(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"

Division of 
Computer Research 
and Technology 

Annual Report of Program Activities 
October 1, 1978 tlirougli Septennber 30, 1979 




sscs^ 



TABLE OF CONTENTS 

Page 

DIRECTOR'S SUMMARY 1 

OFFICE OF THE DIRECTOR 

Office of Scientific and Technical Communications ..,..,. 10 
Research Project Report 

Basic Computer Algorithms for the Applications of Cytology . 17 

Office of ADP Policy Coordination 19 

Office of Administrative Management 23 

LABORATORY OF APPLIED STUDIES 

Summary of Activities 28 

Annotated Project and Activities List 33 

Publications and Presentations List 35 

Research Project Reports 

Mathematical Models and Simulation Programs in Physiology . 38 

Mathematical Models of Binding Equilibria 40 

Computer Aided Analysis of Electrocardiograms 42 

Computer Systems for Nuclear Medicine 45 

Computer-based studies in ultrasonography 49 

Statistical Resea:rch in Clinical Pathology 52 

Mathematical Modeling of Substrate Transport in 

Physiological Environments 56 

Analysis of Coupled Transport and Biochemical Kinetics ... 59 

Computer-based Studies in Pulmonary Pathophysiology and 

Respiratory Disease 63 

Investigation of Hybrid Computing for the Construction of 

■Simulation Models and for the Analysis of Physiologic Signals 67 

Computer Based Analysis and Image Processing in Electron 

Microscopy and X-ray and Electron-Loss Spectroscopy .... 71 

Generalized Mathematical and Computational Methods 74 

Numerical Approosimation Techniques for the Solution of 

Reaction-Diffusion Systems in Biology 77 



COMPUTER SYSTEMS LABORATORY Page 

Summary 82 

Annotated Project and Activities List 88 

Publications and Presentations List 102 

PHYSICAL SCIENCES LABORATORY 

Summary 106 

PSL Projects and Activities FY79 109 

Publications -. Hi 

Research Project Reports 

Theory of Biochemical Separation Techniques 114 

Theory and Measurement of IntermoteouloT Forces 116 

Consulting Services 119 

Correlation Function of Spectroscopy /Laser Light Scattering 122 

Cell Motility and Chemotaxis 125 

Theory and Application of Buolear Magnetic Resonance 

Spectroscopy 127 

Interactions of Erythrocyte Spectrin with Actin 131 

Studies in Mathematics and Statistics 133 

Quantitative Analysis of Electronmicrogra:phs and 

Membrane Cellular Biophysics 135 

LABORATORY OF STATISTICAL AND MATHEMATICAL METHODOLOGY 

Summary 140 

LSM Projects and Activities 147 

Publications • • 152 

Research Reports 

Automated Data Processing of Medical Language ....... 154 

Biological and Visual Shape 157 

Cluster Analysis 158 

Discrete Mathematics and Applications 161 

Linear Methods in Statistics . 163 



n 



Page 

Non-numerical programming techniques and applications . . , 165 

Multivariate Statistical Analysis 168 

Research Topics in Computer Science 170 

Nonlinear Equations 172 

DATA MANAGEMENT BRANCH 

Summary of Activities 176 

Annotated Project & Activities List 180 

Publications 201 

COMPUTER CENTER BRANCH 

Summary 204 

Activities, Services and Facilities 208 

Publications 210 



m 



DIRECTOR'S SUmARY 



The success of DCRT in FY 1979 stands forth in the details of this 
annual report as a tribute to three factors: 

1. Fundamentally sound principles behind the NIH decision to establish 
a "computer" division in 1964; 

2. Continuing high quality of work by members of the DCRT laboratories 
and branches across discipline and organizational lines; 

3. Effective long-term working relationships between DCRT staff and 
scientists and administrators throughout NIH. 

The Director, NIH, established DCRT in 1964 to embody certain principles 
set forth by the report in 1963 from an NIH study of computing and data 
processing at NIH. These concepts included: 

• A central NIH resource to conduct research and technical development 
in computing as well as to provide computing services for NIH. 

• A multidisciplinary organization with several distinct laboratories 
and branches--combining expertise in mathematics, statistics, elec- 
tronic engineering, and a variety of biomedical disciplines — as a 
way to reduce recognized barriers in making a powerful emerging com- 
puting technology an integral part of the NIH programs. 

f A division co-equal within the NIH structure to carry the primary 
responsibility for this new approach to computing at NIH. 

Two Illustrative Stories 

The subsequent history of DCRT and computing at NIH has illustrated the 
validity of these basic ideas. The following two stories are examples 
of long term collaborative efforts possible only under conditions de- 
signed to foster multidisciplinary and multigroup interaction among DCRT 
staff and NIH scientists. 



The first story began in 1969 when the DCRT Laboratory of Applied 
Studies (LAS) conceived a project to apply computer processing techni- 
ques to x-ray-like images recorded by so-called "gamma cameras" from 
minute amounts of radioactive materials administered by the Nuclear 
Medicine Department of the NIH Clinical Center. The DCRT Computer Sys- 
tems Laboratory (CSL) joined LAS to specify the appropriate camera-com- 
puter system since none was available commercially. The first five 
years of this project went largely into developing hardware and software 
systems and mathematical techniques and proving their feasibility. 

Then quite rapidly over 15 months in 1975-76, basic techniques were 
applied by members of the Nuclear Medicine Department and the Cardiology 
Branch to exercise tests in patients with coronary artery disease, dem- 
onstrating a valuable new tool for diagnosis and treatment. Publication 
of their results triggered a nationwide recognition of the value of the 
technique. It is now commercially available from several companies and 
is a routine procedure in major hospitals (e.g., N. Eng, J. Med. 301:94, 
1979). 

Meanwhile, the collaboration between DCRT and the Nuclear Medicine Branch 
also has provided NIH with unexcelled facilities for computer processing 
of dynamic images from the heart, lungs, kidneys and other organs. One 
of the current collaborative projects in this lineage combines expertise 
from the Pulmonary and Clinical Hematology Branches, NHLBI, with DCRT 
and Clinical Center staff in studies of lung disease and its effects on 
pulmonary functions. 

The second story began in 1968 with a decision by the DCRT Division 
Director to develop a computer based dynamic display for chemical struc- 
tures. Initial development of the hardware and software systems again 
took several years in a collaborative project between members of the 
Computer Center Branch (CCB) and the Heuristics Laboratory (HL) in DCRT. 
(HL was abolished in 1974 as the result of a DCRT reorganization caused 
by loss of positions during the 1970-74 Federal employment retrenchments.) 

A number of versatile techniques were exploited in the early and mid 
1970s to provide black and white "skeletal" models of macro molecules 
used in projects by scientists in the DCRT Physical Sciences Laboratory 
(PSL) and the Laboratory of Molecular Biology in NIAMDD. 

The ability to produce stereo images and moving pictures of molecules 
was particularly useful. An atlas of macromolecular structures on 
microfiche made inexpensive stereo views of scores of large molecules 
available to scientists throughout the world, who did not have computer 
systems comparable to the one available to NIH scientists. 

In 1978 the DCRT armamentarium of graphic techniques was expanded with 
new hardware and software to create "space filling" full color models 
from the extensive library of macromolecular coordinates compiled by 
DCRT staff in conjunction with biochemists and x-ray crystal lographers 
around the world. 



In FY79 the new system was used by the staff of other DCRT and NIH 
laboratories for a widening array of chemical and medical images with 
NIDR, NIAMDD, NINCDS, NHLBI, NCI and the Clinical Center. 

Two Later Developments 

This pattern of long-term investment in collaboration to develop and ex- 
ploit a set of basic technical tools appears in two other projects that 
are now beginning to show promise: 

• Data base management expertise developed over the last five years by 
CCB and the DCRT Data Management Branch (Di^lB) has provided the basis 
for collaboration with NIH administrative management divisions. 
Projects to improve NIH material management and accounting systems 
are well underway. 

• A clinical information data base technology was developed over the 
last six years by DMB's working with the Clinical Center's Office 
of Clinical and Management Systems. This has been used by 170 NIH 
clinical scientists to retrieve specific kinds of data from some 

15 million test results recorded by the Clinical Laboratories of the 
Clinical Center. During FY79 the Office of the Director in the 
Clinical Center has moved to use the technology to improve medical 
records in support of patient care by all NIH clinical services; 
collaboration has begun with the DCRT Laboratory of Statistical and 
Mathematical Methodology (LSM) to provide improved facilities for 
building and analyzing subsets of clinical research data. 

The Pervasive Pattern 

Oyer the years multidisciplinary collaboration among members of the DCRT 
has also been an essential part of many "short term" projects in support 
of or in collaboration with members of other NIH Institutes and Divisions. 

A survey this year disclosed 50 projects in each of which members of two 
or more of the six DCRT Laboratories and Branches worked together to 
provide mutually supportive scientific, mathematical, engineering, or 
computing expertise. Indeed this mutual support is an accepted part of 
day-to-day activities. It goes without notice in many annual reports of 
the DCRT staff in much the same way that much of the cross-supporting 
activity among scientists in the NIH Institutes is taken for granted. 

Access to Computing Throughout NIH 

There is another equally important measure-of the validity of the con- 
cepts and principles built into the DCRT. This is the computing done 
independently by NIH staff using the facilities provided by the DCRT. 

In FY79 there were over 3000 registered NIH users of the computer center 
designed and operated by CCB. A study of the patterns of use in FY78 
indicates that about half this $10 million expenditure in FY79 will have 
supported NIH intramural biomedical and biometric programs. The other 



half is distributed across the extramural, collaborative and adminis- 
trative management programs of NIH. 

Indeed, this very mixture of science and administration at NIH was 
recognized as a unique circumstance by the 1963 recommendation to es- 
tablish a separate division for computing at NIH. It proved to be a 
driving force behind the development by CCB of a balanced "Computer 
Utility" that combines facilities for scientific computing and adminis- 
trative data processing. Programs written by CCB systems programmers 
link these distinct components into a single functional system, the 
total power of which was available in FY79 through 1800 remote terminals 
to all users in their diverse working environments, scientific and 
administrative. 

Central Support for Independent Users 

It is well known that even the best computer systems are necessary but 
not sufficient to produce successful computing, particularly for people 
whose primary work is not computers. Some of the other essential in- 
gredients are provided at NIH by additional expertise within DCRT: 
in DMB, in LSM, and in the User Services and Assistance Section of CCB. 

DMB and LSM have developed or adapted a set of data management and 
analysis tools that support the use of the NIH Computer Utility through- 
out NIH. These tools are packages of programs that can be used directly 
for many computer applications. DMB and LSM teach the use of these 
packages in courses given twice each year. They provide advice for 
users and technical support for the packages throughout the year. In 
FY79 these programs on the NIH computers were used hundreds of thousands 
of times. 

The User Services and Assistance Section provides support for program- 
mers and people who want to learn to program. In FY79 they organized 
or conducted 64 computer training courses and seminars for 1300 appli- 
cants. The section handled almost 20,000 calls or visits for assis- 
tance during the year and provided over 80,000 technical manuals and 
other documents to guide programmers and other users of the NIH central 
computing facilities. ^ 

Support for Computing Outside DCRT 

The DCRT role in the successful application of computing throughout NIH 
goes beyond computing at the NIH Computer Center, The 1963 plan saw 
the evolution of smaller computer systems, particularly in the labora- 
tories and clinics of NIH. In FY79, as in previous years, CSL worked 
on more than two dozen projects to put minicomputers or microcomputers 
to work where they were better suited to particular scientific needs 
than a central computing facility. 

In addition to the computing in which DCRT has a role, NIH makes use of 
outside sources for computing resources in both intramural and collab- 
orative programs. In FY79 the DCRT ADP Policy Office reviewed over 450 



Drocurement actions for computer equipment and services throughout NIH. 
Many of these reviews called on expertise in the DCRT laboratories and 
branches for advice. 

Another area of support centers around access to books, journals, and 
reports that provide insight to what has been done in computer science 
and technology and its applications to biomedical science. The DCRT 
Library, an independent "full service" library which began as support 
for the diverse needs of the DCRT staff, now has more active registered 
users from the NIH staff outside DCRT than from the DCRT itself. 

Outside NIH 

The success of computing activities at NIH continued to attract visitors 
and inquiries from across the country and quite literally around the 
world In FY79 they came from every continent (except Antarctica), rrom 
the USSR, Eastern Europe, and Mainland China, as well as more accessible 
areas. Recognition within the Federal government itself came in the 
form of requests for DCRT staff to serve as advisors, consultants, and 
collaborators with other agencies and for access to the NIH computing 
facilites. 

Activities in the Federal Sphere 

The desire of other Federal agencies to benefit from the NIH accomplish- 
ments in computing continues to be a mixed blessing to DCRT. In FY79 
the General Services Administration formally recognized the general 
purpose NIH computer center facility provided by CCB as a Federal Data 
Processing Center for Biomedical and Statistical Computation and the 
CCB Chief, Mr. Joseph Naughton, received the "Man of the Year award 
from the Federal Interagency Committee on Automatic Data Processing. 

Although many other agencies wished to use NIH facilities, DCRT had to 
continue its policy, established in 1974, restricting outside users, 
virtually denying access to new agency applications for use of the _ 
facilities. Despite these limitations, DCRT found itself embroiled in 
the efforts of private computer service firms to obtain business from 
current users of the NIH computing facilities. These efforts included 
Congressional Inquiries and a spate of Freedom of Information requests. 

Computing for NIH Science 

Notwithstanding the facts noted above (half the NIH computer center ex- 
penditures in direct support of scientific activities and the vast bulk 
of CSL work on computer systems for NIH laboratories and clinics) there 
has been concern about computing resources for science at NIH. 

Part of this concern reflects a growing recognition that computing is 
indeed essential for much of biomedical science. The situation in some 
parts of biomedical science has reached the stage stated by Weisskopt 
in his recent article on "Contemporary Frontiers in Physics_ (Science 
203: 240, 1979): "Modern physics without modern computers is 



unthinkable." As Weisskopf noted, "There is a special reason why new 
discoveries were made... namely the exuberant growth of new instrumen- 
tation and the maturing of previous innovations which become ready for 
full exploitation. A partial list of examples includes lasers... and, 
most important, computers." 

But another part reflects the belief that any large expenditures of 
funds deserve examination. This view led the NIH Deputy Director for 
Science to create, in FY78, an NIH Advisory Committee on Computer Usage. 
FY79 was the first full year of its activity. It took a strong position 
to insure that the new Federal Data Processing Center role for NIH did 
not cause NIH scientific use of computers to take second place. 

Procurement of Improved Computing Machinery 

DCRT made some progress on the perennial issue of procuring replacement 
for the aging hardware in the NIH computer utility. An additional year 
of work with the General Services Administration in FY79 produced Re- 
quests for Proposals to replace the existing DtC System 10 and IBM 370 
computers. It is impossible at this point to predict what systems will 
be proposed by vendors, how much time will be needed for the rest of 
the selection and procurement process, and what problems will be en- 
countered in installing the new equipment as an integral part of the 
NIH computing activities. It is clear, however, that newer computers 
already in production and installed elsewhere in the nation can provide 
a significant increase in computing capacity at less cost to NIH users. 

Restricted Employment Levels 

Another perennial issue not forseen in the original plans for DCRT is 
the recurring cuts in DCRT staff. The 1963 report predicted some 600 
employees in the Division by 1968. In fact the allocation of positions 
for DCRT peaked in FY 1969 at 316 and has followed an almost monotonic 
decline ever since. FY79 brought a new loss of 16 positions from those 
budgeted for the year. 

The cumulative effect over 10 years has been to reduce the original 316 
positions to 235. Other organizations have been willing to transfer 
some positions to DCRT to support activities of importance to them, 
primarily the service functions. But these gifts have left the Division 
with a hidden "nutritional" deficiency. The overall DCRT staffing level 
masks starvation in many of the multidisciplinary areas essential for 
computing excellence at NIH. 

The basic principles stated in 1963 still apply. But the emergence of 
new opportunities in areas such as non-numeric computing and image pro- 
cessing will require the addition of some new expertise. There will be 
a need for young scientists to bridge the interface between computer 
expertise and the science of new application areas. How DCRT can modify 
its programs to adjust to these circumstances is not yet clear. Renewal 
is not easily accomplished in periods of retrenchment. 



Another Possible Problem 

An additional small cloud over the future plans of DCRT is the congres- 
sional ly mandated review of all NIH activities under 0MB Circular A-76. 
The history of DCRT and computing at NIH shows that all parts of DCRT 
have functioned as an integral and inseparable part of the NIH core 
activities. It seems inconceivable that parts of the Division could be 
put out on bid as a separate "commercial or industrial activity" with- 
out irreparable harm to the intricate web of functions that have de- 
veloped in computing at NIH over the last decade. But the word that NIH 
intramural scientific research itself faces this very A-76 review makes 
the inconceivable somewhat less unthinkable, but no less threatening 
and potentially destructive. 



A 



b^n/ //, 




Arnold W. Pratt, M.D 
Division of Computer Research 
and Technology 



yi' 



OFFICE OF THE DIRECTOR 




The first month of FY79, DCRT took part in an NIH 
Saienoe Writer's Seminar on Computer Medicine. 
Conversing before the program were (I. to r. ) 
Dr. Thomas Lewis of the Clinical Center with 
DCRT Director Dr. Arnold Pratt and two of his 
staff y Dr. Judith Prewitt and Mr. Martin 
Epstein. 



OFFICE OF SCIENTIFIC AND TECHNICAL COMMUNICATIONS 



William C. Mohler, M.D. 
Chief 



I . SUMMARY 



Function and Scope 

The DCRT office of Scientific and Technical Communications, under the 
direction of the Associate Director, DCRT, includes: 

• 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 its relationship to biomedical research. The 
office handles a wide range of inquiries, produces print and audio- 
visual communications, coordinates special events, and provides 
advice, assistance and educational resources for DCRT staff. 

t The DCRT Library , which maintains a special collection in computer 
science and mathematics, statistics, engineering, information sci- 
ence, and management. The library directly supports DCRT activities 
and is a resource for other NIH staff. It also functions as an in- 
tegral part of the local Washington area network of special libraries, 

t Scientists assigned to this office, working on research and develop- 
ment projects in biomedical data bases, image processing, and de- 
cision analysis. 

Highlights of FY79 Activities 

The DCRT Information Office got a new Information Officer, Patricia 
Orsino Miller, in December. She and Public Information Specialist Mary 
Hodges, the other member of this small office, undertook a widening set 
of activities during the year. 



10 



In addition to over 5,000 varied public inquiries about DCRT and com- 
puting in biomedical activities, a minor flood of Freedom of Information 
requests arrived during the last half of this year. Many of these and 
two Congressional Inquiries through the Office of Management and Budget 
appeared to arise from the desire of a private computer service organi- 
zation to recruit business among users of the NIH Computer Center. 

A new portable exhibit unit, designed by the office, was first used in 
the spring in support of a DCRT Computer Systems Laboratory demonstra- 
tion at the President's Conference on Employment for the Handicapped. 
A few weeks later it proved its "modular, multipurpose" capability by 
appearing with new graphics at the ceremonies establishing the general 
purpose facility of the NIH Computer Center as a Federal Data Proces- 
sing Center for Biomedical and Statistical Computing. 

These two events, arranged by the Information Office, were examples of 
a growing interest outside of NIH in DCRT activities. In addition to 
preparing news articles for local papers, the office responded to re- 
quests from IEEE Spectrum and Smithsonian magazines and a Canadian TV 
station for feature stories, still photography, and 16mm footage describ- 
ing DCRT work. 

But most of the work of the office during FY79 continued to focus on in- 
formation for NIH scientists and others interested in what DCRT does and 
how it can help their work. This meant the hidden labor of revising, 
reprinting, and distributing some 3,000 copies of several brochures de- 
scribing the activities of the division and its laboratories and branches. 
It included arranging and helping other DCRT staff give tours and talks 
for groups who came to the DCRT facilities and presentations about the 
results of DCRT projects. The office produced a simple building direc- 
tory for visitors that can be revised easily over the coming years as 
renovations cause relocation of activities in the Building 12A/B complex. 

To help the division staff prepare and present their own direct personal 
communications, the office introduced a new set of activities as part of 
its longer range DCRT information program. It arranged two staff train- 
ing seminars in effective speaking techniques and use of visual aids and 
it produced two sound-on-slide shows as prototypes for presentations 
that could be tailored to the needs of particular groups. 

During FY79, the Information Officer served on the NIH Information 
Officer's Training Committee, was active in the National Association of 
Government Communicators and was elected vice-president of Women in 
Communications, Inc., a national organization for journalism, broadcast- 
ing, public relations, and communications professionals. 

The DCRT Library operations increased in volume during FY79. The Li- 
brarian, Ellen Chu, and the library technician, Jill Lindau, handled more 
acquisitions, circulation of books, and on-line information searches for 
more registered users than in FY78. 



11 



The increased workload with no increase in staff reflects judicious use 
of automation. A national cooperative network, OCLC, helped identify, 
locate, buy, classify, catalog, and borrow books and a new automated 
"circulation system" went into active use in March. The circulation 
system, developed by the Library and DCRT Data Management Branch, assures 
more accurate records of loaned materials. It provides better informa- 
tion on the status and location of books and documents on loan and gene- 
rates a wide assortment of useful messages and reports for library 
operations. 

As part of the work in coding items for the masterfile of the circula- 
tion system, the library staff reviewed and reclassified the book collec- 
tion to fit changes and expansions that have occurred in the Library of 
Congress classification system since the Library opened over a decade 
ago. Part-time and temporary employees helped with the massive manual 
task of typing in entries for the master files, putting new labels on 
books, reshelving the books according to the new call numbers, and 
affixing small printed labels to about 20,000 catalog cards. 

Addition of an extra 230 feet of shelving and transfer of some bound 
journals to microforms reduced crowding on the shelves. A new Minolta 
450 reader-printer, purchased to handle the variety of these microform 
formats, provided an unexpected benefit for editing some 16mm film de- 
scribing work in the Computer Center Branch. 

The number of on-line bibliographic searches increased over 50% this 
year. MEDLINE and Computer and Control Abstracts were the most heavily 
used of over 80 data bases now directly searchable at the Library's on- 
line terminal . 

The striking change in the pattern of interl i bra ry loans continued. In 
earlier years the Library borrowed more than it lent. Last year it 
broke even. In FY79 the requests from other libraries to DCRT were 
almost twice the requests DCRT made to others. This increase is due to 
the quality of the DCRT Library collection and the service it provides 
and to the active participation of the Library and Librarian in a num- 
ber of formal organizations: 

t The Interl ibrary Users Association of the Washington-Baltimore Area 

• The Metropolitan Washington Library Council 

• FEDLINK (a Federal Library consortium) 

• The national OCLC library users network. 

The Scientific Staff pursued the level of work established in previous 
years. Dr. Judith Prewitt was again yery active outside of NIH orga- 
nizing meetings, working groups, and collaborative projects. She con- 
tinued work with the National Cancer Institute staff on their Diagnostic 
Radiology and Cytology Automation projects. Dr. Shun Chung Wu extended 
his work on the development of computer algorithms for applications to 
image processing in cytology. 

12 



Martin Epstein expanded the documentation of his MEDINQUIRY System and 
some of its uses on the melanoma data base, and he collaborated with 
several scientists outside NIH in developing plans for further applica- 
tions of the system. 

Dr. William Mohler again organized the NIH clinical elective on Compu- 
ters in Clinical Medicine and worked with the Deputy Director and the 
Associate Director of the NIH Clinical Center on projects to use the 
DCRT Clinical Information Utility in condensing medical records, to 
evaluate the Clinical Center's Medical Information System and to explore 
data analysis facilities for clinical research. 

Future Plans 

The activities of this office will continue the major lines followed in 
FY79 and previous years with the following emphases: 

The Information Office plans to: 

t update gradually the content and upgrade the format of its existing 
information publications and add others for complete coverage of 
DCRT activities 

• begin a program to monitor and evaluate the effectiveness of the 
Division's present information program among the NIH Scientific and 
Administrative staff 

t complete and extend its computer-based index of all DCRT scientific 
papers to provide retrieval by author, topic, date, and organiza- 
tional component within DCRT. 

The DCRT Library plans to refine the design and use of its circulation 
system with an emphasis on statistical reports to aid in decisions 
about modifying the library's collection and services to its users. To 
prepare for the impending nationwide change to new Anglo-American cata- 
loging rules, it will study the possible conversions of subject head- 
ings and authority files. 

Plans for the projects of the scientific staff will emphasize some 
efforts toward greater application of the methods developed to date for 
selected biomedical applications in order to evaluate this utility. 



13 



11. PROJECTS 



(The FY79 project reports of Dr. Prewitt were delayed this year and are 
not included in this list.) 

1. Natural Language Access to Clinical Data Bases. M.N. Epstein. The 
objective of this project, begun in FY77, is the development and 
evaluation of a system that will allow physicians access to medical 
data through natural language queries to support both patient 
management and clinical research. The MEDINQUIRY prototype system 
and melanoma application were documented in FY79. Other applica- 
tions were discussed with the National Library of Medicine. 

2. DCRT Library Circulation System . E. Chu (Library), J. Mahaffey (DMB) 
This system, started in FY79, maintains files of records covering 
monographs and documents in the DCRT library collection, users auth- 
orized to borrow them, and transactions on the circulation of items 
from the collection to the users. In 1979 the library completed the 
files on all existing items and users. The system is now fully 
active. Work continued on a document index and statistical reports 
as products of the system. 

3. DCRT Information Program Plan . P. 0. Miller and W. C. Mohler. The 
project began in FY77, became dormant in FY78 and revived in FY79 
under the new information officer. The preliminary program has five 
inter-related parts: Communicative Production and Evaluation, Re- 
source Development, Communication Pathway Development, DCRT Staff 
Education, and Needs Analysis. Activities in FY79 dealt with Com- 
munication Production, with some initial seminars for staff edu- 
cation on communication techniques. 

4. Computer Algorithms for Cytology . S. C. Wu and J. M. S. Prewitt. 
This project started in FY78 to develop new basic computer algo- 
rithms for boundary finding, object extraction, and measurement of 
shape and texture features. Work in FY79 resulted in new techniques 
in several of these areas. 



14 



HI. PUBLICATIONS 



Epstein, M. N., and Walker, D. E.: Natural language access to 
melanoma data base. Proc. of Second Annual Symposium on Computer Appli - 
cations in Medical Care , IEEE Computer Society, 1978, pp 320-325. 

Mohler, W. C: Introduction to Representation of Medical Knowledge. 
Proc. of Second Annual Symposium on Computer Applications in Medical 
Care , IEEE Computer Society, 1978, pp 364-5. 

Prewitt, J. M. S., and Wu, S. C: An Application of Pattern Recog- 
nition to Epithelial Tissues, Proc. of Second Annual Symposium on Com- 
puter Applications in Medical Care , 1978, pp 15. 

Computers: Tools for the Advancement of Medicine , DHEW Pub. No. (NIH)- 
79-1039, 28 pp. 

Computing Resources of the Division of Computer Research and Technology . 
1978, 48 pp. 

Division of Computer Research and Technology; Report of Program 
Activities October 1, 1977 through September 30, 1978 , 180 pp. 

Physical Sciences Laboratory , DHEW Pub. No. (NIH)79-1925, 12 pp. 



15 






/. 




Dr. William C. Mohlev, DCRT Associate Director ^ 
organizing the fifth NIB clinical elective on 
Computers in Clinical Medicine. 



16 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECnUJMBER 



PERIOD COVERED 



October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Basic Computer Algorithms for the Applications of Cytology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES Of PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

Shun Chung Wu, Senior Staff Fellow 



COOPERATING UNITS (if any) 

None 



lab/branch 
Office of Scientific and Technical Communication 



SECTION 



INSTITUTL AND LOCATION 

DCRT/NIH 



TOTAL MANYEARS: 

2.0 



PROFESSIONAL: 

2.0 



OTHER: 



0.0 



CHECK APPROPRIATE BOX(ES) 
C (a) HUMAN SUBJECTS 

n (al) MINORS D (a2) INTERVIEWS 



£1 (b) HUMAN TISSUES 



D (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

(1) A decision tree method has been developed to distinguish the 
Internal pixels of an object with a given chain encoded boundary, 
regardless of the Intensities of these pixels are higher or lower 
than that of background, (2) An objective boundary tracking 
method has been created to select one out of several Isodenslty 
contours. The criterion for selection is based on an optimizing 
objective function of area, shape and Integrated density. 
(3) Several measures of visual texture, based on the average 
local intensity, have been generated. Comparison between this 
texture and the texture based on transition probability matrix 
has been studied on bladder tissue sections. (^) A concavity 
finding algorithm for the chain encoded boundary has directly 
been produced. (5) These algorithms have been applied to the 
images of bladder cancer nuclei, white blood cells and computer- 
ized tomography. 



PHS-6040 
(Rev. 10-76) 



17 



Project Description! 

Objectives! 

Because of the lack of the sound basic computer algorithm, the 
percentage of correct diagnosis made by computer in cytology 
application is lower than that by human expert. Such algorithms 
should include the methods for the finding of true boundary, 
extraction of object, various accurate and meaningful measures 
of the shape and texture features. If the boundary found by 
computer algorithms Is deviated from the true boundary, all 
subsequent feature measures and classifications will be 
subjected to errors and mixed with unwanted noises. For a 
malignant tumor nucleus, the Intensities of pixels Inside the 
nucleus can be either higher or lower than that of background. 
In order to measure the texture of the tumor nucleus, an 
extraction algorithm Is to be developed to distinguish the 
Internal pixels of the nucleus from the outside region. These 
basic algorithms are necessary for successful cytology diagnosis 
by using computer technology. 

Significance J 

The applications of these algorithms to medical Images of white 
blood cells, CT scans and bladder tumor tissue sections indicate 
that not only the quality of the measured features has been 
improved but the percentage of correct classification has been 
increased also. It is very hopeful that the computer can be used 
to do a faster diagnosis than human expert in the practical 
applications. 

Proposed Course* 

At the present time, the decision tree method to extract object 
can be applied only to the chain encoded boundary without 
filaments with 8-neighbor connection. This tree method ought to 
be modified in order to be applied for ^-neighbor chain encoded 
connection. Also, the boundary found by optimizing objective 
function and Isodenslty contours may deviate from the maximal 
edge pixels in certain circumstances. A new approach by combining 
both the method of the optimizing objective function and the 
method of the maximum edge linkage is to be developed. 

Publications! 

Wu, S.C. and Prewitt, J. M.S., Deriving Concavities from the 
Fourier Coefficients and Its Role in Pattern Recognition. Proc. 
of the 30th Ann. Conf. on Engineering in Medicine and Biology, 1977. 
Wu, S.C, Prewitt, J. M.S. and Lehman, J. 1 To Extract a Connected 
Object of Arbitrary Shape from its Background by Decision Tree 
Method, IEEE Comp. Soc. Conf. on Pattern Reco. and Image Pro. 1978. 
Prewitt, J. M.S. and Wu, S.C.i An Application of Pattern Recog- 
nition to Epithelial Tissues. The Second Ann. Symposium on 
Computer Application in Medical Care. 1978, pp. 15« 

18 



OFFICE OF ADP POLICY COORDINATION 



Henry J. Juenemann 
Chtef 



I. Summary of Activities 

Functions 

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

• It serves as a focus for NIH-wide coordination of automatic 
data processing policy matters. 

• It serves as a central NIH point of contact with the Public 
Health Service, the Department of Health, Education and 
Welfare, other HEW Agencies, The General Services 
Administration, and the Office of Management and Budget on 
policy questions and NIH's participation in policy 
development. 

f The Office also provides advice and assistance concerning 
internal DCRT operations and coordinates DCRT's ADP policies 
and activities with those of other agencies. 

Scope 

The role of the office includes: advising the Director of DCRT and 
through him the Director of NIH on ADP policy matters; assisting the 
NIH Division of Management Policy on questions relating to its 
responsibility for administrative and management computer applications; 
reviewing and evaluating proposals from NIH B/I/D/O's for ADP and 
computing related procurements and contracts; directing the develop- 
ment of the annual NIH ADP Plan; representing the NIH in PHS and DHEW 
policy formulation efforts; working with GSA staff on procurements; 
coordinating Interagency Agreements with other Federal agencies that 



19 



use DCRT facilities; and answering inquiries from scientists and 
administrators who are confused hy the whole process. 

Highlights of FY79 Activities 

The annual ADP Plan has changed gradually in the last few years to 
become an expanded version of Exhibits 43A and 43B of budget submissions 
required by A-11. This year the Plan was a part of the FY81 budget 
submission and was again coordinated by this office. This management 
process, involving all B/I/D/O's provides a forward look at projected 
ADP efforts necessary to support NIH research and research management 
programs. The submission showed a total of $51,402,000 FY79 ADP 
related obligations offset by a net of $8 million in receipts from non- 
HEW Federal Agencies. The projected figures for FY81 were $60 million 
offset by an estimated $8.1 million. Projected work years rose from 
652 in FY79 to 688 in FY81 . 

One major highlight of the year was the designation of the DCRT central 
general purpose computer facility as a Federal Data Processing Center. 
This agreement was signed on May 31, 1979 by representatives of GSA, 
HEW, PHS and NIH culminating several years of negotiations aimed at 
recognizing the long record of DCRT's technological leadership in 
Federal computing and our mutually beneficial participation in the 
Sharing Program envisioned by the Brooks Law, The agreement was cast 
in terms assuring that the new role would not disrupt the computing 
support needed by NIH's own research programs. A second form of 
recognition was received by NIH when this office's nomination of the 
Chief of the Computer Center Branch to be Federal "ADP Man of the 
Year" was accepted on behalf of the whole Federal ADP community by the 
Interagency Committee for ADP. Mr. Joseph Naughton was accorded that 
honor on June 12, 1979. 

In the process of monitoring the policy implications of NIH's ADP 
involvements and assuring conformity to existing 0MB, GSA and 
Department regulations, this office reviewed 456 proposals for ADP 
equipment or services during the period. This is more than half again 
over the corresponding figure for last year. This office continues to 
be very fortunate in being able to call on the expertise of the other 
laboratories and branches of DCRT. During the year they were extremely 
helpful in insuring that technical merit is able to be a part of the 
review process. 

The major efforts required to reprocure both the 370 and DECSystem 10 
facilities continued throughout the year. By the beginning of the 
fourth quarter, it appeared that we had reached agreement with GSA on 
a draft of the 370 solicitation to be released to industry for comment. 
Competitive solicitations for supplying NIH standard 30 and 120 CPS 
hard copy terminals were successfully completed as was the procurement 
of Mass Storage Units to augment on-line data storage capabilities. 



20 




One major highlight of the year was the designation 
of the DCRT general purpose oomputer facility as a 
Federal Data Processing Center (FDPC). The May 21 
event was the culmination of several years of 
negotiations, largely the effort of DCRT Assistant 
Director Henry Juenemann (third from left). 




PHS Deputy Assistant Secretary for Health Operations 
Charles Miller, GSA Assistant Commissioner Dr. Robert 
Coyer, and NTH Director Dr. Donald Fredrickson 
affixing their signatures. Five officials- 
representing GSA, HEW, PHS, and NIH— signed the 
FDPC agreement. 



21 



Future Plans 

Federal ADP policies and requirements continue to become more and more 
complex as 0MB, GSA, DHEW and PHS become more and more involved. As 
a result, the Office must spend an increasing portion of its available 
man-hours in attempting to guide NIH policy in productive directions 
and in meeting regulatory requirements. The Office will continue its 
work to spare large numbers of NIH research and research support staff 
members the task of becoming expert in the many nuances of ADP-related 
regulations. However, it is anticipated that the two simultaneous 
full recompetition efforts for both of NIH's main computer systems will 
consume a significant portion of the resources of this office during 
the upcoming year and require extensive assistance from OD and CCB 
Staff. 




One of the signers of the agveement, GSA Commissioner 
Frank Carr^ is seated at the table. Another signer^ 
HEW Deputy Assistant Secretary L. David Taylor, 
addresses officials from the U.S. Congress , NIH, PHS, 
HEW, 0MB, and GSA. 



22 



OFFICE OF ADMINISTRATIVE MANAGEMENT 



L. Lee Manuel 



Chief 



SUMMARY 

Function 

The Office of Administrative Management, under the 
direction of the Executive Officer, provides 
administrative, financial and personnel management for 
the Division's program. The office serves as liaison to 
these functions with the NIH Office of Administration and 
other NIH, PHS and other DHEW offices. 

Scope and Activities 

The office handles the usual range of adiiiini strati ve 
managerial functions for an NIH research division of 
almost 300 people. The Financial Management/Project 
Control Sections maintains 2000 project accounts 
involving 6400 registered users of the DCRT coniputer 
facilities and services. These services total an 
estimated $26,000,000 in FY 78. Requisitions, contracts, 
travel, and training documents are processed by the 
Administrative Services Section and covers a variety of 
procurements of approximately $20 million. Day to day 
activities also cover support in areas of space, 
telephone changes, payroll and other miscellaneous 
administrative subjects. 

Highlights of FY79 Activities 

The Personnel Section continued its mandated three year 
position classification review of all positions and 
reported several to higher authorities for both 
downgrading and upgrading. The Civil Service Reform Act 
required implementing the senior executive service at the 
DCRT level and new evaluation criteria to cover appraisal 
of supervisors and managers in the GS-13 through 15 
1 evel . 

The Financial Management Section was the primary focus 
for development and coordination of basic materials for 
the DCRT portion of the NIH Forward Plan, Zero Based 
Budgetary Workload/Manpower reporting and the normal 
Federal Budget process. 



23 



New requirements mandated by the Secretary, DHEW 
required intensive efforts by the Administrative staff in 
the areas of contracting, procurement, travel and 
consultant services. Most of these requirements called 
for detailed plans to be prepared and monthly and 
quarterly reporting requirements were mandated. Also 
certain delegations of authority were elevated to higher 
levels. All of these actions tended to increase the 
amount of "paperwork" required and further slowed the 
already cumbersome administrative processes. A new EEO 
effort required an in-depth analysis of the DCRT 
workforce. The DCRT Administrative Information Retreival 
System (AIRS) was used to generate our staffing charts 
this year. This eliminated a lengthy typing chore for 
the staff. 

The Project Control Office ran a pilot survey of NIH 
users of DCRT services to determine the extent to which 
these services support major categories of NIH work. 

Future Plans 

The coming year will again include major renovations with 
Buildings 12 and 12A. These renovations will in essence 
complete the long ranged plans developed several years 
ago and allow for adequate physical security for the 
central computer complex. As further guidance on the EEO 
program is received, we may have to initiate new 
monitoring systems to evaluate our performance in this 
area. There w ill undoubtedly be new adrainistrative 
managerial requi redients and modifications of old ones 
forthcoming from the NIH, the PHS, the Department, the 
GSA and the Civil Service Commision. 



24 




ProQeot Contvol Officer Mary Allen and her staff run 
a -pilot survey of the NIH users of DCRT services to 
determine the extent to which these services support 
major categories of NIH work. 



25 



lARnRATORY OF APPLIED STUDIES 



Eugene K. Harris 
Chief 




LAS began research on the develo-pment of adaptive 
finite element schemes which seek optimal mesh 
refinement strategies for the solution of 
reaction-diffusion systems. 



1. SUMMARY OF ACTIVITIES 



Fiinrt.inn*; 

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

• 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 processes, transport 
of substrate to tissues and models of receptor - ligand binding 
kinetics; 

• in collaboration with clinicians, to develop and apply 
mathematical or statistical theory and computer programming to 
improve diagnosis of disease and assessment of treatment; 

• 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 the NIH. 

These primary functions of the Laboratory are carried out 
through two sections: 

1. Applied Mathematics Section (John E. Fletcher, Ph.D., 
Chief). The staff of 6 in this Section includes, at the doctoral 
level, an applied mathematician (Chief), a computer scientist, a 
biomathematician, and a physician-scientist, as well as an M.S.- 
level mathematician, and (part-time) a graduate student in 
mathematics with M.S. degree. 

2. Medical Applications Section (James J. Bailey, M.D., 
Chief). The staff of 6 in this Section includes, at the doctoral 
level, two physician-scientists and an electronics engineer as 
well as two computer systems analysts and a (part-time) gradua.te 
student (M.S. degree) in electronic engineering. 

The Chief, LAS, is a biostatistician. The specific research 
areas of these scientists are described in the individual project 
reports. 



28 



Scope of Work 

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

• biochemists and pharmacologists - at U. of Iowa, Medical 
College of Virginia, and other universities in the U.S. and in 
France, working on models for receptors of drugs or other 
ligands, and the kinetics of enzymes in membranes; 

• physiologists and chemical engineers - in the U.S., U.K., 
Scandinavia, and Germany studying the transport of substrate 
within the microcirculation; 

t clinicians - at NIH in the cardiology, pulmonary branches of 
NHLBI, the arthritis and rheumatism branch of NIAMDD, the medical 
intensive care unit, CC, and the departments of diagnostic radiology 
and diagnostic imaging, CC; 

• clinical chemists and pathologists - at NIH (Clin. Path. Dept., 
CC), elsewhere in the U.S., in the U.K. and in Japan, engaged 
in the collection of and study of reference values in 
laboratory medicine; 

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

During FY 79 LAS staff members participated in various teaching 
and consulting (or advisory) activities: 

J. Fletcher continues to serve as Chairman of the Mathematics 
Department, FAES, and taught ordinary differential equations in 
the NIH graduate school; 

J. Bailey continues as a member of a site-visiting team for 
NHLBI concerned with computer analysis of exercise ECG's. He also 
serves as consultant on common standards for quantitative 
electrocardiography for a public health program sponsored by the 
European Economic Community; 

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 Expert Panel on the Theory of Reference Values 
of the International Federation of Clinical Chemistry; 

E. Hill is an adjunct associate professor of computer science 
at Howard University. 



29 



Highlights of Year's Activities 

The three projects described in this section provide a brief view 

of selected accomplishments in each of the two sections and the Office of 

the Chief. A more complete description of all LAS projects appears 

in the Individual Research Projects Report. 

t Efficient, flexible numerical methods for solving nonlinear 
partial differential equations implemented on NIH 370 
computer system: The mathematical models developed by J. 
Fletcher and B. Bunow of the Applied Mathematics Section to 
describe time- dependent reaction-diffusion and transport 
processes in biological systems consist of nonlinear ordinary and 
partial differential equations. Numerical methods required 
to solve these either do not exist at present or demand very 
large and costly amounts of computer time. During FY 79, Drs. 
Fletcher and Bunow with the aid of M. Bieterman of the AMS, 
have tested and installed on the NIH 370 computer system 
several highly efficient finite element programs for solving 
such equations in two dimensions. Research was begun on the 
development of adaptive finite element schemes which allow 
optimal strategies of choosing time and space meshes for the 
numerical solution of time-dependent reaction-diffusion 
systems. 

In addition, finite difference methods are being explored 
to solve models of capillary transport processes. If 
successful, these procedures will greatly reduce the costs of 
applying the complex mathematical models needed to describe 
time dependent processes found in living systems. 

• Detection of segmental renal artery disease using scintigraphic 
functional maps: Previous work by T. Stibolt and M. Douglas 
(Applied Mathematics Section and Medical Systems Section, LAS) 
and E. Jones (Diagnostic Imaging Branch, CC) has shown a 
significant advance in radionuclide renography through the use 
of computer- generated functional maps, which are now in routine 
clinical use at several research institutions in the United 
States and Europe. During FY 79 new algorithms for 
edge-detection and- masking were designed and incorporated into 
the functional mapping programs. A change in radionuclide 
(from 1-131 to 1-123) has resulted in a 15 or 20 to 1 
improvement in signal-to-noise ratio. With the new algorithms 
and the new radionuclide five dogs were studied following 
surgical ligation of a segmental renal artery. In some of the 
dogs the dynamic effect of the lesion was revealed in the 
functional maps, whereas contrast angiography failed to reveal 
the lesion in any of the dogs. The functional maps will be 
used in FY 79-80 to study large populations of patients with 
renal artery disease and hypertension before and after therapy 
to improve regional renal blood flow. This tool will be a 
sensitive means of following the effects of such therapy. 



30 



• statistical criteria developed for judging the probable 
importance of an observed change between serial blood chemistry 
tests: E. Harris (LAS), G. Shakarji (DMB), S. Brown (Clinical 
Research Centre, Medical Research Council, U.K.): Physicians 
vary considerably in the criteria they use to assess the 
significance of an observed difference between measurements of 
a biochemical constituent in the same individual on two 
successive occasions. These variations depend in large part on 
presumed laboratory imprecision and on the physician's 
experience with that type of patient. However, numerous recent 
studies both at NIH and elsewhere have demonstrated that ^ 
healthy persons vary substantially in both mean and standard 
deviation of results over fixed time periods (e.g., week- 
to-week, or year-to-year). Thus, a given difference 

between two successive lab tests may be medically significant 
in an individual who generally maintains close physiological 
control, but merely random variability in another person whose 
physiology does not closely regulate the constituent measured. 
Using recent data on the distribution of within-person standard 
deviations, methods have been developed for estimating the 
proportion of healthy persons in whom a specified change is 
likely (stated probability) to be at least statistically 
significant. Thus, for the first time, an objective, 
quantitative basis (involving biological as well as analytic 
variation) has been reached for assessing various proposed 
thresholds of acceptable changes over time in the blood 
chemistries of healthy persons. 

Staff Changes in FY 79 

• LAS is pleased to welcome Richard Shrager, mathematician, to 
the Laboratory's Applied Mathematics Section. In addition to 
his independent research, Mr. Shrager is an active consultant 
to and collaborator with many NIH scientists. 

• We are sorry to lose Dr. Edward Hill, computer scientist, who 
has transferred to the Department of Energy after 10 years with 
LAS, and Dr. Bruce Line, senior medical staff fellow who has 
rejoined the Diagnostic Imaging Branch, CC after 3 years with 
LAS. 



31 



Future PI ans 

As noted above, LAS scientists in FY 80 will continue their 
studies of efficient numerical methods for solving complex 
mathematical models. Further, several LAS projects will be heavily 
involved in applications to patient care or in the use of patient 
data. These include the study of functional mapping in renal 
hypertension; collection of data on pulmonary function and 
scintigraphic tests from normal volunteers; testing models of 
pulmonary gas exchange and computer simulations of electrolyte 
balance with data from patients in intensive care units; 
multivariate time series analysis of biochemical measurements in 
large groups of healthy subjects. Finally, we anticipate progress 
in the development of mathematical and physical models to support 
our newly-begun collaborative effort in electron-loss spectroscopy. 




During FY79, LAS staff members participated in various 
teaching and consulting activities. LAS Chief 
Dr. Eugene Harris continues to he a consultant in 
applied statistics to the Food and Drug Administration' s 
Division of Medical Devices and Diagnostic Products. 



32 



II. ANNOTATED PROJECT AND ACTIVITIES LIST 



Clinical research, patient care, epidemiology 

Simulatin n of physiological systems : E. Hill, J. Fletcher, E. 
Harris (LAS); B. McLees (CC): Exploration and testing of 
computer programs simulating responses of physiological systems 
to determine the usefulness of such programs as consultants or 
quality control mechanisms in patient care. 

r.nmputpr aided analysis of plprtrorardiograms : J. Bailey, M. 
Horton (LAS): Separate studies, conducted in collaboration 
with cardiologists and biomedical engineers in U.S. 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. 

Computer systems for diagnostic imaging : J. Bailey, M. 
Douglas, B. Line (LAS); H. Ostrow (CSL); M. Green and others 
(Diagnostic Imaging, CC): Development and application of 
computer systems to such diagnostic imaging activities as 
ECG-gated radionuclide angiocardiography, functional mapping 
and other scintigraphic studies of kidney, brain, heart and 
lung. 

Computer-based studies in ultrasonography: T. Stibolt, Jr., M. 
Douglas, J. Bailey (LAS); B. Maron, J. Gottdiener (NHLBI); S. 
Leighton (BEIB, DRS): Development of computer systems for 
image enhancement and 3-dimensional reconstruction from 
ultrasound data, particularly data derived from scanning of the 
heart and intra-abdominal organs. 

Statistical research in clinical pathology : E. Harris (LAS); 
G. Shakarji (DMB); clinical chemists in California, U.K., Japan 
and elsewhere: Application of variance component and time 
series analysis to description of reference distributions of 
clinical laboratory tests, to serial studies of normal 
biochemistries, and to the design of criteria for recommended 
precision and accuracy of analytic methods. 

Laboratory investigation 

Mathematical modeling of biological processes : J . Fl etc her 
(LAS); A. Spector (University of Iowa): Development and 
application of mathematical models in studies of substrate 
transport in the microcirculation, diffusion processes in 
physiology, and macromolecule-ligand binding equilibria. 

33 



Mechanisms of actlvp transport- biochemical kinetics : B. Bunow 
(LAS); J. Rinzel (NIAMDD); J. De Simone et al. (Medical College 
of Virginia): Experimental and mathematical studies of the 
energy mechanisms for active transport, and of multi-state 
biochemical kinetics in cells and membranes. 

Computer-based modeling of pulmonary gas exhange : Many LAS 
staff members; staff of Diagnostic Imaging (CC), and Pulmonary 
Laboratory (NHLBI). Utilizing scintigraphic and other clinical 
data on lung function to construct sound mathematical and 
computer-based models of ventilation and perfusion in the lung. 

Hybrid computing to analyze physiologi r signals and construct 
«;imiilation models : E. Pottala, A. Mitz (LAS); various I/D 
scientists: Using LAS minicomputer system (MAC-16) for 
hardware simulation of physiologic functions (e.g., retinal 
cell activity) and analysis of analog signals (myogram, 
ultrasonogram, etc.). 

