I/D's in June for training. Central Stores and 
Self Service Stores inventory systems will be 
implemented along with stock requisitioning in October 
1983. 

Design of the new Financial Management System was 
received during March and April 1983. Several changes 
were required, and the system is now being 
programmed by the contractor. Interface requirements 
are being defined, conversion software is being 
designed, and structured test cases are being 
developed. Plans now call for implementation during 
the latter half of FY84. 

Another project that should be of general interest at 
NIH has to do with the common problem of maintaining 
and easily retrieving bibliographic data. To resolve this 
problem DMB has been looking for an inexpensive 
method to store and retrieve personalized bibliographic 
data sets. Uses of the personal computer (PC), 
bibliographic services, and individual bibliographic 
references are being investigated. In a pilot test, DMB 
has been successful in downloading data from the 
Biosciences Information System (BIOSIS) to a PC, 
adding individual references and retrieving both, using 
the inverted techniques provided by a software product 
called SUPERFILE. 

For years now, the Clinical Center Blood Bank has 
been manually preparing antibody identification panels 
to identify those red cells that would be most useful in 
finding compatible blood types for patients with 
unusual' antibodies. In a joint effort. DMB is attempting 
to computerize these accumulated years of knowledge 

24 



to both simplify and standardize the approach. To date, 
red cells to be included in antibody panels are being 
identified, and the Blood Bank and DMB are working 
together to investigate the nature of a panel so as to 
optimize its utility to identify and quantify those features 
that are most desirable. 

For a detailed review of the many other important 
projects in which the Data Management Branch has 
been involved, please refer to the project reports in the 
DCRT FY83 Annual Report. Volume 2. These projects 
are too numerous to highlight in the summary. 

In the area of general support for NIH activities, DMB 
continued to maintain and teach courses on the Inquiry 
and Reporting System (IRS) and MARKIV; to support 
NIH use of Chemical Biological Activities (CBAC) and 
Biosciences Information System (BIOSIS) current 
awareness searches on a biweekly and semimonthly 
basis, respectively; to maintain and distribute the NCI 
Survival System; and to consult with and assist NIH 
programmers and contractors, enabling facile use of 
DCRT computer facilities. 

Future Plans/Trends 

Plans with the ADB call for the development of the 
Market Requisitioning System during the next fiscal 
year with full implementation in early FY85. As usual 
DMB will implement this system in phases to make new 
capabilities available as early as possible. During the 
second half of FY84, DMB plans to start adding the 
inventories such as Clinical Center Pharmacy, Planning 
and Control Branch, Biomedical Engineering and 
Instrumentation Branch, and NIEHS to the ADB. 
Priorities for these inventories have not yet been 
established. 

The Financial Management System will be added to the 
ADB dunng the latter half of FY84, and DMB will begin 
development of the new property system at that time. 

Future CIU efforts will concentrate on improving data 
accessibility. For less complex retrievals by Medical 
Records personnel, a 'user friendly' retrieval assistance 
system will be developed. In a more general user 
sense, software will be developed to support ad hoc 



formulation of retrievals, online definition of output 
formats and electronic delivery of output. As full 
integration of the data base nears completion, the 
classes of data that can be transmitted to the PDP-10 
for subfile creation and analysis will be expanded. 

Its role as a central resource for computer applications 
development throughout the B/I/D's will continue to 
receive pnmary support by DMB. 



Publications 

Data Managentanl Branct) Juty 1M3 

Rodbard. 0, Cote. B. •nd Munaon. PJ Ovvatopmam ol t Fn»n««». S«M 
Teaching. Interactive Statietical Package hx Ana»irs« o< Cmcal Re•earc^ 
DaU Seventh Annual Meeting ol the Soc«ly tor Cofnputef ApiAcatnrw •> 
Me<*cal Care. 1983 

The NIH Admnuntnf Data Bsaa Ju*y 19*3 



551 
2<»9'i03696 
578125<t02 
5783<i089<» 
0<i936<»332 
25822<i851 
218386755 
219<t82378 
1<»5'i82095 
21920'i<i23 
138<»<i93<t8 
215529218 



PANEL 9 1 ANTIGEN PROFI LE 



LLFFJJLL KKJJXDD 

C Moovykkuu ppssqiiP 



+ 000 + + + + 

+ 000 + + + + + + 

+ 000 + + + + + 

+ 000 + + + + 

+ 000 + + + + + + 

+ 000 + 00 + + + 

+ 000 + + + + 

+ 00 + + + + + + 

0+ + + + + + + 

+ + + + + + 

+ + + + + + 



+ + + + ♦♦0 

+ + 0*00*0*0 + 0* 

0000+ o + o* 

^00**0 + + 

0*0***0***0*0* 

+ + + + + + + + + 0* 

0-f + + + + + 0* 

+ + 0+ 0*0* 

+ + + + + + ♦ 

+ + + ♦♦0* 

+ + + * ♦♦0* 



PANEL NOT ACCEPTABLE - DONOR LACKING ANTIGEN E- CANNOT BE FOUND TO REPLACE I38^<i95^8 



In