Image processing in electron-loss spectroscopy : E. Pottala, M. 
Douglas (LAS); K. Gorlen (CSL, DCRT); J. Costa (NIMH); C. Fiori 
(BEIB, DDR) and others: 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. 

Computer research and technique development 

General mathematical and computational methods : E. Hill (LAS); 
R. Shrager (LAS, formerly of LSMM): Study of methods of fitting 
non-linear models utilizing other than least squares criteria. 
Evaluating methods of organizing large data files for rapid 
storage and retrieval. 

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 
(U. of Maryland): Study, development and implementation of 
efficient, flexible numerical methods for the solution of 
nonlinear ordinary and partial differential equations involved 
in modeling important physiological processes. 



34 



III. PUBLICATIONS AND PRESENTATIONS LIST 



Bailey, J.J. , Horton, M.R. : Reproducibility of version 2 of 
the IBM program with and without the serial comparison option. 
The Fifth International Congress of Electrocardiology , 1978 (in 
press). 

Bailey, J.J., Horton, M.R., Goldman, A.P. : Testing of updated 
programs for ECG analysis. Proceedings of the Second Annual 
Svmposiun on Computer Application in Medical Care . IEEE 
Catalog 78CH1413-4, 1978, pp. 606-609. 

Bailey, J.J., Horton, M.R., Goldman, A.P. : Performance of 
updated programs for ECG analysis: comparison of ECAN-E with 
ECAN-D. Proceedings of the Eighth Annual Conference of the 
Society for Computer Medicine . Minneapolis, Minnesota, 1978, 
pp. 2.3.9-11. 

Bailey, J.J. , Horton, M.R. : Type A electrocardiogram data 
bases: purpose and development. Prnrpp di ngs of Internationa l 
Federation in Information Processing TC4 "Optimization of 
Computer-ECG Processing", 1979 (in press). 

Brereton, H.D. , Line, B.R., Londer, H.N. , O'Donnel, J.F. ,-Kent, 
C.H. , and Johnson, R.E. : Gallium scans for staging small cell 
lung cancer. JAMA 240, 666-667, 1978. 

Bunow, B. Chemical Reactions and Membranes I: Linear Analysis, 
J. Thenr. Rinl . 75,51, 1978. 

Bunow, B. Chemical Reactions and Membranes II: Nonlinear 
Analysis, J. Theor. Biol . 75,79, 1978. 

Bunow, B. and Weiss, G.H. How Chaotic is Chaos? 1979. 
Mflthemflticfll Biosciences (in press). 

Bunow, M.R. and Bunow, B. Phase Behavior of Ganglioside- 
Lecithin Mixtures 1979. Biophysical Journal (in press). 

Bunow, B., Line, B.R. , Horton, M.R., Weiss, G.H.: Regional 
ventilatory clearance by xenon scintigraphy: A critical 
evaluation of two estimation procedures. J. Nucl. Med . 20 
703-710, 1979. 

Covacci, Renato: Techniques for the hardware simulation of the 
turtle cone photo responses and extensions of the validity of 
the model to the salamander cone. Proceed del Gruppo 
Nazionaledi Cibernetica e Biofisica . Piza (Italy), April 10, 
1979. 

35 



Crystal, R.G. , Fulmer, J.D. , Baum, B.J. , Bernardo, J. Bradley, 
K.H. , Bruel, S.D. , Elson, N.A. , Fells, G.A., Ferrans, V.J., 
Gadek, J.E., Hunninghake, G.W., Kawanami , 0., Kelman, J. A., 
Line, B.R. , McDonald, J. A., McLees, B.D., Roberts, W.C, 
Rosenberg, D.M., Tolstoshev, P., Von Gal, E., Weinberger, S.E. : 
Cells, collagen and idiopathic pulmonary fibrosis. Lung 155, 
199-224, 1978. 

Douglas, M.A. , Green, M.V. , Ostrow, H.G. : Evaluation of 
automatically generated left ventricular regions of interest in 
computerized ECG-gated radionuclide angiocardiography. 
rnrnpHtprt; in Cardiology . 201-204, September 1978. 

Fletcher, J.E. An Overview of Mathematical Modeling in the 
Microcirculation. Proceedings of 31st ACEMB meeting . Atlanta, 
GA., 21-25 October 1978, 384. 

Fletcher, J.E., On Facilitated Oxygen Diffusion in Muscle 
Tissues, The Biophysical Journal . (in press). 

Fletcher, J.E., Facilitated Diffusion in a Krogh Cylinder 
Model, Prnrppdings of 3rd Internationa l Oxygen Transport to 
Tiq';ijp Cnnfprenr.p (TSOTT) . La Jolla, Calif, July 1979. 

Green, M.V., Brady, W.R., Douglas, M.A. , Borer, J.S., Ostrow, 
H.G. , Line, B.R., Bacharach, S .L., Johnston, G.S.: Ejection 
fraction by count rate from gated images. Jour. Nucl. Med. 
19(8): 880-883, August 1978 

Green, M.V. , Ostrow, H.G. , Scott, R.N. , Douglas, M.A., Bailey, 
J.J., Johnston, G.S.: A comparision of simultaneous 
measurements of systolic function in the baboon by the 
electromagnetic flowmeter and high frame rate, ECG-gated blood 
pool scintigraphy. Circulation fin press), 1979. 

Harris, E.K. and Shakarji, G. : Use of the population 
distribution to improve estimation of individual means in 
epidemiological studies. J. Chronic Diseases. 32: 233-243, 
1979. 

Harris, E.K. : Statistical principles underlying analytic goal- 
setting in clinical chemistry. Amer. J. of Clinical Pathology . 
August 1979. 

Harris, E.K. and Brown, S.S. : Temporal changes in the 
concentrations of serum constitutents in healthy men: 
distributions of within-person variances and their relevance 
to the interpretation of differences between successive 
measurements. Annals of Clinical Biochemistry (in press). 



36 



Harris, E.K. : Review of statistical methods of analysis of 
series of biochemical test results. Annales de Biologie 
36: 194-197, 1978. 

Harris, E.K. : Appropriate use of population and individual 
reference data in patient care. Proceedings of thp <^prnnH 
Annual Symposium on Computer Application in Medical Care (IEEE 
Computer Society), November 5-9, 1978, pp. 408-409. 

Hill, E. : Discrete rational approximation in the L-one and 
L-infinity Norms using the Simplex Method, Proceedings of 
ACM 1978 . December 1978. 

Kernevez, J.P, Joly, G. , Duban, M. C, Bunow, B. , and Thomas, 
D. : Hysteresis, Oscillations, and Pattern Formation in 
Realistic Immobilized Enzyme Systems. .1. Math. Rinl . 7,41, 
1978. 

Kernevez, J.P. , Blanchard, G. , Thomas, D. and Bunow, B. 
Pattern Formation and Wave Propagation in the S-A System, 
1978 Springer Lecture Notes in Mathematics, (in press). 

Line, B.R. , Fulmer. , J.D. , Reynolds, H.Y. , Roberts, W.C, 
Jones, A.E., Harris, E.K. , and Crystal, R.G. : Ganium-67 
citrate scanning in the staging of idiopathic pulmonary 
fibrosis: Correlation with physiology, morphology and 
bronchoal veolar lavage. Amer. Rev. Resp. Pis ., 1978, 118 
355-365. 

Shrager, R. and Hill, E. : Some properties of the Levenberg 
methods in the L-one and L-infinity norms. Mathematics of 
Cnmpiitatinn (in press). 

Vallerga, S. Covacci, R. , and Pottala, E. : Hardware model of 
turtle cone photo responses. Proceed del Gruppo Nazionaledi 
Cibernetica e Biofisica , Piza (Italy), April 10, 1979. 



37 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00035-03 LAS 



PERIOD COVERED 

October 1,1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Mathematical Models and Simulation Programs in Physiology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



E. Hill 



OTHER: J. Fletcher 
E. Harris 
B. MacLees 



Computer Systems Analyst LAS DCRT' 

Mathematician, Chief LAS DCRT 

Statistician, Chief LAS DCRT 

Chief MICU CC 



COOPERATING UNITS (if any) 

none 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT. NIH. Bethesda. MP 20205 



TOTAL MANYEARS: 

.A. 



PROFESSIONAL; 

.4 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



S (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves the identification and classification of 
various simulation programs used in physiology . These programs are being 
examined for utility in a clinical environment, accuracy of representation 
of normal and abnormal physiologic conditions, and the identification of 
areas of defective or missing physiologic relationships. The programs a^re 
being tested and evaluated. The outcome of the evaluation will be used to 
develop a more representive model to reflect the current state of 
knowledge in physiology. 



PHS-6040 
(Rev. 10-76) 



38 



Background and Objectives: 

The objective of this project is to investigate the use of 
simulation programs in physiology as diagnostic and patient 
management aids to physicians and other clinical staff. Several 
biomedical researchers are using the programs to check their 
usefulness in physiology as diagnostic and patient management aids to 
physicians and clinical staff. 

Significance to Biomedical Research: 

Simulations offer the clinician the oppportunity to try a 
proposed or alternative course of medical treatment on an ideal 
(computer) patient, without the attendant risk of injury to the 
actual patient. 

Progress in FY 79: 

The computer simulation of electrolyte and acid-base disorders 
is the first simulation program to be tried in a clinical setting 
at the NIH. This model was chosen because elecrolyte and acid-base 
equilibrium data can be measured. The simulation program has been 
described and demonstrated during the NIH Computers in Clinical 
Medicine Seminar. It was found to be useful as a teaching aid. A 
copy of this program has been implemented for B. MacLees (MICU, CC) 
and other researchers for use in a clinical environment, and as a 
training aid. At the present time, the project is collecting 
comments from the user-researchers about the usefulness of the 
program. 

Proposed Course: 

Feasibility studies are being conducted to determine the 
utility of this program using real patient input from clinical 
settings and using the results as a guide to patient management. 
Pending the outcome of these studies, new simulators will be 
developed to approximate the interdependent pathways in 
physiological structures. This stage of development will require a 
continuing effort, requiring the cooperation of the interested 
clinical laboratories. The refinement of the program will involve 
the testing of the acid-base balance program with existing data on 
electrolyte, blood gas, and pH measurements from patients before 
and after specific therapeutic intervention. Feedback from the 
researchers will be used to develop a more meanful model to 
approximate physiologic structures. 



39 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00005-09 LAS 



PERIOD COVERED 

October 1, 1978 to September 30. 1979 



TITLE OF PROJECT (80 characters or less) 

Mathematical Models of Binding Equilibria 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



J.E. Fletcher 



OTHERS: E. Hill 

R. Shrager 

A. Spector 

L. Hoffman 



Chief, Applied Mathematics Section 

Computer Scientist, AMS 
Mathematician, AMS 

(formerly of LSM, DCRT) 
Professor, Univ. of Iowa 

Medical School 
Professor, Dept. of Microbiology 

Univ. of Iowa 

Medical School 



LAS DCRT 

LAS DCRT 
LAS DCRT 



COOPERATING UNITS (if any) 

none 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MP 20014 



TOTAL MANYEARS: 

.5 



PROFESSIONAL: 

.5 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

n (al ) MINORS n {a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project is concerned with mathematical models of 1 igand- receptor 
or 1 igand-macromolecul e binding studies at equilibrium. Models are 
analyzed for mathematical as well as for biological validity and are 
studied to determine their suitability for fitting to experimentally 
determined laboratory data. The appropriateness of various fitting 
criteria are studied, and general guidelines and computational algorithms 
are designed for interactive model fitting. 



PHS-6040 
(Rev. 10-76) 



40 



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 
1 igands to macromolecules, particularly small ions to proteins. It 
has produced an interactive methodology for the fitting of models 
to data and has developed other computer oriented tools for the 
analysis of data from laboratory experiments. 

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. Such procedures broaden 
biomedical knowledge and add 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 validity 
for the interpretation of current laboratory and clinical experiments 
is therefore essential to the progress of science. 

Progress in FY 79: 

The previous ten years of research in this area is being 
collected in the form of a technical report. This report details 
the various alternative models, graphical presentations of data, 
and algorithms for fitting models to data. New fitting algorithms 
are now available for fitting with other than the least squares 
criteria. Publication is expected in FY 80. 

Proposed Course: 

Applications of existing and new methodology to data analysis 
will continue to be made as they are requested by collaborating 
laboratories. Some new concepts in membrane- receptor studies are 
being considered and experimental studies such as the analysis of 
cholesterol exchange in lipid bilayers are under study. The dominant 
direction of this project will continue to be in the examination of 
the binding and exchange of lipids, fatty acids, and cholesterol in 
laboratory experiments. The level of effort will depend on the success 
of new experimental designs. 



41 



SMITHSONIAN SCI 
PROJECT NUMBER 



NCE INFORMATION EXCHANGE 
Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00002-09 LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978. 



TITLE OF PROJECT (80 characters or less) 

Computer Aided Analysis of Electrocardiograms 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



OTHERS: 



J.J. Bailey 
M.R. Norton 
J. Gottdiener 



Chief, MAS LAS DCRT 

Computer Systems Analyst LAS DCRT 
Chief, ECG Laboratory CB NHLBI 



S. Allen Medical Research Analyst CSL DCRT 

M.E. Womble, Aeromedical Cybernetics Branch, Brooks AFB, Texas 

P. MacFarlane, Medical Cardiology, Glasgow Royal Infirmary, Scotland 



COOPERATING UNITS (if any) 

none 



lab/branch 
Laboratory of Applied Studies 



section 

Medical Applications Section 



institute and location 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARE: 



1.3 



PROFESSIONAL: 

1.2 



OTHER: 



0.1 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

D (al) MINORS D (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



D (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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. 



PHS-6040 
(Rev. 10-76) 



42 



Background and Objectives: 

In the past fifteen years numerous computer programs implemented 
upon a variety of computer systems have been developed for analysis of 
routine ECG's. Computer processing of ECG's has become a sizeable 
enterprise in many parts of the country, including both commercial 
bureaus offering service for a fee and no n- commercial (academic or 
government) centers establishing the capability for themselves. 

Since 1970, LAS in collaboration with the Cardiology Branch (CB), 
NHLBI has studied several programs to determine which, if any, would be 
useful to implement on the NIH campus. As a result of this work, an 
evaluation methodology was evolved which was published in 1974 and has 
become one of the standard references in the field. Although a program 
was selected and implemented in 1974 for daily use at NIH, additional 
programs and other computer systems continue to be evaluated as possible 
improvements to the NIH system with regard to cost and accuracy. Past 
evaluations have included the ECAN-D program (1964), the Mayo-IBM program 
(1968), the experimental IBM program (1971), and the AVA 3.4 (Pipberger) 
program (1975). Guidelines for evaluation of ECG programs which were 
developed by LAS have been adopted by the American College of Cardiology. 

Progress during FY 79: 

A series of 300 ECG's were collected on patients in the Glasgow 
Royal Infirmary, (GRI), Scotland. Clinical documentation of the 
patient's cardiovascular status by non ECG means in the form of enzyme 
studies, cardiac catheter laboratory data, etc. was obtained whenever 
such investigations were warranted, i.e. in most cases. The standard 12 
lead ECG data was analyzed by the IBM program at NIH; modified McFee 
lead (XYZ) ECG data was analyzed by a program developed at GRI. The 
results of these two programs were compared with respect to 
accuracy where clinical documentation exists and with respect to 
cardiologist agreement in all other cases. A manuscript has been 
submitted to Circulation. 

A method, published in 1974, for using interleaved digital samplings 
from the same analog tracing to test reproducibility of programs has been 
extended and applied to the ECAN-E program (1975) and Version 2 of the 
IBM program (1976) (refs 1-3). 

Members of LAS and CSL developed specifications for a minicomputer 
system to be dedicated to ECG analysis which the Clinical Center plans 
to procure. A Request-for-Proposals has been sent out. 

Significance: 

The estimated number of computer-processed ECG's in North 
America in 1971 was 600,000 and in 1978 it was 6.3 million. In 



43 



view of this increase in computer usage, it will become even more 
important to have methodologies and guidelines by which ECG 
computer systems can be evaluated. 

These studies seek to establish the diagnostic limits of ECG itself 
and the degree to which computerized algorithms may achieve these 
limits. Important evaluation methodology continues to be developed, 
which may have a significant impact on the further diffusion of computer 
technology in electrocardiography. 

Proposed Course: 

The testing of ECG programs developed by other organizations has 
largely been completed in FY 79; however, some additional evaluations 
will need to be made in selecting a minicomputer system for the 
Clinical Center. 

Most ECG programs today use a feature extraction scheme which 
supposedly imitates the human reader; however, these may not be the 
optimal features for computer programs. Therefore, further work in 
this area will involve collaboration with the Aeromedical Branch at 
Brooks Air Force Base in investigation of various feature extraction 
schanes, including the use of Karhunen-Loeve (KL) expansions. 

KL functions are orthogonal functions with shapes similar to 
various ECG patterns and are derived from a large collection of ECG's. 
KL expansions have been shown to be useful for noise filtering and 
data compression as well as having potential for a computer-based 
scheme for diagnostic classification. 

Publications and Abstracts: 

Bailey, J.J., Horton, M.R. : Reproducibility of version 2 of the 

IBM program with and without the serial comparison option. The Fifth 

International Congress of El ectrocardioloqy , 1978 (in press). 

Bailey, J.J. , Horton, M.R., Goldman, A.P. : Testing of updated programs 
for ECG analysis. Proceedings of the Second Annual Symposium on 
Computer Application in Medical Care . IEEE Catalog 78CH1413-4, 
1978, pp. 606-609. 

Bailey, J.J. , Horton, M.R., Goldman, A.P. : Performance of updated 
programs' for ECG analysis: comparison of ECAN-E with ECAN-D. Proceedings 
of the Eighth Annual Conference of the Society for Computer Medicine. 
Minneapolis, Minnesota, 1978, pp. 2.3.9-11. 

Bailey, J.J. , Horton, M.R.: Type A electrocardiogram data bases: 
purpose and development. Proceedings of International Fpdpratinn 
in Information Processing TC4 "Optimization of Computer-ECG Processing " , 
1979 (in press). 



44 



SMITHSONIAN SCI 
PROJECT NUMBER 



^NCE INFORMATION EXCHANGE 
;Do not use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00003-08 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Computer Systems for Nuclear Medicine 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI 



J.J. Ba i 1 ey 

M.V. Green 

M.A. Douglas 

T.B. Stibolt, Jr. 



OTHERS: 



B.R. 


Line 


H.G. 


Ostrow 


S.L. 


Bacharach 


A.E. 


Jones 


G.S. 


Johnston 



Chief Med. Appl. Sec 
Ch , Appl . Physics Sec 
Comp. Syst Analyst 
Senior Staff Fellow 

Medical Research Analyst 

Engineer 

Physicist 

Chief, Diagnostic Imaging 

Chief 



LAS 


DCRT 


MM 


CC 


LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


CSL 


DCRT 


NM 


CC 


NM 


CC 


NM 


CC 



COOPERATING UNITS (if any) 

Nuclear Medicine Department, CC, NIH 
Computer Systems Laboratory, DCRT, NIH 



lab/branch 
Laboratory of Appl ied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

3.1 



PROFESSIONAL: 

3.0 



OTHER: 

0.1 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

D (al) MINORS D (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



)Q (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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. 



PHS-6040 
(Rev. 10-76) 



45 



Background and Objectives: 

Since FY 72 LAS and CSL in collaboration with the Nuclear Medicine 
Department, CC have been involved in the acquisition and development 
of several minicomputer systems which gather and process data from the 
scintillation cameras in the Nuclear Medicine Department. 

The objective of this program is continuing 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 FY 79: 

The minicomputer systemsin the Nuclear Medicine Department 
have been upgraded to new, faster CPU's and from magnetic tape to 
disk drives. These hardware upgrades, together with software 
refinements have resulted in greater ease of use, especially by 
the Nuclear Medicine technicians, and improved turnaround, making 
the results more readily available to clinicians. 

There has been continued development of IMAGE, a DEC-10 
based package of image processing routines begun in FY 76; it has 
been enlarged to allow a wider range of options for edge- detector 
methods, smoothing methods, display techniques and new methods 
for background minimization. 

An outgrowth of IMAGE is PSTACK, also a DEC-10 based system which 
allows the application of various image processing algorithms to a 
stack of pictures with a single command. It has been designed to 
handle dynamic scintigraphic images so that spatial data can be 
augmented with temporal data - i.e. processing of a moving picture. 
It also allows dual processing of two picture stacks so that the 
results of two different algorithms can be compared side-by-side, 
visually and quantitatively. Various processing options include: 
background subtraction by value, by percent of maximum, or by variable 
mesh; automatic masking and time function generation; image rotation 
and clipping; linear/nonlinear modification of gray scale; temporal 
and/or spatial smoothing by weighted average, by statistical bounding, 
by median replacement or by iterative replacement; and "sharpening" by 
a selection of convolution operators. 

LAS has acquired a DeAnza VC 2000 image display terminal allowing 
interactive design of image processing applications and generation of 
movies. 



46 



While the main thrust of IMAGE, PSTACK, and the DeAnza terminal 
has been towards processing dynamic scintigraphic images, it is 
anticipated that these methods will also find ready application in 
ultrasound and electron- microscopy images (see reports on diagnostic 
ultrasound and electron-microscopy). 

Automatic edge detection methods, designed and tested by means of 
IMAGE and PSTACK have been successfully applied to radionuclide 
angiograms (ref 1), renal scintigrams, and pulmonary scintigrams. 

During FY 79 for the first time dual radionuclide studies of the 
heart, i.e. angiography (blood pool) and myocardial perfusion with 
thallium - 201, have been accomplished. This allows areas of wall 
motion abnormalities to be correlated with areas of decreased perfusion 
in a semiquantitative way. 

Previous work in FY 76 showed a significant enhancement of 
radionuclide renography by the use of functional maps and since 
FY 77 functional maps have come into routine clinical use in the 
Diagnostic Imaging Branch. A series of 5 dogs was studied before 
and after surgical ligation of a segmental renal artery. Using 
the functional mapping techniques, it was possible in 3 dogs to 
demonstrate the presence of a defect. Defects were not detected 
in any dog with contrast angiography, the current definitive 
technique. This work was presented at the Society of Nuclear 
Medicine Meeting in June. 

The location and degree of severity of inflammatory processes in the 
lung has been studied in several types of disease, using a method for 
quantifying the uptake of radio gallium in the lung (refs. 3-4). 

Significance: 

Scintigraphy is a noninvasive tool which shows considerable detail 
concerning the dynamic function of an organ on a regional basis. 
Computer processing not only enhance.s scintillation images but allows 
quantification of the dynamic function. Real time implementation of 
algorithms on the minicomputers allows the clinical investigator to 
perform repeated studies on patients with exercise or therapeutic 
manipulation, thus better elucidating the nature of the patient's 
pathophysiology. 

Proposed Course: 

Use of automatic edge detection in both end-diastol ic and 
end-systolic angiographic images is being evaluated in collaboration 
with investigators at the University of Ulm, West Germany. Further 
studies of the combined use of myocardial (thallium) images and blood 
pool images to refine detection of regional wall motion and perfusion 



47 



and to determine "absolute" volumes, if possible, are planned. 
Methods for enhancing dynamic myocardial images will be extensively 
investigated, using IMAGE, PSTACK, and the DeAnza terminal. 

Renal scintigraphy with functional mapping will be applied to a 
series of hypertensive patients suspected of having segmental renal 
arterial disease. These patients are to be treated with a new drug that 
has demonstrated the potential of producing local renal arterial 
dilatation. Additionally, some of these patients may undergo renal 
artery dilitation using a newly developed catheter technique. Functional 
mapping will be used in evaluation and follow-up of both patient groups. 

Further work this year will occur in animal models. 
Particular emphasis will be given to renal transplant models and 
immune complex renal disease models. 

In FY 80 the reproducibility of pulmonary scintigraphy and 
its correlation to pulmonary functions will be investigated in a 
series of normal patients (see section on Computer-based Studies 
in Pulmonary Pathophysiology and Respiratory Disease). 

Publications and Abstracts: 

Douglas, M.A. , Green, M.V., Ostrow, H.G. : Evaluation of 
automatically generated left ventricular regions of interest in 
computerized ECG-gated radionuclide angiocardiography. Computers 
in Cardiology. 201-204, September 1978. 

Green, M. V. , Brady, W.R., Douglas, M. A. , Borer, J.S., Ostrow, 
H.G. , Line, B.R. , Bacharach, S .L., Johnston, G.S.: Ejection 
fraction by count rate from gated images. Jour. Nucl. Med. 
19(8):880-883, August 1978. 

Brereton, H.D. , Line, B.R., Londer, H.N. , O'Donnell, J.F. , Kent, 
C.H. , Johnson, R.E. : Gallium scans for staging small cell lung 
cancer. JAMA . 240:666-667, 1978. 

Line, B.R., Fulmer, J.D. , Reynolds, H.Y., Roberts, W.C., Jones, 
A.E., Harris, E.K. Crystal, R.G. : Gallium-67 citrate scanning in 
the staging of idiopathic pulmonary fibrosis: correlationg with 
physiology, morphology and bronchoal veolar lavage. Amer. Rev. 
Resp. Pis. , 118:355-365, 1978. 

Green, M.V. , Ostrow, H.G. , Scott, R.N. , Douglas, M.A., Bailey, 
J.J., Johnston, G.S.: A comparision of simultaneous measurements 
of systolic function in the baboon by the electromagnetic 
flowmeter and high frame rate, ECG-gated blood pool 
scintigraphy. Circul ation (in press), 1979. 



48 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo MOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00043-01 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Computer-based studies in ultrasonography 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES Of PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: 



OTHERS: 



T.B. 


Stibolt, 


Jr. 


LAS 


DCRT 


B.J. 


Maron 




CB 


NHLBI 


J.S. 


Gottdiener 


CB 


NHLBI 


T.H. 


Shawker 




DRD 


CC 


M.A. 


Douglas 




LAS 


DCRT 


J.J. 


Ba i 1 ey 




LAS 


DCRT 


S.B. 


Leighton 




BEIB 


DRR 



COOPERATING UNITS (If any) 

NHLBI 



Radiology 
CB 



CC 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARSi 

1.4 



PROFESSIONAL: 
1.2 



OTHERj 

0.2 



CHECK APPROPRIATE BOX(ES) 
B (a) HUMAN SUBJECTS 

O (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves collaboration of LAS, with the Cardiology 
Branch, NHLBI, the Diagnostic Radiology Department, CC, and Biomedical 
Engineering and Instrunentation Branch, DRR. It is directed toward 
computer-based processing for image enhancement , pattern recognition , 



and 3 dimensional reconstruction from ultrasound data. The principal 
sources of data are wide-angle, phased array echo-cardiography and 
B-mode, gray scale, abdominal ultrasonography. 



PHS-6040 
(Rev. 10-76) 



49 



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, certain 
limitations relating to presence of bone which is completely opaque to 
sound waves and to processing practices make this technique fall short 
of its full potential. Using the computer, it is possible to 
overcome or circumvent many of these limitations. 

Progress during FY 79: 

The cardiology Branch has acquired a Varian Sector Scanner which 
uses a phased array transducer head. This can now be fixed to a 
patient's chest wall using an NIH designed support device. This 
device contains an integral indicator for transducer angle which 
allows acquisition of serial images over 50 degrees in 2 degree 
intervals. Programs to correct the images for beam angulation have 
been written and an abstract describing the technique has been 
submitted for presentation at the American Heart Association 
Scientific Meeting in 1979. 

The Diagnostic Radiology Department, CC with advice and 
consultation from LAS has acquired equipment which gathers digital 
records of B-mode ultrasound images of intra-abdominal organs. The 
Evans and Sutherland display system at DCRT can be used to transform 
echo intensity to gray scale or color values in the image in any 
selected linear or non linear way. In this way texture 
information in areas which are not well seen in ordinary ultra- 
sonograms, such as the pancreas, can be enhanced. 

IMAGE a DEC-10 based package of picture processing routines and 
a recent outgrowth, PSTACK have been developed and updated to allow 
interactive application of picture processing algorithms from the De 
Anza terminal recently acquired by LAS (for further details, see 
Computer-based studies in Nuclear Medicine). 

Significance: ^' 

Patients with hypertrophic cardiomyopathy are at increased risk of 
sudden death. Unfortunately, many of these persons are not diagnosed 
ante-mortun because they are asymptomatic. A reliable technique to 
screen those persons with a family history of hypertrophic 
cardiomyopathy would be of great use since prophylactic drug therapy 
is probably feasible. Unfortunately, many different patterns of 
hypertrophy appear to exist in the population afflicted. Techniques to 
assess the distribution of hypertrophy are needed in those 
patients with uncommon distributions, especially those missed by 



50 



traditional M-mode echocardiographic techniques. Additionally, it is 
likely that the prognosis may differ among the various patterns of 
hypertrophy. The determination of regional wall motion abnormalities 
and other parameters of left ventricular function in patients with 
coronary artery disease with this technique could be a wery important 
adjunct or even supercede such determinations now being accomplished by 
radionuclide angiography. 

B-mode abdominal echography is a widely employed technique in 
evaluation of many disease states involving the abdomen. In some 
instances, useful information is lost during the assignment of grey 
scale value from echo intensity. A tool to allow changing this 
assignment retrospectively in an interactive mode would allow testing of 
this hypothesis. 

Proposed Course: 

Groups of normal volunteers and patients with hypertrophic 
cardiomyopathy or coronary artery disease will be studied using this new 
method of echocardiography. In addition computer-based reconstruction 
of left ventricular anatomy in 3 dimensions from serial 2-dimensional , 
wide angle phased array sector scans will be undertaken. 

It is anticipated that the De Anza terminal will allow IMAGE 
and PSTACK to be applied also to the B-scans of the abdomen. 



51 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER [Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00007-11 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Statistical Research in Clinical Pathology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: E.K. Harris 

OTHERS: G. Shakarji 

G.Z. Will iams 

S. S. Brown 



E. A. Robertson 
T. Yasaka 



Chief, Lab. of Applied Studies LAS DCRT 

Supv. Systems Analyst DMB DCRT 

Director, Institute for Health 

Research, San Francisco, CA 

Clinical Chemistry Service 

Clinical Research Centre 

Harrow, England 

Clinical Pathology CC 

Perfect Liberty Medical Service Department 

Osaka, Japan 



COOPERATING UNITS (if any) 

None 



lab/branch 

Laboratory of Applied Studies 



SECTION 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARSi 

0.6 



PROFESSIONAL: 



0.6 



OTHER: 



CHECK APPROPRIATE B0X{ES) 
S (a) HUMAN SUBJECTS 

D (al) MINORS D (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

In cooperation with Dr. G. Z. Williams and staff of the Institute for 
Health Research (San Francisco), records of some 30 different biochemical 
and hematological tests performed annually over a 4-7 year period on 
several hundred healthy volunteers have been analyzed to test the behavior 
and usefulness of statistical forecasting models described in previous 
reports and publ ications. An initial draft of a manuscript describing the 
results of this work has been completed and is now undergoing review. An 
application of the theory of Empirical Bayesian estimation to improve 
accuracy of individual mean values in epidemiological studies has been 
published. A cooperative study with Dr. S.S. Brown (Clinical Research 
Centre, Harrow, U.K.) of the distribution of within-person variances of 
biochemical constituents in healthy volunteers and its appl ication to the 
interpretation of differences between successive measurements has been 
completed and the report is now awaiting publication. A new study with 
cooperating scientists at the NIH and in Japan has begun on the application 
of multivariate time series theory to short series of biochemical tests. 



PH 8-6040 
(Rev. 10-76) 



52 



Objectives: 

To investigate applications of statistical theory, particularly 
the use of variance components, measures of within-person variation, 
and the theory of discrete time series, to the interpretation of 
clinical laboratory measurements and the evaluation of analytic 
methods. 

Background and Progress during FY 79: 

The database gathered through the health monitoring program of 
the Health Research Institute (HRI), San Francisco, represents one 
of the largest extant collections of serial biochemistries on normal 
volunteers. Hence it offers an unusual opportunity to study the 
suitability of various recently published statistical models and 
analytic methods aimed at detecting step changes and trends in short 
series of biochemical data from healthy persons. During the past 
year, statistical analysis has been completed on over 30 biochemical 
and hematological analytes in several hundred individuals who have 
undergone at least 4 annual examinations, with several weekly 
retests around each time of examination. Results show that a 
nonstationary, random walk model of within-person variation is 
particulary well-suited to the detection of trends in individual 
series. The practical effects of using forecasting ranges based on 
both stationary and nonstationary models are being explored in this 
database. 

Utilizing a smaller database of weekly measurements on 37 
healthy male volunteers over a 5-month period, collected in a 
cooperative study with the Clinical Research Centre, Harrow, 
England, the second of a new series of reports on normal variation 
has been accepted for publication and should appear before the end 
of 1979. This analysis focuses on the heterogeneity of intra- 
personal variation in each of 10 common analytes, and the effects of 
such heterogeneity on the applicability of commonly accepted 
criteria concerning the amount of change between successive 
observations which should trigger investigative medical activity. 

During the summer of FY 79, with the aid of a graduate student 
in biostati sties (Ms. Emily Terwey of the University of Texas Health 
Sciences Center) work has begun on the application of multivariate 
random walk theory to the sequential forecasting of biochemical test 
results in short series from healthy subjects. Data has been 
provided from past records of an extensive ongoing health monitoring 
program in Japan through the kind cooperation of the Director of the 
data processing center for this program in- Osaka. 



53 



Significance to Biomedical Research: 

The definition and estimation of analytic and biological 
variance components provides an essential basis for the objective 
interpretation of clinical laboratory tests in patients and healthy 
persons alike. The development, testing and routine use of 
univariate stochastic models to describe and forecast sequential 
results of laboratory tests in individual cases has 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 should 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). These 
methods of data analysis require the use of standard computer 
program packages as well as construction and implementation of 
special algorithms for computer-based laboratory reporting systems. 
A spin-off from these research efforts has been the development of 
versatile computer systems for storing, updating and retrieving 
serial information on multiple laboratory results for individual 
patients. These systems are currently being employed in the 
Hypertension-Endocrine Branch (HNLBI) and the Arthritis and 
Rheumatism Branch (NIAMDD). These developmental and associated 
consulting activities have expanded the services which DCRT offers 
the NIH clinical community. 

Future Course: 

A final report on application of univariate stochastic models 
to serial data from the monitoring program of the Institute for 
Health Research should be completed and submitted for publication 
early in FY 80. Current cooperative investigations of multivariate 
models will continue during FY 80 with at least a preliminary report 
of results expected to be ready by the spring of 1980. A study of 
the probabilities of false alarm associated with the sequential use 
of these forecasting models is also expected to begin during FY 80. 
The P.I. (E.H. ) has accepted an invitation to prepare an extensive 
review of the general theory of reference values in healthy persons 
and this should be completed during FY 80. As more and more 
clinical pathologists, chemists and practicing physicians become 
interested in the application of time series models to monitoring 
individual patients, a certain amount of time must be devoted to 
suitable "packaging" of the relevant computer algorithms and 
communication of the programs and instructions on their use to 
interested users. Through such means, new cooperative studies may 
arise which cannot now be foreseen. In addition, it is anticipated 
that use of the storage, updating and retrieval programs in NIH 
clinical laboratories will expand and lead to collaborative research 
studies, particularly in the use of serial patient data to predict 
clinical outcome. 



54 



Publications and Abstracts: 

Harris, E.K. and Shakarji, G. : Use of the population distribution 
to improve estimation of individual means in epidemiological 
studies. J. Chronic Diseases . 32: 233-243, 1979. 

Harris, E.K. : Statistical principles underlying analytic goal- 
setting in clinical chemistry. Amer. J. of Clinical Pathology , 
August 1979. 

Harris, E.K. and Brown, S.S.: Temporal changes in the concentrations 
of seriin constitutents in healthy men: distributions of within- 
person variances and their relevance to the interpretation 
of differences between successive measurements. Annals of CI inical 
Biochemistry (in press). 

Harris, E.K. : Review of statistical methods of analysis of series 
of biochemical test results. Annales de Biologie Clinique , 36: 
194-197, 1978. 

Harris, E.K. : Appropriate use of population and individual 
reference data in patient care. Proceedings of t h e Second Annu al 
Symposium on Computer Application in Medical Care (IEEE Computer 
Society) , NoVember 5-9, 1978, pp. 408-409. 



55 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00044-01 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Mathematical Modeling of Substrate Transport in Physiological Environments 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: J.E. Fletcher Chief, Appl ied Mathematics LAS DCRT 

Section 



OTHER: M. Bieterman Mathematician, AMS 
R. Shubert Associate Professor 
Louisiana Tech. Univ. 



LAS DCRT 



COOPERATING UNITS (if any) 

none 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

nCRT. NTH. Bethesda, 



MP 20205 



TOTAL MANYEARS: 



PROFESSIONAL: 

.5 



OTHER: 



CHECK APPROPRIATE BOX(ES) 

D (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Mathematical models of microcir cul ato ry structure and function are 
developed from conceptual models into syste ms of coupled ordinary and/or 
partial differential equations. Metho'^s~'of sol ution 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 microcircul atory physiology and are published 
in the scientific literature"^ 



PHS-6040 
(Rev. 10-76) 



56 



Objectives: 

The objectives of this project are to develop mathematical 
models which can be used to simulate microcircul atory physiology 
and to explain, interpret and/or predict physiologic behavior and 
limits. Such models may lead to a better understanding of basic 
biological processes and suggest new experimental approaches to 
microcircul atory phemomena of biomedical and biochemical 
importance. 

Background: 

The mathematical modeling of substrate supply to tissue from 
the microcirculation has been under study since FY 69. 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 microcircul atory 
geometry, nonlinear oxygen hemoglobin dissociation properties, 
intracellular binding proteins, and substrate dependent metabolic 
rates requires such a detailed description to achieve physiologic 
validity. These models require the numerical solution of a system 
of coupled distributed parameter models which are of a nonlinear 
type. 

Significance to Biomedical Research: 

Such modeling is necessary to examine the state of local tissue 
microcircul atory dynamics since direct measurements are generally not 
possible and microcircul atory function 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. 

Progress in FY 79: 

The contribution of binding proteins to substrate transport in 
tissue was initially considered in FY 78. A new model was developed 
which included the previously neglected effects of flow, capillary 
length and radius, and oxygen dissociation kinetics in capillary 
blood. The resulting mathematical model required nonclassical 
solution techniques. A parametric study af the steady state system 
has been completed and the results have been submitted for 
publication. A new methodology for the extended time dependent 
system of equations is currently being explored. Preliminary results 
with steady-state systems suggest that diffusion facilitation is 



57 



important at low blood flow rates and/or at high metabolic rate 
conditions. The principal investigator is currently participating in 
international conferences which are to examine the various aspects of 
microcirculatory modeling. 

Proposed Course: 

Mathematical techniques for the solution of time dependent 
model equations will be explored for accuracy and cost-effectiveness 
The numerical techniques currently under development will be 
applied to the models describing the complex interactions of 
facilitation, metabolism, and blood-oxygen extractions. Efforts 
will be made to consider rhythmic or periodic conditions which 
relate more closely to the cardiac and skeletal muscle functions. 

Publications and Abstracts: 

Fletcher, J.E. An Overview of Mathematical Modeling in the 
Microcirculation. Proceedings of 31st ACEMB meeting , Atlanta, GA., 
21-25 October 1978, 384. 

Fletcher, J.E., Facilitated Diffusion in a Krogh Cylinder Model, 
Proceedings of 3rd International Oxygen Transport to Tissue 
Conference (ISOTT), La Jolla. Calif, July 1979. 



58 



SMITHSONIAN SCI 
PROJECT NUMBER 



NCE INFORMATION EXCHANGE 
Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00033-03 LAS 



PERIOD COVERED 

October 1, 



1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Analysis of Coupled Transport and Biochemical Kinetics 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



OTHERS: 



B.Bunow 



Senior Staff Fellow LAS 

Applied Mathematics Section 



DCRT 



J. 


Kernevez 


J. 


DeSimone 


D. 


Mi kul ecky 


G. 


Weiss 


A. 


Kaplan 


M. 


Bunow 


J. 


Weinstein 


M. 


Bieterman 



Prof. Univ. of Tech. Compiegne France 
Prof. Medical College of Virginia 
Prof. Medical College of Virginia 
Chief PSL DCRT 

LCM NC I 
Guest Worker LCP NIAMDD 

LTB NC I 
Mathematician, AMS LAS DCRT 



COOPERATING UNITS (if any) 

none 



lab/branch 

Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT, NIH Bethesda, MD. 20205 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 

1.0 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



^ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project investigates two fundamental problems in biology: (1) 
the role of dynamic patterns in embryology and evolution , and (2) 
kinetics of enzymes located in cell membranes . The first area involves a 
demonstration of the role which simultaneous reaction and diffusion might 
play in the formation of biological patterns such as organ shapes and 
surface markings. The second area involves investigation of the effect 
of diffusion limitation by cell membranes on the kinetics of enzymes 
contained in them. Digital computer simulation and numerical solution of 
partial differential equations are the main tools in these investigations. 



PHS-6040 
(Rev. 10-76) 



59 



(1) Dynamic Patterns 

Background and Objectives: 

The objectives of this investigation are twofold. The first is 
a study of dynamic patterns in reaction-diffusion systems obeying 
realistic biochemical kinetics. The objective here is to demonstrate 
that such systems can spontaneously develop patterns of concentration 
variation when unstructured solutions are slightly perturbed. These 
patterns may be stationary, periodic, or may propagate as waves. The 
second objective is to evaluate the hypothesis that such patterns may 
play a role in the formation of adult structures during embryonic de- 
velopment. 

Significance to Biomedical Research: 

Patterns generated by reaction-diffusion processes have been 
implicated in numerous areas of biology, including embryonic 
organogenesis, wound healing, and epilepsy. Several important 
questions are addressed about such patterns. 1) How do the 
patterns depend upon the geometry of the region where they appear? 
2) How stable are the patterns to variations in the shape of the 
region or the physical and chemical parameters of the describing 
equations? 3) How do the patterns change as the region grows? 

Progress in FY 79: 

The kinetics of the enzyme uricase have been shown to produce 
patterns in a one dimensional spatial domain. Stationary, 
oscillatory and propagating patterns have been produced on this 
domain. A computer code using the finite element method has been 
completed which permits study of pattern formation in plane 
regions (two space dimensions) as well as on closed surfaces, such 
as the epithelium of a developing Drosophila egg or imaginal disk. 
Bifurcation analysis of of the describing equations permits us to 
predict where patterns will be found, but not, generally, what 
form the patterns will take. A manuscript summarizing 
accompl ishments during FY 78 has appeared. A second manuscript 
presenting results for FY 79 has been completed and will be 
submitted by the the end of FY 79. 

Future Course: 

During FY 80 a computer code will be developed to model wave 
propagation on two-dimensional regions. This code will be used to 
investigate rotating waves, a phenomena which is observed in 
aggregation of Dictostel iim discoideum as well as in the 
Zhabotinskii-Belusov reaction. Further work will be done on the 
formation of patterns on the egg blastoderm of Drosophila, using 
the finite element code developed in FY 79. 



60 



(2) Kinetics of Enzymes 

Background and Objectives: 

Many enzymes are bound to cell membranes or compartmentalized 
in specialized organelles. In either case, the access to the 
enzyme of reactants, products and cofactors is restricted by the 
relatively low permeability of the membrane structure. The effect 
of these membranes is generally ignored in studies of integrated 
metabolic function. It remains to be determined what effect their 
presence might have on studies of the mechanism of these enzymes or 
on the interpretation of data for the design of specific 
inhibitors. 

Significance for Biomedical Research 

The regulation of cellular electrolyte composition is 
attributed to the sodiim/potassium exchange pump, an enzyme complex 
confined to cell surface membranes. The mechanism of this enzyme 
is an active research area. Among many unanswered questions about 
this enzyme are 1) What is the order of addition of sodium and 
potassium as the enzyme proceeds through the transport cycle? 2) 
Where does metabolic energy couple into the process? Can these 
questions be answered using only data from transport experiments? 

Progress in FY 79: 

The use of the shape of standard plots of enzyme kinetics 
(Lineweaver/Burke, Dixon, etc.) as a reliable indicator of the 
mechanism of enzymes in membranes has been shown to be 
unsatisfactory. Neither net flow nor unidirectional flow 
experiments can discriminate among several biochemically distinct 
models for the site of coupling of metabolic energy into the 
transport cycle. Two manuscripts have been completed. One has been 
submitted, and the second will be submitted shortly. An abstract 
describing current work has been accepted. 

Future Course: 

During FY 80 a dynamical network analyzer, SPICE, will be used 
to simulate temperature- jump experiments on active transport 
models. Methods developed in FY 79 for active transport systems 
will be applied to the modelling of receptor-cyclase interactions, 
in collaboration with D. Rodbard (RR CHI). 

Publications and Abstracts: 

Kernevez, J.P, Joly, G. , Duban, M. C, Bunow, B. , and Thomas, D. : Hysteresis, 
Oscillations, and Pattern Formation in Realistic Immobilized Enzyme 
Systems. J. Math. Biol. 7,41, 1978. 



61 



Bunow, B. Chemical Reactions and Membranes I: Linear Analysis, 
J. Theor. Biol . 75,51,1978. 

Bunow, B. Chemical Reactions and Membranes II: Nonlinear Analysis 
J. Theor. Biol . 75,79, 1978. 

Mikulecky, D.C., and Bunow, B. Distinction among Active Transport 
Models using Dynamic Simulation, Abstracts of the American Physio - 
logical Society , Fall 1979. 

Kaplan, A., Weiss, E. , and Bunow B. Methods for Kinetic Studies 
with Application to Lactate Dehydrogenase from Control and Transformed 
Hepatocytes. Abstracts of XI International Congress of Biochemistry 
1979. (in press) 

Bunow, B. and Weiss, G.H. How Chaotic is Chaos? 1979. Mathematical 
Biosciences (in press). 

Bunow, M.R. and Bunow, B. Phase Behavior of Gangl ioside-Lecithin 
Mixtures 1979. Biophysical Journal (in press). 

Kernevez, J. P., Blanchard, G. , Thomas, D. and Bunow, B. Pattern 
Formation and Wave Propagation in the S-A System, 1978 
Springer Lecture Notes in Mathematics , (in press). 



62 



PROJECT NUMBER (Do MOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00034-03 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Computer-based Studies in Pulmonary Pathophysiology 
and Respiratory Disease 



NAMES, LABORATORY / 


\UD INSTITUTE 


PROFESSIONAL PERSONNEL ENGAGED 


PI: J.J. 


Ba i 1 ey 


B.J. 


Bunow 


B.R. 


Line 


T.B. 


Stibolt 


R.G. 


Crystal 


A.W. 


Nienhui s 


OTHERS: J.E. 


Fletcher 


E.K. 


Harris 


M.R. 


Horton 


B.A. 


Keogh 


A.E. 


Jones 


B.D. 


McLees 


A.R. 


Mitz 



AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
ON THE PROJECT 

Chief, MAS LAS DCRT 

Senior Staff Fellow LAS DCRT 

Medical Research Analyst LAS DCRT 

Senior Staff Fellow LAS DCRT 

Chief PB NHLBI 

Chief CHB NHLBI 

Chief, AMS LAS DCRT 

Chief LAS DCRT 

Computer Systems Analyst LAS DCRT 

Visiting Scientist PB NHLBI 

Chief, Diagnostic Imaging NM CC 

Chief, Critical Care Unit CC 

Engineer LAS DCRT 



COOPERATING UNITS (if any) 

Nuclear Medicine Dept., 
Branch, NHLBI, Critical 



CC, Pulmonary Branch, NHLBI, Clinical Hematology 
Care Unit, CC 



.ab/branch ^ , , . . ^^ . . 

.aooratory of Applied Studies 



SECTION 

Applied Mathematics Section, Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

2.0 



PROFESSIONAL: 

2.0 



OTHER: 



CHECK APPROPRIATE BOX(£S) 
li (a) HUMAN SUBJECTS 

□ (al) MINORS □ (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves a collaborative effort of LAS with the 
Nuclear Medicine Department, CC, the Pulmonary Branch, NHLBI, and 
the Medical Intensive Care Unit, CC. It is directed toward a deeper 
understanding of pulmonary pathophysiology through the constuction of 
computer-based models of pulmonary gas exchange and respiratory 
mechanics and comparisons of model predictions with real patient 
data. 



PH 3-6040 
(Rev. 10-76) 



63 



Background 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 which tended to obscure the true 
nature of lung function. Furthermore, certain nonl inearities inherent 
in the lung system allowed only partial quantitative models and 
sometimes these could only be expressed in the form of nomograms of 
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. 

This program involves the Pulmonary Branch, and the Clinical 
Hematology Branch, NHLBI; the Medical Intensive Care unit, CC; 
the Nuclear Medicine Department, CC; and the Laboratory of 
Applied Studies, DCRT. The objectives include the use of 
computer technology to refine diagnostic methods and to construct 
models for pulmonary gas exchange and respiratory mechanics. For 
example in FY 78 a computer simulation model was developed to 
test the accuracy and precision of various methods for extracting 
ventilation parameters from radioxenon studies of the lung. The 
simulation model showed how the results are affected by 
variations due to counting statistics and revealed how much 
improvement could be obtained through the use of a better isotope 
(xenon 127 instead of xenon 133), by increased dosage over a 
shorter time, and by using more pixels to represent a lung 
compartment (ref 1). 

Also in FY 78 the Kelman procedure which relates gas 
tensions and contents in blood for given values of hemoglobin, 
hematocrit, pH, etc. was revised to incorporate Adair 
binding constants for hemoglobin, to remove unnecessary terms in 
the formulae and also sources of numerical imprecision, and to 
utilize more efficient computer algorithms. 

Progress in FY 79: 

A method using radiogalliim citrate to tag white cells was 
used to study active inflammatory processes in the lung and to 
correlate regions of inflammation with regions of ventilation- 
perfusion abnormalities (refs. 2-4). 

The NIH implementation of the isopleth (Kelman) model for 
blood gases allows global V/Q ratios and respiratory quotients to 
be estimated when both mixed venous and systemic arterial gas 
tensions are known. The data collected on the few patients so 



64 



far has shown some unexpected discrepancies between model 
predictions and laboratory values. Whether this is caused by 
failings in the model or problems with obtaining reliable values 
for gas tension requires further study. 

Also in FY 79 analog devices(physiological signal 
amplifiers) have been set up in the Pulmonary Function Laboratory 
so that pressure volume loops and dynamic compliance data can be 
recorded. At present the data is digitized by the LAS 
minicomputer and the plots and calculations are done with the 
DEC-10 system. It is anticipated that automatic processing of 
this data will be more reliable and far less tedious than the 
traditional manual methods of analysis. Furthermore, this system 
will allow investigators to study the relationship of dynamic 
compliance and other parameters of respiratory function, gas 
exchange, etc. in a large population of patients with varying 
degrees of disease. 

In meetings with LAS, the Pulmonary Branch, and the Clinical 
Hematology Branch the diagnostic and prognostic use of exercise studies 
and the comparison with resting studies has been extensively discussed. 

Significance: 

These computer-based models when combined with data from 
scintillation, cardiac catheter, and pulmonary laboratory studies 
should allow a quantitative description of pulmonary pathophysiology 
on a regional basis. They should allow separation of diseases (e.g. 
bronchitis from emphysema), separation of disease components 
(destructive vs. restrictive vs. vascular), assessment of severity of 
disease component, and prediction of the degree to which each 
component compromises overall pulmonary function. 

In certain diseases of lungs, blood (e.g. hemoglobinopathy), 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 method of assessment which could be more 
objective involves the use of continuous exercise; one can evaluate 
the overall ability of the patient to meet the demands of exercise by 
monitoring his ECG, blood pressure, blood gases and lactate, oxygen 
comsumption, 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. 

Proposed Course: 

In FY 79 a protocol for study of normal volunteers was 
designed and approved. The reproducibility of ventilation- 



65 



perfusion scintigraphy, the sensitivity and spatial 
resolution of regional scintigraphy, and relationship of 
scintigraphic parameters to other measures of pulmonary function 
are to be systematically studied in this population. The 
discrepancies between the isopleth model and measured blood gases 
are also to be further investigated in additional cases. The 
potential use of resting thallium scintigrams to detect right 
ventricular overload in patients with pulmonary and other various 
cases will also be studied. 

The Clinical Hematology Branch has been operating an exercise 
lab; in collaboration with LAS and the Pulmonary Branch they propose 
to implement computer processing of this data, a major aim of which 
will be to define and detect anaerobic threshold in a variety of 
patients as mentioned above. 

Publications and Abstracts: 

Bunow, B., Line, B.R. , Norton, M.R., Weiss, G.H.: Regional 
ventilatory clearance by xenon scintigraphy: A critical 
evaluation of two-estimation procedures. J. Nucl. Med. 20 
703-710, 1979. 

Line, B.R., Fulmer. , J.D. , Reynolds, H. Y., Roberts, W.C, 
Jones, A.E., Harris, E.K. , and Crystal, R.G. : Gallium-67 
citrate scanning in the staging of idiopathic pulmonary 
fibrosis: Correlation with physiology, morphology and 
bronchoalveolar lavage. Amer. Rev. Resp. Pis. , 1978, 118 
355-365. 

Brereton, H.D. , Line, B.R., Londer, H.N. , O'Donnel, J.F., Kent, 
C.H. , and Johnson, R.E. : Gallium scans for staging small cell 
lung cancer. JAMA , 1978 240 666-667. 

Crystal, R.G., Fulmer, J.D. , Baum, B.J. , Bernardo, J. Bradley, 
K.H. , Bruel , S.D. , Elson, N.A. , Fells, G.A. , Ferrans, V.J., 
Gadek, J.E., Hunninghake, G.W., Kawanami, 0., Kelman, J. A., 
Line, B.R. , McDonald, J. A., McLees, B.D., Roberts, W.C, 
Rosenberg, D.M., Tolstoshev, P., Von Gal, E., Weinberger, S.E.: 
Cells, collagen and idiopathic pulmonary fibrosis. Lung 155, 
199-224, 1978. 



66 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00004-08 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Investigation of Hybrid Computing for the Construction of Simulation 
Models and for the Analysis of Physiologic Signals 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI 



E.W. Pottala 



OTHERS: J.J. Bailey 
A.R. Mitz 
I. Yaar 
W.C. Van Arsdale 



Elec. Engineer 

Chief, MAS 
Elec. Engineer 
Vi siting Scienti st 
Pharmacologist 



LAS DCRT 

LAS DCRT 

LAS DCRT 

MNB NINCDS 

DDE FDA 



COOPERATING UNITS (if any) 

Medical Neurology Branch, NINCDS 
Division Cardio-Renal Drug Products, FDA 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 



1.0 



PROFESSIONAL: 

1.0 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
[g (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



^ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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 , 
ultrasonogram, and electromyogram . 



PHS-6040 
(Rev. 10-76) 



67 



Background and Objectives: 

In some simulation models, certain pieces or functions can be split 
off and implemented in hardware circuitry or a set of microprocessors. 
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. And third, some models are so complicated and 
extensive, that their implementation on a large scale digital computer 
is not feasible; whereas 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 FY 76). 

Since FY 72 the LAS Laboratory minicomputer system (MAC 16) has 
been developed and utilized for various projects including (1) the 
construction of physiologic simulation models and (2) the processing of 
physiologic signals. Since FY 72 the system has been interfaced with 
the Marquette tape drive (for routine ECG's from the Clinical Center); 
with the Honeywell 7600 analog tape transport; with a general purpose 
switch-filter network; with a real time spectral analyzer and ensemble 
averager; and with a neural control panel for simulation of neural 
networks (FY 75) and central nervous system subsystems (cerebellum, FY 
76). 

A general advantage of this system is that an investigator can 
automatically pre-process (edit, filter, and digitize) dynamic 
physiologic data so that optimal use of a large scale digital computer 
can be obtained. This was demonstrated in el ectromyo grams from subjects 
on a muscle fatigue protocol studied by the National Institute of 
Occupational Safety and Health (FY 75). This facility was also used for 
electrocardiogram (ECG) and ventricular pressure data in monkeys (FY 76) 
and also ECG data from Glasgow Royal Infirmary (see ECG section). 

Progress During FY 79: 

The Laboratory of Neurophysiology, NINCDS, is studying the 
electrophysiology of the retina of the larvae tiger salmander. The 
Bayl or-Hodgkin-Lamb model of a cone cell requires six compartments 
connected by six linear differential equations and three compartments 
involving sodium and potassium fluxes related by three nonlinear 
differential equations. This model has been implemented in hardware by 
LAS (ref. 1-2). It was also implemented in software on the PDP-10 so 
that the hardware model could be thoroughly checked out. The 
model parameters were adjusted to best fit the actual photoresponses 
(membrane potential) recorded intracellul arly from cones in the retina 
of the larva tiger salamander. Stimuli consisted of flashes and steps 
of light. The responses of the model and the real cone cell were in 
good agreement for 10 millisecond flashes and 0.7 second steps of light 



68 



whose intensities ranged over 2-5 log units with the intensity of the 
un attenuated light at 1.5 X 10 photons per second per square 
centimeter. 

An operating system for the MAC-16 system has been developed which 
provides the user with easy access to compilers and major programs, 
handles I/O for all peripheral devices, and automatically 
allocates system facilities and storage during execution of user 
programs. For example, a program for A/D conversion of dynamic pulmonary 
compliance data was assembled using this new operating system (see 
Section on Pulmonary Pathophysiology and Respiratory Disease). 

The Medical Neurology Branch, NINCDS is studying the use of 
electromyograms to determine muscle fiber conduction velocities and to 
investigate disease states. Preliminary analog to digital conversion 
and spectrum analysis has been performed on the data, to define the 
filtering and amplification requirements of the data collection. 

The Division Cardio-Renal Drug Products, FDA is investigating the early 
detection of cardiac toxicity resulting from drug therapy. Electrocardiograms 
are being used to determine the sensitivity of detection. Programs have 
been written and preliminary analog to digital conversion of the data 
has been done to provide data for the analysis programs. 

Significance: 

The hardware simulation effort has continued to make available to 
the physiologic researcher cost effective ways to investigate and help 
teach basic cellular behavior. 

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 a more sensitive and accurate end-points for drug 
effects. 

Proposed Course: 

The operating system for the MAC-16 will be upgraded for user 
ease. The EMG study will be continued with emphasis on attempting 
diseases states, and additional analysis programs will be written 
for the drug detection study. 

Publications and Abstracts: 

Covacci, Renato: Techniques for the hardware simulation of the 
turtle cone photo responses and extensions of the validity of the 



69 



model to the salamander cone, Proceed del Gruppo Nazionaledi 
Cibernetica e Biofisica , Piza (Italy) , ( in press). 

Vallerga, S. Covacci, R. , and Pottala, E. : Hardware model of turtle 
cone photo responses. Proceed del Gruppo Nazionaledi Cibernetica e 
Biofisica, Piza (Italy) , (in press). 



70 



SMITHSONIAN SCI 
PROJECT NUMBER 



EN 
(D 



CE INFORMATION EXCHANGE 
o NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00042-01 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Computer Based Analysis and Image Processing in Electron 
Microscopy and X-ray and Electron-Loss Spectroscopy 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



OTHERS: 



E.W. Pottala 

J.L. Costa 

C.R. Swyt 

C.E. Fiori 

M.A. Douglas 

K.E. Gorlen 

C.R. Creveling 
E. Silbergeld 



Elec. Engineer 
Medical Officer 
Physicist 
Physical Scientist 

Computer Systems Analyst 

Elec. Eng. 

Chief 

Chief, Neurotox. Sect. 



LAS DCRT 

CN NIMH 

EA NHLBI 

BEIB DDR 

LAS DCRT 

CSL DCRT 

LC NIAMDD 

LP NINCDS 



COOPERATING m\JS (jf any). , , . r. . 

Biomedical Engineering and Instrumentation Branch, DRR, Computer Systems 
Laboratory, DCRT, Clinical Neuropharmacology Branch, NIMH Laboratory of 
Chemistry, NIAMDD, Laboratory of Experimental Atherosclerosis, NHLBI, Section 



g;^/to b' ptox i cology, NINCDS 
Laboratory of Applied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT. NIH. Bethesda, 



MP 20205 



TOTAL MANYEARS: 

2.1 



PROFESSIONAL: 



JUL 



OTHER: 
■ 1 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



a (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves collaboration of LAS with BEIB and several NIH 

It is directed toward the development of computer-based 

and statistical analyses , pattern recognition , and image 



Institutes, 
mathematical 



processing of data, principally X-rays and electron energy loss spectra , 
derived from biological specimens studied in an analytical electron 
microscope. 



PHS-6040 
(Rev. 10-76) 



71 



Background and Objectives: 

BEIB has purchased a 200 kilovolt scanning transmission 
electron microscope which can produce both pointwise spectral data 
(electron energy loss, X-ray) and matrix images containing spectral 
information. It is connected to a KEVEX 7000 spectrometer and with 
various other hardware interfacing to a DEC 11-60 computer system. 
The basic hardware configuration and systems software is being 
developed by CSL along with subroutines to store data and perform 
arithmetic manipulations on various types of images (i.e. to perform 
background corrections and to compare images of mass density present 
in a specimen with those reflecting the distribution of specific 
elements). 

BEIB has planned this research facility in collaboration with 
NIMH, NHLBI, NIAMDD, NINCDS, and DCRT to develop pioneering 
applications of analytical electron microscopy. This can be 
accomplished because: 

1. The high energy beam (200Kev) will provide a significant 
increase in the signal-to-noise ratios of the chemical 
information and should make possible the analysis of hydrated 
specimens. 

2. The assemblage of hardware and computer systems allows rapid 
data acquisition and the complex processing necessary to extract 
chemical information from the raw data - both heretofore 
unavailable at other installations. 

Of particular interest are the electron- energy loss spectra 
(inelastically scattered electrons) which 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 which can 
combine this information into a "true" or corrected elemental map on 
a pixel-by-pixel basis. These models will likely account 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 potentiaKresolution for chemical analyses of the system is in 
the range of 10" to 10" gms. and the spatial resolution is in 
the order of 100 - 1000 square Angstroms. In addition the system can 



72 



construct maps of 1024 x 1024 pixels. Clearly, 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, and ultrasound. 

Progress During FY 79: 

The main event was the arrival of the equipment - the microscope, 
the spectrometer, hardware, and computer. The set-up of the 
hardware and basic systems software is expected to be completed near 
the end of FY 79. 

Significance: 

The ultimate biological goal is to relate structure and function at 
the ultramicroscopic, molecular, and chemical levels. Certain active 
molecules - e.g. enzymes, neurotransmitters, hormones, antibodies, etc - 
can be tagged with appropriate labels - e.g. fluorine - and then 
localized and quantitated within cells by means of this tool. It should 
be possible to determine as well 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 - can also 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 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 
prepared by such means as vacuum evaporation, which are wery thin and of 
known composition. 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 DEC 11-60 system for 
rapid turnaround. 

LAS proposes to undertake, in collaboration with BEIB and the other 
participating Institutes, some of these objectives after the 
combined system is operational. 



73 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00006-09 LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Generalized Mathematical and Computational Methods 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 


E. Hill 


OTHER: 


R. Shrager 




J. Fletcher 



Computer Systems Analyst LAS DCRT 

Mathematician LAS DCRT 

(formerly LSM, DCRT) 

Chief, AMS LAS DCRT 



COOPERATING UNITS (if any) 

none 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT. NIH. Bethesda. MD 20205 



TOTAL MANYEARS: 

.4 



PROFESSIONAL: 

.4 



OTHER: 



CHECK APPROPRIATE BOX(£S) 
D (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



25 (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

The work under this project has been in two general areas: 1) 
generalized optimization techniques for data fitting in the L-one and 
L-infinity norms using the Simplex Method; and 2) file migration modeling 
using optimum network flows in the evaluation of computer system 
performance. In the first area, curve fitting methodology utilizing the 
Levenberg-Marquardt Method for fitting Non-linear Models have been 
implemented as a general Modeling system called an Optimization Laboratory 
(OLAB). Constrained methods are being developed that will contain the 
unconstrained methods as a subset. After completion of formulation and 
testing of the constrained methods the unconstrained methods will be 
replaced by the new constrained methods. The second project area involves 
the analysis of computer systems performance. A methodology for 
evaluating computer system performance has been formulated and described. 



PHS-6040 
(Rev. 10-76) 



74 



Background and Objectives: 

The project on optimization techniques and data fitting is 
concerned with the development of two areas of computer 
applications and mathematics in the biosciences. The first area 
is concerned with nonlinear model fitting and parameter 
estimation. The "standard" methods of least squares have been 
extensively developed for fitting nonlinear models to 
experimentally obtained data with estimation of unknown 
parameters in these models as a basic objective. For many 
applications, the least squares criteron is not appropriate and 
may introduce unnecessary computational difficulties. For 
example, fitting data taken over a wide range of values of the 
independent variables, and fitted without weights, tends to bias 
the results toward the data associated with larger values of the 
independent variables. The problems of ill-conditioning with 
systems having several parameters are also well known. Until 
recently, methods for fitting models with nonlinear parameters 
were not available with other than the least squares criteria. 
This project was concerned with the development of such new 
criteria. 

These methods have been embedded in a generalized modeling 
system called a Mathematical Modeling Laboratory (MLAB). The new 
modeling system is called an Optimization Laboratory (OLAB). 
Currently, OLAB is running using unconstrained algorithms. Ideally, 
the bioscientist would like the capability to constrain the 
parameters in his models to insure that the models are realistic. A 
constraint algorithm is under development to allow the bioscientist 
to fit models with the parameters both unconstrained and with 
constraints. 

The project concerned with evaluating performance of 
computer systems deals with the development of algorithms for 
optimum file migration, and computer system benchmarks. The file 
migration area of research was initiated because of the increase 
in the number of magnetic tapes containing migrated files. The 
benchmark scheme is a method designed to analyze computer system 
data. This method was proposed to assist management in the 
evaluation of the performance of a computer system based on the 
past history of the system. 

Significance for Biomedical Research: 

Data fitting offers the biomedical scientist a methodology 
of determining unknown parameters in biological models. After a 
model has been formulated these techniques are used with 
laboratory or clinical data to study parametric relationships in 
the model . 



75 



The file migration algorithm is designed to reduce the 
nunber of files at certain levels in the computer system. 
Development of a good benchmark is essential for use when the 
system is being analyzed for updating or replacement. These 
techniques can directly reduce the cost of biomedical research 
by reducing the space to store files and aiding management in 
the evaluation of computer system performance. 

Progress in FY 79: 

In FY 79 the Levenberg-Marquardt Method for Nonlinear 
curve-fitting in the L-one Norm and L-infinity Norm have been 
embedded in a generalized modeling system called MLAB which is 
implenented on the NIH PDP 10 computer. Articles describing these 
methods and some applications have been published. The 
interactive computer systems are now available as general research 
tools, and seminars describing their use have been given. 

A file migration algorithm has been formulated using 
network theory. A benchmark scheme has been developed for use 
when setting up a benchmark for computer system performance 
evaluation. Articles describing this work have been published. 

Proposed Course: 

New constraint algorithms will be implemented in OLAB as they 
are developed and tested. Further work on the evaluation of computer 
performance will be subject to section priorities. 

Publications and Abstracts: 

Hill, E. : Discrete rational approximation in the L-one and 
L-infinity Norms using the Simplex Method, Proceedings of 
ACM 1978 , December 1978. 

Hill, E. : Attributes of a file migration model for shared files, 
ACM Computer Science Conference , February 1979. 

Hill E. : Computer systems evaluation. Conference on Information 
Sciences and Systems, The Johns Hopkins University, 
Electrical Engineering Department, March 1979. 

Shrager, R. and Hill, E. : Some properties of the Levenberg method 

in the L-one and L-infinity norms. Mathematics of Computation (in press). 



76 



SMITHSONIAN SCI 
PROJECT NUMBER 



ENC 
(Do 



CE INFORMATION EXCHANGE 
NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT00045-01 LAS 



PERIOD COVERED 

October 1. 1978 to September 1, 1979 



TITLE OF PROJECT (80 characters or less) 

Numerical Approximation Techniques for the Solution of Reaction - 
Diffusion Systems in Biology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: M. Bieterman 

OTHER: J.E. Fletcher 
B. Bunow 
I. Babuska 



Mathematician, AMS 



LAS DCRT 



Chief, AMS LAS DCRT 

Biomathematician, AMS LAS DCRT 
Professor, Institute of Physical 
Science and Technology 
Univ. of Mary! and 



COOPERATING UNITS (if any] 

none 



lab/branch 
Laboratory of Applied Studies 



SECTION 

Appl ied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MP 20205 



TOTAL MANYEARS: 

.9 



PROFESSIONAL: 

.9 



OTHER: 



CHECK APPROPRIATE BOx(ES) 
n (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



B (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 



reac 

soft 

via 

math 

proc 

nine 

prog 

impl 



The fundamental tools needed 
tion-diffusion systems are the 



ware used to solve nonlinearly 
computer. This project provid 
ematical support for ongoing 1 
esses: (1) new, theoretically 
rical solution techniques are 
rams, and (2) existing methods 
emented to solve specific mode 



to compute the solution of model 
mathematical techniques and related 
coupled partial differential equations 

es two interrelated forms of 

aboratory modeling of biological 
sound and computationally efficient 

investigated and developed into computer 
and computer programs are modified and 

Is. 



PHS-6040 
(Rev. 10-76) 



77 



Background: 

The mathematical models describing reaction, diffusion, and 
transport processes in biological and physicochemical systems 
consist of systems of nonlinearly coupled ordinary and partial 
differential equations. Specific mathematical techniques and 
iiLitierical methods needed to treat many of those equations either do 
not exist in the present literature, or require a '^ery large amount 
of computer time, making their implementation economically 
impractical. The development of new techniques, and continuing 
attempts to improve the efficiency of known algorithms are the 
objectives of this research area. 

Significance to Biomedical Research: 

The types of equations being studied are used to model 
substrate transport in the mircrocircul ation and embryologic 
pattern formation. Other biomedical application areas include 
population dynamics of ecological systems, population genetics, 
nerve impulse transmission, and tumor growth, vascularization, 
and necrosis. The effective modeling of physiologic processes 
requires efficient methods for the numerical solution of systems 
of lumped or distributed parameter models. Such models are 
required to describe the complex interacting systems with flow, 
diffusion, and chemical reaction which are basic to normal and 
pathophysiology. Collection and characterization of solution 
techniques also makes them available to the NIH biomedical 
community as research tools. 

Progress in FY 79 

Investigation of numerical methods used to solve coupled 
systems of linear and nonlinear partial differential equations has 
continued. The methods include both finite difference techniques 
and finite element methods. Previously written software was 
upgraded, and a number of new computer programs were added to our 
1 ibrary. 

In the area of microcirculation modeling, finite difference 
schemes are being developed to solve a model which incorporates 
time dependent capillary transport effects. 

Versions of the Bathe-Wilson finite element code were 
implemented on the NIH IBM 370 system. The programs were modified 
to solve time dependent, steady state, and eigenvalue problems 
associated with biological pattern formation on irregularly shaped, 
two dimensional regions. The flexibility provided the user in 
description and discretization of regions, solution accuracy 
obtained, and speed of computation show the finite element method 
to be a useful and productive tool in biological modeling. 



78 



Background studies and initial research have been completed 
for the development of adaptive finite element schemes for the 
solution of time dependent reaction-diffusion systems. The schemes 
are adaptive in that a temporal and spatial mesh refinement 
strategy, based on reliable, local a posteriori error estimates, is 
carried out by the computer during the problem solution. This 
procedure is expected to minimize required computer time, 
especially for models involving the response of systems in a 
nonuniform mediun, on which attention is focussed. 

Proposed Course: 

Mathematical support of ongoing modeling projects will 
continue in the forms into research of adaptive finite element and 
finite difference schemes and further development and application 
of known techniques to specific models. These efforts will 
continue at levels determined by section priorities and manpower. 



79 



COMPUTER SYSTEMS LABORATORY 



Alan M. Demmerle 
CHTef 




Using teahnology originally designed for medical 
telecommunioations projects, CSL members implemented 
a voice output terminal for blind computer programmers. 
Exhibiting the terminal at the 1979 President's 
Conference on the Employment of the Handicapped, 
Perry Plexico (left) and David Songco (back to camera) 
answer questions. 



I. SUMMARY 
Function 

The Computer Systems Laboratory (CSL) identifies and solves problems 
in areas of biomedical research and clinical care where real-time data 
collection, analysis, display and experiment control are required, 
where economic considerations favor a small computer, or where 
proximity of the computing equipment to the work site is important to 
successful solution. 

CSL's approach to each project varies as each problem presents unique 
challenges. The staff: 

analyzes the users specific requirements. 

evaluates alternative solutions with regard to technical 
merit, time to completion, and cost. 

discovers new solutions to automation problems. 

develops special components to meet user's needs. 

refines systems to adapt to research progress. 

The activities of CSL center in these principal areas: applications 
in a clinical environment, applications in laboratory research, and 
consultation with researchers in need of computer expertise. These 
activities are carried on with approximately 30 employees, two-thirds of 
whom are electronic engineers, and the remainder computer scientists 
with diverse backgrounds including some with backgrounds in medicine, 
biology, and chemistry. 

Scope of Work 

Over the past ten years, CSL has worked with nearly all of the 
Institutes on a wide variety of requirements. Major effort has been 
expended upon a few large projects that have required an investment of 
ten to twenty- five man-years. In fact, some of these long-term 
projects have no distinct ending because of continuing modification 
to accommodate new research protocols. 

Most of the projects this year were started in earlier years, or in 
some cases grew out of projects started earlier. One significant 
factor affecting the ratio of old to new projects is the decreasing 
size of our Laboratory. As the number of personnel decreases the 
progress on existing projects slows and we are seriously restricted in 
our ability to start new projects. 

Each year, CSL also has a number of relatively small jobs, ranging in 
size from one man-week to a half man-year. These involve consulting 
with the intramural and extramural programs in the areas of automated 
data collection, display and analysis, data transmission, certain 



82 



aspects of biomedical sensor design and all aspects of real-time 
computer system design. 

Intramural consultation often is followed by development of hardware 
and software systems; extramural consultation activities are primarily 
advisory. The consulting activities are valuable to us in that they 
help keep us aware of other activities in biomedical engineering, 
but require only a minimal commitment of manpower resources. 

Highlights of the Year's Activities 

During FY79, CSL continued to utilize minicomputer and microcomputer 
based technology in support of clinical care and laboratory research 
programs of widely varying scope and duration. Three projects of 
recent origin are clinically oriented but nevertheless reflect much 
of the diversity of CSL work and fully illustrate the impact of CSL 
contributions to the biomedical community. 

The Cardiac Scintillation Probe Project illustrates CSL's commitment 
to identifying and implementing new applications of computers to 
clinical research at the NIH. It also indicates the impact that the 
new and still emerging microprocessor technology is having on 
biomedical instrumentation. 

This pioneering effort began in FY77 as a spin-off of a large, 
elaborate diagnostic system developed at NIH several years ago. That 
system combines a scintigraphic camera, a computer, a display, and 
other equipment to analyze heart function. Whereas the camera 
system is large and not easily moved, the Cardiac Scintillation 
Probe was developed to provide a direct, quick, and noninvasive 
means of measuring left ventricular (LV) function. These attributes 
offer a physician the potential, at the bedside, for examining 
acutely ill patients prospectively for predictors of a worsening 
condition. 

The instrument consists of a scintillation probe on an adjustable 
arm affixed to a cart that can be easily positioned at the bedside. 
The cart contains probe electronics and a microcomputer. The instrument 
is used to acquire scintillation data by positioning the probe over the 
left ventricle (LV) of a patient who has been injected with an 
appropriate radioisotope. The microcomputer program organizes these 
data into a composite LV time-activity curve, corrects the curve for 
the effects of background radiation, and calculates ejection fraction, 
an important measure of cardiac performance. Both the time-activity 
curve, which is proportioned to LV volume, and the calculated ejection 
fraction are displayed in real-time. 

During the current year, the instrument was further developed for use 
in NHLBI's cardiac catheterization laboratory. It is used, in 
conjunction with LV catheterization, for simultaneous LV volume and 
pressure determinations. LV pressure-volume measurements can be made in 
real-time so that pressure-volume loops and other displays can be 



83 



generated, and parameters such as LV compliance can be determined 
while a patient is still in the laboratory. The advantages of this 
method, compared to conventional approaches, are striking. For 
example, LV volume determinations resulting from measurements made on 
cineangiographic films of the heart can normally be obtained only hours 
or days following the catheterization procedure. 

Current work includes the development of methods for making back- 
ground corrections simultaneously with LV volume measurements. We 
also plan to develop a new high efficiency probe which will allow 
statistically reliable measurements in much less time than is currently 
required. It will also enable beat-to-beat evaluation of LV volume, 
something which is not now possible. 

The Medical Intensive Care Unit Patient Monitoring Project involves 
automated collection, analysis and display of data from the recently 
established Medical Intensive Care Unit (MICU). This unit is admin- 
istered by the newly formed Department of Critical Care Medicine in 
the NIH Clinical Center and receives critically ill patients from 
clinical programs of NIH. The MICU comprises a 5-bed ward area, a 
2-bed special study area, a pair of isolation beds, and a vascular 
research laboratory. 

The primary objective of the Project is to combine continuous automated 
data logging of patient vital signs (including respiratory functions) 
with vascular catheterization data. The MICU goals require the use 
of three minicomputers and a microprocessor. 

The bedside patient monitoring system consists of Hewlett-Packard 
physiological monitoring components at the bedside and a central 
Hewlett-Packard minicomputer system. Bedside 16-button keyboards 
and 12-inch CRT displays allow the medical staff to interact with 
the computerized monitoring system. Terminals located at the nurses' 
station and at three other convenient sites provide for the entry of 
textual information, as well as control and display of data. 

The physiologic signals to be monitored include heart rate; systemic 
arterial, pulmonary arterial, and pleural pressures; body temperature; 
urine output; respiration rate; airway pressure; and airway flow, 

A second Hewlett-Packard minicomputer provides for data acquisition 
and subsequent data processing within the vascular research laboratory, 
A physiologic monitoring console containing eight signal conditioners, 
an eight channel oscilloscope, an eight channel strip chart recorder, 
and a seven channel analog tape recorder will be used to acquire 
physiologic data during pulmonary vascular and peripheral vascular 
catheterization procedures. On-line data collection and analysis is 
controlled from within the catheterization laboratory using a specially 
designed keyboard and a CRT display. This system is magnetic tape 
compatible with the bedside monitoring system. 



84 



Program development for above described computer systems will be 
accomplished with a third Hewlett-Packard minicomputer. This program 
development system will also support data base management software. 

Pulmonary gas exchange will be evaluated through the use of a Cemetron 
medical mass spectrometer permanently located in the computer room. 
The mass spectrometer's inlet line is connected to each patient's 
airway by means of electronically controlled solenoid valve leading 
from a common gas manifold. Long small-bore tubing runs from each 
solenoid valve to a T-adapter on the respirator's delivery tubing. 

Data from this mass spectrometer will be interfaced to the Hewlett- 
Packard data management system, in order to merge pulmonary function 
and gas composition data with the main patient data files. In addition, 
the incorporation of airway pressure and flow monitoring into the 
Hewlett-Packard systems, through modifications to be installed by 
CSL staff, will allow the computation of net O2 consumption and COp 
production values. 

The final components to be added to the commercial patient monitoring 
and data management system are IMED computer-controlled volumetric 
infusion pumps. Integral microprocessors allow these infusion pumps 
to engage in two-way communication with the monitoring system. 
Carefully developed and well-proven algorithms for continuous intra- 
venous infusion therapy will provide rate control of these infusion 
pumps, based on the values of the physiologic parameters to be con- 
trolled. 

The Biomedical Image Analysis Project is expected to provide an image 
analysis capability appropriate to a wide-range of CSL activities and 
NIH requirements. 

This project had its origins in the Computerized Radiotherapy project 
which seeks to use computerized tomography (CT) images to calculate 
better radio therapy protocols. The Biomedical Image Analysis Project 
has taken advantage of the powerful DCRT Evans and Sutherland (E and 
S) Image System to develop, and test the suitability of algorithms to 
achieve improved organ and tumor identification and contouring. The 
algorithms include both classical and experimental techniques for 
producing contrast enhancement, edge detection, contour extraction, 
contour following, contour coordinate data compression and three- 
dimensional representation. While many of these algorithms achieve 
improvement over currently existing clinical techniques, the approaches 
considered to be the most promising use discriminating features 
derived a priori from large populations of data representing normal 
and diseased organs. 

The image analysis effort on CT scans attracted the interest of 
scientists engaged in other research programs. Two projects, an 
NHLBI atherosclerosis study and an NINCDS neuro-anatomy modeling 
project, evolved from this interest during the past year. 



85 



Investigators in the Laboratory of Experimental Atherosclerosis, 
NHLBI study the topological variation of disease along arteries of 
experimental animals fed high cholesterol diets. The arterial segments 
are removed, opened, pinned out, stained, and photographed. 

CSL was asked to provide a way for determining statistical measures 
of the disease involvement from the pictures of these sections. We 
developed algorithms to normalize sections by mapping them into 
standard templates and to extract areas of disease involvement as 
well as total section area. These programs are now operational. 

Investigators in the Laboratory of Neuro-Otolaryngology, NINCDS, 
asked CSL to collaborate on techniques to visually explore the anatomy 
of the auditory system of the mammalian brainstems. Data representing 
neuronal cells and their spatial relations in one superior olivary 
complex of a cat brain were derived manually from sections of the 
complex. 

These data were transferred to the DCRT E&S System via magnetic tape 
and intercomputer transmission and transformed and displayed in color, 
each color representing a particular class of neurons. The primary 
display for each section was composed of small colored squares each 
representing the neurons counted in an 80 micron square area. Next 
software was developed to interpret all the data as stacked planes of 
80 micron colored cubes and to view any arbitrary planar slice through 
this block model. Images generated by the block model may be trans- 
ferred back to a local Neuro-Otolaryngology Laboratory computer for 
printout on their Versatec plotter. 

A 16mm color movie entitled "Atlas of the Cat Superior Olivary 
Complex" was produced. This movie shows the computer generated images 
representing the original 61 sequential transverse sections followed by 
sequential reconstructed images representing horizontal and para- 
sagittal sections. 

Users of the block model feel it provides insight into the anatomy 
of the superior olivary complex. We planned to improve and extend the 
usefulness of the model as a research tool. 

Future Plans 

Our future depends upon our ability to maintain our size, and meet 
the challenge of wisely setting priorities in an environment of 
increasing demand for our serivces. For the past several years, CSL 
has suffered personnel cuts, and FY79 is no exception. These recent 
cuts emphasizes the problem of setting priorities in the work we do; 
in selecting from among the new projects presented to us, and the 
continuation of old ones. 

Perhaps, the most difficult choices to be made concern the level of 
effort required to support some of our projects. The timing of 
completion of our projects is often critical; delays reduce the value 



86 



of a project. In essence, our projects are timed to the interests 
of the scientists we support, and to the activities of the instrumen- 
tation industry. If we are too slow the scientists move on to other 
work and we fall behind the fore-front of technology. In addition the 
"critical mass" concept applies to some of our larger projects where 
a number of people, representing several specialities are required. 

Thus, the priorities cannot always be established only on a deliberate 
assessment of value of the project. It must take into account other 
factors including available talent, previous commitment, project size, 
and timing constraints. 



87 



II. ANNOTATED PROJECT AND ACTIVITIES LIST 



Clinical Research, Patient Care Projects 

Computerized Radiation Therapy, NCI, RO : CSL, in collaboration with 
the Radiation Oncology Branch, NCI, has developed a computer system now 
in operation in R.O. to use the detailed contour and density information 
available from computer assisted tomography to improve radiation 
treatment planning. Our system allows the radiotherapist to review 
scans of a tumor area at a video terminal . 

The therapist may quickly vary the presentation of the image to 
emphasize a particular structure, or magnify an interesting feature. 
Outlines of the tumor or vital organs, in some cases drawn automatically 
by the computer, are then processed by software algorithms procured 
from Atomic Energy of Canada, Ltd. to provide sophisticated individual- 
ized treatment plans. Finally, the calculated dose distribution may 
be overlayed on the scan for evaluation of the plan. 

During the past year, CSL has added improved display hardware to the 
computer system, and written a large collection of programs to allow 
display and manipulation of both grey-scale and true color images of 
the scans. Also, programs were written to permit manipulation of the 
displayed scans (sector selection, magnification, zooming, etc.), to 
computer designated areas and distances, and to aid the therapist in 
planning treatment. For example, it is now possible to display lines 
of isodose in unique colors, overlayed on the original CT scan. A 
large, much faster disc was added to the system to accomodate the 
large amount of data generated by clinical usage. 

Current development includes calculation of dose off the central beam 
axis along the third dimension of the tumor, an investigation of 
ultrasound as an instrument for treatment planning, psuedo-optimization 
of treatment planning, and procurement of tumor registry software. 

Medical Intensive Care Unit Patient Monitoring Computer System, CC, DCCM : 
During this reporting year, a five-year contract was initiated with the 
Hewlett-Packard Corporation for the acquisition of a commercially devel- 
oped computerized patient monitoring system for the Clinical Center's 
new Medical Intensive Care Unit. The monitoring system's primary 
function will be the periodic data logging of vital signs from the nine 
bed care unit. 

A separately procured medical mass spectrometer, with an attached micro- 
computer, will provide the patient monitoring computer with information 
on respiratory gas composition. Pulmonary vascular studies will be 
carried out using a computerized vascular research subsystem, and a 
development computer subsystem will allow the testing of new applications 
programs in support of ongoing research protocols and will provide for 
data-base management. 



88 



The monitoring system's installation requirements were identified and 
the Medical Intensive Care Unit's construction specifications were 
modified appropriately to ensure compatibility. Construction of the new 
care unit proceeded as the computerized monitoring system arrived. 
Following completion of the care unit, installation of the computer 
system was initiated and the first patients were admitted to the unit 
a week later as the bedside monitoring equipment became functional. 

A survey of commercially available intravenous infusion pumps was 
carried out in order to choose a pump for interconnection to the 
computerized monitoring system. A pump was selected containing an 
integral microprocessor that allows bi-directional communication with 
the monitoring system. 

CSL expects continued involvement during the forthcoming year. 

Cardiac Intensive Care Unit Patient Monitoring Computer System, NHLBI, SU : 
For the past several years CSLand S.U. of NHLBI have been collaborating 
in the development of a system to collect, format and display vital 
data from post surgical cardiac patients. The annual report from previous 
years elaborates on the goals and characteristics of this system. This 
year a particularly significant accomplishment has been the acceptance 
of computer output, by the NIH Medical Records Committee, for inclusion 
into the permanent patient chart. Also this year an alarm generating 
software package, which produces messages when monitored parameters 
exceed preset limits, or when certain abnormalities in the parameters 
are detected, was completed. Future plans include a formal evaluation 
of the operation of this system and study of a next generation system 
for implementation in new NHLBI care facilities. A reduced level of 
effort is anticipated during the coming year. 

Computer Systems for Nuclear Medicine (Previously Reported as Cardiac 
Scintillation Probe), CC, NM and NHLBI, IR CB : CSL has continued 
development of its Cardiac Scintillation Probe System begun in 1977. 
This non-imaging ECG-gated scintillation probe, when used in conjunc- 
tion with left ventricular (LV) catheterization, permits simultaneous 
quantification of the variation of LV volume and pressure. It combines 
a new high efficiency scintillation probe with computerized data 
acquisition, processing, and display to produce a time-activity curve 
in real-time. When the probe is properly positioned over the left 
ventricle, this time-activity curve is a measure of change in LV 
volume. By simultaneously measuring LV activity and LV pressure, 
parameters such as LV compliance can be continuously monitored. This 
year we implemented the capability for performing real-time pressure- 
volume measurements in the cardiac catheterization laboratory for drug 
intervention studies. These results were presented to the Nuclear 
Medicine Society in June 1979. During the coming year, we expect to 
achieve further increases in probe efficiency so as to permit beat-to- 
beat analysis of ventricular function. 



89 



We have also continued our consultation and support for the imaging 
systems located in the Nuclear Medicine Department. This involved 
assessing computer requirements with respect to the continuing increase 
in the number of clinical and research studies. This year we began 
working toward a distributed processing system (OS) in Nuclear 
Medicine by inter-connecting two imaging systems. During the coming 
year, the DS system will be expanded so that all imaging systems are 
connected together in a network. For relatively little expense, the 
DS system increases the capabilities and utilization of diagnostic 
imaging by allowing peripherals to be shared and minimizing scheduling 
conflicts. We are also collaborating with Nuclear Medicine and the 
Cardiology Branch is purchasing and implementing a portable scinti- 
graphic camera and associated computerized imaging system for use in 
Cardiology's animal laboratory. 

Automated ECG Processing, CC and NHLBI, IR CB : The Clinical Center's 
ECG Service requested assistance in 1977 in specifying and acquiring an 
on-line ECG system to assist staff physicians in analyzing and 
reporting current ECG's and comparing serial ECG's. CSL and LAS have 
reviewed available literature, field tested commercial systems, and 
followed DOD's Tri-Service Medical Information System procurement 
effort. A system was specified in 1978 but procurement was post- 
poned by the Clinical Center. The 1978 draft request for proposals 
(RFP) containing functional specifications for a minicomputer-based 
ECG system was revised and upgraded in 1979 to include, among other 
things, greater emphasis on research data processing and interchange of 
ECG tracings and reports. The RFP is currently awaiting action by the 
Research Contracts Branch; funding of the system is expected to be 
accomplished with FY79 or FY80 money. 

Pulmonary Branch Project, NHLBI IR : The Pulmonary Branch of the 
National Heart, Lung and Blood Institute has asked that CSL participate 
in an intensive collaboration with two primary goals: 

1. Automation of the various test procedures performed by PB, 
thereby obtaining increased accuracy and faster turnaround from 
test performance to useful clinical results. 

2. Development of a patient database for storage and retrieval of 
all test results generated by PB such that they can subsequently 
be used to follow an individual patient, or to evaluate the 
potential clinical significance of findings across patient 
populations. 

This project is anticipated to be a significant and time-consuming 
endeavor, requiring CSL manpower resources .that are not available at 
this time. An effort to recruit a special expert who will work in this 
area for the anticipated project duration of two to three years is now 
underway. In the meantime, CSL is providing limited consultation to PB 
to ensure that equipment procurements and data processing steps taken now 
will be consistent with future needs. 



90 



Clinical Pathology Department Project, CC : A general -purpose micropro- 
cessor-based instrument interface is being developed in conjunction with 
the Clinical Pathology Department, CC. The intent of this project is to 
link a variety of analytical instruments to the Clinical Pathology 
computer in a similar fashion. This will include the interfacing of 
instruments not previously connected to the computer, as well as the 
replacement of some existing interfaces. The new interfaces provide the 
means of processing specimens independently of predetermined load lists, 
local preprocessing and edit checking of data prior to transmission to 
the computer, and a mechanism for rapid certification of results for 
priority (stat) requests. These enhancements were not possible with the 
existing techniques, 

A prototype unit has been implemented for a Coulter Counter, an instru- 
ment which analyzes blood for seven different parameters. This device 
was installed and placed in operation in May 1979. Additional units are 
planned for a second Coulter Counter, a blood gas analyzer, and a 
Technicon SMAC automatic analyzer. 

B i o medical Image Analysis, NCI, RO and CC, PR, and NHLBI, IR and NINCDS , 
LNO : During the past year, CSL has participated in several collaborative 
intermural biomedical research projects having image digitization, trans- 
mission, storage, retrieval, processing and analysis as common require- 
ments. The DCRT Evans and Sutherland Picture System (E&S) has been 
serving as the host computer in this work. We are working toward making 
a "utility" for picture processing which currently features: 1) tape 
compatibility between EMI CAT Scanners and the DCRT (E&S) System, 
2) tape compatibility between an Optronics densitometer (located at 
Johns Hopkins University, used by NHLBI) and the DCRT E&S System, 3) a 
library of image processing and analysis software, much of which is 
interactive, 4) a 16mm color movie facility for documenting colored 
3-dimensional displays. 

Progress in specific collaborative intermural projects is reported below: 

Members of the Laboratory of Experimental Atherosclerosis, NHLBI are 
studying the topological variation of disease along the arterial tree of 
animal specimens. Images of stained arterial sections taken from 
various experimental animals are currently digitized by means of a 
microdensitometer for input to the DCRT E&S System. The images are 
mapped into standard templates and the area of disease involvement is 
determined. Most of the processing software is operational and current 
efforts are directed toward digitizing, processing, analyzing, and 
presenting results related to approximately 400 images. 

The Laboratory of Neuo-Otolaryngology, NINCDS has requested the develop- 
ment of a 3-dimensional block model of the cat brain stem from which 2- 
dimensional images and shaded surface drawings representing any chosen 
plane may be reconstructed. The block model is now operational and a 
16mm color movie illustrating its use has been produced. 



91 



In collaboration with the Radiation Oncology Branch, NCI and Diagnostic 
Radiology, CC, several programs have been written to improve upon the 
use of CAT scan data in radiation treatment and diagnostic procedures. 
These programs, most of which are interactive include contrast enhance- 
ment, edge detection, edge extraction, data compression and 3-dimensional 
reconstructions. 

We expect to continue work, possibly at a reduced level of effort, during 
the coming year. 

Flow Microfluorometry Cell Sorter (FMF). NCI, I. LP and NIAID, LMI : 
Since FY75 CSL has provided engineering, system integration, and software 
support necessary to meet the data acquisition, data display and analysis 
needs of investigators using Flow Microfluorometers (FMF's) at NIH. 

In FY77, the popularity of cell analysis using the shared computer based 
NCI FMF system increased considerably with the result that the existing 
computer system could no longer handle the workload. In April 1978, the 
installation of a second computer (PDP-11/34) provided each of the two 
NCI FMF instruments a dedicated computer. Also in FY78, major emphasis 
was placed on improving existing and implementing new support software. 
Some of the display and analysis programs have been enhanced to reduce 
the time spent by researchers at the computer console by providing auto- 
matic sequencing through data files. This feature is accomplished by 
preselection of parameters that are automatically applied to one or more 
data files without operator intervention. 

New statistical calculations were added in FY79. Also, a large part of 
a comprehensive user's manual was written to provide operating and instal- 
lation instructions for CSL written software, and to facilitate effective 
use of the computer in the FMF environment. 

A PDP-11/34 system that was ordered in April 1978 by NIAID, IMI, was 
delivered in December 1978. A contract was awarded to a commercial 
vendor to fabricate from CSL plans a FACS II/PDP-11 interface for the NIAID 
system. Because of room renovation delays and modifications to the FACS II 
FMF, the NIAID FACS II instrument-computer system has not yet been 
integrated although most of the software has been successfully tested on 
the PDP-11/34 system. 

Evaluation of existing and anticipated requirements for the two FACS FMF 
systems has resulted in CSL recommending that software currently running 
under the RT-llM single user operating system undergo conversion to 
function under the RSX-llM real-time multiprogramming operating system. 
Planning for this conversion is already underway. We also recommended 
replacement of the NCI PDP-11/40 computer with PDP-11/34 and additional 
memory and data storage capabilities. Currently, we are also working on 
minor improvements to the applications program (under RT-U) for LASL, 
the second NCI FMF svstem. 



92 



The NIAID FACS system has the recommended equipment for the software 
conversion. CSL expects to convert the existing application programs for 
FACS, and to design and implement under RSX-llM a more sophisticated 
data acquisition program that will allow considerable record keeping to 
be done by the computer rather than the experimenter. 

In the forthcoming year, CSL plans to continue adding automatic 
sequencing capabilities to the display and analysis programs, complete 
the User's Manual and begin implementation of RSX-llM software. 

• Laboratory Investigation Projects 

Distributed Laboratory Data Acquisition and Control System, NIAMDD : 
A Distributed Laboratory Data Acquisition and Control System (DLDACS ) i s 
being implemented for NIAMDD, in Building 2, as a replacement to the 
Laboratory Computer System developed here in CSL ten years ago and 
which has now become overloaded and technologically obsolete. The new 
system will consist of a network of remote microcomputers connected in a 
star configuration through a communications processor to a central data 
processing computer. The remote microcomputer handles all of the real- 
time data acquisition requirements and provides 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. Utilizing serial multiplex hardware, 
it is capable of communication with up to twelve remote microcomputers 
and places incoming files on a queue to be transferred over a parallel 
DMA channel to the central computer. The new system for data collection 
and processing is designed to allow laboratories connected to the old 
system to be converted with minimum disruption to ongoing instrument 
use. 

The prototype element of this network was installed this year. It 
connects the C-118 Spectrophotometer to the old H-516 computer via the 
remote microprocessor (DEC LSI-11), using the RT-11 operating system and 
a communications processor (also a DEC LSI-11). 

Much of the software written for this first element in the network can 
be used for all elements of the network, including that for the communi- 
cations of data, and for data acquisition. Thus, the remaining 
elements of this network will require less development time. 

Presently, the "central data processing computer" is a Honeywell DDP-516 
but in conjunction with the conversion to a distributed system the H-516 
will be replaced by a Digital Equipment Corporation PDP-11/70. Major 
microcomputer components required to convert all laboratories currently 
serviced by the H-516 to the distributed system have been ordered. 
Software and interface hardware for tv/o of these units is currently under 
development. We estimate that six to nine man-months will be required 
to replace each of the remaining three instruments and we expect to 
continue working on this project during the coming year. 

93 



Scanning Electron Microscope, NIAMDD. LCP and NCI, LP : Following the 
recommendations of CSL, Drs. Hagins (NIAIDD) and Banfield, NCI, 
initiated procurement of a combined energy-dispersive (EDS) and wave- 
length-dispersive (WDS) X-ray analysis system for their scanning electron 
microscope. A single vendor did not market a combined EDS-WDS system. 
However, the supplier of the EDS system. Tracer Northern, agreed to 
integrate their PDP-ll/34-based system with a WDS system manufactured 
by ETEC. The final system will perform analyses as well as acquire and 
transmit data to the DECsystem-10 for further computation. After a 
number of delays on the part of the WDS vendor, it now appears that 
delivery of the entire system is imminent. No further involvement by 
CSL is anticipated. 

Selected Ion Recording gc-ms Data System (Previously Reported as Mass 
Spectrometer), NIMH, LCS : The SIRS project has been a collaborative 
effort between the NIMH Laboratory of Clinical Science and CSL to 
produce a minicomputer based data acquisition and control system for 
magnetic sector and quadrupole mass spectrometers which would 
facilitate high volume use of the spectrometers for quantitative 
analysis and be capable of being modified and enhanced by the user. 

Despite signigicant difficulties with the spectrometer, especially in 
focusing specific ions reliably to within 0,1 mass number, the SIRS data 
system was completed this year, and is now being used routinely by 
investigators to perform quantitative analysis with a LKB-9000 gc-ms 
instrument. SIRS has also been expanded to work with quadrupole 
spectrometers. 

A poster session reporting the availability of SIRS and describing 
its general characteristics was presented at the Spring Meeting of the 
American Society for Mass Spectroscopy in Seattle, Washington. 

No further work by CSL is planned. 

Analytical Ultracentrifuge Data System, NHLBI, IR : Design and develop- 
ment work began in May 1979, on a microcomputer-based data acquisition 
and communications system for Dr. James Osborne, Jr., Metabolism Branch, 
NHLBI. This system will facilitate the acquisition and processing of 
information from an analytical ultracentrifuge and a circular dichroic 
spectropolarimeter that are the primary tools Dr. Osborne's group use 
to investigate the interactions between human lipoprotein subunits. 
Although, the microcomputer system will perform some of the processing, 
the complex nonlinear analyses necessary for the characterization of 
interacting systems will be carried out under MLAB on the DECsystem-10. 
CLINK, the PDP-ll/PDP-10 communications software package jointly 
developed by CSL and CCB, will be used to. perform the data transfers. 

Microanalysis Facility, BEIB : BEIB is undertaking the establishment of 
a microanalysis facility which will conduct electron energy-loss 
spectrometry studies with investigators of several institutes at NIH and 
NIMH. These studies will involve computer acquisition and processing 



94 



of spectral data, investigation of techniques for computer control of 
the electron microscope, and investigation of some image processing 
and pattern recognition algorithms. 

The facility will consist of two electron microscopes and a PDP-11/50 
computer system. The first microscope, a 200 KeV Scanning Transmission 
Electron Microscope (STEM) was delivered in May 1979, and is equipped 
with 1) a spectrometer for performing Electron Energy Loss Spectrometry 
(EELS), 2) a lithium-drifted silicon (Si(Li)) detector for gathering 
Energy-Dispersive X-ray Spectra (EDS), and 3) detectors for measuring 
forward-scattered, back-scattered, sample, and secondary electron 
currents. 

The second microscope, a 50 KeV electron microprobe, will also be 
delivered during 1979. This microscope will also be equipped with a 
Si (Li) detector and electron current detectors. In addition, it will 
have three detectors for collecting Wavelength-Dispersive X-ray 
Spectra (WDS). 

The PDP-11/60 computer system, purchased in FY78, will be used to 
acquire and process data from the two microscopes. The computer system 
was made operational in April 1979 and a Kevex Model 7000 Analytical 
Spectrometer was interfaced to it. The Kevex 7000, in conjunction with 
the 11/60, is capable of gathering and processing spectra from the 
Si(Li) and EELS detectors on the 200 KeV STEM. 

CSL is currently involved in designing and procuring the hardware and 
software necessary to interface the 11/60 to the 200 KeV STEM. It is 
planned to equip the 11/60 with a microcomputer-based (LSI-11) front- 
end processor to acquire the four electron current signals at up to 
10,000 samples/second-signal, and also acquire the EELS data and control 
beam position. All of this data will be transferred to the 11/60 
for storage and processing. An Analogies AN5400 data acquisition 
subsystem has been ordered to permit the 11/60 to control portions of 
the electron spectrometer and monitor the electron beam voltage, 
beam current, and lens currents and temperatures. 

In conjunction with studies invo'lving wavelength-dispersive x-ray 
spectra, CSL has been assisting BEIB in developing software for resolv- 
ing the overlapping peaks which occur in these spectra by using a 
modification of the Simplex method to optimize the fit of Gaussian 
functions to the x-ray peaks. 

Only the beginning phases of this project have been dealt with to 
date. It is likely that the data acquisition and control aspects alone 
will require the contributions of two or three CSL staff members for at 
least another two years. 

Microcomputer-Based Controller for Evoked Response Experiments, NIMH, LBEB 
A DEC LSI-U microcomputer has been configured to operate as a program- 
mable 16 channel pulse generator under the control of an existing labora- 
tory computer system, a PDP 11-40. Prompted by the user, the host 

95 



computer, via special communications software, downloads the program and 
precomputed data files to the microcomputer satellite. The host 
computer is then free to proceed with data collection and reduction tasks 
while the satellite controls various aspects of the evoked response 
experiments. The satellite requires no console, operating system, 
auxiliary storage or user intervention. Any changes in the timing 
relationships of the pulses can be effected by the host computer. This 
project is complete except for final integration of the satellite into 
the experiment and final documentation. Only minimal effort is 
expected to be required during the coming year. 

Spectrophotometer Data Logger, NIDR : A microprocessor-based data logger 
was developed to record digital data from a spectrophotometer onto 
digital cassette tape. Installation of this device was accomplished in 
July 1979. The user controls the timing and quantity of the sampling, 
and also the recording of special control blocks via switches 
associated with the Data Logger. The digital cassette tapes may be 
played back to NIDR's computer by a commercially procured playback 
terminal. The format of the tapes is identical to that of cassettes 
created by other CSL data logging instruments used by NIDR, notably the 
Radiation Counter Data Recorders, thereby allowing a single cassette 
tape terminal to be used for playback of data recorded on a variety of 
CSL developed data loggers. 

In conjunction with this project, CSL has been organizing, refining and 
standardizing a software library and the hardware requirements for 
collecting data, recording tapes and interfacing to various laboratory 
instruments (see, for example, Amino Acid Analyzer Data Logger elsewhere 
in this report). 

Amino Acid Analyzer Data Logger, NCI, SURG : During the summer of 1979, 
CSL developed a data logger that records data from a Beckman amino acid 
analyzer onto digital cassette tape. The data is formatted in a manner 
consistent with other CSL data loggers and with commercial playback 
units. The Data Logger is connected between the amino acid analyzer and 
its teleprinter and is transparent to both. This instrument includes 
hardware features and software procedures that have been developed by 
CSL for other data logging applications such as the Spectrophotometer 
Data Logger described elsewhere in this report. Those hardware and 
software elements were developed with special consideration given to 
the possibility of a variety of future applications. 

Radiation Counter Data Recorder (Previously reported as CSL Data 
Recorder), DCRT, CSL : The Radiation Counter Data Recorder is a device to 
log data from a rediation counter (liquid scintillation counter, gamma 
counter, etc.) and write it onto digital magnetic cassette tape. The 
Data Recorder connects between a radiation counter and its teleprinter 
and is transparent to both devices. The format of recorded data is 
suitable for playback to either the DECsystem-10 or the IBM 370 via a 
commercially available cassette tape data terminal. Over twenty 
Radiation Counter Data Recorders have been fabricated. Installation of 
these units in laboratories of NIAID, NICHD, NIDR, NIAMDD, and NCI was 

96 



completed early this fiscal year. In anticipation of a continuing 
demand within the NIH community, a request for proposals (RFP) was 
prepared this year for outside fabrication and maintenance of the Data 
Recorders. The RFP is currently awaiting procurement action; initial 
funding will be in part by CSL and in part by the various laboratories 
scheduled to receive the initial production quantity of about 25 
instruments. Other FY79 activity on this project included the 
development of test procedures and instrumentation for the contractor 
to use in testing Data Recorders for quality assurance purposes. Some 
additional effort will be required during the coming year for monitoring 
the contract and guiding the installation of new Data Recorders. 

NIEHS Computer Facility, NIEHS : The Biometry Branch of the National 
Institute of Environmental Health Science, located in Research Triangle 
Park, North Carolina, is responsible for providing computer facilities 
to meet the needs of the Institute. These needs include scientific and 
statistical computation, simulation, laboratory data processing, and 
some business data processing. 

In FY78, CSL conducted a study of the current and anticipated future 
data processing needs of the NIEHS. Based on the recommendations 
resulting from this study, NIEHS purchased and installed the suggested 
computer system, which became operational in October 1978. CSL also 
performed an extensive survey of available telecommunication multi- 
plexors and concentrators which could be used to improve telecommuni- 
cations access to the DCRT Central Facility by NIEHS, but this solution 
was abandoned when NIEHS was offered the opportunity to share a tele- 
communication service that was to be extablished by the Bureau of 
Health Statistics. 

Bioassay Information System, NCI. DCCP; VRB, SAF; FDA, NCTR : The 
National Cancer Institute has an interagency agreement with the National 
Center for Toxicological Research for the implementation of NCTR's 
Research Support System as a computer based information management 
system for NCI's Bioassay Program. CSL has, since FY77, served as a 
consultant to NCI in evaluating various specialized hardware and soft- 
ware components of this system, and has provided consultation to NCTR 
on specifications for microcomputer-based, programmable data acquisition 
terminals. During FY79, proposals were received from prospective 
vendors for these terminals in response to a previously issued RFP, 
and are currently being evaluated. CSL is participating in the 
evaluation. Ultimately, about 600 of the terminals, representing a 
value of several million dollars, will be purchased. Over 400 will be 
used in laboratories doing carcinogenesis testing for NCI. 

Closely allied with this effort is an association between CSL and the 
Small Animal Section of DRS which began this yedir . The SAS has need for 
an information management system for its small animal breeding activities, 
and eventually for supporting long-term animal holding experiments. CSL 
is assisting DRS staff in evaluating the applicability of the NCTR 
system to this work. Based on a CSL recommendation, this project was 



97 



suspended for the latter half of CY79 pending the outcome of NCTR's 
terminal procurement and initial testing of the system for NCI. 

Laboratory of Immunology System, NIAID : This year, CSL has expanded the 
capability of a laboratory data acquisition system which we developed 
several years ago. The improved system provides simultaneous acquisition 
storage, and processing of instrument data from multiple instruments, as 
enhanced system has been completed, documented, and operational since 
December 1978; however, some minor improvements and refinements are 
expected to be made next year. 

Laboratory of Chemical Biology System, NIAMDD : At the request of Dr. C. 
B. Anfinson of NIAMDD, a peak detecting device for use with a high 
pressure liquid chromatograph was developed. A microcomputer was used to 
implement algorithms to detect peaks from the instrument's spectropho- 
tometer output, mark beginnings and ends of peaks on a chart recorder, 
and control the advance of a fraction collector so that separated 
material from the chromatographic column may be collected. The unit has 
been installed and in use since February of 1979. 

Potentiometric Titration Controller, NHLBI, IR LC : Work has continued 
on the microcomputer system for automated electrodic potentiometry. 
This year the system has been used by the Laboratory of Cell Biology, 
NHLBI to study the redox potentials of E. Coli cytochromes, resulting in 
the development of new techniques for the resolving of spectra vs. 
voltage. A new multi-channel spectrometer has been delivered and is 
presently being integrated into the system. With this new device under 
microcomputer control, the time for acquiring complete spectra will be 
reduced to milliseconds instead of seconds. This is of considerable 
importance in some experiments because of the rapid change in some 
parameters and the associated difficulty of maintaining the required 
experimental conditions. Support for this project is expected to 
continue during the coming year. 



• Biomedical Communications Projects 

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 
helped plan the 1978 Conference at Stanford University, and edit the 
Conference Proceedings. Reduced CSL involvement in the 1979 Conference 
is anticipated, but to date neither the veQue of the Conference nor the 
role of CSL have been determined. 

CSL Development System : A PDP-11/70 computer system was purchased in 
1978 to be used to support hardware and software development for similar 
systems with which CSL is involved. This support includes media 
conversion, cross-generation of operating systems for smaller target 



98 



systems, maintenance and development of hardware and software for 
systems that are unavailable due to their use in production, and 
prototype computer network software development. 

The system was delivered, installed and made operational in October 
1978. A second large capacity disk drive which had been previously 
leased for use on the DCRT DECsystem-10 was purchased and added to the 
system in March 1979. A communications multiplexor which will permit 
remote access to the system will be installed later this year, as will 
an electrostatic printer/plotter, which will be used to support an LSMM 
project concerned with techniques for presentation of mathematical 
and statistical information. 

Miscellaneous Projects : A limited number of projects engaged in by 
CSL are not susceptible to classification in the aforementioned 
categories. Four such projects are reported below: 

Library Automation Project, DRS, L : This project was initiated during 
the current reporting period, in response to a request from the NIH 
Library to investigate further automation of major library functions. 
Working closely with the Library staff, CSL has made a thorough study 
of current activities in the Library to determine those aspects of 
their operations that could benefit from automation using computers. 
We then visited several other libraries in various states of automation, 
interviewed manufacturers of turn-key library automation computer 
systems and one Bovernment designed system. From this data a report was 
produced addressing the feasibility, cost, impact and preferred 
method of automating the NIH Library using computer technology. We 
expect in the coming year to collaborate further with the NIH Library, 
to issue an RFP for the principal elements of the automation system, 
and to develop those elements which are not available commercially. 

Voice Output Terminal for the Blind : Using technology originally 
designed for medical telecommunication projects, a voice output 
terminal has been implemented for use by a blind computer programmer. 
A microcomputer combined with a speech synthesizer form the basis of a 
linkage module between a standard alphanumeric computer terminal and 
any available host computer. Transparent to the host computer, all 
data routed to the terminal is intercepted in the microcomputer and is 
converted to phonetic codes using English rules of pronunciation. Data 
edit features are offered to alter the speed and translated content of 
speech output; data review options are available to permit the 
repetition or spell-back of poorly understood text output. A proto-^ 
type terminal has been field tested by a blind programmer for over six 
months and has significantly increased his productivity. The terminal 
was exhibited at the 1979 annual meeting of the President's Committee 
on Employment of the Handicapped in Washington, D.C. Present plans 
call for the hardware development of low-cost portable units (under 
$3,000 for components) and the software design for additional voice 
output devices with more general application for the visually impaired. 



99 



In Vitro Information System, NCI, DCCP : The National Cancer Institute 
conducts an in vitro testing program under the Carcinogenesis Bioassay 
Program. In vitro and submammalian assays provide information as to the 
genetic potential of chemicals; many chemicals which are mutagenic are 
also carcinogenic. The development of a battery of in vitro assays to 
serve as a screening method for selection of candidates for carcinogenic 
testing in vivo is therefore desirable, since the latter are much more 
time-consuming and expensive. 

In the interest of improving the quality and timeliness of data from 
the In Vitro Program, CSL was asked to investigate the applicability of 
source data collection techniques to the acquisition of mutagenesis 
data. A study was performed in early 1979, leading to the finding that 
source data collection was both feasible and desirable, and recommending 
that a contract be let for a pilot program development utilizing 
programmable, microprocessor-based, intelligent terminals. 



100 



He 

•a 

01 
4J 

in 

01 

> 
t— c 
q: •— 1 
(J 
Q ^ 



oinOLnooincsjurtO ^a-oo^ 
r- ir> «j- m ro t— I — 



roco(TiLnooo<*uf>vDO.— Lno 



0) 

.o 
E 

0) 



O S- 



CSJC\JC\l<NJtX)v£>^OJC\J^O 



C\J^O*:t C\JCSJ^«:3-v£)C\JCsJC>J«:|-^£> 



m CM CM r— 



<_) 


(U OJ 


LlJ 


•■->T3 


-5 


O lO 


O 


I- (U 


a: 


D. _1 



OOOO NN NNJNO 

T3T3-OXXXXT333 333X 

>>>i>ii— ■— .— r— >,00 OOOi— 



O O O O C3 

o u u o u 



O O O O N 
(_) U (J u +J 



o o o 
u u o 



XXXXXXXXX3-OXXX 

Q.a.o-Q_Q.a-a-0-a.ooooa-ci.a_ 



o 

O i-i 



CO 






o 

z 



o 



























o 






z 


























£ 


Q- 

_J 




,_, 


Q. 

CO 






^ 


en 






























CO 


Ol 






c/) 


















SQ 








+j 


«> 




_I 


a: 






Q 


f, 
















U 








(/) 






3: 








u_ 


























>>o 




z 


■o 








ca 
















o: 








00 


Z « 






01 






oO 


— 1 




























00 




J^ 








3: 












z> 












^ 


oO tJ 


(J 


E 


o 






c^ 


z 






a. 






ZU1 




•• 








o 


z 


_J 


Ol 


> 






o 


Q 






o 






o 












s. 


o 




4-> 


UJ 


>— _1 




<_) 








o 






(J - 




CD 








+J 


Q - 


* 


CO 




o; <_) oo 




Q 


ct t. 






o 




O 


QH-i 




—1 








c 




IE 


>^ 


s- 


Q z o 






►1 01 


01 








Of 


00 




IE 








o 


< 0) 


S OOCQ 


o 






r. 


Z r- 


o 




e 






"_1 




z 








o 


"H +J 




*— I 


t- 


z - •> 




1 — 1 


f— 


c 




1—4 




«K 


(J 3: 














Z i- 


z 


10 LU 




u s- 




<-> Q 


• o 


Ol 




o 




1— 1 


<JZ 




oS 








OS 


o 




■MCQ 


s- 


. 0) 01 




z — < 


>> i- 


i- 




z 




o 
















.Crt 


E 


m 


Ol 


S- 01T3 




^< 


CT)<-> 


01 — 








z 


•• at 




<-> 








c 


01 


OJ 


Q " 




0) CD t- 






o c 


<*- T3 










+J +J 




O 








o 


Q..— 


•4-> 


c/l 


I— CO 


cn o o 




EZ 


■— o 


C C/) c 




^ 




• 


■^ 'r- 














O 1— 


CO 


oici: 


O UJ 


CJ1_I u 




OJ 


o o 


o o: •'- 




01 




>, c c 




«t 








*^ 


<J Ol 


>, 


DIQ 


i- CO 


O 01 




4J - 




tj Q .— 




4-> 




0.:= = 




01 








•1— 


00 O 


00 


3 


■M _l 


—itaoL 




CO >,CQ C 


CO 




(/) 




10 






c 


t-H 






« 


o 




M- • 


c 


•4-1 


>> 


>,I31 


o 


>> -■— 




>> 




u 


01 0) 






CO 


u 






I- • 


en 


■■- >, O - 


fO (O (O 


4-> 


OO o 




C31 C 


E OO 




0) 


s- s- 




Ul 


-J 


<_) 




3 


u >, 


c 


1- ■!-> o 3: 


■M Q 4-> 


•r- 


r— 


m +J 


o o .— 


0) 






f 


lO IT) 


o 


CO 


a: 






cr 


•.- ;_ 




+J •!- 


s 


n «o 




c o 


u <o 


r- •»- to 


4J 


c 


0) 


4-) 


O O O 


0) 


z 


* 


_l 


O ^ +J 


T3 


C 1 — 


•o —■ 


O S- Q 


•f— 


o c 


•r- 1_ 


O ■!-> C 


(/I 


o 


^— 


o 






o 




>>ln < 


<u 


I- 


01-.- 


0) z 


01 


o 




E +-> 




>, 




■M 




01 01 


•« 


o 




cn<_) 




C E 


o 


o o 


CO 


S- N S. 


(O 


+-> E 


01 •.- 




OO 


4^ 


•r— 


■o 


> > 


01 


u 


-C 


O 




«tl 


o o 


u 


(O lO 


lO - 


Ol >) OJ U- 


(O S 


-C 1— 


s» o S^ 




<o 


1— 


nj 




c 


CL. 


o 






•M 


1. i. 


OJ 


i-Ll. 


OQ CO 


4Jr- 4J 






CJI 


03 -LJ Ol 


4-1 






cc 


wi m 


•1— 




c 


"o 


CO 


lO 


■>-> o 


a: 


-M 


4-> 


0) lO c 


1- 


s- 


u 


O 3 1— 


c 


s- 


■!-> 




c c 


O 13 


rs 


4= 


■(— 


O 


O 3 




1— (fl 


S_ C 


E C 3 


Ol 


o t- 


M- — 


ct 




o 


U 


■o 


01 OJ 




(_> 


s- 


4J 


CO 




O) .— 


c 


=3t- 


01 01 


o<: o 


4~> 


l(- o 


o s- 


C ■!-> 




If- 


0) 


0) 


•tJ 4-> 


-o 


LU 


CO 


lO 


>>-o 


r— 14- 


o 


(/) 


4-> E 


*-> o 


3 


c 


4-> 


•1- >, 3 


CL 


c 


••-5 


tM 


c c 


a> 






CI. 




0) 


Ul o 




r- >, 3 -r- 


O T3 


a. 


•—>,>> 01 


1- a. o 


t—t 


o 






2: -a 


>1 




lO 


•M 


s. 




(0<— 


CL S. 


^ -r- C 




s- 


1. E 


CO (O 4-1 






i- 


I- 






0) 


s- 


r^ 


c 


3 


nji u 


■o 


U (O 


E 01 


CL O O 


Q 


>. o 


o o 


i- U 3 


01 


o 


Q. 


(U 


i— u 


i- 


+-> 


lO 


(o <; 


J3 


C -1- H-l 


o; 


•■- C 


O Q. 


o <: ■■- 


o 


lO +J 


+J ■.- 


01 JD O 


> 


i- 




4-) 


(O (O 


la 


n3 


c 


u 




•I— 


•r- Z X 


4J 


•!-> lO 


U X 


1- 4-> 




{/) (O 


<o ••-> 


4J •.- 


01 


4J 




3 


O -r- 


0) 


E 


o 


•f— 


0) 


S- 


C _1 


<J 


>>o 


O UJ 


+J O lO 00 


CO s- 


s. c 


3 _j 01 


C3 


•f— 




C3. 


■•- T3 




o 




c 


cn+J 


c 2 


<u 


r- S- 


i- 


O C -r- 


3: 


lO o 


O 01 


Q. U 




> 




E 


X> S- 


u 


4~i 


r— 


•r- 


(C 


LO 


lo o 




lO U 


o 


OJ •»- 73 


UJ 


O JD 


J3 -M 


E 3: ■.- 


_J 






o 


oj to 


3 


3 


3 




E 




O f— 


<u 


C •»- 




Q- E "O 




■1- ro 


<o o 


O •— O 00 


c 




O Z CJ Z < Q- o 




O 


CO Ll_ 


crxz ^ 


00 « o; z 


CO —1 


_r Q. 


<J> z > o 





lO 
3 

(/) "O 

01 •>- 

<n > 

10 •>- 

^ -a 

u c 

s- T- 

3 

CL O 
4-) 

a 
4J -a 

01 

01 en 

3 S- 

•a ro 

..= 

•— o 

(0 
+J 4-1 

o o 



en 


01 






1- 




-C 


lO 




4-1 


-C 




cn 


1.1 




T3 






01 


c- 




01 


t 




U 






X 


Wl 




01 


s 




4-1 


4-1 




01 


• r— 


> 


CT 




v> 


TJ 


4-> 


4-) 


-1 


I- 


tJ 


CO 


O 


01 




D. 


••-o 


_J 


a. 


o 


1/1 


-I 


i- 


<-J 


CO 


o. 



101 



III. PUBLICATIONS AND PRESENTATIONS LIST 

Bacharach, S., Ostrow, H., Green, M., Borer, J., and Johnston, G. 
Distributed Processing in Medical Imaging, Proceedings of the International 
Conference on Processing in Medical Imaging , Paris, France, July 1979. 

Bacharach, S., Green, M., Borer, J., Ostrow, H., and Johnston, G. 
A Computer System for Clinical Nuclear Cardiology. Proceedings of the 
Second Annual Symposium on Computer Application in Medical Care . IEEE 
Computer Society, Long Beach, California, November 5-9, 1978, pp. 50-55. 

Barden, L., Tate, R., Schultz, A., Markey, S., and Oliver, J.: 
Selected Ion Recording Data System. Proceedings of the American Society 
for Mass Spectrometry 27th Annual Conference on Mass Spectrometry and 
Allied Topics , Seattle, Washington, June 3-8, 1979 (in press). 

Douglas, M. , Green, M., and Ostrow, H.: Evaluation of Automati- 
cally Generated Left Ventricular Regions of Interest in Computerized ECG- 
Gated Radionuclide Angiocardiography. Computers in Cardiology , Stanford, 
California, September 12-14, 1978, pp. 201-204. 

Green, M. , and Ostrow, H.: A Comparison of Simultaneous 
Measurement of Systolic Function in the Baboon by Electromagnetic Flow- 
meter and High Frame Rate ECG-Gated Blood Pool Scintigraphy. Circulation , 



(in 



August 1979 (in press) 

Green, M., Brody, W., Douglas, M., Borer, J., Ostrow, H., Line, 
B., Bacharach, S., and Johnston, G.: Ejection Fraction by Count Rate from 
Gated Images. Journal of Nuclear Medicine , Vol. 19, No. 8, August 1978, 
pp. 880-883. 

Martino, R. , and Risso, W., Jr.: An Arterial Blood Pressure Pre- 
processor Using a Combined Analog and Digital Signal Processing Method. 
Proceedings of the Seventh New England (Northeast) Bioengineering Conference , 
March 1979, pp. 267-270. 

Ostrow, H., Allen, S., Bacharach, S., Green, M., and Borer, J.: 
A Gated Probe System for Real-time Creation of Pressure Volume Loops. 
Journal of Nuclear Medicine , Vol. 20, June 1979, p. 610. 

Plexico, P., and Kempner, K.: Microcomputers in Medicine. 
Infotech International, Ltd., Maiden, Berkshire, England (in press). 



102 




since FY75, CSL has provided engineering , system 
integration, and software support to meet the 
data acquisition, data display, and analysis needs 
of NIH investigators using Flow Micro fluorometers . 
Evaluating some enhancements are Donald Jansen, 
John Powell, and Arthur Schultz. 



103 



PHYSICAL SCIENCES LABORATORY 



George H. Weiss 
Chief 



« XJ 





jyp. James Fevvetti works on developing new methods of 
nualear magnetic resonance spectroscopy (NMB) and on 
applying NMR to the study of small proteins. In the 
past year, experimental studies have heen conducted 
on the interaction of iron with hleomycin, an 
important anticancer compound. 



I . SUMMARY 
Function 

The Physical Sciences Laboratory has three principal functions: 

* to carry out research in the physical sciences in order to under- 
stand biological phenomena in terms of physics and chemistry 

to develop the 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 six 
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. 



• 



t 



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. These collaborative 
projects are done jointly with approximately twenty five other 
investigators including two major projects with data being generated 
by off-campus scientists. 



Highlights of the Year's Activities 

In the past year progress was made in most of the PSL continuing projects, 
and a start was made in two new areas. 

Dr. Stephen Brenner has begun an experimental study of the interaction 
spectrum with actin. This work is carried out in the laboratory of 
Dr. E. Korn, NHLBI. So far the work has been devoted to purifying sheep 
erythrocyte spectrin and rabbit skeletal muscle actin, and studying 
their interaction. Spectrin is thought to play a dominant role in 
determining the shape of red blood cells. 



106 



A second new project is one on the quantitative analysis of electron- 
micrographs initiated by a newcomer to the PSL, Dr. Nachum Gershon. 
Electron micrographs of different systems of biological interest are 
being furnished by collaborators both on and off the NIH grounds. The 
systems include viral ly infected cancer cells and insulin receptors 
on rat adipocytes. 

Dr. Parsegian continued a systematic study of forces between phospholipid 
bi layer membranes immersed in solution together with experimental 
collaborators at Brock University, led by Professor Peter Rand. He 
discovered a new class of forces that appears to dominate the interaction 
of biological membranes at distances of less than 30 A. These forces 
to the work of removing water from cell surfaces. A related part of 
this project is making mechanical measurements of the deformability 
of membranes. Preliminary results in this area differ from those that 
one would expect from studies of phospholipid monolayers. Dr. Gingell 
of the Middlesex Medical School, who has been collaborating with Dr. 
Parsegian on studies of forces in biological structures, has shown that 
cells can be held to surfaces by long-range electromagnetic forces which 
act at thousands of Angstroms distance. 

Dr. Parsegian 's project has led to the development of powerful new 
theoretical and experimental tools that allow one to study physical 
forces at distances of the order of tens of angstroms. There is considerable 
potential for studying different biological systems with these tools, 
such as protein aggregation, interactions within hemoglobin, and gelation. 
These projects await additional personnel to implement presently available 
methods. 

Dr. Nossal continued to develop theoretical and experimental techni- 
ques for using laser systems to measure different parameters in biologi- 
cally interesting materials. The most recent work in this field has 
been in the measurement of elastic coefficient in gels. Much of this 
work has been performed with Dr. R. Gelman, NIDR. Dr. Nossal, together 
with Dr. J. Gladner, NIAMDD is completing a study of fibrin gels to 
find a relation between the mechanical strength of blood clots and the 
interchain crossl inking of the constituent proteins. 

Dr. Weiss continued working with Dr. W. F. Caveness, NINCDS, on the 
analysis of a large data base of medical data on head injured veterans 
of Vietnam. Several studies of this data have been completed. These 
include the occurrence of post traumatic epilepsy, the consequences 
of using different materials for cranioplasties, and the origin and 
persistence of aphasia. Different injury parameters could be related 
to the onset of aphasia, but no reliable prediction could be made of 
the disappearance of aphasia based on what is known about the injury. 

Dr. Weiss was appointed to the Board of Editors of the Journal of 
Statistical Physics 



107 



Future Plans 

All of the research projects of the laboratory will continue. We see 
no change in direction or emphasis within the coming year. 

The experiments on forces operating in membranes will continue, and 
there is a possibility that a new series of measurements will be made 
on forces between proteins. 

A study of the physical properties of clathrin baskets will be under- 
taken by Dr. Nossal. This investigation will help to elucidate the 
structure of these biologically important molecules as well as the 
measuring hydrodynamic parameters. 

Investigators at the University of Connecticut will undertake experiments 
on the acceleration of sedimentation equilibrium experiments. The theory 
for this technique was developed by Dr. Weiss in collaboration with 
Professor Yphantis at the University of Connecticut. 




The Physical Soienaes Laboratory has a oonibined program 
of research projects intenmal to the laboratory and 
collaborative projects with scientists at NIH and at 
other institutions. Here visiting Israeli scientist 
Br. Nahum Gershon (left) and PSL Chief Dr. George Weiss 
(right) discuss a study with James Kiefer (seated). 



108 



II. PSL PROJECTS AND ACTIVITIES FY 79 



Theory of Biochemical Separation Techniques . George H. Weiss, PSL, 
in collaboration with Professor D. A. Yphaantis, University of Connecti- 
cut. This is a joint theoretical and experimental project to develop 
techniques for interpreting measurements made by such methods as ultra- 
centrifugation and electrophoresis. There was little activity on this 
project in the last year except for some tests of different methods 
of inverting sedimentation equilibrium data to measure molecular weight 
distributions. 

Theory and Measurement of Intermolecul ar Forces . V. Adrian Parsegian, 
PSL, in collaboration with Professor P. Rand, Brock University and Dr. 
D. Gingell, University of London. This project consists of laboratory 
investigations based on theoretical techniques developed by Dr. Parsegian 
for measuring forces that operate at short distances. In the past year 
the project participants have discovered and measured a new force in 
the interaction of phospholipid membranes at distances of 30A or less. 
This force has been attributed to the work required to squeeze water 
molecules out from between the membranes. 

Consulting Services . George H. Weiss, PSL. This project includes 
elements of epidemiology and the development of mathematical techniques 
applied to specific problems of NIH researchers. A considerable amount 
of time was spent on several studies of a large data base on head-injured 
Vietnam veterans related to the occurrence of posttraumatic epilepsy, 
aphasia, the treatment of severe head injuries, and the use of different 
cranioplasty materials. A study was completed on interpolation in com- 
puterized tomography and a new collaboration was begun on the develop- 
ment of a theory to interpret experiments on DNA fractionation. 

Correlation Function Spectroscopy/Laser Light Scattering . Ralph J. 
Nossal, PSL, in collaboration with Drs. R. Gelman, NIDR, R. Bonner, 
DRS, and J. Gladner, NIAMDD. This project comprises laboratory investi- 
gations and development of theory for interpretation of experiments. 
In the past year Dr. Nossal developed a theory to enable one to measure 
elastic coefficients of soft biological gels. These were used in a 
study of elastic moduli in polyacrylamide gels. 

Cell Motility and Chemotaxis . Ralph J. Nossal, PSL, in collaboration 
with Dr. L. Lipkin, NCI. This project is a combination of laboratory 
experiments and the development of theoretical models to elucidate 
factors that influence cell locomotion and chemotaxis. Experiments 
are currently in progress on the response of neutrophils to different 
chemical factors produced by stimulated lymphocytes. Algorithms and 
computer programs have been developed to automate cell tracking experi- 
ments. 

Theory and Application of Nuclear Magnetic Resonance Spectroscopy . 

Dr. James A. Ferretti, PSL, in collaboration with Drs. E. Becker, NIAMDD, 

R. Highet, NHLBI, G. Weiss, PSL and G. Marshall, Washington University. 



109 



This project consists of laboratory experiments and concurrent develop- 
ment of theory for the application of NMR to elucidate the structure 
and function of biologically interesting chemicals. In the past year 
experimental studies have been conducted on the interaction of iron 
with bleomycin, an important anticancer compound, and on bradykinin 
and gramicidin S. Similar investigations are continuing. Further 
analyses of the techniques of NMR are being carried out with a view 
to measuring spin-lattice relaxation times as accurately as possible 
in a fixed time. 

Interactions of Erythrocyte Spectrin with Actin . Stephen L. Brenner, 
PSL,in collaboration with Dr. E. Korn, NHLBI. This is a laboratory 
investigation of the properties of a molecule believed to be important 
in determining the shape of red blood cells. Different forms of spectrin 
exist. This year was devoted to finding the elastic properties of these 
variant forms. 

Quantitative Analysis of Electronmicrographs and Membrane Cellular 
Biophysics^ Nachum Gershon, PSL, in collaboration with Dr. B. Bowers, 
NHLBI, L. Jarett, and R. Smith, Washington University. This project 
is a methological one using data gathered in different laboratories 
to determine conf igurational properties of proteins on cell surfaces. 
In the past year several studies were initiated, computer programs were 
written, and algorithms developed. An example of a project now under 
way is a joint study to determine whether the surface of phagocytosing 
cells has a different protein content than does its interior. 

Studies in Mathematics and Statistics . George H. Weiss, PSL, James 
E. Kiefer, PSL, in collaboration with R. J. Rubin, National Bureau of 
Standards and I. G. Darvey, Sydney University. This project includes 
several unrelated investigations in numerical analysis and applications 
of mathematical problems in chemistry. A study was completed on 
determining the best technique for accelerating the convergence of 
slowly convergent Fourier series. Further investigations were made 
of the properties of random walks relevant to the description of polymer 
configurations. Some work was done on optimal design of enzyme kinetic 
experiments. 



110 



III. PUBLICATIONS 



Becker, E. D., Ferretti, J. A., Gupta, R. K., and Weiss, G. H.: The 
choice of optimal parameters for measurement of spin lattice relaxation 
times. II. Comparison of saturation-recovery, inversion-recovery, 
and fast inversion-recovery experiments. Jl. Magn . Resonance (in press). 

Becker, E. D., Ferretti, J. A. and Gambhir, P. N.: Selection of opti- 
mum parameters for pulse Fourier transform NMR. Analytical Chemistry 
(in press). 

Benovic, J. L., Ferretti, J. A. and Gupta, R. K.: Carbon-13 NMR and 
metal binding properties of Bleomycin. Biophysical J. 21: 199a, 1978. 

Blumenfeld, D. E., and Weiss, G. H.: Curve fitting the probability 
distribution of acoustic noise form freely flowing traffic. Transp . 
Res. 12: 111-114, 1978. 

Blumenfeld, D. E. and Weiss, G. H.: Statistics of delay for a popu- 
lation of drivers with step and distributed gap acceptance functions. • 
Transp . Res . 12: 423-429, 1978. 

Blumenfeld, D. E. and Weiss, G. H.: The effects of gap acceptance 
criteria on merging delay and capacity at an uncontrolled intersection. 
Traf . Cont . & Engin . 20: 1-5, 1979. 

Bonner, R. F., Bowen, P., Bowman, R. C. and Nossal, R.: Real-time 
monitoring of tissue blood flow by laser Doppler velocimetry, "Pro - 
ceedings Electro-Optics/Laser '78 Conference" , pp. 539-550, Industrial 
and Scientific Conference Management, Inc. Chicago, Illinois, 1978. 

Brenner, S. L., Parsegian, V. A., and Gingell, D.: The effects of 
image forces on double-layer interactions. J. Phys. Chem. 82: 1727- 
1731, 1978. 

Brooks, R. A., Weiss, G. H., and Talbert, A. J.: A new approach to 
interpolation in computed tomography, J.. Comp . Tomog . 2: 577-585, 
1978. 

Bunow, B., Line, B. R., Morton, M. R., and Weiss, G. H.: Regional 
ventilatory-clearance by xenon scintigraphy: A critical evaluation 
of two estimation procedures. J.- Nucl. Med , (to appear). 

Caveness, W. F., Meirowsky, A. M., Rish, B. L., Mohr, J. P., Kistler, 
J. P., Dillon, J. D., and Weiss, G. H.: The nature of post traumatic 
epilepsy. Arch. Neurol . 50, 545-553, 1979. 

Chew, E., Weiss, G. H., Brooks, R. A., DiChiro, G.: Effect of CT noise 
on detectability of test objects. Am. Jl^ Roentgen . 13: 681-685, 1978. 



Ill 



Cowley, S., Fuller, N., Rand, R. P. and Parsegian, V. A.: Measurement 
of repulsion between charged phospholipid bilayers. Biochemistry 
17: 3163-3168, 1978. 

Dishon, M., Weiss, G. H.: Numerical inversion of Mellin and two-sided 
Laplace transforms. J. Comp . Phys . 28: 129-132, 1978. 

Ferretti, J. A. and Marshall, G. R.: Carbon-13 properties of Angioten- 
sin-II. Biophysical J. 21: 79a, 1978. 

Ferretti, J. A., Marshall, G. R., and Gupta, R. K.: Carbon-13 relax- 
ation parameters of small peptides, "Proceedings of the 20th Collogque 
AMPERE, Tallinn (1978)," p. XXX, North-Holland, Amsterdam. 

Ferretti, J. A. and Marshall, G. R.: Field dependent carbon-13 relax- 
ation studies on peptides. I. Aggregation of Angiotension-II, Biophysi- 
cal ^. (in press). 

Gail, M. H., Weiss, G. H., Mantel, M., and O'Brein, S. J.: A solution 
to the generalized birthday problem with application to allozyme 
screening for cell culture contamination. J^. Appl . Prob. (to appear). 

Gelman, R. A. and Nossal, R.: Laser light scattering from mechanically 
excited gels. Macromolecules 12: 311-316, 1979. 

Gelman, R. A., Gladner, J. A. and Nossal, R.: The rigidity of fibrin 
gels as measured by quasielastic light scattering. Biopolymers (to 
appear). 

Gershon, N.: On the effect of crossl inking on the attachment of membrane 
proteins to the cytoskeleton. Cell Surface Events in Cellular Regulation, 
Elsevier-North Holland, (to appear). 

Gershon, N., Dempsey, A., and Stackpole, C: Analysis of local order 
in the spatial distribution of cell surface molecular assemblies. 
Exp . Cell Res ., (to appear). 

Gupta, R. K., Becker, E. D., Ferretti, J. A., and Weiss, G. H.: A 
variable perturbation method for nuclear spin-lattice relaxation measure- 
ments. jLl Magn . Resonance (in press). 

Gupta, R. K., Ferretti, J. A., and Caspary, W. J.: Carbon-13 NMR 
studies of the structure of the iron-Bleomycin complex. Biophysical 
J. 25: 236a, 1979. 

Kiefer, J. E., Parsegian, V. A. and Weiss, G. H.: Some convenient 
bounds and approximations for the many body van der Waals attraction 
between spheres. J. Coll. Int. Sci. 67: 141-153, 1978. 



112 



Lis, L. J., Rand^^R. P. aiji^ Parsegian, V. A.: Measurement of the 
adsorption of Ca and Mg to phosphatidyl cholines bi layers in 
Bioelectrochemistry : Ions , Surfaces and Membranes . American Chemical 
Society, Washington, D, C, Ed. M. Blank, 1979. 

Meirowsky, A. M., Caveness, W. F., Rish, B. L., Dillon, J. D., Mohr, 
J. P., Kistler, J. P., and Weiss, G. H.: Definitive care of cerebral 
missile injuries crossing the midline. Mil . Med , (to appear). 

Nossal, R.: A theory of quasielastic laser light scattering by polymer 
gels. J. Appl . Phys . 50: 3105-3112, 1979. 

Parsegian, V. A., Fuller, N. and Rand, R. P.: Measured work of deform- 
ation and repulsion of lecithin bi layers. Proc. Nat. Acad. Sci. 
(USA) (in press). 

Parsegian, V. A.: Competitive forces influencing the interaction of 
biological cells. Proceedings of the Feb. 1979 Workshop on Interfacial 
Phenomena, University of Washington, Seattle, Ed. J. Berg (in press). 

Parsegian, V. A. and Gingell, D.: Determination of the repulsive 
force sufficient to prevent adhesion of red cells to planar surfaces. 
Experimental Cell Research (in press). 

Rish, B. L., Dillon, J. D., Meirowsky, A. M., Rish, B. L., Mohr, J. 
P., Kistler, J. P., and Weiss, G. H.: Cranioplasty: A review of 
1030 cases of penetrating head injuries. Neurol , (to appear). 

Weiss, G. H., Gupta, R. K., Ferretti, J. A., and Becker, E. D.: The 
choice of optimal parameters for measurement of spin lattice relaxation 
times. I. Mathematical formulation. J. Magn . Resonance (in press). 

Weiss, G. H.: Transport equations with quadratic nonlinearities. 
Sep. Sci . 14: 243-246, 1979. 

Weiss, G. H.: A note on occupation times of random walks. J. Stat. 
Phys. (to appear). 

Weiss, G. H., Rubin, R. J.: Internal coordinate of span-constrained 
random walks. J. Stat. Phys. (to appear). 



113 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space] 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

Z01 CT 00014-12 PSL 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Theory of Biochemical Separation Techniques 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: G. H. Weiss, Chief, Physical Sciences Laboratory 

Other: D. A. Yphantis, Professor of Biology, University of Connecticut 



COOPERATING UNITS (if any) 

None 



lab/branch 



Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 

Division of Computer Research & Technology, NIH, Bethesda, MP 



TOTAL MANYEARS: 



iLl 



PROFESSIONAL: 



_aj_ 



OTHER: 



JUL 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

Q (al) MINORS D (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



[H'(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

The project explores the use of different mathematical methods to aid 
the interpretation of Biochemical separation experiments by such techniques 
as ultracentrifuqation and electrophoresis . Some exploratory work was 
done on the use of reqularization techniques for deriving molecular weight 
information from sedimentation equilibrium experiments on inhomoqeneous 
proteins . 



PHS-6040 
(Rev. 10-76) 



114 



Only a small effort was expended on this project in the past year. Our 
efforts to apply regularization methods to the derivation of the moments 
of molecular weight distributions seemed to indicate that only the lowest 
moment can be obtained with any degree of confidence. 

Keyword Descriptors : Ultracentrifugation, equilibrium sedimentation, 
regularization techniques. 

Publications : 

Weiss, G. H.: Transport equations with quadratic nonlinearities. Sep . 
Sci. 14, 243-246, 1979. 



115 



SMITHSONIAN SCI 
PROJECT NUMBER 



ENCE 
(Oo 



NFORMATION EXCHANGE 
NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00026-04 PSL 



PERIOD COVERED 

October 1, 1978 to Septpmbpr 30, 1979 

TITLE OF PROJECT (80 characters or leA) 



Theory and Measurement of Intermolecular Forces 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 

PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: V. A. Parsegian, PSL.DCRT 

G. H. Weiss, PSL,DCRT 

D. 0. Tinker, University of Toronto 

J. E. Kiefer, PSL,DCRT 

Others: R. P. Rand, Brock University 

L. Lis, Brock University 

S. Cowley, Brock University 

M. McAlister, Brock University 

N. Fuller, Brock University 

D. Gingell, Middlesex Medical School.. . 



COOPERATING UNITS (if any) 



None 



lab/branch 



Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 



Division of Computer Research & Tec h nolofly, NIH, B e thesda, MB. 

MANYEARS: '^ PROFESSIONAL: WHER: 



TOTAL 



-2,^ 



-0^ 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



Jg (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project aims to understand the role of intermolecular forces 
in biological phenomena. A major topic has been the measurement 
of forces between phospholipid bi layer membranes immersed in 
water. We are also measuring intermolecular forces between 
lipids in the same membrane. 

We have discovered a new class of "hydration" forces that appear 
to dominate the interaction of biological membranes at distances 
less than 30a. These forces are independent of electrical charge 
on the membrane surface and drop off exponentially with a decay 
constant about the dimension of a water molecule. It is these 
forces, reflecting work of removing water from the cell surface, 
that seem to control physical contact between membranes. 

We have measured the deformability of membranes and succeeded 

in forcing aggregates of molecules to go through phase transitions 

while measuring the work of creating the transition. 



PHS-6040 
(Rev. 10-76) 



116 



Theory and Measurement of Intermolecular Forces 

With Professor Peter Rand of Brock University we continue to 

make measurements of forces between phospholipid bi layer membranes. 

This includes several estimates of the van der Waals force between 

bodies in water. The results published so far are a series 

of systematic studies on the physical properties of cell membrane 

lipids. We hope to extend these then to direct observations 

of forces between natural cell membranes. We have been able also 

to measure forces between molecules within the same membrane. 

This finding has opened up new means to determine the mechanical 

properties of artificial and natural membranes. 

It is clear now that, upon their close approaches, the dominant 
force between membranes that are stable in water is due to the 
work of removing water from between their approaching surfaces. 
This forces dies off exponentially and is independent of electric 
charge on the membrane surface. In natural systems it poses 
a final and imposing barrier to contact between different membranes. 

We now have mechanical measurements of the deformability of 
phospholipid bi layer membranes. The results differ markedly 
from what is expected from the study of phospholipid monolayers 
and will provide critical information for testing models of membrane 
stability and structural transition. 

In one particular instance we have observed the adsorption of 
charged particles (ions) to membrane surfaces by measuring the 
effect of that adsorption on electrical forces between membranes. 
The force measurement then becomes a measure of the reactivity 
of ion with surface as it depends on electrical potential of 
the reacting surfaces, distance between membranes, and the ionic 
conditions of the bathing medium. Alkaline earth ions bind 
strongly to all phospholipid membranes with a specificity that 
depends on ion type as well as packing of the phospholipid 
molecules. 

With Dr. David Gingell of London, England, we have been studying 
the interactions of red cells with each other and with artificial 
materials. Guided by the theory of forces as developed in this 
laboratory, we have devised and performed experiments demonstrating 
that cells can be held to surfaces by long-range electromagnetic 
forces, which have now been shown to act at thousands of Angstroms 
separation. As a result of these studies we now suggest that 
long-range association is not a useful concept for studying contact 
between cell surfaces in physiological saline since the predicted 
minimum is less than the likely distance of protrusion of glycoproteins 
from the cell surface. The weakness of long-range forces far from cell 
contact is such as to render them unable to confer mechanical 
stability at long distances. 



117 



Publications 

Cowley, S., Fuller, N., Rand, R. P. and Parsegian, V. A.: 
Measurement of repulsion between charged phospholipid bilayers. 
Biochemistry 17, 3163-3168, 1978. 

Kiefer, J. E., Parsegian, V. A. and Weiss, G. H.: Some convenient 
bounds and approximations for the many body van der Waals attraction 
between spheres. J. Coll. Int. Sci. 67, 141-153, 1978. 

Brenner, S. L., Parsegian, V. A., and Gingell, D.: The effects 
of image forces on double-layer interactions. J. Phys. Chem. 
82, 1727-1731, 1978. 

Lis, L. J., Rand, R. P. and P|»j;segian, V. A.: Measurement of 
the adsorption of Ca and Mg to phosphatidyl cholines bilayers 
in Bioelectrochemistry : Ions , Surfaces and Membranes . American 
Chemical Society, Washington, D. C, Ed. M. Blank, 1979. 

Parsegian, V. A., Fuller, N. and Rand R. P.: Measured work 
of deformation and repulsion of lecithin bilayers. Proc. Nat. 
Acad. Sci. (USA) (in press). 

Parsegian, V. A.: Competitive forces influencing the interaction 
of biological cells. Proceedings of the Feb. 1979 Workshop 
on Interfacial Phenomena, University of Washington, Seattle, 
Ed. J. Berg (in press). 

Parsegian, V. A. and Gingell, D.: Determination of the repulsive force 
sufficient to prevent adhesion of red cells to planar surfaces. 
Experimental Cell Research (in press). 



118 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00022-12 PSL 



'BcWer 1, 1978 to September 30, 1979 



PERIOD COVERED 



TITLE OF PROJECT (80 characters or less) 

Consulting Services 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



G. H. Weiss, Chief, Physical Sciences Laboratory, PSL,DCRT 



COOPERATING UNITS (if any) 

W. F. Caveness, M.D., Chief, LEN, NINCDS, R. A. Brooks, Ph.D., SN, NINCDS 

I. M. Chaiken, Ph.D., LCB, NIAMDD, K. W. Kohn, Ph.D., LMPH, NCI 

B. W. Bunnw, Ph.D., LAS, DCRT 



lab/branch 



Physical Sripnrp«; I ahnrat.nry 



SECTION 



INSTITUTE AND LOCATION 



Division nf Cnmpiiter Rpsparrh R Tprhnnln 



ft 



y, Rpthesda, MD PflPOR 



TOTAL MANYEARS: 



n.8 



PROFESSIONAL: 



SLl. 



THER: 



JLX 



CHECK APPROPRIATE BOX(ES) 

□ (a) HUMAN SUBJECTS 

□ (al) MINORS □ (a2) INTERVIEWS 



n (t>) HUMAN TISSUES 



JS (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Members of the PSL provide consulting services to scientists and physicians 
at NIH in different areas of applied mathemtics and the physical sciences. 
A large segment of this effort is devoted to a study of the effects of head 
injury in veterans of Vietnam.. We have continued our study of interpolation 
error in computerized tomography . We have examined experimental data and 
suggested further experiments with af f i n i ty chromatography to discover the 
causes of discrepancies from theoretical predictions. We have worked on 
combinatorial problems that arise from a method of fractionating DNA by 
alkaline elution. 



PHS-6040 
(Rev. 10-76) 



119 



Consulting Services 

The PSL has provided computer services and statistical assistance to 
Dr. W. Caveness, NINCDS for the analysis of data on 1220 head injured 
Vietnam veterans. In particular a study of post traumatic epilepsy 
in these veterans showed that the rate of occurrence of this sequela 
was approximately the same as that for veterans of world wars 1 and 
2, and the Korean war. This occurred in spite of advances in treatment 
particularly the prophylactic use of anticonvulsants. Furthermore, 
the pattern of onset agrees with the historical data. A second study 
of the data related to the use and effectiveness of different cranioplasty 
materials. We found that the risk of complications was significantly 
reduced if cranioplasty was delayed for at least a year after the initial 
operation. Furthermore no significant differences were detected that 
could be attributed to the different materials used in cranio-plasties. 
A study of the origin and persistence of aphasia as well as correlation 
with motor deficits is presently being completed. We have found several 
factors that can be correlated with the onset of aphasia, but none have 
so far appeared that can be associated with its disappearance. 

A study with R. A. Brooks, NINCDS on discretization and interpolation 
errors in computerized tomography is nearing completion. We have focussed 
attention on obtaining an exact reproduction of a reconstructed delta 
function using Fourier techniques that we developed earlier. Using 
these techniques we have been able to demonstrate the origin of Moire 
patterns found in earlier empirical reconstructions. 

We have investigated, in some depth, the theory underlying affinity 
chromatography for Dr. I. M. Chaiken, NIAMDD, to try to explain some 
discrepancies between his experimental data and the theory. A tentative 
physical explanation has been found, but it must be examined further 
experimentally. Recently we have started to work on several combinatorial 
problems for Dr. K. W. Kohn, NCI, required for the design of experiments 
of fractionation of DNA strands by alkaline elution. 

Together with several members of LAS, DCRT, we have developed a theory 
of noise in lung scanning. The theory allows one to choose experimental 
parameters that lead to small values of bias and variance. 

Keyword Descriptors: Head injuries, post traumatic epilepsy, 
cranioplasty, aphasia, interpolation errors, discretization errors, 
computerized tomography. Moire patterns, affinity Chromatography, 
fractionation, lung scanning. 



120 



Publications 

Brooks, R. A., Weiss, G. H., Talbert, A. J.: A new approach to 
interpolation in computed tomography, jJ. Comp . Tomog . ^2, 577-585, 
1978. 

Chew, E., Weiss, G. H., Brooks, R. A., DiChiro, G.: Effect of CT noise 
on detectability of test objects. Am. J. Roentgen . 13, 681-685, 1978. 

Gail, M. H., Weiss, G. H., Mantel, M., O'Brein, S. J.: 
A solution to the generalized birthday problem with application to 
allozyme screening for cell culture contamination. J^. Appl . Prob . 
(to appear). 

Bunow, B., Line, B. R., Morton, M. R., Weiss, G. H.: Regional 
ventilatory-clearance by xenon scintigraphy: A critical evaluation 
of two estimation procedures. J^. Nucl. Med , (to appear). 

Meirowsky, A. M., Caveness, W. F., Rish, B. L., Dillon, J. D., Mohr, 
J. P., Kistler, J. P., Weiss, G. H.: Definitive care of cerebral 
missile injuries crossing the midline. Mil . Med. (to appear). 

Caveness, W. F., Meirowsky, A. M., Rish, B. L., Mohr, J. P., Kistler, 
J. P., Dillon, J. D., Weiss, G. H.: The nature of post tranumatic 
epilepsy. Arch. Neurol . 50, 545-533 (1979). 

Rish, B. L., Dillon, J. D., Meirowsky, A. M., Rish, B. L., Mohr, 

J. P., Kistler, J. P., Weiss, G. H.: Cranioplasty: A review of 1030 

cases of penetrating head injuries. Neurol , (to appear). 



121 



•SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do MOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAl RESEARCH PROJECT 



PERIOD CaVERED 



PROJECT NUMBER 

ZOl CT 00021-08 PSL 



er 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 chtracteps or less) 

Correlation Function Spectroscopy/Laser Light Scatteri 



ng 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



R. J. Nossal, Ph.D., Research Physicist, PSL, 



DCRT 



COOPERATING UNITS (if any) 

J. Gladner, Ph.D, 
R. Bonner, Ph.D.. 
R. Gelman, Ph.D., 
H. T. Pretorius, 



.Laboratory of Biophysical Chemistry, NIAMDD 
Biomedical Engineering & Instrumentation Branch 
Laboratory of Biochemistry, NIDR ' 

M.D., Clinical Endocrinology Branch, NIAMDD 



DRS 



lab/branxh 

Physical Sciences Laboratory 



SECT I 01. 



INSTITUTE AND LOCATION 

Division of Computer Research & Tprhnni 



aicx 



\^h:^^^* BothoGda, Md 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 

0.8 



-0^ 



CHECK APPROPRIATE BOx(£S) 

□ (a) HUMAN SUBJECTS 

□ (al) MINORS D (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



K (c) NEITHER 



SUMMARY OF WORK (200 .orts or less - underline keywords) 



Experimental and theoretical studies have been performed to 
J^^f '°P l aser inelastic light scattering methods for studying 
biological ^ejs and other materials. The technique is being 



ique 
Studies 



used to examine the strength of fibrin clots. 

are being performed in order to understand how laser Doppler 

techniques can be used to measure capillary blo^J TTow in tissues 

"baskp'^.'^r'^r °' I'' structure ^fb?iiT^ti?iUi d^veS ciathi ' 
baskets" has been started. 



PHS-6040 
(Rev. 10-76) 



122 



Correlation Function Spectroscopy/Laser Light Scattering 

The primary objective of this project is development of laser 
inelastic light scattering techniques for performing rapid and 
precise measurements on biological systems and materials. In 
principle, any process giving rise to refractive index fluctuations 
can be monitored. The instrument which we have constructed 
has been used to measure diffusion coefficients of macromolecules, 
swimming speed distributions of motile microorganisms, elastic 
moduli of gels and blood flow in capillaries. 

During the past year we improved a scheme for measuring elastic 
coefficients of dilute polymer networks and soft biological gels 
developed by us in the past year. A new theory was developed to 
account for internal energy dissipation by the polymer lattice. Several 
collaborative studies utilizing this new technique were undertaken. 
Experiments on polyacrylamide gel models were performed, and the manner 
in which elastic moduli depend upon parameters such as temperature and 
concentration was determined (with R. Gelman, NIDR). Fibrin gels were 
studied in order to obtain insight into the relationship between the 
mechanical strength of blood clots and the nature and extent of interchain 
crossl inking of protein constituents (with J. Gladner, NIAMDD). 

Experiments also have been undertaken to examine how laser Doppler 
techniques can be used to measure tissue blood flow. Dr. R. Bonner 
(BEIB) is developing a laser Doppler flowmeter for clinical use, 
and we have collaborated on experiments utilizing synthetic flow 
models to determine relationships between measured spectra and 
such variables as blood density, flow rate, and back-scatter 
illumination from surrounding tissue. A mathematical theory for 
data interpretation has been derived. 

Keyword Descriptors: 

Laser light scattering, macromolecules, diffusion coefficients, 
correlation functions, gels, blood flow. 

Publications: 



Gelman, R. A. and Nossal, R.: Laser light scattering from 
mechanically excited gels. Macromolecules 12 , 311-316, 1979. 

Nossal, R.: A theory of quasielastic laser light scattering by 
polymer gels. J. Appl . Phys . 500 , 3105-3112, 1979. 



123 



Bonner, R. F., Bowen, P., Bowman, R. C. and Nossal, R.: Real-time 
monitoring of tissue blood flow by laser Doppler velocimetry. 
Proceedings Electro-Optics/Laser '78 Conference , pp 539-550, 
Industrial and Scientific Conference Management, Inc. Chicago, 
Illinois, 1978. 

Gelman, R. A., Gladner, J. A. and Nossal, R.: The rigidity of fibrin 
gels as measured by quasielastic light scattering. Biopolymers 
(to appear). 



124 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00017-07 PSL 



PERIOD COVERED 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Cell Motility and Chemotaxis 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI; 



R. J. Nossal, Research Physicist, PSL,DCRT 



COOPERATING UNITS (if any) 

L. Lipkin, M.D., Image Processing Unit, DCBD, NCI 



lab/branch 

Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 

Division of Computer Research & Technology, NIH, Bethesda, MP 



TOTAL MANYEARS: 

0.2 



PROFESSIONAL: 

0.2 



OTHER: 







CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



(b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project has been undertaken to study various aspects of 

cell locomotion and chemotaxis . Analytical expressions to quantitate 

capillary migration (MIF) assays have been derived. New procedures 

for measuring macroscopic coefficients of cell migration are 

being developed, including computer assisted tracking techniques. 

Studies of the manner in which lymphokines affect the migration 

of individual leukocytes are in progress. 



PHS-6040 
(Rev. 10-76) 



125 



Cell Motility and Chemotaxis 

This study relates to cell locomotion and chemotaxis. Recent 
emphasis has been on examining certain immunologic aspects of 
leukocyte migration. 

As part of this project, a general mathematical theory for 
interpreting results of capillary migration assays for cellular 
immune sensitivity (MIF tests) was derived. However, various 
basic parameters of leukocyte movement which are necessary for 
quantitating the assay are not well known. Consequently, collabora- 
tive experiments now are being performed with Dr. Lewis Lipkin 
(DCBD/NCI) which involve studying the response of neutrophils 
to various chemical factors ("lymphokines") produced by stimulated 
lymphocytes. Specialized measurement techniques have been devised, 
an example of which is a scheme where occupation number fluctuations 
are analyzed to determine mobility coefficients of migrating 
cells. Also, algorithms and computer programs have been developed 
to adapt an automated microscope system for cell tracking 
experiments. This instrument will be used to examine the behavior 
of neutrophils when responding to chemoattractants. Prototype 
measurements have been performed which now are being analyzed. 

Publication 

None 



126 



PROJECT NUMBER 



ENCE INFORMATION EXCHANGE 
(Do not use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00025-04 PSL 



PERIOD COVERED 

October 1. 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Theory and Application of Nuclear Magnetic Resonance Spectroscopy 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 
Other: 



James A.Ferretti, Ph.D., Research Chemist, PSL,DCRT 

E. D. Becker, Chief, Laboratory of Chemical Physics, LCP.NIAMDD 

G. R. Marshall, Professor of Physiology, Department of 

Physiology and Biophysics, Washington University School 

of Medicine, St. Louis, Mo. 

R. J. Highet, Laboratory of Chemistry, NHLBI 

G. H. Weiss, Chief, Physical Sciences Laboratory, PSL,DCRT 



COOPERATING UNITS (if any) 

Laboratory of Chemical Physics, NIAMDD 



lab/branch 



Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 

Division of Com put er Research & Technology. NIH. Bethesda. MD 



TOTAL MANYEARS: 



1.5 



PROFESSIONAL: 



_li 



OTHER: 



CHECK APPROPRIATE BOx(ES) 
n (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

The purpose of this project is to develop new methods in nuclear magnetic 
resonance spectroscopy and also to apply NMR to the study of small proteins. 
In particular, development of the correlation method of obtaining NMR 
spectra is of special interest. An experimental and theoretical study 
of interference effects in correlation spectroscopy has been undertaken. 
Saturation effects in correlation NMR are currently being studied. 
Investigations of the solution conformation, of derivatives of angiotensin , 
bradykinin and bleomycin are in progress. In these systems we have 
demonstrated the importance of the contribution of internal motion 
to the relaxation behavior. 



PHS-6040 
(Rev. 10-76) 



127 



The magnetic field strength and temperature dependences of the spin- 
lattice relaxation times, T,, spin-spin relaxation times, T^, and 
nuclear Overhauser enhancement (NOE) factors in small peptides are 
continuing to be studied. Studies on Angiotension-II and a heptapeptide 
analog at pH 4.3 have recently been completed and similar studies 
at pH 8.3 are ongoing. In the course of this pilot investigation on 
angiotensin-II, it was necessary to develop a mathematical model for 
analyzing the field strength dependence of the T, values in terms of 
the overall and internal reorientational correlation times. Such know- 
ledge is useful in understanding interactions at the receptor site. 
Use of the model in this fashion yields 1^ and NOE values which are 
in good agreement with those which are experimentally observed. These 
results have permitted us to establish the nature of the molecular 
associations in solution, estimate the size and shape of the various 
peptides and also to quantitate paramagnetic ion-peptide distances 
in the complexes. Similar studies have been carried out also on bleomycin, 
bradykinin, and gramicidin S. From the temperature and field dependences 
of the T, and NOE values, it has been possible to estimate the relative 
free energies of activation for both overall and internal motions. 

We have completed a study on the interaction of Fe (II) on bleomycin. 
Bleomycin is a glycopeptide whose iron complex has been shown to be 
effective against a variety of human neoplasms. The formation of this 
complex is required for efficient, oxygen-dependent degradation of 
DNA by bleomycin. The first aspect of this study was to characterize 
the spin state of the iron and to demonstrate that the complex is para- 
magnetic. For this purpose we carried out both optical and magnetic 
resonance measurements to show that the ion in the complex at pH 5- 
6 is predominantly in the high spin state. With these results and 
on the basis of field dependent relaxation studies on the carbon atoms 
of bleomycin and the complex at various molar ratios of iron, we determined 
the metal-carbon distances for various atoms in the glycopeptide. 
These results enabled us to propose a mechanism of action of the iron 
complex on DNA which involves intercalation of the bithiazole rings and 
where the oxydation of the iron generates free radicals of oxygen which 
induce rupture and strand scissoring of the DNA. 

We have initiated a study on a novel form of vitamin 8,^ (cobalamin). 
This new isomeric form of vitamin B12, which is achieyea by substituting 
the benzyimidazole base by a less bulky group like 0H~ or CN , is present 
as an impurity in all commercially available cobalamins. Preliminary 
studies suggest that it is at least as active biologically as the major 
constituent of vitamin B-jp. Spectra on the new form of cobalamin have 
been obtained and analyz^a and relaxation studies as well as other 
spectroscopic studies are under way to characterize the structure and 
conformation of the molecule. 



128 



The study of NMR correlation spectroscopy is continuing. We have 
completed both the experimental and theoretical aspects of the study 
of the effects of driving the nuclear spin system into its non linear 
response region. We demonstrate the usefulness of the correlation 
technique for effective flip angles which approach 90 . We have also 
investigated the consequences of the causality principle on the relations 
between absorption and dispersion mode signals obtained in correlation 
experiments with particular emphasis on the distribution of information 
between the two signals. 

We are proceeding to analyze additional aspects of NMR spin-lattice 
relaxation time, T-i, studies determined by various methods. In one 
previous work we showed that for the problem where Ti and the equilibrium 
value of the magnetization are the only unknowns, the fast inversion- 
recovery technique is often the method of choice. Recently we have 
considered two new techniques which also incorporate a rapid recycling 
of pulse sequences. The first method is a variable flip angle 
perturbation technique. The major advantage to this method is that 
the data may be plotted on a linear scale with uniform variance. A 
second method, which is a variant on the fast inversion-recovery methods, 
gives precisions in T-. which are approximately as good as those obtained 
by the normal fast inversion-recovery method. In addition, this method 
offers some additional advantages in the elimination of effects due 
to systematic errors in the effective flip angle parameters. 

Publications : 

Ferretti, J. A. and Marshall, G. R.: Carbon-13 properties of Angiotensin- 
II. Biophysical J. 21, 79a, 1978. 

Benovic, J. L., Ferretti, J. A., and Gupta, R. A.: Carbon-13 NMR and metal 
binding properties of Bleomycin. Biophysical J. 21, 199a, 1978. 

Ferretti, J. A., Marshall, G. R., and Gupta, R. K.: Carbon-13 relaxation 
parameters of small peptides, "Proceedings of the 20th Collogque AMPERE, 
Tallinn (1978)," p. XXX, North-Holland, Amsterdam. 

Weiss, G. H., Gupta, R. K., Ferretti, J. A. and Becker, E. D.: The choice 
of optimal parameters for measurement of spin lattice relaxation times. 

I. Mathematical formulation. J. Magn . Resonance (in press). 

Becker, E. D., Ferretti, J. A., Gupta, R. K. and Weiss, G. H.: The choice 
of optimal parameters for measurement of spin lattice relaxation times. 

II. Comparison of saturation-recovery, inversion-recovery, and fast 
inversion-recovery experiments, J. Magn . Resonance (in press). 

Ferretti, J. A. and Marshall, G. R.: Field dependent carbon-13 relaxation 
studies on peptides. I. Aggregation of Angiotension-II, Biophysical 
J. (in press). 



129 



Gupta, R. K., Becker, E. D., Ferretti, J. A. and Weiss, G. H.: A variable 
perturbation method for nuclear spin-lattice relaxation measurements. 
J. Magn . Resonance (in press). 

Gupta, R. K., Ferretti, J. A., and Caspary, W. J.: Carbon-13 NMR studies 
of the structure of the iron-Bleomycin complex. Biophysical J. 25, 
236a, 1979. 

Becker, E. D., Ferretti, J. A., and Gambhir, P. N.: Selection of optimum 
parameters for pulse Fourier transform NMR. Analytical Chemistry 
(in press). 



130 



PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00040-01 PSL 



PERIOD COVERED 

October 1. 1978 to September 30. 1979 



TITLE OF PROJECT (80 characters or less) 

Interactions of Erythrocyte Spectrin with Actin 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: Stephen Brenner, Research Chemist, PSL,DCRT 



COOPERATING UNITS (if any) 

E. D. Korn, Lab. Cell Biology, NHLBI 



lab/branch 



Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 

Division of Computer Research ft Terhgnlngy, NTH^ 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
□ (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



gl (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Sheep erythrocyte spectrin and rabbit skeletal muscle 
actin have been purified and their interactions studied by viscometry 
and ultracentrifugation. It has been found that the tetrameric 
form of spectrin crosslinks F-actin filaments to form a ^el and 
that spectrin dimer binds to, but does not crosslink, F-actin. 
Neither phosphorylation of the spectrin, nor dephosphorylation, 
was found to affects its interaction with actin. A high molecular 
weight oligomeric complex of spectrin, -actin, and several minor 
proteins, has been isolated from sheep erythrocyte ghosts. 
We have found that this complex can induce the polymerization 
of G-actin. Studies on the mechanism of action of this complex 
are in progress. 



PHS-6040 
(Rev. 10-76) 



131 



Interactions of Erythrocyte Spectrin and Actin 

Spectrin and actin are major components of the cytoskeletal network 
that is believed to be a determinant of erythrocyte shape and 
deformability. The primary objective of this project is to determine 
how spectrin and actin interact to form this cytoskeleton and to examine 
possible control mechanisms for regulating erthrocyte shape via modulation 
of spectrin/actin interaction. 

Spectrin is purified by gel filtration chromatography of low ionic 
strength extracts of sheep erythrocyte ghosts. This yields a herero- 
dimer of a 240,000 and 220,000 dalton polypeptides. The dimer is 
in simple equilibrium with a tetramer, although inter-conversion is 
between dimer and tetramer is \fery slow at low temperatures due to 
an unusually high activation energy. Equilibrium mixtures of dimer 
and tetramer are formed at 30-37 and the species separated by gel 
filtration and stored on ice. We have studied the interaction of 
spectrin dimer and tetramer with both G-actin and F-actin. 

Contrary to published reports we find that highly purified spectrin 
does not induce the G F polymerization of actin although we have 
isolated an oligomeric spectrin/actin couple from sheep erythrocyte 
ghosts which does induce actin polymerization. We find that when 
small amounts of spectrin tetramer is added to F-actin a highly viscous 
gel is formed indicating of crossl inking of the actin filaments. 
Spectrin dimer is totally ineffective as a ckossI inker, although ultra- 
centrifugation studies with radiolabelled ( P) dimer show that it 
does bind to F-actin. This data suggests that each spectrin dimer 
possesses a single F-actin binding site with the tetramer, therefore, 
having two F-actin binding sites. Since the tetramer is believed 
to be the vr\_ vivo form, we suggest that the cytoskeleton may be composed 
of a lattice work of actin filaments crossl inked by spectrin tetramer. 

Spectrin is a phosphoprotein and it has been suggested that a kinase/ 
phosphatase couple might control the interaction of spectrin with 
actin. We have investigated this possibility in detail, both by phosphoy- 
lating spectrin with a cAMP-independent kinase isolated from sheep 
erythrocyte ghosts (1 mole phosphate incorporated/mole spectrin) and 
by deohosphorylating spectrin labelled jm situ by incubating ced cells 
with P-labelled inorganic phosphate (83%) of protein-bound P released 
by alkaline phosphatase treatment). We find the interaction of spectrin 
tetramer and dimer with F-actin is independent of the state of phosphoylation 
of the spectrin. 

Work in progress includes quantitation of the spectrin/actin binding 
constants and stoichiometry and studies of the mode of action of the 
oligomeric complex which induces actin polymerization. 



132 



SMITHSONIAN SC 
PROJECT NUMBER 



ENCE INFORMATION EXCHANGE 
(Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00024-04 PSL 



PERIOD COVERED 



October 1. 1978 to Septpmhpr 30, IQyq 



TITLE OF PROJECT (BO characters or less) 

Studies in Mathematics and Statistics 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: George H. Weiss, Chief, Phys. Sci. Lab., PSL,DCRT 

Other R. J. Rubin, Senior Scientist, NBS 

I. G. Darvey, Visiting Scientist, PSL, DCRT 

J. E. Kiefer, Research Mathematician, PSL, DCRT 

D. E. Blumenfeld, Lecturer, University College, London 



COOPERATING UNITS (if any) 

None 



lab/branch 

Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 



Division of C omputer Research & Te chnolo 



w. NIH, B e th e sda 



TOTAL MANYEARS: 



-CLZ- 



PROFESSIONAL: 



JX^ 



-&^ 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Several unrelated investigations are included in this project. We have 
completed some work on a comparison of weighting schemes for deriving 
parameters from kinetic experiments on systems that obey Michaelis - Menten 
kinetics. We have nearly completed a study of span-constrained random 
walks that serves as models for the configurations of polymer chains . 
A third study was in the area of acceleration procedures for speeding 
the convergence of Fourier series. 



PHS-6040 
(Rev. 10-76) 



133 



studies in Mathematics and Statistics 

We have completed a study of different forms of weighting used in 
the interpretation of data collected from kinetic experiments on systems 
that follow Michael is-Menten kinetics. We have compared approximately 
twenty different forms of weighting and showed, by simulation techniques, 
that there is little difference between weighting schemes when the 
linearized form of the Michael is-Menten equations is used, but there 
are definitely preferred weights when data is analyzed directly. 

Together with Dr. R. J. Rubin we have developed the theory of span 
constrained random walks in an attempt to explain simulation experiments 
on the configurations of polymer chains. The observation has been 
that two qualitatively different configurations can occur depending 
on the size of chain. This observation has now been confirmed by 
our theoretical calculations. 

Our work on the acceleration of Fourier series was motivated by earlier 

applications of numerical methods for inverting Laplace transforms. 

There we found that the resulting series showed incredibly slow convergence 

in many instances. We have found extremely good results using an 

iterated summation by parts. This technique is as good or better 

than the currently recommended algorithm for accelerating the convergence 

of Fourier series. 

Keyword Descriptors : Least square weights, Michaelis-Menten reactions, 
polymer chains, random walks, acceleration of convergence, Fourier 
series, Laplace transforms. 

Publications : 

Blumenfeld, D. E., Weiss, G. H.: Curve fitting the probability distri- 
bution of acoustic noise form freely flowing traffic. Transp . Res . 
12, 111-114, 1978. 

Blumenfeld, D. E., Weiss, G. H.: Statistics of delay for a population 
of drivers with step and distributed gap acceptance functions. Transp . 
Res. 12, 423-429, 1978. 

Blumenfeld, D. E., Weiss, G. H.: The effects of gap acceptance criteria 
on merging delay and capacity at an uncontrolled intersection. Traf. 
Cont . & Engin . 20, 1-5, 1979. 

Bunow, B. Weiss, G. H.: How chaotic is chaos? Chaotic and other 
"noisy" dynamics in the frequency domain. Math. Biosc. (to appear). 

Dishon, M., Weiss, G. H.: Numerical inversion of Mellin and two-sided 
Laplace transforms. J. Comp . Rhys . 28, 129-132, 1978. 

Weiss, G. H. A note on occupation times of random walks, J. Stat. 
Rhys, (to appear) . 

Weiss, G. H., Rubin, R. J.: Internal configurations of span-constrained 
random walks. J. Stat. Rhys, (to appear). 

134 



SMITHSONIAN SCI 
PROJECT NUMBER 



ENC 
(Do 



E INFORMATION EXCHANGE 
NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00041-01 PSL 



PERIOD COVERED ^^ ^„^„ 

October 1, 1978 to September 30, 1979 



TITLE OF PROJECT (80 characters or less) 



(Quantitative Analysis of Electronmicrographs and Membrane Cellular 
Biophysics. 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 
Other: 



N. Gershon, Ph.D., Visiting Scientist, PSL,DCRT 

L. Jarett, Head, Division of Laboratory Medicine 

Washington University School of Medicine, St. Louis, MO. 

R. Smith, Washington University School of Medicine, 

St. Louis, MO. 

B. Bowers, LB, NHLBI 



COOPERATING UNITS (if any) 

None 



lab/branch 

Physical Sciences Laboratory, DCRT 



SECTION 



INSTITUTE AND LOCATION 

Division of Computer Research & Technology, Bethesda, MP 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 

1.0 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Methods of quantitating electron micrographs of protein particles on 
membranes were developed and applied to the analysis of cancer cells 
infected by viruses, and to jnsulin receptors in rat adipocytes . 
Mobility of cell membrane proteins and their interaction with cytoskeletal 
elements. 



PHS-6040 
(Rev. 10-76) 



135 



In the past few months we have been mainly concentrated working on the 
following subjects. 

(1) Quantitative methods in electron microscopy. We have developed 
quantitative methods to analyze electronmicrographs of biological systems. 
The methods includes digitization of micrographs and computational analysis 
of their contents (e.g. protein particles on membranes). The systems 

that have been analyzed so far are the following: 

(a) For a virally infected cancerous cells we have found that the viral 
glocyproteins are distributed in a nonrandom fashion on the cell surface 
even at non-permissive temperatures where the viral buds cannot be 

formed. This work was done in collaboration with Dr. A. Demsey. 

(b) In rat adipocytes we have examined insulin receptors. We are 
interested in elucidating the mechanism of insulin action on cells. 
Using ferritin-insulin it is possible to visualize insulin binding sites. 
Electronmicroscopy shows that the insulin binding sites, which are mostly 
located on the glycocalyx, are aggregated into discrete groups. Cytochalasin 
B, a pharmacological agent and a potent glucose transport inhibitor 

is found to disrupt groups of insulin binding sites while not interferring 

with its other biological activities. The quantitative analysis shows 

that binding sites separated by distances of the order of 300° -400a 

are likely to be separated by cytochalasin B. This finding suggests 

that besides the functional linkage of the insulin and glucose transport 

system, the latter machinery requires a number of groups of molecules 

to be aggregated where the distance between consecutive insulin binding 

sites is of the order of 300A-400A. An additional detailed analysis 

shows that the distances between individual binding sites in pairs 

are about the same in adipocytes with or without the treatment of cytochalasin 

B. This work is done in collaboration with Dr. L. Jarett and Dr. R. 

Smith. 

(c) Studies of membranes of phagocytosing cells has been initiated. 
The analysis will show if the internal regions of the membrane have 

a different intramembraneous particle content from the cell surface, 
or they are in a different aggregation state. The study of phagcotosis 
has its valuable impact on the understanding of cell function and structure 
especially in circulating cells of the immune system. This work is 
done in collaboration with Dr. B. Bowers. 

(2) Mobility of membrane proteins and their interaction with 
cytoplasmic components. Membrane proteins can interact with various 
components inside the cell, e.g. cytoskeletal elements. We have looked 
for possible physical mechanisms which would account for the attachment 
of membrane proteins to cytoskeletal filaments e.g. by entanglement 

or by polymerization of cytoskeletal elements around aggregated membrane 
proteins. These studies might shed light on how signals are transferred 
through membranes to cell interiors. 



136 



Keyword Descriptors : Electron microscopy, digitization of micrographs, 
cell surfaces, rat adipocytes, insulin binding sites, cytochalasin B, 
phagocytosis, membrane proteins, cytoskeleton. 

Publications : 

N. Gershon, On the effect of cross! inking on the attachment of 
membrane proteins to the cytoskeleton. Cell Surface Events in Cellular 
Regulation, Elsevier-North Holland, p. 163 (1979). 

N. Gershon, A. Demsey, C. Stackpole, Analysis of local order in the 
spatial distribution of cell surface molecular assemblies. Exp . Cell 
Res. , (to appear). 



137 



LABORATORY OF STATISTICAL AND MATHEMATICAL METHODOLOGY 



James E. Mosimann 
Chief 




LSM staff interact with all NIH institutes and with 
staff in other Federal agencies outside HEW. In a 
consultation session (I. to r.), LSM Chief Dr. James 
Mosimann and Statistical Software Section Head Ray 
Banner sit with Dr. William Blackwelder of EIA14DD. 



I. SUMMARY 

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. The laboratory has 17 
full-time 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 program 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 eight mathematicians, computer scientists 
and programming aides. 

f The Statistical Methodology Section (SMS) works closely with 
the Statistical Software Section. Four individuals who work 
under the direction of a mathematical statistician 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. Five 
individuals work under the direction of a computer systems 
analyst who specializes in computational linguistics. 

Scope of Work 

LSM staff interact with all NIH institutes and with staff in other 
federal agencies outside HEW. Fiscal year 79 was LSM's fifth year as 
a separate entity within DCRT. The volume of its computational and 
consultation services continued to expand while its research 
activities were maintained at about the same level as the preceding 
year. 



140 



Highlights of the Year's Activities 

Computation . A major part of LSM activity is the offering of 
statistical and mathematical program packages to the NIH user 
community. LSM accepts responsibility for evaluation of new program 
packages and their suitability for NIH. When LSM does offer a 
package to the NIH community, LSM makes three basic commitments: 

• The maintenance of the package, with adequate documentation, 
through NIH computer system changes, package updates and 
corrections. 

• The rapid response to queries concerning user access to a 
package program including job control language and program 
parameters. 

• The assistance in Interpretation of results. 

During this year, as in the past year, following program packages and 
programs were maintained by SSS of LSM. 

BMD, BMDP, Biomedical Computer Programs, UCLA SPSS, Statistical 
Package for the Social Sciences, SPSS, Inc. SAS, Statistical 
Analysis System, SAS Institute, Inc. PSTAT, Princeton Statistical 
Package, Princeton University IMSL, International Mathematical and 
Statistical Libraries, IMSL 

Inc. MSTATl, Collection of Mathematical and and Statistical 
Programs, 

DCRT. 

SSS added one additional package, SCSS, during fiscal 79 year. SCSS 
is an interactive version of SPSS. 

The effort expended in the commitment to maintain these packages is 
considerable. During this year every package went through at least 
one major update. The effort expended in response to queries 
concerning package access is also considerable and requires 
continuous availability. During the year over 4,000 calls were 
responded to by SSS staff alone. Two courses were taught on each of 
the SPSS, SAS, and BMDP packages. 

The use of program packages continues to show considerable Increase. 
Statistical package use over the past five years is illustrated in 
Figure 1. The average accesses per month of all the statistical 
packages rose from almost 9000 during fiscal 78 to over 12000 in 
fiscal 79. For the third year in a row SAS experienced 
the largest Increase of any of the packages. SAS averages almost 
7000 accesses per month, up from 4000 per month in fiscal 78. The 



141 



I 
I 
I 

I 
I 






X 
X 
X 



r 



■o 
• 
■f 
1. 
o 
a 
a 



9 



X 
X 



• 1 

I. a 

(1 

u. u 



X 
X 



X 
X 



o 

JC 
■*> 
c 
o 

E 

1. 
a 
a 

n 

« 
n 



X 
X 



I 

^, oiuor^eo 
I 
+ ZO>I~«0 

+ ooi-r~«o 

I 

•I- «nuia.r^oo 

I 

I 

I 

+ ->32r«.«o 

+ EO-r-oo 
I 

I 

+ E'<OCi>.aO 

4- u.tussr^<o 

I 

••• -xzr^to 

I 

+ OUior^r^ 

I 

+ ZO>r~t^ 

I 

+ oo>-r^r» 

I 

I 

+ •oor^r^ 

I 

I 

I 



^ 

•o 
a 
■♦' 

L 
O 

a 

a 

n 

n 

w 

Ol 

re 

j£ 
u 
re 
a 

9 

•o 

3 

u 

c 



o 
c 

m 
or 
o 

o 



•f<a.Q£i^r^ o -< 




1 E c 




+ E<0Cf>.f^ 


o 




1 


c 
o 




1 


■*> 




+ OUJOr^<o 


• 




•f zo>r><e 


lO 




1 


• 
1. 




••■ miua.!^^ 


re 


, 


1 


3 


c 


+ OO^xS 


•♦' 


o 


1 


H- 




■f ->3-ir^>« 


O 


•t' 


j 


to 


u 


■»• ->3Zf^<0 




a 


1 


ft 


«n 


••■ E* Vi^o 


n 




1 


u 


a 


••■ ■<a,Bii^-o 


*^ 


u 


1 


+> 


c 


+ E<oer>« 


n 


01 


+ u.uj(Qr><0 


■p 


u 


1 


re 


<n 


+ -xzt*^ 


-f 




1 


«o 


L 


••• oiuur^in 







1 


« 


^ 


♦ zo>r^in 


X 


3 


1 


•»< 


a 


■f oot-rs'in 




E 


1 


>> 


o 


■^ <nuja.r>in 


^ 


o 


+ oor^m 


■D 


"0 


1 


e 


c 


1 


4^ 

1. 


re 


+ ->3zr~in 


o 


n 


1 


a 


u 


•«■ E<>-c^in 


Q. 




1 


3 


4^ 


* <Q.O£r^ir) 


n 


re 


1 
1 


*i 


E 
o 


1 


a 


£ 


1 


Ol 


A* 


1 


re 


re 


1 


■X 


E 


I 


u 


O 



-+*•»•■•••♦• + + • 



■»■*■¥■¥*■*■*■■*■■*■■*■■¥■ 



■■*■■¥■*■■¥■*■■¥■¥*■¥■¥•¥+• 



in t- 
I1J0£Z 

<nujo 



^^ooo^^oo^eeo^ocso^oooo^oooe^eeeoeoo^ 

oinoinoinoineinoinoinoinomoinemoinomomomoirteinoirt 
•Or^r^<e>omin<a-<riOK>(M(M>-^oo^CK««Or^rv«^inin<>'4-ioiocsi(M^^ 



142 



average number of accesses per month for SPSS increased from 3900 to 
4900, an increase of over 55%. The average combined accesses of the 
BMDP and BMD packages rose slightly from 750 accesses per month in 
fiscal 78 to around 850 this year. As an example of a package used 
for specialized purposes, PSTAT averaged 35 accesses per month, up 
from 20 average accesses per month in fiscal 78. The main programs 
in MSTATl averaged 180 accesses per month, which is slightly above 
the usage in fiscal 78. The subroutine usage of MSTATl, which can 
only be estimated, is much higher than the main program usage. 
Accesses to the IMSL package cannot be counted, but it is estimated 
that usage increased during fiscal 79. 

In addition to the statistical packages discussed above, the 
Biomathematics and Computer Science Section maintains the PDP-10 
interpretive program MLAB for bi ©mathematical modeling at NIH. This 
package was designed and implemented by BCS staff, and has been sent 
to many computer installations here and abroad. Statistics taken at 
NIH for nine working days during May showed an average of 31 distinct 
MLAB users each day, with 96 distinct users over the period. During 
fiscal 79, the documentation for MLAB was expanded. An earlier 
edition of the MLAB Reference Manual was updated and replaced by two 
new documents, a concise Reference Manual and an Applications Manual. 
An extensive Beginner's Guide was completed and will be distributed 
as soon as the printed copies are received. A Primer was obtained 
from EPA and distributed to NIH users with changes appropriate to 
MLAB usage at NIH. BCS research results were incorporated into new 
MLAB features, including an improved search method for curve-fitting, 
an operator for interpolation on surfaces in 3-dimensional space, and 
improvements in hidden-line 3-dimensional graphics techniques. 

BCS continues support of C-LAB, a computer package for pattern 
recognition and cluster analysis developed by a BCS staff member. 
New documentation for C-LAB was distributed at the beginning of the 
fiscal year. Minor improvements and corrections were incorporated 
into C-LAB, and compatabil ity with MLAB was maintained. 

BCS staff members taught courses in curve-fitting methods in Mlab in 
the Fall and Spring, and clustering methods in Clab (Fall). The new 
manuals cited above facilitated the presentation of the course 
material . 

BCS assumed responsibility for the Unified Generator Package, a 
software package for generating S/370 assembly language programs for 
data base management, during the fiscal year. This package was 
formerly supported by the Data Management Branch, DCRT, and was 
transferred to BCS support because the program designer joined the 
BCS staff. A course was taught on the Unified Generator Package 
during the Fall session, and a paper was given at the Washington ACM 
Meeting of 12/78, describing the package. The package was maintained 
and improved during the fiscal year. 



143 



A new version of the REDUCE package for symbolic manipulation of 
algebraic formulas was received from the University of Utah, and it 
is currently being tested. A software package called PROLOG for 
symbolic and logical data processing was obtained from the University 
of Edinburgh. It serves as a computational tool for BCS research, 
and is currently being evaluated for possible support by the section. 

BCS discontinued support of the program package MODELAIDE during the 
fiscal year, when the BCS staff member who designed the package 
transferred to the Laboratory of Applied Studies. 

Consultation . The diverse nature of LSM consulting is indicated by 
the projects and activities listed in Part II. 

The pattern of LSM consultation activity shifted slightly to more 
computer-oriented consultation in fiscal 79. Estimated percentages 
during fiscal 78 and 79 are given below for comparison: 

1978 1979 
t Mathematical, statistical and computer 
computer science advice with 
limited computer use 10% ,5% 

• Mathematical or statistical advice 

with considerable computer use 50% 55% 

• Computational advice alone 40% 40% 

The continued availability and use of general -purpose statistical and 
mathematical packages like SAS and MLAB has maintained the large 
component of computer use in these figures. 

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. For example, for the past three years LSM has 
collaborated with Dr. R. Webber of the Clinical Investigations 
Branch, NIDR, on the uses of symmetric axis geometry in describing 
biological shape. One finished study has discussed new shape 
invariants in the development of the human mandible. The 
collaboration has involved not only new research in symmetric axis 
geometry by an LSM staff member, who orginated the concept, but the 
preparation of innovative computer programs, by LSM and CIB staff. 
These programs permit interactive examination of shape descriptions, 
and the generation of shapes from descrifttions. 

Many consultations which involve considerable LSM effort do not 
involve new research in computer science. For example, the Unified 
Generator Package, supported by LSM, was used extensively for a 
project of Dr. H. Guttman, OD, OPPE, NHLBI. Her data on researchers 



144 



associated with NIH grants were collected for use in program 
management. The Unified Generator Package was used to create the 
software for maintenance of the data base and for the retrieval of 
information from it. The author of the Unified Generator Package, an 
LSM staff member, provided assistance in its use, as well as in the 
use of the updating, reformatting and reporting systems which were 
generated. 

An example of a long-term effort involving statistical analysis is 
LSM's study of complications of dialysis in collaboration with Dr's. 
G. Hirschmann, R. Wineman, and M. Wolfson, all with or formerly with 
the Artificial Kidney-Chronic Uremia Program of NIAMDD. 

A number of statistical consultations involve intensive effort over a 
period of several weeks. One example would be a study of 
factors related to case of adopted infants by Dr. R. Klein, Social 
and Behavioral Sciences Branch, NICHHD. LSM provided the analysis 
which used the general linear model. A careful study of the 
hypotheses being tested and their substantive meaning was done. 

Research . BCS research included projects in computer science, 
biomathematics and general mathematical methods. A computer science 
project for developing heuristic tools for symbolic and logical data 
processing included applications to computational linguistics and 
computer generation of symbolic, closed-form solutions to 
differential equations by applying Laplace transform methods. 
Research in clustering methods continued, including adaptation and 
improvement of current methods for generation of minimal spanning 
trees, and optimal selection of variables for linear regression using 
branch and bound methods. Study of computer storage and retrieval 
methods is continuing. 

Research in the "symmetric axis" method of shape description was 
continued, with most effort directed towards extension of the 
mathematics: (1) to 3-D objects for application to data from a 
variety of 3-D scanners, and (2) to the direct extraction of 
descriptions from gray scale pictures, such as X-rays. The BCS 
investigator in these studies recently presented the results at an 
NSF workshop in Philadelphia on three dimensional representation of 
objects. 

Research in general mathematical methods continued in the areas of 
curve-fitting by the sum of absolute difference (L-one) and maximum 
of absolute difference (L-infinity) criteria, and in methods for 
analyzing inclusion relations between vector spaces. 

In SMS research on multivariate analysis and simultaneous statistical 
inference was performed. Study on size and shape methods continued, 
with emphasis on the mathematical transformations connecting the 
commonly occurring size and shape variables. Applications are to 



145 



biological data in the form of ratios or proportions. Research also 
was initiated in the general linear model for the analysis of unbal- 
anced data. Such data are of common occurrence at NIH. In the area 
of simultaneous inference, a procedure for multiple ratio estimation 
was extended to the many sample, unequal covariance, case. 

In MIS, their system for the storage and retrieval of pathology 
information was applied to pathology reports. These natural language 
reports were encoded automatically by computer into SNOP (The 
Systematized Nomenclature of Pathology) as well as into ICDO (The 
International Classification of Diseases for Oncology). Research was 
continued on the construction of semantically structured dictionaries 
for pathology. Studies were also continued on rules for automated 
morphosemantic segmentation of medical compound words derived from 
Greek and Latin, and for paraphrasing them in English and French 
medical languages. Work continued on the construction of medical 
microglossaries for use in small computers. 

Future Plans 

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



146 



II. LSM PROJECTS AND ACTIVITIES 
Major LSM consulting activities of the year included: 

t Clinical research, patient care, epidemiology 

Hemodynamic and plasma catecholamine responsis to hyperthermic cancer 
therapy in humans . Young Kim, AS, CO. Cancer patients treated by 
induction of hyperthermia under thiopental and fentanyl anesthesia, 
respond with attenuated hemodynamic changes compared with those 
reported in normal volunteers. Measured plasma catecholamine at 
hyperthermic condition showed evidence that sympathetic nerve 
response to hyperthermia. Statistical procedures used include 
multiple regression analysis, ANOVA, and descriptive statistics. 

Predicting the Onset of Huntington's Chorea . C. Cox, NINCDS, CNB. 
Neuropsychological measurements, biochemical data, and scan data was 
collected on 85 subjects including normals, persons at risk, and 
patients with H. C. (A person whose parent succumbed to H. C. is 
thought to have a 50% chance of contracting the disease). Advice and 
help was given with data management and the use of discriminant anal- 
ysis and repeated measures ANOVA programs. The discriminant analyses 
indicated that a certain subgroup on some of the tests. 

Triglyceride and HDL Cholesterol . E. J. Schaefer, IR MD, NHLBI. 
This year the Statistical Analysis System was heavily used as we 
looked at the effects of several diet types on lipoproteins. A major 
diet effect was observed when diets such as low fat, high 
carbohydrate or high polyunsaturated fat were constrasted with ad lib 
diets in normals and high blood fat subjects. Further work was done 
in investigating sex and age effects on lipoproteins in normal and 
dyslipoproteinemic subjects. Results support the earlier conclusion 
of major sex differences. 

Management of epidemiology data . T. Mason, EEB, NCI, and N. 
Shusterman, ORI. ORI is a consulting firm with an NCI contract to 
manage a number of epidemiology data bases, involving diverse areas 
such as breast cancer detection or carcinogens in water supplies. 
ORI uses the Unified Generator Package, maintained by LSM, to create 
updating and reformating systems used to maintain these data bases. 
LSM provided assistance for effective use of the Unified Generator 
Package. 

Plasma melatonin in humans with affective disorders. A. Lewy, J. 
O'Steen, LCS, NIMH. Daily, round-the-clock measurements of plasma 
melatonin were taken from patients with affective disorders and 
normal volunteers. LSM assisted by designing and testing procedures 
for automatic generation of graphical displays of data, using MLAB. 

147 



Test score analysis . A. Lipsky, LPP, NIMH. Scores were obtained 
from a sample population using the Leyton Inventory Test Scale. LSM 
designed and tested procedures for automatic generation of graphical 
displays of data, using MLAB. 

Childhood behavior problems . C. Edelbrock, LDP, NIMH. LSM assisted 
in preparation of 142 graphs displaying clinical data on childhood 
behavior problems, using MLAB. This data is part of a monograph 
survey to be published. 

Encoding of autopsy reports from an American city . L. Thomas, LP, 
NCI. Several thousand autopsy reports were automatically encoded by 
computer into SNOP and ICDO, for a study of the use of autopsy data 
from a large city (Atlanta, GA). 

• Laboratory investigation 

Schistosoma Japonicum : Divergence Between Egg Deposition in the 
Feces in Rabbits Infected with the Japanese and Philippine Strains. 
A. W. Cheever, NIAID. There has been a renewed interest in the 
rabbit as a model of Schistosoma japancium infection since it has 
been demonstrated that in rabbits hepatic vascular lesions are very 
similar to those in Symmer's fibrosis in man. Eighty infected 
rabbits, 67 exposed to the Japanese and 13 to the Philippine strain, 
were examined 8 to 66 weeks after exposure to 50 to 750 cercarial. 
SAS was used to analysis the data. 

Circadian rhythms of protein production in the rat eye . A. Goldman, 
LVR, NEI. Experimental injections of two radioactive labels into 
live rats were assessed by separation of retinal homogenate using gel 
electorphoresis and measurement of product radioactivity by 
scintillation counter. Labels were injected at different times of 
day to access circadian influence on production levels. LSM provided 
assistance in design and testing an MLAB procedure to process 
scintillation counter paper tapes, perform elementary statistical 
computations, and generate graphical displays. 

Phagosome measurement in the rat eye . A. Goldman, LVR, NEI. Rats 
were subjected to surgical and drug treatments, to study the effects 
upon the retinal renewal process. Phagosomes in pigment epithelium 
tissue samples were counted using light microscopy. LSM provided 
assistance in design and testing of MLAB procedures to check data 
base for errors and compute elementary statistical analysis of 
selected experiments. 

Mouse cataract study . P. Kador, LVR, Ntl. The "Philadelphia" mouse 
strain develops hereditary cataracts visible 5 to 6 weeks after 
birth. In this study, Philadelphia mice were measured for cataract 
development and lens electrolyte levels were compared to levels in 
normal controls. LSM assisted with MLAB procedures for processing 
experimental data and computing elementary statistical measures. 



148 



Redox studies of cytochrome . R. W. Handler, 0. H. Setty, LB, NHLBI. 
LSM is contributing to revisions of a paper on voltage titration of 
cytochrome spectra, and to biomathematical modeling of an intense, 
irreversible release of acid from E. Coli membranes under electrical 
stimulation. Oxygen uptake of reconstituted membranes is also under 
study, for which LSM provided software and mathematical analysis for 
digitization and evaluation of strip charts. 

Oxygenation of whole blood . R. M. Winslow, IR, NHLBI. Experiments 
show that pH changes in whole blood can be modeled by simple scaling 
of oxygen pressure. Imai (JBC 1975) announced that hemoglobin pH 
does not have this property. LSM is providing biomathematical 
analysis for appropriate models, to include explanation for 
experimental findings of nontypical hemoglobin saturation curves. 

Hemoglobin saturation analyzer . R. Berger, G. Dossi , TD, NHLBI. LSM 
is providing biomathematical analysis for design of an analyzer used 
in measuring biological effects of altitude. Relative concentrations 
of four states of the oxygen-combining site of hemoglobin are 
computed by measuring absorbance at several wavelengths and solving 
related linear equations. A minimum variance technique is being used 
to select an optimal set of wavelengths. Design criteria require 
simplicity, accuracy and reliability, for projected uses on a Mt. 
Everest expedition, in SKYLAB II, and for a study of Peruvian Indians 
living at 15000 foot elevation. 

Regularization of ultracentrifuge data . M. Lewis, BEI, DRS. The 
applied mathematics of regularization is currently an active field, 
with many available methods of regularization effective for some 
cases of the general problem. LSM is advising on the available 
methods of damping and cross-validation, to design a method tailored 
for the particular class of ultracentrifuge data obtained by the 
investigator. 

Modeling of Chemical Kinetics . A. Thakur, LTB, NCI. Advice was 
given on the statistical analysis of chemical kinetics modeling, 
including the calculation of confidence intervals and tests of 
hypotheses. 

DNA Pattern Matching . M. Bina, NIAID. Consultation was given on DNA 
string matching to detect regions which are clipped out of the 
corresponding RNA. 

Immunological binding study . J. Dower, NCI. LSM assisted in 
modeling and curve-fitting data obtained in immunological binding 
experiments. 

Ultracentrifuge data model ing. M. Johnson, CE, NIAMDD. Alternative 
mathematical models for ultracentrifuge experiments were considered. 
LSM provided advice on curve-fitting procedures and modeling in MLAB. 



149 



Enzyme kinetic studies . J. Harmon, D, NIAMDD. LSM provided advice 
on modeling and solving differential equations in MLAB, for enzyme 
kinetics experiments. 

Diffusion model . A. Weinstein, IR, OD, NHLBI. A steady-state model 
of paired differential equations was analyzed. LSM assisted in the 
analysis and debugging of MLAB procedures for obtaining numerical 
solutions. 

Insulin binding experiments . P. De Meyts, D, NIAMDD. Models of 
insulin binding with negative cooperativity were analyzed. LSM 
provided advice on modeling and curve-fitting in MLAB. 

• Program management and adminstration 

Investigator career profile study . H. Guttman, OD, OPPE, NHLBI. 
Data on investigators associated with NIH grants have been collected 
to be used for program management purposes. Included are evaluation 
of the effects of past NIH-supported training programs, and 
projections of current and future training programs based upon 
national needs and currently available researchers. The Unified 
Generator Package, supported by LSM, has been used to create the 
software for maintaining this data base and retrieving information 
from it. LSM provided assistance in the use of the Unified Generator 
Package, in the use of the generated updating, reformatting, and 
reporting systems, and in the training of users of the generated 
systems. 

NIH training support . W. S. Batchelor, OD, NIH. Data consists of 
the number of predoctoral and postdoctoral trainees supported by each 
Institute, classified by disciplines or medical specialities and by 
fiscal years. LSM advised on graph-theoretical computations for 
analysis of this data using MLAB, and assisted in designing and 
testing a procedure to output the results. 

Index's for the DCRT Library Document Collection . E. M. Chu, OD, 
DCRT. Existing linguistic concordance programs were adapted to 
produce key word indexes on author and title words of documents which 
are shelved by an accession-document number. 

t Computer research and technique development 

SLANG (Structured Language Compiler) . R. Magnuson, DMB, DCRT. SLANG 
is a language processor designed to enable users to write 
block-structured IBM 370 assembly language source code. LSM assisted 
in its development by participating in design decisions, detecting 
errors during the implementation phase, and editing the 
documentation. 



150 



Chemical toxicity study . T. Hopfinger, R. Potenzone, Case Western 
Reserve University. Clustering techniques relating toxicity to 
chemical structure were performed. Several small sets of compounds 
(one including 32 nitrosaminos) showed excellent clustering 
properties, so that toxicity for a compound of each such class could 
be predicted from the compound's structure. A set of 385 compounds 
with diverse structures showed poor clustering properties. LSM 
provided cluster analysis using C-LAB. 

• LSM research projects 

Automated Data Processing of Medical Language . M. G. Pacak, LSM, 
DCRT. Work on further development of a semantical ly structured 
medical lexicon to be used for the creation of a medical data base 
for NCI. 

Multivariate Statitical Analysis . J. E. Mosimann, LSM, DCRT. The 
study of multivariate statistical methods for the analysis of data 
which take the form of ratios or proportions. 

Biological and Visual Shape . H. Blum, LSM, DCRT. Development and 

application of a new geometry of biological shape which gives a 

natural and efficient description to a variety of biological objects 
at vastly different levels of complexity. 

Cluster Analysis . M. Shapiro, LSM, DCRT. Research into optimum ways 
of clustering using the minimal spanning tree algorithms. 

Discrete Mathematics and Applications . G. A. Hutchinson, LSM, DCRT. 
Inclusion relations between vector spaces and related problems 
concerning modules over rings were studied. 

Linear Methods in Statistics . J. D. Mai ley, LSM, DCRT. Family 
confidence limits for ratios of sample means from multivariate normal 
distributions were extended beyond the original unequal covariance 
data sets. 

Nonl inear Equations . R. I. Shrager, LSM, DCRT. 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. 

Research Topics in Computer Science . G. D. Knott, LSM, DCRT. The 
development of flexible and efficient storage and retrieval 
algorithms. 



Non-numerical programming techniques and applications . L. M. Norton 
LSM, DCRT. Using both general purpose and special purpose 
programming systems, techniques for computer processing of non- 
numerical data are developed and evaluated by implementing small -to- 
medium-scale research projects in selected application areas. 



III. PUBLICATIONS 



Blum, H. , and Nagel , R.: Shape description using weighted symmetric 
axis features. Pattern Recognition , 10:167-180, 1978. 

Cole, B. R., Shapiro, M. B., and Rodbard, D. : Chromatography and 
Electrophoresis Analysis System, in Electrophoresis 78 . Elsevier 
North Holland. 1978. 

Habbersett, M. C, Shapiro, M. B., Bunnag, W. , Nishiya, I., and 
Herman, C: Quantitative analysis of flow microfluorometric data for 
screening gynecologic cytology specimens. J. Histochem and Cytochem , 
27(l):536-544, 1979. 

Evarts, R. P., Brown, C. A., and Atta, G. J.: The effect of 
hydroxyl amine on the induction of mammary tumors by 7,12- 
dimethylbenz[a]anthracene. Experimental and Molecular Pathology , 
30:337-348, 1979. 

Hutchinson, G., and Czedli, G. : A test for identitites satisfied in 
lattices of submodules. Algebra Universalis , 8:269-309, 1978. 

James, I. R. , and Mosimann, J. E.: A new characterization of the 
Dirichlet distribution through neutrality. Annals of Statistics , in 
press. 

Knott, G. D.: Fixed-bucket binary storage trees. Proc. 12th Hawaii 
International Conference on Systems Science , 36-48, University of 
Hawaii, 1979. 

Kamel , I. A., Elwi, A. M. , Cheever, A. W., Mosimann, J. E., and 
Danner, R.: Schistosoma Mansoni and S. Haematobium Infections in 
Egypt. IV. Hepatic lesions. The American Journal of Tropical 
Medicine and Hygiene . 27:931-938, 1978. 

Lieblick, A. K., Symmes, D., Newman, J. D., and Shapiro, M. B. : 
Development of the Isolation Peep in Laboratory Bred Squirrel 
Monkeys. The Journal of Animal Behavior , in press. 

Marimont, R., and Shapiro, M. B.: Nearest Neighbor Searchers and the 
Curse of Dimensionaltiy. J. Institute of Math, and Applications , in 
press. 

Mosimann, J. E. , and Malley, J. D. : Size and shape variables. 
Multivariate Methods in Ecological Work , L. Orloci, C. R. Rao, and W. 
M. Stiteler, eds., p. 17, in press. International Co- operative 
Publishing House, Fair! and, MD. 

Mosimann, J. E. , and James, F. C: New Statistical Methods for 
Allometry with Application to Florida Red-winged Blackbirds. 
Evolution, 33:444-459, 1979. 



152 



Mosimann, J. E. , Malley, J. D. , Cheever, A. W., and Clark, C. B.: 
Size and Shape Analysis of Schistosome Egg-Counts in Egyptian Autopsy 
Data. Biometrics . 34:341-356, 1978. 

Norton, L. M.: A program generator package for management of data 
files - the input language. Proceedings, 1978 Annual ACM Conference , 
p. 217-222, Washington, D. C. 

Pacak, M. G., and Dunham, G. S. : Computers and Medical Language. 
Medical Informatics , 4(l):13-27, 1979. 

Webber, R., and Blum, H. : Angular invariants in developing human 
mandibles. Science, in press. 




The use of program packages continues to show considerahle 
increase. The average accesses per month of all the 
statistical packages rose from almost 9, 000 during FI78 
to over 12,000 in FY79. 



153 



SMITHSONIAN SCIENCE INFORMATION EXCHAN 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00001-08 LSM 



PERIOD COVERED 

Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Automated Data Processing of Medical Language 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: M. G. Pacak 

A. W. Pratt 

OTHER: G. Dunham 

S . Harper 

M. DeMeyts-Graitson 



Computer Systems Analyst LSM DCRT 

Director DCRT 

Computer Programmer LSM DCRT 

Computer Programmer LSM DCRT 

Guest Worker LSM DCRT 



COOPERATING UNITS (if anj 



LAB/BRANCH 

Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 

Medical Information Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, 



NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

2.5 



PROFESSIONAL: 

2.5 



OTHER: 



0.0 



CHECK APPROPRIATE BOX(ES) 
J (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



£] (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

The major objective of the project is the continuation of work on further 
development of a semantically structured medical lexicon to be used for the 
creation of a medical data base for NCI. 

The operational system for information storage and retrieval of pathology data 
is going to be used for the creation of a medical data base mentioned above. 
Further improvements of the present encoding system on the morphological , 
syntactic and semantic level will be studied. 



PHS-6040 
(Rev. 10-76) 



154 



Project Description: 

1) Several thousand autopsy reports from Altanta, GA, were encoded into SNOP 
and ICDO for National Cancer Institute (NCI) . Some minor changes in the 
encoding system to improve its efficiency are under consideration. 

2) The computer program for the segmentation and interpretation (paraphrasing 
rules) was completed for - ITIS forms and is being tested. The program will 

be used for the development of a generalized system for automated morphosemantic 
segmentation and interpretation (paraphrasing algorithm) of medical compound 
words. A similar system is under development for French. 

It appears that the utility of automated morphosemantic analysis and corresponding 
paraphrasing can be adapted for use with other suffix forms such as - ECTOMY - 
EMIA etc. If so, the lexicon of morphosemantic constituents may become a 
practical tool for automated processing of medical terminology in English as well 
as in other languages. It was found that the ratio of similarities in modern 
national medical languages is very high for medical word forms derived from 
Greek and Latin. Using the computer as a tool, it might be possible to 
construct a computer-oriented multilanguage medical lexicon which utilizes these 
similarities between medical terminologies in different countries. 

The program for paraphrasing procedures was written in a modified version of 

the programming language PROLOG, which was developed at the University of Marseille. 

Lexicographic work continued on the construction of a lexicon for pathology. 
This lexicon was extracted from "Systematized Nomenclature of Pathology (SNOP) . 
It contains parts-of-speech information, and when ever possible, the word is 
listed with its allomorphs. The goal of the research is to reduce the size of 
the lexicon and generate automatically linguistic algorithms for processing 
of morphosyntactic and morphosemantic data which are derived from the microglossary 
for pathology. 

The formal description of a syntactic and semantic parser for medical language 
data processing is being completed. The parser will be tested on medical data 
in the near future. 

The concordance program was used to create (produce) title and author indexes 
for the document collection of the DCRT library. 

The programming of a frame system for medical linguistic research was completed. 
Frames are displayed at a CRT terminal from which an operator may select or 
generate in response. Responses are captured and stored for practical information 
needs, and responses determine which new frames are called up. 

The system has been designed to test the hypothesis that the form (structure 
and relationships of frames) in which complex data is most easily and naturally 
elicited from a medical user in a frame system. The system is closely related 
to the syntax of the noun phrases which are used by the user to express these 
data in natural language. Collaboration with a clinical research group is being 
sought. 

The tape copies of computer-based medical dictionaries (SNOP, ICDO) were made 
available to several universities. Their distribution to selected requesters 
will continue. 

155 



Future Efforts: 

Continuation of research studies in medical language at present level 
(morphology, syntax, semantics). Comparative study of structures of computer- 
oriented medical dictionaries (SNOP, SNOMED, ICDO, SNODERM) is under consideration. 
Paraphrasing procedures for medical terms derived from Greek and Latin will be 
tested. 

Increased effort to obtain a version of the encoder for implementation on a 
small computer system. 

Testing of frame system on medical data. 

Preparation of a tutorial report on logical problems in medical data bases. 

Publications: 

Pacak, M. G. , and Dunham, G. S.: Computers and Medical Language, Medical 
Informatics , 1979, Vol. 4, No. 1, pp. 13-27. 



156 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00012-07 LSM 



PERIOD COVERED 



Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Biological and Visual Shape 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: H. Blum 
OTHER: M. O'Connor 
R. L. Webber 



Res. Gen. Phys. Scientist LSM DCRT 
Mathematician LSM DCRT 

Chief, Clin. Invest. Branch CIB NIDR 



COOPERATING UNITS (if any) 

DMG, CIB, NIDR 



lab/branch 
Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 

Biomathematics and Computer Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

0.9 



PROFESSIONAL: 

0.9 



OTHER: 



0.0 



CHECK APPROPRIATE BOX(ES) 
C (a) HUMAN SUBJECTS 

(a1 ) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



£] (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project develops and applies a new geometry of biological shape that gives 
a natural and efficient description to a variety of biological objects at 
vastly different levels: chromosomes, cells, organs, organisms, etc. 

Applications are to (1) automation of shape analysis for diagnosis and taxonomy, 
including those issuing from new devices such as the tomographic scanners, (2) 
the description and understanding of organ and organismic development , and (3) 
the psychology and neurosphysiology of shape processes in vision . 



PHS-6040 
(Rev. 10-76) 



157 



Project Description: 

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 these processes for biological and medical purposes. 

The methods employed stem primarily from a new geometry conceived by the 
principal investigator. It is based on the notion of growth as the primitive 
process. It is being applied to a variety of problems, both to develop new 
mathematics and computer science in new biologically relevant directions and 
to uncover and clarify biological processes taking place. The applications 
include cell and tissue description from light microscopy, shape description 
of developing cells and organs, chromosome description, visual psychophysics 
and visual neurophysiology. 

New computer capabilities and directions have emerged from this work and 
continue to do so. A general program for extracting these descriptions is 
now available at DCRT. The program is currently being rewritten to allow it 
to be adapted to other computers. The mathematics for extending this geometry 
to 3-D data, such as will be coming from the variety of new scanners, has been 
developed to a necessary degree. Computer programs for implementing 3-D 
analysis are being developed. In addition, mathematics for extracting these 
descriptions from gray scale data (for example, slides and X-rays) is now 
being developed. Computer programs for implementing these are also planned. 

This geometry has been applied to the study of growth and development of the 
human mandible. New shape invariants and growth constraints have been found. 
Results from this work have been submitted for publication. 

Publications: 

Blum, H., and Nagel, R. : Shape description using weighted symmetric axis 
features. Pattern Recognition, 10:167-180, 1978. 



158 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this spact) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 
ZOl CT 00008-05 LSM 



PERIOD COVERE^^^^ ^^ ^^^^ ^^^^^^^ ^^^^^ 3^^ ^^^^ 



TITLE OF PROJECT (80 characters or Itss) 
Cluster Analysis 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: M. Shapiro 



Research Mathematician 



LSM DCRT 



COOPERATING UNITS (if any) 



lab/ BRANCH 

Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 



Biomathematics and Computer Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 

1.0 



OTHER: 

0.0- 



CHECK APPROPRIATE BOx(ES) 
i; (a) HUMAN SUBJECTS 

(al) MINORS Q (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



S (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Cluster Analysis and related work has been in three areas: 

1. Continual updating of the C-LAB system for Cluster Analysis. 

2. Research into optimum ways of clustering using the minimal spanning tree 
algorithms. Comparison of these methods with standard hierarchical clustering 
techniques. 

3. Initial development of programs and methods for optimum selection of 
variables in linear regression using branch and bound methods. 



PHS-6040 
(Rev. 10-76) 



159 



Project Description: 

Objectives: 

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

Methods : 

The development and testing of algorithms based on the latest published 
research and extensions to it. 

Significance to Biomedical Research: 

Pattern recognition techniques are now being widely used on biomedical 
data for classifying objects, finding relationships between variables, 
and for processing biological images. These applications of artificial 
intelligence has led to both automatic processing and a better under- 
standing of data. 

Proposed Course: ' 

A wider range of Cluster Analysis algorithms will continue to be developed 
and applied. 

Publications: 

Habersett, M. C, Shapiro, M. B., Bunnag, W. , Nishiya, I., and Herman, C: 
Quantitative Analysis of Flow Microf luorometric Data for Screening 
Gynecologic Cytology Specimens. J. Histochem and Cytochera. , Vol. 27, 
No. 1, pp. 536-544, 1979. 

Cole, B. R., Shapiro, M. B., and Rodbard, D.: Chromatography and 
Electrophoresis Analysis System, in Electrophoresis 78. Elsevier 
North Holland. 1978. 

Marimont, R. , and Shapiro, M. B.: Nearest Neighbor Searches and the Curse 
of Dimensionality. Accepted for publication in J. Institute of Math, and 
Applications. 

Lieblick, A. K. , Symmes, D., Newman, J. D., and Shapiro, M. B.: Development 
of the Isolation Peep in Laboratory Bred Squirrel Monkeys. The Journal of 
Animal Behavior, in press. 



160 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do MOT use this spice) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 
ZOl CT 00011-05 LSM 



PERIOD COVERED ^ ,„,„ 

Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Discrete Mathematics and Applications 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



G. A. Hutchinson 



Research Mathematician 



LSM DCRT 



COOPERATING UNITS (if any) 



LAB/BRANCH 

Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 
Biomathematics and Computer Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

O.A 



PROFESSIONAL: 

O.A 



OTHER: 



0.0 



CHECK APPROPRIATE BOX(ES) 
□ (a) HUMAN SUBJECTS 

n (al) MINORS D (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



^ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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



PHS-6040 
(Rev. 10-76) 



161 



Project Description: 

Objectives: 

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. 

Methods Employed and Major Findings: 

Modules over a ring (a mathematical concept generalizing vector spaces 
and commutative groups) were studied using the computational methods 
and algebraic logic reported in previous fiscal years. Some progress 
was made in extending to related problems the previously-developed 
computational methods for analyzing inclusion relations between subspaces 
of vector spaces. A study of the ring similarities which imply that 
two different rings lead to the same theory of inclusion relations for 
their modules has progressed. The major finding is that only rings with 
prime power characteristic need be considered, in order to resolve the 
problem for arbitrary rings. Both studies above are continuing. Revision 
for publication of the logical theory described in the previous fiscal 
year is continuing. 

Significance to Biomedical Research and the Program of the Division: 

General purpose mathematical techniques and computer programs implementing 
them are made available to the biomedical research community. 

Proposed Course: 

Previous work in vector spaces and logic appropriate to algebraic theory 
will be revised and extended. 

A proposed computer science study involving graphical display of mathematically- 
oriented text and diagrams will be assessed for feasibility. 

Publications: 

*Hutchinson, G. and Czedli, G.: A test for identities satisfied in 
lattices of submodules. Algebra Universalis 8, 269-309, 1978. 

(*Reported as in press in a previous fiscal year.) 



162 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
IHTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00039-02 LSM 



PERIOD COVE^E^D^^ ^^ ^^^g ^^^^^^^ gept. 30. 1979 



TITLE OF PROJECT (80 characters or less) 

Linear Methods in Statistics 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: J. D. Malley 



Staff Fellow 



LSM DCRT 



COOPERATING UNITS (if any) 

None 



lab/branch 

Laboratory of Statistical and Mathematical Methodology (LSM) 

SECTION 

Statistical Methodology Section 



IMSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

1.5 



PROFESSIONAL: 

1.5 



OTHER: 

0.0 



CHECK APPROPRIATE BOX(ES) 
Ij (a) HUMAN SUBJECTS 

n (al) MINORS D (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



£] (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Linear methods in statistics continue to be studied, with the general linear 
model serving as a point of departure. Family confidence limits for ratios of 
sample means from multivariate normal distributions were extended beyond the 
original one-sample result to many-sample equal/unequal covariance data sets. 

Also, in the area of the general linear model, extensive study was made of the 
hazards and appropriate procedures needed for unbalanced data, which is perhaps 
the most frequently occuring type of data to which the linear model is applied. 



PHS-6040 
(Rev. 10-76) 



163 



Project Description: 

The overall objective of this project is the study of linear methods in 
statistical analyses. During the past year, in addition to family confidence 
limits for ratios of normal means, and work in automorphism groups and 
algebras, the linear methods were utilized in the study of log linear models 
for contingency tables. 

Publications: 

Mosimann, James E. and Malley, James D. (1979). Size and shape variables. 
Multivariate Methods in Ecological Work , L. Orloci, C. R. Rao, and W. M. 
Stiteler, eds., pp. 17 (In press). International Co-operative Publishing 
House, Fairland, Maryland, U.S.A. 



164 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00047-01 LSM 



PERIOD COVERED 

Oct. 1, 1978 through Sept. 30. 1979 



TITLE OF PROJECT (80 characters or less) 

Non-numerical programming techniques and applications 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: L. M. Norton 
Other: M. G. Pacak 

M. D. Graitson 



Research Mathematician LSM DCRT 
Computer Systems Analyst LSM DCRT 
Guest Worker LSM DCRT 



COOPERATING UNITS (if any) 



lab/branch 
Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 

Biomathematics and Computer Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, Md. 20205 



TOTAL MANYEARS: 

0.8 



PROFESSIONAL: 

0.8 



OTHER: 

0.0 



CHECK APPROPRIATE B0X{ES) 
~ (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



□c(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Using both general-purpose and special-purpose programming systems, techniques 
for computer processing of non-numerical data are developed and evaluated by 
implementing small-to-medium-scale research projects in selected application 
areas. In the period covered by this report, projects have been initiated in 
the areas of computational linguistics and differential equations . Programming 
systems utilized include PROLOG (a language for logic or rule-based programming ) 
REDUCE (a language for algebraic symbolic manipulation) , and IBM 370 assembly 
language (for efficient processing of larger volumes of data) . 



PHS-6040 
(Rev. 10-76) 



165 



Project Description: 

Objectives: 

To provide programming tools for non-numerical data processing and, 
simultaneously, to obtain research results in applications areas. 

Methods: 

Programming systems and programming techniques are evaluated by 
implementing selected research projects utilizing them, thus concurrently 
performing independent research, and assessing the suitability and 
effectiveness of the programming systems and techniques. 

Major Findings: 

The University of Edinburgh PROLOG system for logic and rule-based programming 
appears to be quite powerful and straightforward to use. It provides a 
natural context for experimentation with and evaluation of heuristic 
programming techniques. Application experience in the subject area of 
computational linguistics has shown that relatively sophisticated routines 
can be implemented quickly and easily. A routine to analyze and paraphrase 
compound terms from medical terminology has been developed and is currently 
being used in a research project. In addition, a particularly powerful 
form of parser, one for "augmented transition net" grammars, is very 
readily implemented using PROLOG. 

A second research project dealing with word fragments of medical terminology 
(both English and French) has been implemented in IBM 370 assembly language 
because of the volume of data involved. As might be expected, the routines 
are very efficient, at a cost of increased programming time. Both this 
and the preceding computational linguistics project are not completed yet, 
but the appropriateness of the programming techniques has been established. 

The University of Utah REDUCE system for algebraic symbolic manipulation is 
being investigated by means of an effort to obtain closed form solutions to 
ordinary differential equations with constant coefficients using inverse 
Laplace transforms. This investigation is so far inconclusive, due partly to 
the inferior nature of the documentation available for the REDUCE system. 
Productive use of a system such as REDUCE requires the establishment of 
appropriate mathematical results on which to base computational techniques 
for symbolic manipulation (just as computational techniques for numerical 
analysis are grounded in theory involving error analysis, convergence, etc.). 
As an example, the optimum form of a table of Laplace transforms to be used 
with REDUCE differs from the form published for manual use. A preliminary 
version of such a table has been derived as part of this research. 

Significance to Biomedical Research: 

Development of effective programming techniques for computational linquistics 
and algebraic symbolic manipulation will make it possible to advance the 
state of the art in automated processing of medical information, such as 
medical records, pathology reports, etc., and other biomedical research data. 

165 



Proposed Course: 

The PROLOG and REDUCE systems will continue to be used and evaluated to 
determine their suitability for non-numerical data processing, and if 
suitable, to identify applications on which to use them. Techniques 
for heuristic programming and for symbolic mathematical computation 
will be developed and evaluated in the context of these systems. The 
ongoing computational linguistics projects will be pursued simultaneously 
for their own sake. 



167 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 
ZOl CT 00013-05 LSM 



PERIOD COVERED 

Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Multivariate Statistical Analysis 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: J. E. Mosimann 

OTHER: M. V. Ratnaparkhi 

J. D. Malley 

D. R. Ratcliffe 



Chief LSM DCRT 

Visiting Associate LSM DCRT 
Staff Fellow LSM DCRT 
CSIRO, Brisbane, Australia 



COOPERATING UNITS (if any) 



None 



lab/branch 

Laboratory of Statistical and Mathematical Methodology 



SECTION 

Office of the Chief, LSM, DCRT 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NIH, Bethesda, MD 20205 



TOTAL MANYEARS: 

1.2 



PROFESSIONAL: 
1.2 



OTHERi 
0.0 



check APPROPRIATE BOX(ES) 
G (a) HUMAN SUBJECTS 

□ (al) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



:□ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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



Multivariate statistical methods (size-shape methods) for analyzing ratios 



which follow a lognormal distribution have been developed. Exact statistical 
tests were developed and applied in two biological studies concerning the 
distribution of schistosome eggs in man at autopsy and morphological measure- 
ments of birds. Publications on these studies appeared during this year. Work 
on the theoretical meaning of size-shape concepts for statistical distributions 
continued, with an extensive study of the transformations connecting the 
commonly occuring size and shape variables (Ratnaparkhi, Mosimann, and 
Ratcliffe) . The relationships among the various definitions of size and shape 
variables in the literature were studied (Mosimann and Malley) and a summary 
of the relationships prepared (see publications) . 



PHS-6040 
(Rev. 10-76) 



168 



Publications: 

James, I. R. , and Mosimann, J. E.: A new characterization of the Dirichlet 
distribution through neutrality. Annals of Statistics , in press. 

Mosimann, J. E., and Malley, J. D.: Size and shape variables. Multivariate 
Methods in Ecological Work , L. Orloci, C. R. Rao, and W. M. Stiteler, eds., 
p. 17, in press. International Co-operative Publishing House, Fair land, MD. 

Mosimann, J. E., and James, F. C: New Statistical Methods for Allometry 

with Application to Florida Red-winged Blackbirds. Evolution, 33:444-459, 1979, 

Mosimann, J. E., Malley, J. D., Cheever, A. W. , and Clark, C. B.: Size and 
Shape Analysis of Schistosome Egg-Counts in Egyptian Autopsy Data. Biometrics , 
34:341-356, 1978. 



169 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER [Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 
ZOl CT 00009-05 LSM 



PERIOD COVERED 



Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Research Topics in Computer Science 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: G. D. Knott 



Computer Specialist 



LSM DCRT 



COOPERATING UNITS (if any) 

None 



lab/branch 

Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 



Biomathematics and Computer Science Section 



INSTITUTE and LOCATION 

Division of Computer Research & Technology, NIH, Bethesda, Md. 



20205 



TOTAL MANYEARS: 

0.6 



PROFESSIONAL: 

0.4 



OTHERj 
0.2 



CHECK APPROPRIATE BOX(ES) 
H (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



E (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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

Currently, an exhaustive survey of storage and retrieval methods is underway. 
This has resulted in several promising hybrid approaches. 

Optimal item orderings in split hashing schemes and certain interesting algebraic 
characterizations of fixed permutation open addressing methods are also being 
studied. Research on trie methods is also being conducted. 



PHS-6040 
(Rev. 10-76) 



170 



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. 

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

Optimal item orderings in split hashing schemes and certain interesting 
algebraic characterizations of fixed permutation open addressing methods 
are also being studied. 

Research on trie methods, which involves storing items in trees so that the 
path to the item is determined by its key, is underway as well. 

A new hybrid method combining binary storage trees and tries has been 
developed. This scheme leads to an interesting statistical application 
as well. This work was reported at a conference this year. 

Publications: 

Knott, Gary D.: "Fixed-Bucket Binary Storage Trees", Proc. 12th Hawaii 
International Conference on Systems S cience , pp. 36-48, University of 
Hawaii, Jan. 1979. 



171 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 

ZOl CT 00010-03 LSM 



PERIOD COVERED 

Oct. 1, 1978 through Sept. 30, 1979 



TITLE OF PROJECT (80 characters or less) 

Nonlinear Equations 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: R. I. Shrager 

OTHER: G. D. Knott 

E. Hill 

J. E. Fletcher 



Mathematician LSM DCRT 

Computer Specialist LSM DCRT 

Mathematician LAS DCRT 

Research Mathematician LAS DCRT 



COOPERATING UNITS (if any) 
LAS, DCRT 



lab/branch 
Laboratory of Statistical and Mathematical Methodology (LSM) 



SECTION 

Biomathematics and Computer Science Section 



INSTITUTE AND LOCATION 

Division of Computer Research and Technology, NTH. Bethesda. Md. 



20205 



TOTAL MANYEARS: 

1.0 



PROFESSIONAL: 

1.0 



OTHEfii 
0.0 



CHECK APPROPRIATE BOX(ES) 
U (a) HUMAN SUBJECTS 

□ (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



□C(c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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. 



PHS-6040 
(Rev. 10-76) 



172 



Project Description: 

Objectives: 

To develop methods for solving nonlinear equations frequently encountered 
at NIH. 

Methods: 

A continuing effort is made to create methods or extend existing methods to 
solve problems in a host of NIH applications, and to house those methods in 
accessible computer programs or routines. Modelaide and MLAB are two examples. 

Major Findings: 

Two revisions of the paper "Non-linear Curve-fitting in the L. and L^ Norms", 
with Edward Hill as co-author, have been submitted to Mathematics of Computation. 
The revisions involved choosing test problems, making runs, and tabulating 
results. No new theoretical results were involved. These methods are now 
coded in SAIL, and a special version of MLAB called OLAB is currently being 
checked, in which L, or L^ norms may be used. 

There appear to be some cases in which the step size control of the MLAB 
differential equation solver chooses extremely small steps. Some heuristic 
is being sought for treating thses cases. 

Significance to Biomedical Research: 

These methods are now being applied to problems in human metabolism, cell 
growth, chemical kinetics, and spectral analysis (UV, IR, CD, ORD, NMR, ESR) . 

Proposed Course: 

As the methods are proved in test and practice, they will be incorporated into 
easy-to-use systems like MLAB, and as a result, the systems themselves should 
evolve to do more useful work with less human and machine effort. 



173 



DATA MANAGEMENT BRANCH 



J. Emmett Mard 
CKTeT 




DMB staff design and oveate aomputer-hased data management 
systems for speaifio users. The above plan reflects the 
program modules of a system developed hy DMB's Applied 
Systems Programming Section for NHLBI. 



I. SUMMARY OF ACTIVITIES 



Functions 



The Data Management Branch (DMB) provides advice and assistance to 
research investigators, program officials and administrators through- 
out 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 for 
specific users and train those users. They also teach courses about 
some data management and programming tools, provide advice on data 
management techniques to NIH programmers and serve as consultants on 
implementation by contractors. Finally, DMB creates and maintains 
general purpose, user-oriented programming tools to speed building and 
improve operation of specific applications systems. 

In FY79 DMB again worked on more than one hundred projects involving 
virtually every Bureau, Institute and Division of the NIH. Almost 30 
percent were projects for applications in patient care, clinical 
research, or epidemiology; about 13 percent for laboratory research 
areas; over 41 percent for program direction, management and 
administration; just over 3 percent for biomedical communications and 13 
percent for development of data processing and analytic tools. 

About 26 percent of the projects required less than 80 hours each of DMB 
work and another 42 percent less than 800 hours each. The partial 
project list in Section II is intended to provide a reasonable sampling 
of the breadth of DMB involvement in the overall NIH mission. The 
following items highlight a few areas of particular interest. 



176 



FY79 Highlights 

As we complete the second full year of our efforts to develop a 
comprehensive data base for NIH administrative data, the impact on both 
Materiel and Financial Management has been impressive. 

The Materiel Management System (MMS) is capable now of (1) handling all 
purchases, receiving and payment functions at point of transaction, (2) 
making all entered data available to all central NIH organizations, (3) 
providing both regular and exception reports in on-line and batch modes, 
(4) automatically preparing file reference and reminder information for 
out-of-sequence events and (5) generating purchase order documents and 
tracking purchase order number assignments. We have identified several 
enhancements to the existing system which will be implemented in stages 
as they are completed during the next four months. 

The Central Accounting System (CAS) progress has been a bit slower in 
its gradual movement toward a data base environment. The primary reason 
for this is the existing CAS which must be maintained and has required 
modifications to integrate its processes with the MM data base. The 
necessary modifications to the existing CAS have been completed. The 
MMS is automatically generating all relevant obligation, accrual and 
payment transactions to the CAS and the CAS is capable of processing and 
feeding back actual payment data to the MMS. During this MMS/CAS 
integration process we have been able to identify the proposed 
methodology for handling the CAS data base environment and we are now 
designing and preparing to implement the first version of a complete 
MM/CAS data base. 

To demonstrate the utility of the data base approach to NIH 
administrators, a pilot project was implemented with the NCI. This 
system can track actual obi igations-to-date entered in MMS against each 
basic purchase agreement and indefinite delivery contract. The software 
was developed for general use and with only minor modification can be 
used by any program official. Currently the DCRT Information Office is 
planning a briefing on this system for the AO's. 

In July, 1978, the Clinical Support Section hired an unusual computer 
technician with the understanding that he perform all of his work from 
his bed at home. Rick Pilgrim is a quadraplegic, who after extensive 
assistance from the Virginia State Rehabilitation Service learned to 
program by way of a voice data entry system. When his new section head 
hired Rick, he also assumed the task of adapting this voice entry system 
to function with the NIH IBM-370 system. 



177 



A collaborative effort between the Clinical Support and the Software 
Support Sections resulted in a vocabulary, a training program and an 
interface program which enables Rick's voice communication with the NIH 
computer facility. Although this effort was initiated primarily for 
Rick Pilgrim, we are now examining alternative voice entry hardware and 
potential new applications for this approach. 

Immediately after implementing the Cardiology System for the NHLBI in 
FY78 the benefits of combining its data with that of the Cardiac Surgery 
System became apparent. Research analyses relating to the same patient 
or group of patients could be accomplished much more quickly and 
comprehensively and there also was some potential for a reduction in 
file maintenance. After obtaining agreement between the Cardiology and 
Surgery Branches, the DMB created the combined file in early FY79. 

On June 1, 1978 the Data Management Branch began full support for the 
MarklV File Management System. MarklV is structured to be used as a 
function of the system design phase and provides the full range of file 
maintenance, retrieval, computation and report generation in both batch 
and on-line environments. The introduction of this system at the NIH 
should result in a significant reduction in the time required to move 
from design to actual implementation of complete file management systems 
for many NIH computer users. 

The initial version of a standard package for on-line, interactive 
scintillation counter data reduction was completed this year and several 
organizations have already begun to use it. This package, called SCINT, 
is designed to be used by the non-programmer and can be used as a 
comprehensive tool for analyzing experimental data. It also interfaces 
easily with other more powerful systems for analyzing experimental data 
such as BRIGHT and MLAB. The addition of SCINT complements the standard 
data logger approach taken by the Computer Systems Laboratory. Thanks 
to the availability of both techniques scintillation counter users can 
now record, review and analyze results in an accurate and facile manner. 



178 



Future Plans 

The Materiel/Financial Management data base effort in the coming year 
will provide (1) a complete source data entry procedure for basic 
purchase agreements and indefinite delivery contracts, (2) an invoice 
entry /accounts payable subsystem, (3) an improved requisition tracking 
methodology, (4) additional departmental contract information system 
data, (5) an automatic receiving function, (6) additional data display 
functions for the BID's and (7) full data base update of open document 
records for materiel management transactions. 

Voice answerback and voice data entry will be studied for possible 
application to current and future application developments within the 
DMB. 

During FY80, the Clinical Information Utility will (1) expand its data 
coverage, (2) improve data access, display and analysis functions and 
(3) be considered for possible interface with other clinical data 
processing systems. 



179 



II. ANNOTATED PROJECT & ACTIVITIES LIST 



The list below does not include a number of small new projects and 
modest revisions and additions to existing systems. These become 
literally to numerous to mention in an annual report although each is 
clearly important to the client and requires careful work by the DMB 
staff. 



• Clinical Research Patient Care Epidemiology 



The Clinical Information Utility System is an archive for data recorded 
by computers in the Clinical Laboratories of the Clinical Center and by 
its Medical Information System. The CIU has evolved over the last seven 
years under the design and operation of the Clinical Support Section, 
DMB. 

The Section made improvements to the CIU during this fiscal year. 
First, the Bone Marrow and Microbiology Subsystems were redesigned and 
developed to reduce the costs of both maintaining and retrieving from 
their data bases. The cost reductions for these operations were 
approximately 60 percent. Secondly, the CIU Inverted File was expanded. 
Branch and Institute were added as indices to the file for all the 
organizations which admit patients in the Clinical Center's Hospital. 
The expanded indices make it possible for a retrieval to be done by 
organization, sex, race, test results, body source, specimen and 
diagnoses. At the request of the Deputy Director of the Clinical Center 
work began to develop chronological summary reports of all clinical 
laboratory tests for hospitalized patients. 



An Qn-Line Data Management System for NIMH Patient Care called DMS was 
completed during the year. DMS programs are general programs which 
execute under TSO (Time-Sharing Option) on the IBM System 370. The 
programs are interactive; i.e., they interpret and respond to commands 
entered by the user at a computer terminal. 



180 



DMS is different from many other facilities in several ways: (1) A data 
dictionary is used to contain all file-dependent information (e.g., 
variable names, types, lengths, locations, legal values, formats, etc.). 
Consequently, DMS programs are independent of user files. (2) Programs 
are available to prospective users to initialize the system without 
consultation with DCRT personnel. (3) Master files have' a direct 
organization so that records may be accessed individually. (4) User 
profiles may be defined to restrict access to variables in files. 

DMS was designed for data management with small scientific data files. 
It can be used to enter data, update data, retrieve data, and perform 
statistics. Data can be entered in free format or fixed format, 
directly from a terminal or indirectly from a data set. Individual 
records can be added, changed, or deleted. Data is tested for legality 
on entry, and all transactions are posted to a log file. Data can be 
retrieved and listed at the terminal or written to an on-line data set; 
the conventional arithmetic, relational, and logical operators can be 
used to define sets of records to be retrieved. Statistics can be 
performed on-line or in the batch mode using SAS; DMS will set up SAS 
jobs from specifications entered by the user. 



The Cancer Sur vival System was originally developed in the late 1960 s 
to support the End Results in Cancer studies of NCI. Maintenance and 
improvement of the system is now the primary goal. Catherine Creede of 
the Scientific Applications Section is responsible for the continuing 
support of this system. 

A number of operating system related problems which surfaced during the 
past year were corrected. 

During FY79 the system was sent to the College of Medicine at the 
University of Iowa, the Virginia State Tumor Registry, the Medicare 
Bureau in Baltimore, Md., and the Department of Statistics at George 
Washington University. 



The data base for Work-Able Cancer Patients Employment Studies, NCI 
involves pooling of numerous sub data bases prepared by nve contractors 
in different locations for tumor registry. Each contractor data base 
includes survey questionnaire responses from cancer patients and their 
employers. Vigorous editing of the data revealed some inconsistencies 
among contractors with regards to number of options in questions, type 
of questions asked and type of codes to determine certain values. The 
contractors were notified and necessary adjustments were made to 
standardize the data. The process of merging the data sets is underway. 
During FY80 analysis will be completed to determine variation in 
patient/employer responses by geographical area. 



181 



The Adriamycin Toxicity Study (Analysis of Risk Factors for the 
Development of Adriamycin (A) - Induced Congestive Heart Failure) 
was begun in FY77 for the Cancer Therapy Evaluation Program, 
NCI. Dr. Van Hoff was the Principal Investigator and was supported by 
Dennis George and Peter Basa of the Documentation and Systems Support 
Section. During FY79 data entry and initial analysis for U.S. patients 
was completed. Further analysis of the data precipitated the 
development of three papers for presentation and publication by Dr. Van 
Hoff. The system was used to provide information in both tabular and 
graphic form for dosage levels and schedules of administration by age 
group, sex and race. Dr. Van Hoff left the NIH during the FY and 
anticipated future work may be delayed on this project pending 
agreements between him and the NCI. 

Reports were produced to provide initial analysis of designated risk 
factors such as previous cardiac diseases, tumor type(s), therapy 
(adriamycin and/or concomitant drugs/radiation) and patient condition at 
start of adriamycin treatment. 



In FY78 the Cardiology Branch Data Processing System became operational. 
Prior to that the Heart Surgery Branch Data Processing System had 
been operational for a few years. It became apparent that the heart 
surgery and cardiology data were complementary and involved the same 
processing approaches. It was therefore decided to develop a Combined 
Cardiology/Heart Surgery Data Base . This would provide a complete 
chronological record or activity of NHLBI Cardiology and Heart Surgery 
Branch patients. This effort involved logically merging the two data 
bases and developing and implementing computer programs that duplicated 
the capabilities of the two separate systems but provided the advantage 
and simplicity of a single system and data base. Retrievals from the 
common data base are enhanced since selections can now be logically 
based on both surgery and cardiology data. The combined system became 
operational in FY79 and has proven to be an asset to NHLBI physicians 
and researchers. 



In support of the Applied Systems Programming Section in FY79 for the 
Cardiology System , John R. Parks of the Software Support Section created 
tailored generator software. Special versions of TRANSACTGEN and 
REPORTGEN were written permitting generation of many source COBOL 
subroutines. 



182 



Giant Foods is conducting a " Foods for Health " program in cooperation 
with NHLBI. Stores in Washington participating in the program are being 
compared with a control group of stores in Baltimore to see if 
customers' buying habits are changing in response to the campaign. Data 
for fourteen food groups are being analyzed. Diane Feskanich, 
Scientific Applications Section, has developed a method of reformatting 
the data as recorded on tapes received from the stores and preparing it 
for input to a time series/analysis of variance program. She has 
produced summary statistics and scatter plots which provide a visual 
representation of buying habits over time. 



During FY79 the Analysis of Systemic Lupus Erythmatosis (SLE) Nephritis 
Patients System was developea to store chemistry ana therapy data on all 
SLE patients. These patients participated in the Arthritis and Rheumatism 
Branch studies of SLE, from November 1968 to present. Patient data is 
stored by categories; i.e.. Dating of Clinical Manifestation, Therapy, 
Objective Information, Lab Data, Complications, etc. Each category of 
information contains sub-categories of information. For example, under 
the "Objective Information" category, the sub-categories include: skin 
biopsy, parotid flow rate, lip biopsy, parotid scan, brain scan, bone 
marrow, etc. The system also allows multiple entries of all items. 

The clinical chemistry data, urine data and the hematology data for each 
participating patient are supplied by the Clinical Support Section. 

A subsystem was designed and implemented to format and verbally describe 
the content of each patients' updated file. This provides an easily 
available reference for doctors. 

Analysis is underway on a group of patients with SLE participating in 
the immunosuppressive trials and assigned to receive either prednisone 
only or the combination of prednisone and cyclophosphamide. The 
variables that will be examined include serum creatinine, creatinine 
clearance, 24 hour proteinuria, DNA binding and C'3. 



Throughout FY79 George Roberts of the Applied Systems Programming 
Section has worked closely with Dr. Knowler in the Southwest Field 
Studies Section, NIAMDD, to systemize the several separate steps 
involved in his analysis of increased incidence of retinopathy in 
diabetic patients with elevated blood pressure. The Diabetic 
Retinopathy Data Analysis System has been quite successful and the 
procedure is being examined to determine if it can be expanded to 
evaluate more and different variables from different data files. 



183 



The Study of the Incidence and Prevalence of Kidney and Urinary Tract 
Diseases in the Armed Forces was begun in 1974 for the Kidney Disease 
Program, NIAMDD, to evaluate the occurrence, morbidity and mortality of 
kidney and urinary tract diseases in an effort to determine research 
needs. The principal investigator is Dr. Nancy Cummings. The project 
leader for the system is John A. Haggerty, Jr. and the principal 
programmer is Darius Georg of the Documentation and Systems Support 
Section. Data for the study was made available by the Air Force, Army 
and Navy. There has been difficulty in obtaining correct data from the 
Anny. We received our fourth set of data from the Arny during FY79 and 
ran a series of reports. The data involved additional diagnosis codes 
and covered the years January, 1971 through June, 1977. The original 
data requested was for the period of January, 1971 through December, 
1973. Present plans are to prepare reports using the Army data. After 
this is completed a decision will have to be made to either request the 
additional data from the Air Force and Navy or eliminate the additional 
data received from the Army when comparing data for the three services. 
Evaluation of reports prepared using Army data are being made by Colonel 
Hodder of the Anny who has become a collaborating principal investigator 
of the study. 



The Kidney Transplant Histocompatibility Study was organized in 1972 to 
determine the efficacy of tissue matching in human kidney transplants. 
The primary objective of the study is to examine the relationship 
between HL-A matching and the success of renal transplantation. Since 
August, 1974, the Biometrics Division of the Naval Medical Research 
Institute has, under contract to NIAID, served as the Statistics and 
Data Management Center for the study. Some preliminary analysis of the 
data has taken place. The study is in the process of being transferred 
from the Naval Medical Research Institute to the NIAID. In November of 
1978 a KTHS Data Analysis Protocol was prepared, which at present, 
lists 48 proposed areas of analysis. These are to be performed by means 
of frequency tables, actuarial life table analysis, and multivariate 
analysis based on application of the Makeham model developed by Drs. 
Homer and Bailey of NMRI and included as a procedure in the Statistical 
Analysis System. Catherine Creede of the Scientific Applications Section 
is involved in the programming support of the analysis mandated by the 
protocol. 



184 



In collaboration with Dr. Eugene Harris, Chief of the Laboratory of 
Applied Studies, DCRT, George Shakarji conducted an analysis of a set of 
data from a 1975 Study of Chemistry Data from Survey Participants by the 
College of American Pathologists. Analysis of variance and 
non-parametric ranking tests were conducted on calcium, glucose and 
chloride test values to determine trends among survey participants. 



In collaboration with Dr. Eugene Harris, Chief of the Laboratory of 
Applied Studies, DCRT George Shakarji continued work on the Use of 
Statistical Models of Mithin-Person Variation in Long-Term StlTdies of 
Healthy Individual sT This included analysis of 14 biochemical and 7 
hemotological test values obtained from 115 healthy men and 47 healthy 
women who participated in a community-wide program of biochemical 
profiling and health surveillance in San Francisco. 

In an effort to determine patterns, forecast ranges were computed using 
statistical models to evaluate the validity of projected hypotheses. 
Trend data among tests were also compared to determine whether 
consistent patterns were detectable. 



In collaboration with Dr. Eugene Harris, Chief of the Laboratory of 
Applied Studies, DCRT, George Shakarji conducted the Temporal Changes in 
the Concentration of Serum Constituents in Healthy Men: Distributions of 
Mith-in Person Variances and their Relevance to the Interpretation of 
Differences Between Successive Measurements Study . The study involved 
37 healthy males, aged 23 to 57 years, each of whom contributed 19, 20 
or 21 weekly blood specimens (mean 20.2) during a five month period. 
Long-term analytic variation was eliminated by storing the serum 
specimens at -20 degrees C and then analyzing them all together in a 
single run on a Vickers M-300 multi-channel analyzer. Analytic variance 
within the run was monitored in a conventional manner. The data base 
comprised approximately 7400 assay results. 

During FY79 programs were written to perform time trend analysis using 
the mean-squared successive differences approach. While models have 
been developed to define patterns of the data, unexpected fluctuations 
have occurred. We are now attempting to both determine the causes of 
and predict a pattern for these fluctuations. 



185 



In collaboration with Dr. Eugene Harris, Chief of the Laboratory of 
Applied Studies, DCRT, George Shakarji examined the Use of the 
Population Distribution to Improve Estimation of Means in 
Epidemiological Studies . The properties of some statistical 
calculations which would help to improve the accuracy of observed values 
in clinical trials as estimates of the true values of individuals. The 
procedure is simply that of replacing the observed value by its weighted 
average with the group mean. 

The Empirical Bayes' approach, applied in this study, aims at improving 
the accuracy of estimating an individuals true value by a weighted 
average of a single measurement of the individual and the mean of such 
measurements over the entire group of individuals sampled. 




At the request of Br. Griff Ross, Deputy Director of the 
Clinical Center, work began to develop chronological 
summary reports of all clinical laboratory tests for 
hospitalized patients. Dr. Ross (right) explains his 
need to reduce the size of medical records through 
computerization, to DMB Clinical Support Section Head 
William Vincent. 



186 



• Laboratory Investigations 

The In Vitro Information System (IVIS) was developed to process 
information about In Vitro carcinogenesis tests conducted under the 
auspices of the Division of Cancer Cause and Prevention of the National 
Cancer Institute. The sponsor was Dr. Virginia Dunkel and the principal 
investigator was Mary Linhart. Joseph Campbell of the Software Support 
Section designed the system. The system provides for the collection, 
maintenance, analysis and reporting of In Vitro data. Data for the 
system will come from contracting laboratories on input forms with both 
descriptive and results data. The input forms will be converted into 
machine readable format and be processed on the Division of Computer 
Research and Technology (DCRT) computing facility using the proprietary 
MarklV system installed in February, 1978. Development of a system 
using MarklV was completed and turned over to the user during FY79. The 
MarklV system was also used to develop and implement the mammalian study 
for the NCI and was turned over to the contractor during the year. 



In FY79 the Clinical Studies Section in the Laboratory of Viral 
Carcinogenesis, NCI asked the Data Management Branch to analyze their 
current data processing capabilities and approaches to see how they 
could be improved upon by way of a newly designed computerized 
Seroepidemiology Data Processing System . 

Judy Mahaffey of the Applied Systems Programming Section assumed 
responsibilities for the study. The results of the analysis along with 
a recommended systemized approach to regaining control of the data 
processing and making the data much more accessible and usable was 
presented to NCI in the form of an analysis report/proposal. The 
report/proposal was well -received and the system design ideas accepted. 
It is anticipated that Mrs. Mahaffey will be responsible for 
implementation of the major portion of the proposed system. 



Late in FY79 the Data Management Branch was asked to design a 
computerized Wild Asian Mouse Breeding Colony Data Processing System for 
maintaining data tor a taci Mty which breeds and raises a rare, valuable 
strain of wild Asian mice originally acquired in Vietnam. The system is 
being designed for the Laboratory of Viral Carcinogenesis, Division of 
Cancer Cause and Prevention, NCI by Vivian Pel ham of the Applied Systems 
Programming Section. The system will maintain virtually all data 
collected on the animals. Information from the system will aid in mate 
selection, etc. The system is at present in the initial design phase. 
Development will continue into FY80. 



187 



The Genetic Marker Tracing System for the Laboratory of Immunogenetics, 
NIAID, involved the analysis, design and creation of a system to provide 
the capability for tracing genetic markers within a colony of rabbits. 
The principal investigator for the system was Dr. John Sogn, LIC/NIAID; 
the Documentation and Systems Support Section with Dennis George as 
project leader and Jane Blessley assisting, provided the analysis, 
design and creation of the system which was turned over to the user 
during the year. 



The Data Management Branch was asked by the Veterinary Resources Branch, 
DRS in the Poolesville Installation to design a computerized Canine 
Breeding Colony Data Processing System to assist that facility with 
tneir recora Keeping ana work scnedunng. The proposed system should 
produce many reports to be used by VRB personnel to help them with their 
daily work. The project was taken over by Roger Dai ley of the Applied 
Systems Programming Section. Pre-design analysis has been completed and 
a design proposal is being prepared. 



The Reproduction Research Branch of NICHD employs various physical/ 

chemical methods of protein separation in its work. It is possible to 

obtain, across time, quantitative data on the position, bandwidth, 

peaks, etc. of the proteins under study. In January of 1977, work began 

on the Chromatography and Electrophoresis Peak Analyzer System for 
automating the analyzer ot this data. 

Working closely with Dr. David Rodbard, RRB, Brian Cole of the 
Scientific Applications Section developed an interactive system for 
fully automatic peak detection and characterization, non-linear curve 
fitting of overlapping Gaussian distributions with an exponential 
baseline, quantitative measuring of the resolution between any two peaks 
for optimizing the conditions of resolution, and cut-point purity and 
yield projections useful in preparative fractionation. All this has 
been designed into a program which can interface with a non-computer 
oriented user, with a degree of automation and ease and convenience of 
operation never before achieved in a peak detection program. 

During FY79, this system (designated "CEPHAS") has been enthusiastically 
received by a growing number of research scientists at NIH and copies of 
the system have been sent to researchers at MIT, Texas A&M, the 
University of Washington, Boston University Department of Chemistry, 
Baylor College of Medicine and the University of Maryland. Dr. Rodbard 
with Mr. Cole's assistance, has written an article which illustrates the 
value of the computer system, to be published in the international 
technical journal STERIODS. 



188 



Now that a data logger is available to standardize liquid scintillation 
output for easy input to the computer, Brian Cole, of the Scientific 
Applications Section, is developing on the PDP-10 the Liquid 
Scintillation Counter Package System which can answer the needs of a 
reasonably large percentage of scientists who use liquid scintillation 
counter reduction. 

During FY77-78 Mr. Cole produced a working version of an interactive 
self-teaching system which handles most of the problems encountered in 
Chromatography and Electrophoresis. It provides both simple, fast 
graphs of experimental results at the user's display terminal or 
teletype (for quick review and as an aid in planning the next phase of 
an experiment), and high-quality graphs produced offline by the computer 
center's Calcomp plotter (suitable for publication). During FY79 work 
has been completed on expanding the system to provide channel-ratios, 
dual-label experiment calculations and other basic scintillation counter 
reduction. 

Documentation for the system was completed and the system was announced 
to the users of the NIH PDP-10 in April, 1979. Several people at NIH 
have successfully used at least some portion of the system. People 
consulted during the development of the system have been contacted again 
to assess the need for further development of the system at this time. 
Mr. Cole is currently in the process of training new users of the 
system. 



189 



• Administrative and Program Management 



In 1978 when the Data Management Branch began the New Developments in 
the Central Accounting System they assumed responsibility for the NIH 
Central Accounting System; two areas were identified to immediate 
attention. System testing required use of the production software to 
verify accuracy of new additions or changes to old procedures and there 
was no standard document for the year end processing procedure. Both of 
these areas left the maintenance team and the system unnecessarily 
vunerable to error. 

In FY79 the Financial Support Section (FSS) designed and established a 
small test model of the Central Accounting System (CAS). Test system 
master files were created from a representative sampling of the various 
types of records in the CAS. Current CAS report programs have been 
modified and catalogued to provide step by step reporting of the actions 
of the test system. The test system provides the Financial Support 
Section with the ability to fully test proposed changes and additions to 
CAS without interrupting the on-going activities of the financial 
management at NIH. 

The Financial Support Section prepared a manual to be used as a basic 
reference for personnel involved in the fiscal year end closing of the 
NIH Central Accounting System. The manual contains a calendar of events 
to occur each day with notations describing special handling. This 
calendar can be reused year after year with only minor changes. The 
sequence of events does not change. It also has chapters to describe 
each of the processes required at year end. Each chapter contains a 
narrative description of the process and the programs to be executed. 
There are also file descriptions, run and restart instructions and 
modifications, if necessary from year to year. 



The Financial Support Section (FSS) received over 30 requests from the 
DFM for Modifications to the Central Accounting System to modify 
processing and to provide new or modified reports. TlTe FSS collaborated 
with the Program Support Section, CCB, in identifying potential areas 
for improvement in daily, weekly and monthly procedures. Each of the 
recommended improvements were analyzed and, where feasible and 
reasonable, were implemented. In addition, the FSS reviewed its 
requirements for improved maintenance and use of on-line (dedicated), 
card and tape storage facilities. This review resulted in several 
modifications which improved both the maintenance function and the 
efficiency of daily processing. A summary of these activities follows. 



190 



Recurring production jobs were reworked to save on cost and time and to 
eliminate the possibility of errors. This process included: 

(1) Changing the blocksize of files to reduce 
required dedicated disk space. 

(2) Combining functions of two programs into one 
program. 

(3) Reordering steps in the jobs according 
to files being used. 

(4) Replacing tapes with on-line disk space 
for passing data between steps. 

(5) Renumbering the steps in the jobs using 
unique and meaningful numbers. 

When the laser printer became available, a study was conducted by The 
Division of Financial Management and the Financial Support Section to 
determine current reporting needs. It was decided that the daily 
status reports should be printed on the laser printer. Other reports 
to be kept for historical purposes were to be produced on microfilm 
or microfiche. The number of hard copy reports was greatly reduced. 
In FY79 v/e changed 3 reports to microfilm, 25 to microfiche and 26 
to laser printer. This effort significantly reduced the special 
handling required for CAS output. 

The Financial Support Section addressed several other problem areas 
in the Central Accounting System and made the following modifications. 

(1) Setup our own procedure library. 

(2) Put programs on WYLBUR under project 
initials and thus eliminated the need 
for punched card maintenance. 

(3) Established use of DSSUBMIT to get jobs 
run without conflict. 

(4) Replaced parameter cards with one parameter 
data set for all steps within a job. 

(5) Arranged to have Treasury accept labeled 
tapes from NIH. 

Some other changes to the CAS were: 

(1) Added Vendor data base to CAS where 
needed to verify EIN's. Removed the file 
GEN. VENDOR and all references to it. 

(2) Added an output file to CAS to report 
to the Materiel Management System all 
payment activity. 

(3) Added Wylbur data set to the monthly 

travel system programs to permit the Division 
of Financial Management Travel Section to flag 
travel advance records on which they have 
received repayments not yet posted to the CAS. 

(4) Reformatted the two column vendor name and 
address report to three columns. 

191 



The Office of the Secretary, DHEW, requested that the NIH provide 
Special Reports From the Central Accounting System for a study on 
commissions, ihis required a review or an od ligations and 
disbursements made with CAS since FY73. The Division of Financial 
Management (DFM) requested assistance and Clare Hoover of the Financial 
Support Section assembled information from the historical tape files of 
the CAS and produced the necessary disbursement reports. Several 
additional reports were provided to DFM to verify the reported 
information. A system for Quarterly Reporting of transportation payment 
data to the General Services Administration was designed for the 
Division of Financial Management by the Financial Support Section and 
implemented starting with the first quarter of the fiscal year. The 
system provides DFM with a file containing the total amount of payments 
made on Transportation Schedules by NIH to each Employer Identification 
Number (EIN) during the quarter being reported. Two reports generated 
by the system help the Division of Financial Management staff to answer 
questions from carriers concerning payments which were made to them by 
NIH and to combine records on the payment file if two EIN's in CAS 
equate to one carrier. 



The Financial Support Section is responsible for providing an Interface 
between the Materiel Management System and the CAS at NIH. At the 
beginning or hY/y, MMb was generating transactions to record obligations 
in CAS. By the end of the year CAS was also processing commitment, 
accrual and payment transactions from MMS and work has been started on 
projects involving advance payments, partial payments and an invoice 
entry system. 

During FY79 the Central Accounting System (CAS) was modified by the 
Financial Support Section to accept accrual and payment transactions 
from the Materiel Management System (NWS). In addition several other 
modifications were required to (1) duplicate the timing of the on-line 
MMS system in the CAS batch processing mode, (2) notify MMS of payment 
and other transactions which bypassed it for various reasons and (3) 
accept commitment transactions as a function of end of month reporting. 

Verification of accurate processing was provided by the new CAS test 
system and all integration appears to be functioning well in the 
production mode. 

The Financial Support Section reviewed the specifications for batches of 
transactions from MMS to verify that the values of the data elements 
within the records were compatible with the demands of CAS. The CAS 
Test Model was useful in this review. 



192 



The Materiel Management Billing System provides a method for billing 
service and supply fund users for their Materiel Management service 
charges. It runs at the beginning of each month using the CAS Monthly 
History files from the previous month as input. A billing tape is 
created covering Central and Self-Service Stores, contracts and small 
purchases. The system creates and updates year-to-date files containing 
niBiiber of procurement actions, total dollar value of procurements and 
dollars charged for processing these actions. Six reports are generated 
by the system for use by the Service and Supply Fund staff and some for 
distribution to the BID's. 

There have been frequent small changes to the system this year. Service 
and Supply Fund management has requested that the system be run, to 
reflect activity thru the 22nd of each month. It will be necessary to 
change the source of information used to run the system and some 
programs will have to be changed. 



The Materiel Management System is a computer based NIH-wide system 
designed and implemented for the NIH Office of Administration by the 
Data Base Applications Section in collaboration with the Data Base 
Administrator, DMB. The project leaders are Marvin Katz and Ron Wicks. 

The system operates primarily in an on-line environment utilizing data 
base management technology, namely IBM's Information Management System 
(IMS). In addition to the on-line functions, the system provides for 
periodic batch-processing functions to meet operational, accounting and 
management control requirements. 

FY79 saw significant expansion of the Materiel Management System. Phase 
I, procurement control, successfully went into production last year. In 
early February Phase II was placed into operation. This phase permits 
on-line entry of receiving data as well as the capability of processing 
requests to produce reports in support of order payment. The total 
cycle from order entry to receiving and finally payment was thus 
completed with the flow of accrual and disbursement transactions to the 
Central Accounting System. 

An exception reporting system was also implemented to improve control of 
operations in the receiving and accounts payable areas. Included in 
these reports is the automatic generation of overdue receiving letters 
to the vendors. 



193 



The principal analyst/programmer for the Materiel Management System 
(MMS-Query and Reports) is Jane Blessley of the Documentation and 
Systems Support Section. The original goal of this project was to 
determine the feasibility of using MarklV as an economical alternative 
to IMS Inquire for reporting information from the ^WS data base. MarklV 
has proven to be very cost effective when compared to Inquire. 
Currently, MarklV is used to provide most of the ad hoc reports produced 
from the MMS data base as well as 11 daily, 5 weekly and 3 monthly 
reports. In addition MarklV is also used to provide ^WS data to the 
Contract Information System, Document Numbering System, Requisition 
Tracking System and the BPA/IDC Balance System. Future plans to 
optimize the production of reports from the MMS system will involve the 
use of MarklV in combination with other programs. 



During FY79, Catherine Creede, Scientific Applications Section, 
completed the Workload Statistics System which provides an analysis of 
the workload of the Accounts Payable Section, DFM. The workload 
statistics programs provide a number of tabular reports which display 
quantities, equivalent units and man-hours for 15 types of invoices 
received and processed by APS during a monthly period. These reports 
are statistical analyses by work-station, team, unit, section and 
commercial unit. In addition to the printed reports, a new balance 
table is produced which is used as input to the next month's run. DFM 
may run reports monthly, weekly and on request. 



Ronald Edwards of the Scientific Applications Section wrote the Leave 
Liability Reconciliation program to generate information needed to 
reconcile the general ledger unfunded accrued leave and reserve for 
leave liability for annual leave and compensatory time against the 
subsidiary civil service and commissioned officers records. The 
Division of Financial Management will run the program on an as needed 
basis. Results are recorded on microfiche. 



Darius Georg of the Documentation and Systems Support Section is working 
with the principal investigator. Dr. Rhode of the 0/D office in 
developing the Bioscientists Information System to create and maintain 
files of information on people qualified to serve as members of NIH's 
Initial Review Groups (IRGS) etc. This information will be used by the 
Office of the Director and other BID's. 



ia4 



The principal analyst/programmer for the Committee on Academic Science 
and Engineering (CASE) Reports project is Dan us beorg oT the 
UdCUttlfehtatldH and systems support Section. As a result of the 
recommendations of the CASE, these reports have been produced annually 
since 1965. The reports are prepared for the Division of Research 
Services, Office of Program Planning and Evaluation (OPPE). They 
summarize DHEW awards to Institutions of Higher Education, Health 
Professional Schools, Non-profit Hospitals, Non-profit Research 
Institutes and Operating Foundations, and Research and Development 
Centers. The data for FY79 has all been received, edited and balanced, 
and copies sent to the National Science Foundation (NSF). The program 
developed during the last fiscal year to match previous years data with 
the incoming data was used on all data this year and proved to be very 
successful in that a number of reporting component organizations were 
alerted to gross discrepancies and were able to recheck their records 
and either resubmit new or amended data. Final reports should be 
completed during this fiscal year. 



Darius Georg of the Documentation and Systems Support Section developed 
the System for Controlling and Monitoring Complaints of Discrimination 
at NIH tor the Division or Lquai opportunity, me data Dase contains 
information on formal and informal complaints of discrimination at the 
NIH to: (1) provide statistics associated with processing complaints in 
a more timely fashion and (2) enable the user to more closely monitor 
status of complaints. The system was implemented during the year for 



formal complaints, 



The Office of Program Analysis, DRR, requires an Integrated Information 
Management System for handling program data. In FY79, DRR requested a 
complete systems redesign. They would like more capabilities than the 
original design allows (e.g., interactive querying, plots, etc.) plus a 
more manageable update/retrieval scheme, better correlation of 
scientific and administrative data and shorter response time for filling 
requests. 

Sigurd Knisley of the Scientific Applications Section, began the overall 
systems analysis last year. During FY79, systems design and 
implementation have been pursued. A number of data handling systems 
were explored and MarklV was selected as offering the best combination 
of capabilities for solving DRR's information handling needs. Meetings 
have been held with DRR program managers to determine their specific 
requirements and retrieval specifications have been written for those 
reports which have been identified. The file structures have been 
devised and data elements defined. Conversion from the old "Integrated 
Management System" format to MarklV format was done. MarklV retrievals 
based on the converted files have been supplied. 



195 



Plans for FY80 include supplying report capability for coordinated 
files, exploration of interactive retrieval and design, and 
implementation of data collection capability. 



The computerized Review and Evaluation Branch Grants Information 
Processing System was designed and implemented for the Division of 
Cancer Grants, NCI, by Penny Brogan of the Applied Systems Prograimiing 
Section. The effort has extended over a four year period. The system 
provides for (1) data input from IMPAC and CRISP as well as data entered 
by RAEB, (2) file maintenance for extramural funded and unfunded grants 
and contracts, (3) several reports and (4) flexible retrievals which 
make it possible to select on scientific codes as well as on many other 
variables. 

With knowledge gained through her experience with the NCI RAEB system, 
Mrs. Brogan has been able to consult with and advise two other 
institutes on improving their internal grants and contracts management 
systems. 



Peter Basa of the Documentation and Systems Support Section provided 
analysis, design and implementation of a computerized Rental Information 
System to replace the DRS manual inventory control and accounting 
system. The purpose for establishing the system is to provide better 
control of rental equipment inventory and more timely reports. The 
system has been operational since early FY79. 



The initial phase of this automated Survey of Biohazardous Materials 
S ystem for DRS was designed to print a questionnaire that was sent to 
all of the NIH labs. Data was collected for each employee who is 
working with any biological agents, tissue cultures, etc. listed in the 
questionnaire. An on-line data collection program has been written and 
tested. All reports requested were completed during the year. This 
project was done by John A. Haggerty and Charles Twigg of the 
Documentation and Systems Support Section for the Principal 
investigator. Dr. John Irwin of ESB/DRS. 



196 



Charles Twigg of the Documentation and Systems Support Section is 
working closely with both the sponsor, Mr. Ralph Van Wey of the OD/DRS 
Office and the principal investigator, Mr. Zoon of the RSB/DRS. The 
Radiation Safety Control System for DRS will be composed of the 
following subsystems: 

(1) Inventory and Bioassay 

(2) Training 

(3) Laboratory Survey and Airborne Release 

(4) Waste Processed and Activity Balance 

(5) Film Badges 

The total system should be an integrated system capable of supplying 
data from one subsystem to another in order to produce desired reports. 

The analysis and programming for the Inventory and Bioassay subsystem 
has been completed. The system was put into production as of March 1, 
1979. 

Analysis and programming of the training subsystem will not proceed 
until the Inventory and Bioassay subsystem has become operational. The 
Laboratory Survey and Airborne Release/Waste Processed and Activity 
Balance subsystems have not yet been defined. 



197 



• Biomedical Communications 



In FY79 the Library Circulation System , designed by Judy Mahaffey of the 
Applied Systems Programming Section for the Division of Computer 
Research and Technology Library, became operational. The system 
provides the DCRT Library with a computerized method of inventory 
control. This in turn helps to improve circulation and monitor usage of 
"individual" items. The system pays particular attention to accuracy 
and status of circulation items, usage statistics and generation of 
overdue notices. 



The Scientific Applications Section has continued support of the current 
awareness search for Chemical Biological Activities (CBAC). The 
Selective Dissemination of Information (SDI) Service is still offered 
free of charge to all researchers at NIH and is run bi-weekly as tapes 
are received from Chemical Abstracts Service in Columbus, Ohio. 
Retrospective requests are being referred to the on-line service, 
TOXLINE, available at NLM. 

DMB continued support of the current awareness search of Biosciences 
Information System (BIOSIS). Three times a month tapes are received 
from the Biological Abstracts Service and information is disseminated to 
the NIH conmunity thru the same vehicle as CBAC. 

The Reference and Bibliographic Services Section of the NIH Library has 
been the primary contact with the NIH researcher wishing to search this 
data base; profiles are submitted through the library to DMB for current 
awareness searching, and to NLM for retrospective searching. 

As the data base suppliers updated their tape formats and added new data 
elements for searching during FY79, Sigurd Knisley made the necessary 
changes to the system programs. 



198 



• Computer Research and Technical Development 



During the year the Clinical Support Section developed a Voice 
Recognition and Response System which is used by a handicapped 
programmer to write digital computer programs. When a speaker makes an 
utterance the speech is digitized and the result is compared with all 
known digitized speech patterns in the minicomputer. If a match is 
found the character set associated with the utterance is inserted into a 
line of programming code. Once the line of code is completed, it is 
transmitted to the Central Computer System for storage and further 
processing. 



In FY79 the Voice Input Compiler , designed by Robert A. Magnuson of the 
Software Support Section became operational. Rick Pilgrim's program for 
voice input to WYLBUR was written using this compiler. This is an RMAG 
implemented compiler that works with the Interstate Electronics Voice 
Data Entry System. The user is permitted true symbolic entry which is 
compiled into the rigid, absolute format required by the Voice Data 
Entry System. 



In FY79 the SLANG (Structured Language) Compiler , designed by Robert A. 
Magnuson of the Software Support Section became operational. Designed 
to assist programmers writing powerful structured programs for the IBM 
370, the SLANG compiler generates block structured assembly language 
code. An extremely efficient machine language program, SLANG is coded 
in itself. 



In FY79 the Project RMAG: Arithmetic Subroutine , designed by Robert A. 
Magnuson of the Software Support Section had new features added. Two 
new operators, JDAY and JDATE, were implemented. These permit powerful 
date calculations involving the current calendar system. 



199 



In FY79 the Project RMAG: Symbolic Logic Retrieval system, designed by 
Robert A. Magnuson of the Software Support Section had new features 
added. Users can now have upper and lower case descriptions in their 
history files. Also the inclusion of the JDAY and JDATE operators in 
the AR relations permit powerful date calculations involving the current 
calendar system. 



In FY79 REPORTSLR . designed by John R. Parks of the Software Support 
Section became operational. This is a special version of REPORTGEN that 
gives SLR (Symbolic Logic Retrieval) users flexible report generation 
capability. 



In FY7 9 REPORTGEN, designed by John R. Parks of the Software Support 
Section had new features implemented: (1) Added use of RMA31 file read 
subroutine which provides the ability to read edit format files and 
files with RECFM = F, FB, V, VB, or U. (2) Error handling improved. 
(3) Added use of RMA20 (Arithmetic Subroutine) to provide JDAY/JDATE 
calculations. 



In FY79 the entire staff of the Software Support Section provided 
necessary support for the many and varied section products. This 
support included program maintenance, customer assistance and the 
teaching of formal DCRT courses on these products. The supported 
products included RMAG, SLR, Logic Subroutines, Arithmetic Subroutine, 
SLANG, REFORMATGEN, REPORTGEN, TRANSACTGEN, Standardized Update, Voice 
Input, IRS and MarklV. 



200 



III. PUBLICATIONS 

Cole, B.R., Shapiro, M.B. and Rodbard, D.: Chromatography and 
Electrophoresis Analysis System (Cephas): Computerized Peak Detection 
and Characterization. In Catsimpoolas, N. (Ed.): Electrophoresis '78 . 
New York, Elsevier-Holland, 1978, pp. 79-92. 

Harris, E.K. and Shakarji, G. : Use of the Population Distribution to 
Improve Estimation of Individual Means in Epidemiological Studies. In 
J. Chron. Pis. Great Britain, Pergamon Press Ltd., 1979, pp. 233-243. 

Rodbard, D. , Cole, B.R., Murakami, T., and Strott, C: Computer 
Analysis of Concentration Profiles: Automated Peak Detection, 
Characterization, and Estimation of Molecular Size. Steriods, 1979. 



201 



COMPUTER CENTER BRANCH 



Joseph D. Naughton 
Chief 




Oli-ver Movton, Earnet Weltman, and others staff the 
PAL (Progranmer Assistance and Liaison) Unit, which 
handled 19, 750 calls ov visits for user assistance 
dialing the ijeacr. 



I . SUMMARY 

Functions 

The DCRT Computer Center Branch designs, implements, and operates the 
NIH Computer Center, a powerful network of modern computers and commun- 
ication facilities. The nucleus of this network is composed of two 
large multi -computer subsystems, the IBM System 370 and the DECsystem- 
10, each having unique capabilities. Communications facilities link 
these two subsystems and connect them by telephone lines to terminals of 
various types located in research laboratories and administrative of- 
fices throughout NIH. The computing and communications equipment 
(systems hardware) are controlled, balanced, and complemented by a very 
complex set of computer programs (systems software) designed and 
implemented by the Center or acquired from other sources and adapted to 
meet the requirements of the NIH research program. 

Services and facilities provided by the Center include a variety of 
programming languages (e.g., FORTRAN, COBOL, PL/I, SAIL, Assembly 
Language), a data base/data management system (IMS), and a comprehensive 
library of utility programs. For users with terminals there are also 
interactive systems such as WYLBUR, TSO, and the timesharing services of 
the DECsystem-10, which facilitate creation, submission and output of 
jobs and permit direct interactive computing (using FORTRAN, BASIC, CPS, 
APL, MLAB and other languages). The Center also provides many facil- 
ities for the output of textual and graphic information on paper and 
microfiche. Programs to show two or three dimensional pictures on 
cathode ray tube displays and "sketch pads" for advanced graphic 
projects such as involving macromolecular structures are available. 

This network of computers and variety of services are provided by the 
139 members of the Computer Center Branch. Making the total system 
function smoothly requires professional, technical and administrative 
support people. The (51) professional people include computer 
specialists, programmers, and system analysts with widely varied 
backgrounds, including biophysics. The members of this group provide the 
expertise to design and implement the complex computing systems. They 
provide extensive training and information services, and assist users in 
problem diagnosis; they also maintain and schedule recurring production 
applications. The (78) technicians operate the computer systems and 
plotters and provide data entry services. The remainder of the staff 
provides the necessary administrative support. 

Scope of Work 

The Computer Center operates the NIH Computer Utility 24 hours a day, 
providing services to over 6,000 users, including research scientists 
and program managers from all NIH Bureaus, Institutes, Divisions and 
offices. Twenty-nine other Government agencies use these services. 



204 



During the past year, the IBM System 370 Facility continued to grow, 
exceeding the previous year by 10%. (Figure 1 is a chart of workload 
growth since 1967.) At the year's end the Center was handling over 
430,000 tasks per month. This includes 10,000 mainstream jobs a day 
processed by the Utility, with 80% of them completed within one hour. 
The source of many batch jobs were the 5,200 interactive sessions per 
day conducted from some 1800 remote terminals. 

The DECsystem-10 use grew 10% over the previous year (Figure 2 is a 
chart of use since 1972). This growth has nearly saturated the CPU 
capabilities of the system, causing degraded response at times. The 
14,000 user sessions per month used 14,000 connect hours resulting in 
the use of 400 CPU hours per month. 

Research continues in the areas of: text editing, application utiliza- 
tion of mass storage devices, display of biomedical objects, computer 
networking, and communications. 

Highlights of the Year's Activities 

The highlight of the year was the establishment of the Computer Center's 
general purpose facility as a Federal Data Processing Center for Bio- 
medical and Statistical Computation. The agreement to become a Federal 
Center was made under the authority given to GSA by Congress to estab- 
lish computing centers that take advantage of equipment pools and allow 
multi-agency sharing. Such sharing enables small and new agencies to 
have immediate access to a wide range of tested, specialized computing 
techniques at reasonable costs. The NIH users of the Center also gain 
since the sharing brings with it significant cost advantages usually 
reflected in lower rates for computing services. Sharing of computer 
resources enables other agencies to benefit from software developments 
at NIH for similar computational needs. 

A mass storage device was delivered late in the year. The mass storage 
will solve several of the Computer Center's operational problems and 
will lead to a comprehensive, hierarchical data storage system to meet 
the diverse demands of the Computer Center users. 

The Computer Center completed its drive to provide the NIH user with 
better interactive terminals. Two types of hardcopy terminals were 
selected to augment the NIH7000 CRT terminal. One new terminal, the 
Alanthus T1222, is a matrix printer that operates at 120 characters per 
second. It can be used as a standalone terminal and will function with 
any NIH terminal system. This terminal has the added feature that it 
can be configured to provide hardcopy output for the NIH7000 CRT. The 
second terminal, the CleanType 45, is a "quality print" hardcopy ter- 
minal. It operates at speeds of either 30 or 45 characters per second 
and can produce very legible output in a variety of type fonts. These 
selections complete a three year task of specifying, negotiating and 
selecting interactive terminals that best fit the needs of biomedical 
research and its administrative support. 

205 



Effective, fast, data processing is always a function of online data 
storage. Throughout the past few years, space to store online data has 
been a scarce commodity. During the past year, new disk drives were 
installed on the DECsystem-10 that provided a 33% increase in online 
data storage. At the same time, the disk units on the 370 were modified 
with the effect of doubling the capacity of public online data storage. 
This extra capacity linked with the mass storage facility should satisfy 
the NIH demand for data storage. 

The popularity of the Computer Center Training Program led to an 
extensive reorganization of the program. The main thrust was to make 
training available on a year round basis, thus providing more convenient 
times for students. There are now four three-month sessions per year 
rather than two four-month sessions. The self study and audio visual 
courses were expanded into additional areas in an effort to broaden the 
scope of courses offered to users. 

The surface display of the Computer Center Branch graphic system im- 
proved its sphere-drawing algorithm. Scientists are now able to draw 
10,000 to 15,000 atoms in one pass, instead of the previous limit of 
1800 atoms. The success of the surface display for macromolecular 
research is evident: over 10,000 stereo images were drawn last year, 
many for publication. 

After several years of intense negotiations, the GSA and NIH came to 
agreement about the contents of the RFPs for replacement of the 
DECsystem-10 and System 370 equipment. As the year ended, the DEC-10 
RFP was about to be released and the System 370 document was being 
commented upon by industry through the pre-release mechanism. 

Future Plans 

The second floor of Building 12B finally became available for Computer 
Center personnel occupancy. The long delayed reconstruction of the 
second floor of Building 12 and the first floor of Building 12A began. 
The reconstruction projects will provide additional modern, physically 
secure space for computers and new user output areas featuring locked 
boxes as well as improved security for output. 

New WYLBUR remains on the horizon. It was hoped that it would reach the 
users this year but early next year is a more realistic date. The final 
sections of the document preparation modules are being debugged. The 
documentation is nearing completion and the final software and hardware 
plans needed to insure a smooth transition from Old to New WYLBUR are 
being finalized. 

The coming year will see the facilities of the newly-arrived mass 
storage unit offered in multiple ways to the user. Eventually, the mass 
storage and the software developed for it will handle online data sets, 
small data sets currently on magnetic tape, and migration. A hierar- 
chical data storage system will become a reality with this device. 



206 



For some time, the Computer Center has been operating page printers that 
produce output of consistently high quality. The full capabilities of 
this device will be offered to users in the upcoming year. Expanded 
character sets, user provided character sets, forms design, and forms 
flashing are the more important features that will be introduced. 

The past year saw the installation of a digitizing densitometer. 
Extensive work on the software for the system should be completed and 
the densitometer put into production during the coming year. 

The Computer Center will complete a major telecommunication advance 
early next fiscal year. A very comprehensive study of the NIH Computer 
Center communications needs was followed by a complete survey of the 
state of the art communications equipment that would meet the Center's 
needs. The equipment envisioned had to have improved switching capa- 
bility, substantial capacity for growth, diagnostic testing facilities 
and coherent design flexibility to adapt to new and innovative data 
communications services. An "all data" Centrex system was selected. 
Final cutover to the new switch is scheduled for December, 1979. At 
that time, the old stepper switch that has served NIH for ten years will 
be retired. All data communication services will then be centralized on 
the new Centrex switch designed for data traffic, the first switch ever 
designed for this purpose. The new Centrex switch will provide the Com- 
puter Center with greatly improved reliability and more cost effective 
data communications. 

Because of the great interest in the surface display system, a number of 
unsolvable macromolecular structure problems were proposed by various 
collaborators. The focus of the problems seems to be on the need for an 
integrated energy minimization capability. This problem area is still 
in the analysis stage and should emerge into programming and experimen- 
tation at the beginning of the next fiscal year. 

In order to bring the results of the macromolecualr surface display 
system to the largest possible audience, an information dissemination 
project similar to the AMSOM atlas has been planned and is in the imple- 
mentation stage. When finished, this project will provide teachers of 
biochemistry with a tool for the stereoscopic presentation of the 
structural aspects of proteins and nucleic acids. 

Most important to NIH, the coming year should bring the results of the 
RFPs written for the selection of new equipment for both the DECsystem- 
10 and the System 370. Selections and awards may be made and possibly 
late next year the first of the next generation hardware will arrive and 
be put into productive use. With the award of the contracts, the Com- 
puter Center will be able to concentrate on offering the users improved 
tools for biomedical and statistical computing as well as the finest in 
turnaround and support. 



207 



II. ACTIVITIES, SERVICES and FACILITIES 

DECsystem-lO Improvements. DECsystem-10 users finally got relief from 
the online disk storage shortage when new, higher capacity disk drives 
were installed. These new drives will provide a 33% increase in public 
data storage capacity. 

SAIL, one of the more popular programming languages for the DECsystem- 
10, was enhanced. The new facilities included improved random access 
facilities, new packages to handle input/output, software interrupts, 
interprocessor conmunications and non-blocking input/output. 

Use of the link between the DEC-10 and 370 was such that a second link 
between the systems was installed. The number of files transmitted has 
more than doubled. CLINK, the communications link from the DEC-10 to 
small laboratory computers, was further improved to handle. a wide range 
of laboratory computers. Over 1,000 usages per month were recorded for 
CLINK. 

The DECsystem-10 announced the availability of PASCAL. PASCAL was 
designed so that simple, powerful and consistent rules encourage the 
programmer to create structured programs with few errors. One important 
feature of the language is the ability of programmers to create data 
types. The compiler further checks that the types are being used in 
consistent ways and thus many errors are spotted. 

The task of generating finished books, manuals, papers, letters and 
other documents was simplified with the introduction of PIUTO - an easy 
to use documentation language. The system includes commands for docu- 
ment sectioning, creating tables of contents, footnotes, one and two 
level indices and cross referencing. Hyphenation, chart generation, 
formatted and quotations are also available. PLUTO will allow even the 
novice to produce finished documents with a minimum of effort. 

System 370 Improvements . The Center's users benefited when performance 
monitoring showed that region size could be increased without adversely 
affecting performance. The region size was doubled to 1 million bytes 
for all job classes. The new region size will permit larger programs to 
be run without segmenting or use of overlays. 

The third and fourth IBM 3800 page printers were added to the system. 
They will allow the system to handle the increased output on standard 
8-1/2 X 11 inch paper. But more importantly to the user, there is 
sufficient capacity to allow the Computer Center to announce and support 
additional capabilities of the page printer. Additional character sets, 
user defined character sets, and special forms will be offered to users 
by the end of the year. 

The scarcity of online storage space was alleviated as the disk drives 
were updated. The online public space was doubled in size. The Com- 
puter Center hopes to keep ahead of user demand by proper use of the 
online disk space combined with the installation of the mass storage 
system. 

208 



A step forward in the area of data security was taken with the announce- 
ment of new tape data integrity protection facilities. The new facili- 
ties reduce the incidence of problems such as accidental overwriting of 
tape data sets and assure that jobs access only tapes whose use is 
properly authorized by the owner of the data. 

Terminal Improvements . Two new hardcopy terminals were made available 
and delivered to users. Culminating a three year effort, a high speed 
hardcopy terminal and a quality hardcopy terminal were selected and 
deliveries to users started during the year. The user now has a com- 
plete range of terminals available for his use at the NIH Computer 
Center. 

Rate Reductions . The fiscal year started with a 10% rate reduction 
offered by both systems. Additionally, the night time discount rate for 
370 services was increased to 35%. System utilization kept increasing 
(some of it no doubt caused by the rate reductions) faster than the 
projected growth rate. This led to the second rate reduction of the 
year. In May 79, the 370 system reduced its batch processing rate by an 
additional 20% to $.80 a machine unit, and increased its night time dif- 
ferential up to 50%. The DECsystem-10 reduced its basic charge an ad- 
ditional 17% to $1.25 a machine unit. Also, the DEC-10 night dif- 
ferential was doubled to 50% and extended to include Saturdays and 
Sundays. Terminal connect time for the DEC-10 was reduced 1/3 to $2.00 
per hour. The economies of scale and continuing growth both contribute 
to reduce the cost of biomedical and statistical computing at the NIH. 

Procurement . After years of intense negotiating with GSA, several RFPs 
were launched or are about to be launched. An RFP resulting in the 
selection and award of a contract for the mass storage device was com- 
pleted late in the year. The RFP for the replacement of the DECsystem- 
10 was released to the vendors late in the year. The RFP for the re- 
placement of the 370 reached a significant milestone as the document 
entered the prerelease stage, that is, industry was asked for comments 
on the RFP before its formal issuance. 

User Education and Assistance . The Computer Center training program was 
extensively reorganized during the year. Instead of two four-month 
terms per year, four three-month sessions were offered. This made 
training available year-round and provided more convenient times for 
many students. Self-study courses were expanded to include audio visual 
courses and computer assisted courses. Applications for training during 
the year numbered 1317, but only 889 of these could be accepted. There 
were 64 sessions of 37 different courses given during the year. 

The theory that by educating the user you increase his understanding and 
reduce his problems, was fortified by the fact that Programming Trouble 
Reports researched and answered during the year decreased to 2300. User 
Services also applied over 4200 system fixes during the year and in- 
stalled 35 new releases of current software packages. There were 19,750 
calls or visits for assistance during the year. 



209 



The Technical Information Office maintained its high level of respon- 
siveness to users with the aid of a newly automated user profile system 
operating under IMS. There were 79,000 individual pieces of technical 
documentation in support of the NIH Computer Utility mailed to the 6,080 
users registered for the automatic documentation service. An additional 
5,574 requests for documentation were satisfied in response to over-the- 
counter and telephone requests. 

III. PUBLICATIONS 

Feldmann, Richard J., Bing, David H., Furie, Barbara C, Furie, Bruce: 
Interactive Computer Surface Graphics Approach to study of the Active 
Site of Bovine Trypsin, Proceedings of the National Academy of 
Science USA. Vol. 75, No. 11, November 1978. 




The Computer Center operates the NIH computer utility 
24 hours a day. Computer operators like Robert Harris 
(center) provide services to over 6, 000 users including 
research scientists and program managers from all NIH 
Bureaus^ Institutes, Divisions and offices. 



210 




211 



Sess i ons 



17 



in 

c 

rd 

o 






DECsystefn-1077 
CCB/DCRT/NIH 




KAIO 



KIIO duml KIIO 



-L. 



_U 



_l_ 



-L. 



-L. 



Jun 79 



«5 vi vj *i v< ^4 « 
♦"^ ♦-» <'/ ♦^ <'/ ♦■9 
^^ -t? -i? -V -^r -^f 



* *i «^ «i 
*'/' *-9 *'/ ♦■» 

-Jf ^ ''^ 



>> 






-*# -** ^ 



FIGURE 2 



212 



SMITHSONIAN SCIEMCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00036-03 CCB 



PERIOD COVERED 



TITLE OF PROJECT (80 characters or less) 

Graphic System for the Display of Biochemical and Biomedical Objects 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

P.I. Richard J. Feldmann, CCB, DCRT, Computer Specialist 

Others: Thomas K. Porter, CCB, DCRT, Computer Specialist 
Charles R.T. Bacon, CCB, DCRT, Computer Specialist 



COOPERATING UNITS (if any) 

None 



lab/branch 

Computer Center Branch 



SECTION 



INSTITUTE AND LOCATION 

DCRT, NIH, Bldg. 12A, Room 3009. Bethesda. 



MD 20205 



TOTAL MANYEARS: 

1.25 



PROFESSIONAL: 

1.25 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) . , . , . 

Implementation of graphics system for the respresentatTon and manipulation of 
biochemical objects is being done at various levels of hardware, system 
software and user software. Work on the link between the DECsystem-10 and the 
graphics system was brought to completion. Modifications to the surface 
generation algorithm resulted in the removal of a limit of 1800 displayable 
atoms. The system can now do hidden sphere calculations for between 10,000 
and 15,000 atoms. Programs were developed for the interactive modeling of up 
to 300 atoms at a time. Modeling of larger structures is handled by scooping 
out sections within the 300 atom limit. This programs have been applied to 
the modeling of drug-protein interactions and drug-nucleic acid interactions. 

Implementation of a digitizing microscope densitometer is proceeding slowly at 
the system software and hardware testing levels. 



PHS-6040 
(Rev. 10-76) 



213 



SMITHSONIAN SCI 
PROJECT NUMBER 



ENCE INFORMATION EXCHANGE 
(Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOICT 00046-01 CCB 



PERIOD COVERED 



TITLE OF PROJECT (80 characters or less) 



Dissemination of Macromolecular Surface Representations 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

P.I. Richard J. Feldmann, CCB, DCRT, Computer Specialist 
Others: None 



COOPERATING UNITS (if any) 

David R. Sing, Center for Blood Research, Boston Mass. 



lab/branch 

Computer Center Branch 



SECTION 



INSTITUTE AND LOCATION 

DCRT, NIH, Bldg. 12A, Room 3009, Bethesda, MD 20205 



TOTAL MANYEARS: 

0.5 



PROFESSIONAL: 

0.5 



OTHER: 







CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 

D (al) MINORS n (a2) INTERVIEWS 



D (b) HUMAN TISSUES 



D (c) NEITHER 



SUMM/^BY OF WORI^ (200 vords or less - uaderline keviords) . ^. ^ . , j ^u * 

Collaboration with NIH and NlH-supported scientists has showed that macro- 
molecular surface representations, which are generated on the Surface Display 
system are very important in understanding molecular function. The clearest 
perception comes from looking at the images in stereo. After using the Sur- 
face Display system for about one year and looking at a number of macromol- 
ecule, it seemed that the images could be used by a large population of bio- 
chemists. A stereo sample was developed and sent out. The response to the 
sampler was 'i/ery strong. Comments came back that these stereo images would be 
an important tool in teaching the relationship between macromolecular struc- 
ture and function. A search was made for a medium of presentation and the 
result was an inexpensive stereo viewer. The stereo images for the student 
are formatted 7 per card and the package will have 7 cards. The teaching unit 
will consist of 100 stero slides of which 7 x 7 = 49 for the students are a 
subset. Depending on user response, we will consider doing other, perhaps 
more specific, packages in this medium. 



PHS-6040 
(Rev. 10-76) 



214 



PUBLIC HEALTH SERVICE - NATIONAL INSTITUTES OF HEALTH 

DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 

Report of Program Activities 
October 1, 1977 through September 30, 1978 



ANNUAL REPORT 

TABLE OF CONTENTS 

Page 
DIRECTOR'S SUMMARY ^ 

Office of Scientific and Technical Communications -----___3 

AppUccLtLoyv:^ oi ?^IV I V^ZIV^ I G^kVH 8 

BZaddeA Canczn. Imagz Ptoce^h-s-lyig -----------____]] 

ChaAactz/Uzation 0|f CzlLi (^A.om E^eru,^ A^p-0iaXz6 -------I4 

CoyvbLitting Ac.tivi,tiz^ -------------_______] 5 

HcutuAol Languagz Acc2^4 to CtiyvicaX Vata. S<xi2yi ------- -19 

Tumon. GfioMtk Co.iue.i -----------_-_______ _22 

Office of ADP Policy Coordination __-______.23 

Office of Administrative Management -------------- -26 

LABORATORY OF STATISTICAL AND MATHEMATICAL METHODOLOGY 

Summary ------------------__________ _28 

Research Project Reports 

kaZomatiid data P'toceyi^ing o{, ^ddlccLl Languagz ------- .42 

BA.olcg-lcxLl and <JJ^iu.al Slvipz ---------------- .44 

I'^oiete Ma,tkzma.tic^ and Apptizatiom, ----------- -45 

LinzoA Mzth.od.i> in Statti-tia ---------------- -43 

MattivoAAJVtz StaXZittizaZ Ayuatyili --_-.-_____ _ .50 

Hontinza.x Eqaatiom, -------■--------_--_- .52 

PaJXzAn RzcognLtLon -------------------- s^ 

Rz^!^QjLlzh Top-icb in CompuXzfi Scizncz ------------ -S6 



Page 
DATA MANAGEMENT BRANCH 

Summary ---------------------------- .53 

COMPUTER SYSTEMS LABORATORY 

Summary ---------------- -_____. .37 

COMPUTER CENTER BRANCH 

Summary ----------------------------104 

Research Project Reports 

A^a6 oi Maa^omolzcuZO'i StAactuAS. on. MccA-o^-cc^e. (AiMSOM) - - - 114 

GKapklz Sy.i>tm f^on. tkz V<j>play 05 E4.0ckmA.ccil and ----- 115 
E-iamzdical. Objtct>i 

LABORATORY OF APPLIED STUDIES . , 

Summary ----------------------------117 

Research Project Reports 

Ami£t/4-ci o(i CoapZzd Tfia.yvbpo^'vt and Ziockm-icat KA.m2.tici, - - - 124 

CompuXe^n. Aldtd AncLlyi>-li, 05 Ele-ctAocoAdiog-tajm^b -------127 

Compo^eA-boxied ^odztlng oj^ ?uImona.n.u G(Xi Exchange and - - - 130 
Re.ip-o'ifXtO'tw Mzckaytlcii 

CompujtZyi SyJitmi> ^on. VAjigno-itlc Imaging ----------134 

G2n2AaI. MathmcLtical and CompuXatLona-l. CottaboKotivd - - - - 138 
E^(i0AX6 

Inve^bt^gation 0^ Hybrid ZompuJung ^qx tkz Con,i.txuction - - - 141 
o{^ SAjnulation ^iodeZi, and ^on. thz kn£ilLji>ii> 0^ Pky^io- 
log-lcat S-lgnaLi, 

"AaXWmaJilcaX. WodzLlng 05 biological P1ac'^i>4^i - -144 

)AaALkmaJtical Modeti and SlsnuJjation Ptcg-^om^i In Physiology- - 149 

Statiiitical RuQxotch In CUnlcaZ Patkology ---------152 



n 



Page 
PHYSICAL SCIENCES LABORATORY 

Summary ---------------------_-____i55 

Research Project Reports 

C2IZ MotlLLty and ChejnotaxA^i) -------________] 52 

CaniLLttLng SeAu^ce6 ---------------_____] 54 

CofUizlation Function SpzcJyio^aopy/LcuizK Light --_____] 57 
ScaJXeAi^ng 

StacLie^ in Mathzmatici and StatLiticJi --------___] 59 

Tkzofiy and Apptication o{^ HucIqjx/i MagmtLc R^.i,onanc^ - . . ]J2 
Sp^cLtxo6copy 

TkzoAy and Meaiateme/u!: oq IntzAmolzcula^t Fo^Lce/i ------ 174 

Tkno-ry 0^ 'Siockim-ical S^paAat^-on Tzch.nique.i> -------- ]yj 



m 



DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 
October 1, 1977 - September 30, 1978 

DIRECTOR'S SUMMARY 

During FY ' 78 the laboratories and branches of the DCRT again carried out their 
several, complementary functions very well: 

° The Physical Sciences Laboratory advanced its research on mathematical 
theory and practical instrumentation to explain biological phenomena 
in term of chemistry and physics at the subcellular molecular levels. 

° The Laboratory of Applied Studies extended its collaborative projects 
with scientists, at NIH and elsewhere, with particular emphasis on 
clinical research applications, centering around mathematical models 
of function in health and disease. 

° The Laboratory of Statistical and Mathematical Methodology had increased 
use (more than 12,000 calls per month) of the statistical program packages 
it supports for NIH and received more requests for consultative help in 
statistics, mathematics and computer science. 

° The Data Management Branch continued its widespread support to scientists 
and administrators throughout NIH, working with the staff of about 80% 
of the B/I/Ds during the year; DMB also undertook important new responsi- 
bilities in support of some central NIH administrative systems. 

The Computer Systems Laboratory embarked on several new projects develop- 
ing mini and micro computer systems for NIH laboratories and clinics and 
with the Computer Center Branch it developed new communication links 

systems to the central NIH facilities. 



from these kinds of 



" The Computer Center Branch helped more people (some 6,000 registered 
users) to do more work (over 4 million accountable jobs) at lower rates 
on improved NIH central computer systems. 

The reports of the DCRT laboratories, branches and offices are self-contained 
and speak for themselves. A copy of the detailed FY'78 announcement of DCRT 
computer resources provided for NIH scientists and administrators is enclosed 
with this annual report to complete the archive. There are several long-term 
trends worth noting in those reports. 

First is the continued growth of computing throughout NIH laboratories, clinics 
and offices. This comes from the addition of some new users each year and from 
the refinements made by existing users to make fuller use of their existing 
computing applications. In many cases these refinements take advantage of 
advances in hardware or software technology. 



Second is a growth of intellectual linkages among the DCRT staff and their 
counterparts in the programs of NIH and other research organizations. The [ 
result is better communication between the DCRT staff and those organizations! 
and better use of computing in those programs. One facet of this phenomenon 
is the new Advisory Committee on Computer Usage established by the NIH Deputy r 
Director for Science early in FY 1978. But more important is the growth in l 
the number and productivity of research projects that cross disciplinary and 
organizational boundaries. 

The third trend is a growth in administrative complexity of computing in a ' 
Federal agency. The General Services Administration has mandated competitive 
reprocurement of both the DEC System-10 and the IBM 370 segments on the cen- | 
tral NIH computer facility. The GSA has also limited the NIH ability to up- I 
grade existing systems with newer, less costly, more effective equipment prior 
to the competitive reprocurement. 



The result will be at best some delay in obtaining the benefits of new tech- 
nology for thousands of NIH users of the central systems. The worst case 
could be a return to the circumstances facing NIH in 1965, when it began to 
convert all of its computer software to entirely new kinds of computers. It 
is difficult to predict accurately the effect on NIH of any conversions among 
the markedly more complex computer systems of the late 1970s. 

Three other events in FY '78 may also influence the course of the reprocurement 
of central NIH computer systems. The Office of Assistant Secretary for Manage 
ment and Budget, DHEW, unilaterally initiated the purchase of some of the IBM 
370 hardware currently leased by the NIH Computer Center. The ability to up- 
grade the existing NIH central computing resources will be even more sharply 
limited during the recompetition and prior to the installation of new hardware 

In April, the General Accounting Office published a report titled "The Federal 
Information Processing Standards Program." This exhorts the President, the 
OMB, the GSA and the Secretary of Commerce to tighten up the Federal standards 
program. There are supposed to be 29 such standards; only 10 of these have 
been developed since 1965. A burst of Federal activity to develop new stan- | 
dards over the next year or two may confuse or confound the NIH computer ' 
reprocurement as it moves forward during that period. 

Finally, a study of Federal data processing undertaken as part of the Adminis-[ 
tration's Reorganization Initiatives in the OMB is now coming to completion 
It, too, is reported to recommend various new actions. Many people expect » 
Representative Jack Brooks to hold hearings on the study. J 

The prospect for computing at NIH is therefore one of guarded optimism. The 
trends within DCRT and NIH toward better computing and better coirmuni cation | 
about and understanding of computing are encouraging. They can improve the ** 
record of excellence created over the last decade. The effect of the trends 
elsewhere, in the GSA, GAO, OMB and other parts of DHEW, will certainly com- 
plicate life for the DCRT. It remains to be seen whether they will detract 
from the quality of computing at NIH. 



October 30, 1977 to September 30, 1978 

NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 

1 . DCRT 

2. OFFICE OF SCIENTIFIC AND 3. William C. Mohler, M.D . 

TECHNICAL COMMUNICATIONS CFief 

I . SUMM.ARY 

Function 

The DCRT office of Scientific and Technical Communications, under the direc- 
tion of the Associate Director, DCRT, includes: 

° The DCRT Library , which maintains a special collection in computer 
science and mathematics, statistics, engineering, information science 
and management. The library directly supports DCRT activities and is 
a resource for other NIH staff. It also functions as an integral 
part of the local Washington area network of special libraries. 

° The DCRT Information Office , which serves as the focus for the 
Division's program to provide information about DCRT activities to 
the rest of NIH and to the public. 

° Scientists assigned to this office, working on research and develop- 
ment projects in biomedical data bases, image processing and decision 



analysis. 



Highlights of FY78 Activities 



The DCRT Library continued approximately the same level of circulation services 
for its DCRT and NIH users as in previous years. It had an increase in the 
number of requests for interlibrary loans and for the first time lent more than 
it borrowed from other libraries. The Copyright Law that went into effect in 
January required changes both in interlibrary loan operations and in record 
keeping and notification for photocopying at NIH. 



Use of automated information retrieval services increased slightly. 
Library currently has access to 29 data bases for direct searching, 

,jj,-4..;„„ „£ 17 J^j--, K,,-^^ ^^„™ one 4.U,-- -,-4. I ,• I 



The 
. The 
addition of 17 data bases from BRS this past year has increased our coverage 
considerably. MEDLINE, Computer and Control Abstracts, and OCLC have been 
the most heavily used. These data bases are available through the National 
Library of Medicine, Bibliographic Retrieval Services, Inc. (BRS) and OCLC 
through membership in FEDLINK. The Library is also a registered user to order 
individual searches with the NASA Information Center, Defense Documentation 
Center (DOC), and the Smithsonian Science Information Exchange (SSIE). 

In the middle of FY78 the library staff began work with the DCRT Data Manage- 
ment Branch on design and implementation of a new automated circulation system. 



The new system will include features to improve library services. The most 
important feature will be automatic data checks for inconsistent and erroneous 
terminal input. The system will also save time in updating the records of cir-l 
culating materials and in generating overdue notices and reports. A new fea- 
ture will be collection of statistical data on all monographs and documents to 
facilitate timely purchase of multiple copies to assist in weeding consider- 
ations of seldom used materials. 

The process of creating this system included a complete review of cataloging 
on all library monographs, as well as creation of input records for the master 
files. This large task was aided by two temporary summer employees. 

During FY78 the DCRT librarian served as the Federal Representative to the 
Metropolitan Washington Library Council, a member of the Executive Board of 
the Interlibrary Users Association of the Washington/Baltimore area, a member 
of the NIH Library Committee, a member of the Federal Library Committee of the 
BRS users group and a member of the FEDLINK/OCLC Executive Advisory Committee. 
She was editor of the lUA Directory and Subject Index in FY78. 

The DCRT Information Office continued the progress begun in FY77 under its 
new Information Officer, but in the middle of the year Mr. Hudson accepted 
the call to become the full time minister of his church. The major emphasis 
during FY78 was on creation of new printed materials telling about DCRT activi- 
ties and the resources available to aid NIH scientists and administrators. 

The scientific staff was increased by the return of Martin Epstein from an 
assignment at the Medical School of the University of California, San Fran- 
cisco. Dr. Prewitt, who received her Ph.D. from the University of Uppsala in 
FY 78, continued her work with Dr. Wu on new methods for image processing, 
pattern recognition and decision analysis. She was a consultant and advisor 
to several NIH Institutes and Divisions and to other Federal Agencies, partic- 
ularly the National Science Foundation. She was also very active on IEEE com- 
mittees on Image Processing, worked on the Board of Directors of the Biomedical 
Engineering Society and served as editor and reviewer for several publications 
in the field of biomedical image processing. 

Dr. Mohler reorganized the NIH Clinical Elective for Medical Students on Com- 
puters in Clinical Medicine which had not been given since 1974. 

Future Plans 

The activities of the office will continue along the lines followed in FY78. 

The Library plans to cope with its "space problem" by gradually converting 
some of its journal holdings to microforms. This substantial increase in its 
microform collection will require purchase of a microform reader/printer. 
When file construction for the new automated circulation system is complete, 
the library will create a new set of printed indices for its document col- 
lection and will consider the prospect of converting the card file for mono- 
graphs to a printed format. 



Plans for new activities in the Information Office are awaiting recruitment of 
a new DCRT Information Officer. 

Mr. Epstein will pursue the work he carried out in California on the design and 
implementation of clinical research data bases with emphasis on interactive 
"natural languages" query systems to make them more usable by clinical scien- 
tists. 

Dr. Prewitt's plans for the work on biomedical image processing and decision 
analysis include a larger effort on the collaborative projects in development 
of criteria of malignancy for breast tumor cell aspirates and bladder cancer 
tissue sections. 

Project List 

1. PEEP/DECIDE/GRAPH . J. M.S. Prewitt, S.C. Wu, OSTC 

PEEP/DECIDE/GRAPH is an interactive programming system written in SAIL, 
the Stanford Artificial Intelligence Language, and operating on the 
DCRT PDP-10 computer. PEEP is designed for picture processing appli- 
cations, DECIDE is intended for algorithmic decision-making and ex- 
ploratory data analysis, and GRAPH has capabilities for two and three 
dimensional graphics. PEEP/DECIDE/GRAPH has been implemented as a 
single system with a large library of image processing, feature extrac- 
tion and decision-making algorithms. We are maintaining it for public 
use as well as using it for our own research on blood cells, breast 
aspirates and bladder epithelium. 

2. Characterization of Breast Aspirate Cells . J. M.S. Prewitt, OSTC; 
B. Stenkvist, J. Holmquist, Uppsala. 

This long term collaborative project to analyze and characterize all 
images from human breast aspiration biopsies has identified the utility 
of four analytic features (nuclear length of major axis, density and 
entropy) as distinguishing among five all type categories in breast 
aspirate material stained with Papanicolau Stain. Studies are in 
progress with materials using other stains. 

3. Bladder Cancer Tissue Analysis . J. M.S. Prewitt, S.C. Wu, 

The long term study of nuclei in bladder tissue was reactivated in 
FY78. Preliminary studies were made on two sets of tissue. 

4. Tumor Growth Curves . J. M.S. Prewi tt, -OSTC 

Previous work on a generalized tumor growth curve equation was extended 
during FY78. 

5. Natural Language Access to Clinical Data Bases . M.N. Epstein, OSTC 

The objective of this project is the development and evaluation of a 
system that will allow physicians access to medical data through 
natural language queries to support both patient management and 
clinical research. 



A prototype system has been implemented for a small data base on 
malignant melanoma. The physician can input queries in English that 
retrieve specified data for particular patients or for groups of 
patients satisfying certain characteristics, that perform simple cal- 
culations, that allow scanning of the data base, and that assist in 
identifying relationships among attributes. 

5. DCRT Library Circulation System . E. Chu, Librarian 

Examination in FY77 of alternative hardware/ software improvement to 
the old system led to decision to redo it using the NIH Computer Cen- 
ter facilities rather than as a stand-alone system. The Librarian 
and DMB staff designed and implemented the new system The Library 
staff are in the process of putting the data into the master files. 



7. DCRT Information Program . ■ 

This project, begun in FY77, developed several new brochures describing 
DCRT activities, set up distribution rack in the new Building 12A lobby. 
The project halted temporarily when the Information Officer switched 
careers in the middle of FY78. 



Publ ications 

Chu, E. M.: (Ed): Inter! ibrary Users Association Directory and 
Subject Index. Washington, D.C., Interlibrary Users Association, 1978, 43pp. 

Epstein, M.N., and Kaplan, E.B.: Criteria for Clinical Decision 
Making. In Schneider, W. and Sagvall-Hein, A.L. (eds.): Computational 
Linguistics in Medicine. North-Holland Pub. Co., 1977, pp 35-44. 

Prewitt, J. M.S.: Some Applications of Pattern Recognition and 
Image Processing to Cytoloty, Cytogenetics and Histology. Doctoral Disserta- 
tion, Department of Computer Science, Uppsala University and Department of 
Clinical Cytology, University Hospital, Uppsala, Sweden, April, 1978. 

Prewitt, J. M.S.: Reconstruction Software for Computerized Tomog- 
raphy. Proceedings of Computer Software Applications Conference 1977, IEEE 
Press, 1977, pp 285-292. 

Prewitt, J. M.S.: Interactive Decision-Making for Picture Proces- 
sing. Proceedings of the 1977 IEEE Conference on Decision and Control. 
IEEE Press, 1977, pp 373-379. 

Prewitt, J. M.S., Barber, A. and Wu, S.C: An Application of 
Pattern Recognition to Histology. IEEE Computer Society Conference on 
Pattern Recognition and Image Processing. Chicago, Illinois, June 1978, 
pp 499-506. 

Prewitt, J. M.S., and Mendelsohn, M.L.: Analysis of Cell Images. 
Reprinted in Machine Recognition of Patterns, A.K. Agrawala (Ed.) Wiley, 
New York, 1977, pp 362-280. 

Wu, S.C, and Prewitt, J. M.S.: Deriving Concavities from the 
Fourier Coefficients and Its Role in Pattern Recognition. Proceedings of the 
30th Annual Conference on Engineering in Medicine and Biology, 1977. 

Wu, S.C, Prewitt, J. M.S., and Lehman, J.: To Extract a Connected 
Object of Arbitrary Shape from its Background by Decision Tree Method. IEEE 
Computer Society Conference on Pattern Recognition and Image Processing. 
Chicago, Illinois, June 1978. 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00030-03 OSTC-C 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

Applications of PEEP /DECIDE/GRAPH 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

Judith M.S. Prewitt, Ph.D. 

S.C. Wu, Ph.D. 

Dr. K. Kent, National Heart and Lung Institute 

Dr. Sadek Hilal, Neurological Institute, Columbia-Presbyterian Hospital, N.Y. 



COOPERATING UNITS (if ar:y) 

Neurological Institute, Columbia-Presbyterian Hospital, N.Y. 
National Heart, Lung, and Blood Institute 



lab/branch 
Office of Scientific and TerhmVal Cnn 



SECTION 



nniratinn 



INSTITUTE AND LOCATION 



DCRT/NTH 



TOTAL MANYEARS: 



PROFESSIONAL: 

.50 



OTHER: 



0.05 



CHECK APPROPRIATE BOx(ES) 
Ijc(a) HUMAN SUBJECTS 

□ (al ) MINORS u (^2) INTERVIEWS 



□ (b) HUMAN TISSUES 



□ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

peep/decide/graph is an interactive programming system written in SAIL, the 
Stanford Artificial Intelligence Language, and operating on the DCRT PDP-IO 
computer. PEEP is designed for picture processing applications, DECIDE is 
intended for algorithmic decision making and exploratory data analysis , and 
GRAPH has capabilities for two and three ' dimensional graphics . The system 
was originally written for the NCI Bladder Cancer Image Processing Project 
but is of wide utility. We will continue to make PEEP /DECIDE/GRAPH available 
for public use. 



PHS-6040 
(Rev. 10-76) 



ether Applications of PEEP /DECIDE/GRAPH 

In the PEEP system and appendages to it, objects can be obtained from 
optical density histograms using several approaches. This option is 
available in global and local form. Derivatives, Laplacians, Hueckel 
and other edge detection operators can also be applied. A large 
library of feature extraction algorithms has been built. 

These features encompass numeric descriptions of size, shape, content, 
contrast, comparison, texture and orientation. Examples of such feat- 
ures are area, average brightness, diameter, average chord, entropy, 
kurtosis and skewness of the optical density histogram, integrated 
optical density, variation in brightness, variation in chord length, 
variation in diameter, Fourier coefficients of a boundary, bending en- 
ergy, medial axis transforms, major axis and minor axis. 

DECIDE enables the performance of parametric and non-parametric re- 
cognition of objects. Linear and quadratic discriminant analysis form 
the basis for parametric recognition. Coefficients of the best dis- 
criminant function can be learned from exemplary objects using a learn 
command. New test objects can be classified into a learned category 
using the classify command. The user has control over various para- 
meters of the learning process. For example, likelihood maximizing 
discriminant functions can be developed. Non-parametric recognition 
is performed by the cluster command. Again, many variations of data 
normalization, cluster merger strategy and cluster type emphasis are 
available. 

GRAPH allows the user to create and display graphic structures in two 
and three dimensions and to use different vantage point for viewing 
them. 

1978 Activities 



The three subsystems were merged into one and PEEP data formats were 
changed to be compatible with MLAB. In addition, the Fourier analysis 
of curves and GRAPH have been incorporated into MLAB. 

In collaboration with Dr. K. Kent at the National Heart, Lung and 
Blood Institute, a project was continued in order to make use of the 
decision making capabilities of DECIDE. An on-line expandable data 
retrieval system had been designed for the purpose of compiling data 
on the natural history of patients with coronary artery disease. Data 
on 200 to 300 patients over a period of two years will be collected 
and entered by means of a special questionnaire. A general purpose 
input program STOW for composing questionnaires leading to data struct- 
ures compatable with DECIDE has been written and tested. Decision 
making logic will be developed so that a prospective diagnostic 
scheme can be obtained. From time to time graphic display is neces- 
sary. This project is similar but more elaborate than another National 
Heart, Lung and Blood Institute collaborative program designed to study 
pre and post-operative factors indicative of surgical risk. 



Other Applications of PEEP /DECIDE /GRAPH 



In collaboration with Dr. Sadek Hilal of the Neurological Institute, 
Columbia-Presbyterian Hospital, New York, PEEP was used to apply automated 
pattern recognition to EMI computed tomography scans of the human brain. 
A computer decision tree method was devised to extract tumorous and 
edemous regions in the brain scans. We are awaiting receipt of a data 
tape containing 20 sets of brain scans with different types of tumors for 
analysis. It is anticipated that the new texture measures in PEEP 
will be diagnostic. 

Future Plans 

The PEEP /DEC IDE /GRAPH interactive programming system is very large, 
when loaded in PDP-10 core it occupies 100,000 words. The system will 
be maintained in its present form for public use. A conservative 
policy will be maintained with respect to adding new image processing 
algorithms to PEEP because of its size. New routines will be maintained 
in parallel with PEEP. 

The collaboration with the Neurological Institute will containue. New 
boundary definition procedures will be applied to organ location and 
tumor localization. New texture measures will be tested to see if they 
correlated with tumor type. 

Little activity is expected on the NHLBI project. 

Publications 

Wu, S.C. and Prewitt, J.M.S. 

Deriving Concavities from the Fourier Coef f icients ■ and Its Role in 
Pattern Recognition. Proceedings of the 30th Annual Conference on 
Engineering in Medicine and Biology., 1977. 

Wu, S.C, Prewitt, J.M.S. and Lehman, J. 

To Extract a Connected Object of Arbitrary Shape from its Background 
by Decision Tree Method. IEEE Computer Society Conference on Pattern 
Recognition and Image Processing. Chicago, 111. June, 1978 ,pp . 352-353 . 



10 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00029-03 OSTC/OD 



PERIOD COVERED 

October 1, 1977 - September 30, 197? 



TITLE OF PROJECT (80 characters or less) 

Bladder Cancer Image Processing 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

Judith M.S. Prewitt, Ph.D. 

S. C. Wu, Ph.D. 

Gilbert H. Friedell, M.S., Chief of Pathology, St. Vincent Hospital 

Enrique A. Soto, M.D., St. Vincent Hospital 



COOPERATING UNITS (if any) 

Pathology Department, St. Vincent Hospital, Worcester, Mass. 

Im ag e Processing T , ahnrarnry, Jet Propu l s i on Laboratory , Pasad e na, Calif, 



LAB/BRA:,Ch 

Office of Scientific and Technical Communication 



SECTION 



INSTITUTE AND LOCATION 
DCRT/NIH 



TOTAL MANYEARS: 

1.00 



PROFESSIONAL: 

.95 



OTHER: 



0.05 



CHECK APPROPRIATE BOX(ES) 
~ (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



3 (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK 



(200 words or less - underline keywords) 



Sumr.ary 

This project involves the application of digital computer technology to the 
diagnosis of disease based on the microscopic morphology of stained tissue 
sections. Techniques of digital image processing are applied to delineate 
nuclei in digital images ot human urinary bladder epithelium and to com- 
pute mathematical characterizations of various morphologic attributes such 
as size, shape, texture and tissue architecture. Techniques of statistical 
decision theory and cluster analysis are applied to the problem of ob- 
jectively characterizing tissue sections and devising classification and 
grading systems for them based exclusively on quantitative morphometry and 
biochemical composition. 



11 



PHS-6040 
(Rev. 10-76) 



Bladder Cancer Image Processing 



Objectives 

The long range goal of this project is to quantitatively characterize the 
epithelium of human urinary bladder from scanned digitized images of 
stained sections using digital computers and to develop a data-directed 
taxonomy for the range of tissues from normal to invasive carcinoma. 

1978 Results 

This project, inactive for a while, was reactivated with the change in 

programming staff in FY77. 

The data base was not altered and consisted of images of tissue sections. 
Tissue sections were prepared on microscopic slides at St. Vincent 
Hospital, Worcester, Mass. and stained with hematoxylin. They were 
scanned at the Jet Propulsion Laboratory, Pasadena, Calif, so that the 
resulting images were at 630X, sampled at half micron intervals and 
rendered in 256 linear gray levels. The material was routine clinical 
preparations of variable and sometimes mediocre quality. The absorption 
peak for the tissue sections was determined to be at 570NM and all scans 
were made at both this wave length and in white light. In one experiment 
with tissue sections stained with gallocyanin chromalum, the response in 
the entire visible range was assessed. 

The PEEP/DECIDE/GRAPH system was used to analyze the digitized images. 
(In fact, the project was the impetus for implementing that system, which 
is discussed in a separate project report). All of the object extraction 
methods were applied to digitized images of tissue sections. They met with 
varying degrees of success and depending on the quality of the tissue 
sections, one or another method may be necessary. 

Two morphologically distinct tissue sections were chosen for in-depth 
study. Both yielded to thresholding for obtaining nuclear images. The 
entire armamentarium of PEEP features was extracted on approximately 20 
nuclei from each tissue. Linear and quadratic discriminant analysis and 
cluster analysis were used to learn each tissue section as a category. 
Specimens from each tissue were then classified with the following result 
of demonstrated internal consistency: nuclei in one tissue type were over- 
whelmingly more like each other than like cell nuclei from other tissue 
types. 

A new algorithm for shape assessment, based on a complex number represen- 
tation of arc length along a boundary, was developed. All concavities can 
be found and the area between each one and the subject curve measured. In 
addition, the orientation of the object under study can be obtained from 
the Fourier coefficients. The technique 'was applied to blood cells and 
bladder epithelium nuclei for test purposes. 

A new method of boundary following by using several threshold values 
within a range of gray levels iteratively and optimizing a function has 
been developed and successfully applied to finding the boundaries of nuclei. 
The starting threshold is the point of reflection of the left half of the 
optical density histogram to the first mode with an upper limit being set 
to the range of optical densities for the dynamic search. At each thresh- 
old the area is calculated and a linear function of two boundary features 
taken at successive threshold. The maximum of this function over all 

12 



Bladder Cancer Image Processing 

thresholds is chosen as the threshold to find the boundary. 

The two morphologically distinct tissue sections also were subjected to 
a new decision tree method to separate the nuclei from background since 
many nuclei have some internal pixels with lower gray level than at the 
boundary. Using 20 nuclei in each tissue quadratic discrimination using 
the features nuclear area, nuclear density, nuclear absorbance, standard 
deviation of nuclear density, entropy of nuclear density, and three' 
texture measures, 100% correct classification of nuclei as coming from 
the parent tissue was obtained. A cluster analysis using the same 
features had a 6% error rate. 

Significance 

The data directed classification of tissue sections might well be an im- 
provement over current subjective and often dubious decisions. The dif- 
ficulty of the undertaking should not be under emphasized. Classification 
using algorithms may lead to greater objectivity, public verif lability, 
and greater consistency. There is always the opportunity for discovering 
new significant differences in optical properties between papillomas and 
papillary carcinomas using the digital computer. 

Future Plans 

At the present time, the project has been reactivated and will remain so, 
especially if suitable tissue section material becomes available. The same 
techniques will be applied to cervical biopsies paired with cytology- 
which will be obtained from the University Hospital, Uppsala, Sweden. 

Publications 

Prewitt, J. M.S., Barber, A. and Wu, B.C. 

An Application of Pattern Recognition to Histology. IEEE Computer 
Society Conference on Pattern Recognition and Image Processing. 
Chicago, 111. June 1978. pp. 499-506. 



13 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00028-03 OSTC/01 



PERIOD COVERED 

October 1, 1977 - September 30, 197i 



TITLE OF PROJECT (80 characters or less) 

Characterization of Cells from Breast Aspirates 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

Judith M.S. Prewitt, Ph.D. 

Bjorn Stenkvist, M.D. , University Hospital, Uppsala 

Jan Holmquist, Ph.D., University Hospital, Uppsala 



ITS (if any) 

t of Clinical Cytolc 



COOPERATING UNIl 

Department or Limicai Uytoiogy 
University Hospital, Uppsala, Sweden 



LAB/BRAuCH 

Office of Srienf-jfir and Tprhniral rnnrmnm' p^M" nn 



SECTION 



INSTITUTE AND LOCATION 



TOTAL MANYEARS: 

2. on 



PROFESSIONAL: 



? . n n 



OTHER: 



0.00 



CHECK APPROPRIATE BOX(ES) 

□ (a) HUMAN SUBJECTS 

□ (al) MINORS □ (a2) INTERVIEWS 



a (b) HUMAN TISSUES 



□ (c) NEITHER 



SUMMARY OF WORK (200 «ords or less - underline keywords) 

Summary 

This project involves the application of digital computer technology to the 
diagnosis of disease based on microscopic morphology of cell specimens. 
Techniques of digital image processin g are applied to delineate objects of 
interest in digital images of human breast aspirates and to compute mathem- 
atical characterizations of various morphologic attributes such as size, 
shape and texture. Techniques of statistical decision theory and cluster 
analysis are applied to the problems of classifying or identifying individual 
Cells and defining a grading system for smears. 



14 



PHS-6040 
(Rev. 10-76) 



Characterization of Cells from Breast Aspirates 

The objectives of this project is to characterize images of cells from 
human breast aspiration biopsies, and to distinguish malignant from 
benign cells. Aspiration biopsy is used widely in Europe but not in 
the United States. 

Biopsy material was prepared at the Department of Clinical Cytology, 
University Hospital, Uppsala Sweden, stained with Papanicolau stain and 
scanned there, digitized and recorded on magnetic tape. The Uppsala 
scanning microscope formerly had 7/10 micron square aperture and had a 
moving stage and recorded 256 gray levels. The current model uses a 
256 element linear photodiode array. Using the PEEP system, cell images 
were displayed and thresholded. This generated objects which were sub- 
jected to feature extraction using the PEEP/DECIDE system. Eighteen 
numeric characterizing features were extracted for each of 100 cell 
nuclei in five cell categories. Examples of these features are area, 
average brightness, average chord, entropy, kurtosis and skewness of the 
optical density histogram, density, variation in brightness, variation 
in chord length and variation in diameter. 

Each of four categories of benign cells was distinguished from the 
category of malignant cells. Using quadratic discriminations, com- 
binations of as few as four features and in many cases a single feature 
allowed perfect discrimination between pairs of cell categories. Useful 
features included skewness of the optical density histogram, nuclear 
area, length of major axis, density and entropy. 

A manuscript, "Computer Assisted Identification of Cells in Needle 
Aspirates of Mammary Tumors", has been presented at MEDINFO'77 and 
published in the Proceedings. 

Significance 

The practical significance of this research lies largely in develop- 
ments in the automation of cytologic examinations and in the quanti- 
tative, objective investigation of the correlation of microscopic 
morphology with the prognosis of disease. The project will be con- 
tinued both in the United States and Sweden using more carefully 
scanned cells prepared with gallocyanin chromalum. 

Future Plans 

We are awaiting the arrival of data tapes from Sweden containing 
10,000 digitized images of breast tumor cells stained with gallocyanin 
chrome alum and scanned with a new linear photodiode array scanner at 
0.5p for analysis. We are also awaiting -the arrival of data on the 
relative grading of 100 smears for purposes of analysis by multi- 
dimensional scaling. In addition, we plan to use cluster analysis to 
detect patterns or trends in epidemiological data accompanying the 
cell images. 



15 



Consulting Activities 

Project Description, Objectives and Significanc e 
The field of image processing in medicine is in a state of rapid 
flux as a result of new techniques such as computerized tomography and 
ultrasonography. These consulting activities are necessary to help take 
my experience and expertise gained in the mathematics and computer 
programming of image processing to other groups and to bring back in- 
formation about progress in areas outside of DCRT. 

Activities in 197 8 

.member of the Board of Directors of the Biomedical Engineering Soc- 
iety and Associate Editor of Computer Graphics and Image Processing, 

•reviewer IEEE Transactions, Computer Programs in Biomedicine, Computer 
Graphics and Image Processing, 

•steering committee of the Dahlem Conference on Biomedical Pattern 
Recognition, 

•organizer, Short Course on an Introduction to Biomedical Image Pro- 
cessing for October 1978 ACEMB meeting, 

.membership in the NSF inter-Agency Panel on Medical Physics, the NSF 
Automation Research Council, and chairmanship of the NSF Committee on 
an Ultrasonic Tissue Signiture Library. 

.work on the IEEE Pattern Recognition and Machine Intelligence Group 
and its subcommittees on Image Processing and Scene Analysis, Data Bases 
and Biomedical Pattern Recognition, 

.steering committee and treasurer 1979 IEEE Computer Society Con- 
ference on Image Processing in Radiology, 

.organizer, session on Computerized Tomography, 1978 IEEE Conference 
on Patteim Recognition and Image Processing, 

.the Division of Cancer Biology and Diagnosis of the National Cancer 
Institute as a member of the Diagnostic Research Advisory Committee, 

.the National Heart, Blood and Lung Institute, Division of Research 
Resources, RANN, The Division of Computer Research and the Division 
of Engineering of the National Science Foundation, 

•consultant NCI/DCBD Computerized Transaxial Tomography Project, 

.ad hoc reviewer NCI/DCBD Dose Reduction in Computerized Tomography 
project, 

.NCI representative to Standard Protocol for Evaluation of Imaging 
Techniques in Cancer Diagnosis project, 

•faculty DCRT Computers in Clinical Medicine Elective. 

a National Visiting Lecturer for the Society for Industrial and 
Applied Mathematics (SLAM) on topics related to image processing and 
decision analysis and Visiting Lecturer and Scientist, Department of 
Computer Science and Numerical Analysis and Department of Clinical 
Cytology, University of Uppsala, Uppsala, Sweden. 

Future Plans 

It is anticipated that the same level of consulting activites will be 

maintained in fiscal 1979 with the same federal agencies, and profes- 



17 



Consulting Activities 



sional societies receiving advice and assistance. In addition, 
I will be on the advisory board of the new IEEE Transactions on 
Pattern Analysis and Machine Intelligence. 

Publications 

Prewitt, J.M.S. 

On Some Applications of Pattern Recognition and Image Processing to 
Cytology, Cytogenetics and Histology. Doctoral Dissertation, 
Department of Computer Science, Uppsala University and Department of 
Clinical Cytology, University Hospital, Uppsala, Sweden, April 1978. 

Prewitt, J.M.S. 

Reconstruction Software for Computerized Tomography. Proceedings of 
Computer Software Applications Conference 1977, IEEE Press, 1977, 
pp. 285-292. 

Prewitt, J.M.S. 

Interactive Decision-making for Picture Processing. Proceedings of 
the 1977 IEEE Conference on Decision and Control, IEEE Press, 1977 
pp. 373-379. 

Prewitt, J.M.S. and Mendelsohn, M.L. 

Analysis of Cell Images. Reprinted in Machine Recognition of Patterns, 
ed. by A.K. Agrawala. Wiley, New York, 1977, pp. 362-380. 



18 



:p\u-i:<i-'.:T r,r 



Pr,'.J-OT !ai;Vbi.K (Oo fiOr i-'-c this ^j'.ice) |iL-..Tii, tUUGAT !('!;, A:;!' ■.■.tLi" 

r.JBLIC f-tALTII '.ERVlCt 

NOTice or 



ARt 



^iiOJLOT ;.li:'b:.K 



unRA.uRAL RESEARCH r-.ojEOT ZOl CT 00038-01 OSTC/OD 



PER 1 00 COVERED 



October 1, 1977 to SeptemberJO^ 1_97_8_ 
Natural Language Access to Clinical Data Bases 



:;av£C;, labo-;ato-,y a::d i;iSTiT;)TE aff iliatioi.s, a'.d titles of pi(i;,cipal i'.vfstigatcrs a!:o all othlr 
Tk ore;. SIGNAL ?£i>s::.:;ll engagel. o;; the froject 

Martin N. Epstein, M.S., Research Programmer, DCRT 



OPERATING u:i 1 ..(if j-y) 

SRI International, Menlo Park, California 
University of California, San Francisco, California 



jLAO/BRAIXH 



Office of Scientific and Technical Communication 



;ct!o;< 



liiSTiTUTE a;:o LC3;Tio:i 
Division of Computer Research and Tech nolo gy 

o"Ti-:i-i;;' 



rCTAL MAuYEARV: 

.9 



FF:i;FtG3IO;!AL: 

.9 



CHECK APFRCFRIATE SO,X(ES) 
'J (a) i'UMA:; SUbjLGTS 

rj(il) MI;:CR^ ": (a-) 1 flTFF'.' i l.-,S 



G (b) h'L'VAf; TI?SUEi 



□ (c) i;eith:t, 



SU'^VRY OF WOi^K ;,.-00 words or let,: 



under 1 i ■ 



:y:.vrcb, 



The objective of this research is the development and evaluation of a system, 
called MEDINQUIRY, that will allow physicians personal access to patient 
medical data through natural language queries to support both patient manage- 
ment and clinical research. 

A prototype MEDINQUIRY system has been implemented for a small data base on 
malignant melanoma. The physician can input queries in English that retrieve 
specified data for particular patients or for groups of patients satisfying 
certain characteristics, that perform simple calculations, that allow 
browsing through the data base, and that assist in identifying relationships 
among attributes. 



19 



Objectives: 

1. Develop a clinical data base system that will allow clinicians personal 
access to data on specific diseases using queries formulated in English. 

2. Collect and maintain in machine readable form clinical and histological 
data on patients with malignant melanoma and cancer of the colon for use 
in the system. 

3. Study the use of the system by physicians. 

4. Evaluate the effects of using the system to aid the physician in patient 
management and in studying the clinical course of the disease more precisely. 

Methods: 

A system called "LIFER", developed at SRI International, provides a facility that 
allows a language interface builder (in this case the PI) to specify rules re- 
quired to process English language query requests. These rules are used to create 
and modify the grammar, lexicon, and language interface functions. These tasks 
can be accomplished on-line and is one of the major attractions of the approach 
pursued. Special purpose facilities available include an ellipsis capability to 
process incomplete data, a spelling corrector, and a paraphrase mechanism for 
relatively unsophisticated users to add new rules to the system. A requestor can 
follow a line of inquiry to test a particular hypothesis by entering a sequence 
of requests that depend on each other. 

Findings: 

1. The methods applied provide a stepwise facility for determining the 
classes of questions that physicians desire to ask of a clinical data 
base. 

2. Through use of the English language access facility, hypotheses on mela- 
noma that have been established in the literature can be confirmed and 
new hypotheses can be tested. 

3. The availability to physicians of on-line English language access has 
encouraged a more active interest and participation by physicians in the 
data acquisition process, in completing missing values in the data base, 
and in actively pursuing the information in the data base. 

4. MEDINQUIRY has been demonstrated to physicians who are specialists in 
melanoma, other cancers, and a variety of other diseases. The reactions 
have been uniformly positive, even though th-e existing data base is not 
large enough to provide significant results. The ability to formulate 
queries in ordinary English and to have the results displayed immediately 
is extremely appealing. The facility for correcting spelling errors 
received special commendation. 



20 



5. As originally collected, the melanoma data base was incomplete and 
contained many errors. An extensive effort was required to extract 
the data elements from the primary records, to edit the attributes 
and values, and then to create the data base. 

Proposed Course: 

1. Extend the capabilities of MEDINQUIRY as an experimental system with 
an expanded data base on melanoma. 

2. Create a second data base for Cancer of the Colon to test the generality 
of the concepts developed for melanoma data access. 

3. Incorporate a knowledge base in the system to reflect physician defined 
higher order concepts and thus reflect the current state of medical 
understanding of the disease. A knowledge base also would contain rules 
for understanding the context of requests and developing appropriate 
prompts and paraphrase capabilities that show a greater degree of under- 
standing of the medical domain. 

4. Evaluate the use of the system by physicians and study the requirements 
for adapting the system for different diseases and on other computer 
systems. 

Potential Significance to Biomedical Research: 

1. Aid the physician in clinical practice in planning therapy and assessing 
prognosis for his/her patients. 

2. Assist the clinical investigator in identifying and studying the relation- 
ships among data attributes. 

3. Study the potential applications of such systems in the conduct of clinical 
trials. In particular, they may be used to assess the efficacy of thera- 
peutic protocols. 

4. Provide convenient access to medical data bases to allow physicians to 
browse through their data and more adequately study the clinical course 
of a disease. 

Publ ications: 

Epstein, M.N., and Kaplan, E.B.: Criteria for Clinical Decision Making. In 
Schneider, W. and Sagval 1-Hein, A.L. (eds.): Computational Linguistics in 
Medicine. North-Holland Pub. Co., 1977, pp. 35-44. 



21 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



Z01 CT 00032-02 OSTC/OD 



PERIOD COVERED 
October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

Tumor Growth Curves 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



Judith M.S. Prewitt, Ph.D. 



COOPERATING UNITS (if any) 



LAB/' BRANCH 

Office of Scientific and Technical Communication 


SECTION 


INSTITUTE AND LOCATION 
DCRT/NIH 


TOTAL MANYEARS: 

.05 


PROFESSIONAL: 

.05 


OTHER: 

0.00 



CHECK APPROPRIATE BOx(ES) 
H (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



g (c) NEITHER 



SUMMARY OF WORK (200 woras or less - underline keywords) 

A manuscript on tumor growth curves is being revised. A new second order 
non-linear differential equation for monotonic growth is proposed. This 
equation contains all classic models such as the Gompertz, logistic and 
exponential as special cases, and introduces an infinity of intermediate 
curves with varying inflection points. The solutions to the growth 
equation are obtained in closed form, and can be expressed in a time-free 
form which allows inter-comparison of the shape of different growth curves 
independent of age and dependent only on tumor extent. Numerical solution 
of equations to fit the general growth curve to data is unique. This 
equation should facilitate comparing data from different laboratories and 
provide an intrinsic characterization of the growth phenomena. 



22 



PHS-6040 
(Rev. 10-76) 



July 1, 1977 through June 30, 1978 

NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 

1 . DCRT 

2. OFFICE OF ADP POLICY 3. Henry J. Juenemann 

COORDINATION Chief 

I. SUMMARY 

Function 

The Office of ADP Policy Coordination, under the direction of the Assistant 
Director of the Division, has two closely related function. It serves as: 

a focus for NIH-wide coordination of automatic data processing policy 
matters . 

a central NIH point of contact with the Public Health Service, the 
Department of Health, Education and Welfare, and other HEW agencies, 
the General Services Administration and the Office of Management and 
Budget on policy questions and NIH's participation in policy development. 

Scope 

The Office provides advice and assistance about internal DCRT operations 
and coordinates DCRT ' s ADP policies and activities with those of other 
agencies. This includes advising the Director of DCRT and through him 
the Director of NIH on ADP policy matters, assisting the NIH Division of 
Management Policy on questions relating to its responsibility for 
administrative computer applications, reviewing and evaluating proposals 
from NIH B/I/D/O's for ADP and computing procurements and contracts, 
directing the development of the annual NIH ADP Plan, representing the 
NIH in PHS and DHEW policy formulation efforts, working with GSA and OMS 
staff on procurements, 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. 

Highlights of the Year's Activities 

The annual ADP Plan, required by DHEW was again completed. This 
management process, involving all B/I/D/0 provides a forward look at 
projected ADP efforts necessary to support NIH research and research 
management programs. It showed a 53 million dollar, 654 man-year ADP 
program in place during Fiscal Year 1973 growing to 58 million dollar and 
676 man-years by Fiscal Year 1979. 

One major accomplishment of the office during the year was the development 
of an easy to use set of instructions for NIH staff offices to use when 



23 



ordering teleprocessing services from the private sector. GSA had 
established a mandatory program for this but the GSA instructions contained 
ambiguities and were not easy to use in the NIH environment. Therefore, in 
conjunction with Procurement Branch, a set of instructions specifically 
tailored to NIH needs were developed to ease the burden on NIH. 

In the process of monitoring the policy implications of NIH's ADP involvements 
and assuring conformity to existing 0MB, GSA and Department regulations, this 
office reviewed 300 proposals for ADP equipment or services during the period. 
This is a 70% increase over the corresponding period last year. 

A total reprocurement of the IBM System 370 component of the DCRT/NIH 
central computing facility has been mandated by GSA. The strategy for the 
recompetition was largely worked out with GSA during the year. Extensive 
discussions with GSA and the Department finally resulted in GSA allowing a 
more reasonable approach than they had been willing to allow over the past 
few years. The solicitation document is now being prepared. Due to the 
complexities of the NIH central 370 utility, that effort will undoubtedly 
stretch well past the close of FYSO and will consume significant levels of 
highly skilled and scarce technical talent from the Division. 

Later in the fiscal year, in spite of objections from the Division, the Office 
of the Secretary directed that two of the four central processing units in 
the central IBM 370 System, which are subject of the total reprocurement 
effort, be converted from lease to purchase. 

Another total system reprocurement is simultaneously being conducted for the 
Division's scientific time sharing DECSystem 10. 

The competitive solicitation for editing disply terminals to be available 
NIH-wide which GSA had approved last year was finally completed after more 
than two and half years delay caused by disagreements with GSA. Competitive 
solicitations for two other types of terminals were under way as the end of 
the year approached. 

In reviewing proposals for contracts or in processing equipment acquisitions, 
this office is very fortunate in being able to call on the consultive 
expertise of the other laboratories and branches of DCRT. During the year 
they were extremely helpful in insuring technical merit is part of the review 
process . 

Future Plans 

Federal ADP policies and requirements continue to become more and more 
complex as 0MB, GSA, DREW and PHS become more and more involved. As a 
result, the office must spend an increasing portion of its available man- 
hours in attempting to guide NIH policy in productive directions and in 
meeting regulatory requirements. The office will continue to work to spare 
large numbers of NIH research and research support staff members the task 
of becoming expert in the many nuances of ADP-related regulations. However, 
it is anticipated that the two simultaneous full recompetition efforts for 

24 



both of NIH's main computer systems will consume most of the resources of 
this office during the upcoming year and will in addition, require extensive 
assistance from OD and CCB staff. 



25 



October 1, 1977 through September 30, 1978 



NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



1. DCRT 
2. OFFICE OF ADMINISTRATIVE MANAGEMENT 3. L. Lee Manuel 

I. SUMMARY 



Chief 



Function 

The Office of Administrative Management, under the direction of the 
Executive Officer, provides administrative, financial and personnel 
functions to support the Division's programs. The office serves as 
liaison to these functions with the NIH Office of Administration and 
other OD, NIH offices. 

Scope of Activities 

The office handles the usual range of administrative managerial functions 
for an NIH research division of almost 300 people. The Financial Manage- 
ment/Project Control Sections are currently tracking 2000 project accounts 
involving 6000 registered users of the DCRT computer facilities and 
services. These services have grown to an estimated 530,000,000 in FY 78. 
Requisitions, contracts, travel, and training documents are processed by 
the Administrative Services Section and covers a variety of procurements 
of approximately $20 million. ^ 

Highlights of Years Activities 

The Personnel Section, in compliance with Civil Service Commission and 
DHEW directives, continued the position-classification review of all 
positions. Also the conversion of position descriptions to the new Factor 
Evaluation System began to be implemented. 

Major space and renovation continued for Building 12 and 12A. The 
Administrative Services Section implemented new procedures in the areas of 
payroll processing, check distribution, and procurement documents. It 
also did a major review of DCRT staff identification codes for the revised 
NIH mailing key system for document distribution. DHEW initiatives in the 
areas of grants, contracts and consulting services required the prepara- 
tion of numerous detailed recurring reports and plans for the current 
fiscal year and will continue into FY 1979. 



26 



Future Plans 

The coming year will include major moves and other renovations within 
Buildings 12, 12A and 12B. These will allow consolidation of some 
currently scattered DCRT staff and more unified services for DCRT clients. 
There will, undoubtedly, be new administrative managerial requirements and 
modifications of old ones forthcoming from the NIH, the PHS, the Depart- 
ment, the GSA and the Civil Service Commission. These requirements 
will be implemented within the framework of DCRT program policy. 



27 



Oct. 1, 1377 through Sept. 30, 1978 

,'MTIOMAL IMSTITUTES OF HEALTM 
DIVISIOfJ OF COMPUTER RESEARCH AND TECHMDLOGY 

Sunmary of 3ranch Activities 1. PORT 

2. LABORATORY OF STATISTICAL AND 3. Janes E. MosJ-^ann 

flATMEf'ATICAL METHODOLOGY Chief 

' . ■ I . SUMMARY 

Func t i on 



I 
I 



The Laboratory of Statistical and Mathenatical Methoddgy (LSM)I 
conbines research in ma thema t i ca 1 statistics^ mathenatics, 
computer and information science, with collaboration and servicel 
in these areas to NIH researchers and administrators. There are| 
four sections in LSM: 

. The Statistical Softv/are 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 nrograri packages. 
Three specialists in scientific programming are led by a 
computer systems analyst whose specialty Is statistics. 

. The Biomathemat i cs and Computer Science Section (BCS), 
directed by a mathematician, performs Independent research 
and provides consultation and collaboration in the 
specialties of its eight mathematicians, computer scientists 
and programming aides. 

. The Statistical Methodology Section (SMS) v/orks closely with 
the Statistical Softv/are Section. Four individuals who work 
under the direction of a mathematical statistician provide 
b los ta 1 1 s 1 1 cal 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. Five 
individuals work under the direction of a computer systems 
analyst who specializes in computational linguistics. | 

Scope of Work 

LSM staff interact v/ith all NIH Institutes and with government 



28 



agencies outside HE'W. Fiscal year 78 was LSf*'s fourth year as a 
separate entity within DCRT. The volume of its computational and 
consultation services continued to expand while its research 
activities were maintained at about the same level as the 
preced i ng year . 

Highli,';hts of the Year's Activities 

Computat ion . A major part of LSM activity is the offering of 
statistical and mathematical program packages to the fJ I H user 
community. LSM accepts responsibility for evaluation of new 
program packages and their suitability for NIM. When LSM does 
offer a package to the NIH community, LSM makes three basic 
commi tmen ts : 

. The maintenance of the package, with adequate documentation, 
through NIH computer system changes, package updates and 
correct ions . 

. The rapid response to queries concerning user access to a 
package program Including job control language and program 
parame ters . 

. The assistance in Interpretation of results. 

During this year, as in the past year, SSS maintained the 
following program packages and programs: 

BMD, Biomedical Computer Programs, UCLA 
3MDP, Biomedical Computer Programs, P-series, UCL^ 
SPSS, Statistical Package for the Social Sciences, SPSS, Inc. 
SAS, Statistical Analysis System, SAS Institute, Inc. 
PSTAT, Princeton Statistical Package, Princeton University 
IMSL, International Mathematical and Statistical Libraries, l"SL 

i nc. 
MSTATl, Collection of Mathematical and and Statistical Programs, 

DCRT. 

The effort expended In the commitment to maintain these packages 
is considerable. During this year every package went through at 
least one update. The effort expended in response to queries 
concerning package access Is a>so considerable ani reouires 
continuous availability. During the year over i+,000 calls were 
responded to by SSS staff alone. Two courses were taught on each 
of the packages SPSS, SAS, and one on 3f!DP. 

The use of program packages has shown considerable increase this 
year over fiscal 77. The average accesses per month of all the 
statistical packages rose from 5000 during fiscal 77 to almost 



29 



9Q0G in fiscal 78. For the second year In a rov; S4S experienced 
the largest increase of any of the packages. SAS averages almost 
i+OOO accesses per month, up from 2000 per month in fiscal 77. 
The average number of accesses per month for SPSS increased from 
2600 to 3900, an increase of 50% over the last year for SPSS. 
The average combined accesses of the BflPP and Bf'D packages 
decreased from 900 accesses per month in fiscal 77 to around 750 
accesses per month this year. Bf'.DP was up slightly from 500 
average accesses per month in fiscal 77 to 550 average accesses! 
per month this year, v/hile Bf''D went dovyn from 350 average per' 
month to 100 average accesses per month in the same period of 
time. As an example of a package used for soecialized purDoses,! 
PSTAT averaged 20 accesses per month, down from 30 average! 
accesses per month In fiscal 77. The main programs in f!STATl 
averaged 150 accesses per month, which is about the same as I ni 
fiscal 77. The subroutine usage of f-'STATl, which can only be| 
estimated, is much higher than the main program usage. Accesses 
to the IflSL package cannot be countei, but it is estimated thatg 
usage Increased during fiscal 73. I 

The 3 ioma themat I cs and Computer Science Section maintains t\;o 
computer packages for b I oma thema 1 1 ca 1 modeling at "i\l\: MLAB onl 
the PCP-10 and MODELAIDE on the S/370. Both packages were* 
developed by BCS staff members. MLAB Is widely useJ at ,'IIfl and 
In many other locations In the U.S. and abroad. it has beenl 
used by hundreds of NIM biomedical researchers, with only! 
occasional contacts with BCS computer specialists needed in most 
cases. ViODELAIDE Is not so vndely used, but has greater 
capabilities than MLAB for analysis of very large models. During 
the fiscal year, flLAB was Improved by addition of new 
curve-fitting control strategies, and addition of hldden-line|. 
suppression to the three-dimensional graphics facilities. A newl 
reference manual for MLA3 was distributed near the end of last 
fiscal year, and the next edition of the reference manual Is now 
in preparation. A beginner's guide to MLA1 Is in preparation,! 
with distribution expected within a few months. ^n introductory* 
course for MLAB was taught twice during the fiscal year. A 
number of nev/ features are currently under development for MLAB, I 
including contour map graphics and fast Fourier transform! 
capab i 1 i ty . 

BCS continues support of C-LA3, a .computer package for patternl 
recognition and cluster analysis developed by a BCS staff member. 
A number of new analysis algorithms have been added, and existing, 
methods Improved. C-LAB compatibility with MLAE has been! 
maintained. A course in C-LA3 v;as taught during the fiscal year." 
Revision of C-LAS documentation Is now In progress. 

BCS also supports the University of Utah computer package REPUCE,! 
for symbolic manipulation of algebraic and differential equation 



30 



i 



f ornu 1 as . 

BCS staff menbers also taught courses in the SML progranning 
language and in curve-fitting nethods. 

Consu 1 tat ion . The diverse nature of LSM consulting is Indicated 
by the projects and activities listed in Part II. This list is a 
sample of activities and is not exhaustive. 

The pattern of LSM consultation activity remained similar to that 
observed in the previous fiscal year. Estimated percentages 
during fiscal 78 are shown belov/: 

Mathematical or statistical advice 

v/i th limited computer use 10?^ 

Mathematical or statistical advice 

with considerable computer use 50"^ 

Computational advice alone kO?i 

The continued availability and use of genera 1 -purpose statistical 
and mathematical packages like SAS and f1L\3 has maintained the 
large component of computer use in these figures. f-li S 
activities, which involve relatively little consultation, are not 
represented in the above percentages. 

As in previous years there v;as considerable variation in the 
amount of time required for an LSM consultation. Some very brief 
consultations are very successful, and are brief precisely 
because there Is a known answer to the s ta t I s t i ca 1 /ma thema t I ca 1 
question posed. Other consultations involve extensive time and 
Stat i s t i cal /mathema t i ca 1 research as well. For example, the 
collaboration over the past three years with Dr. A. Cheever, 
Laboratory of Parasitic Diseases, NIAIO, on the study of 
schistosomiasis continues actively. This consultation has 
provided an excellent data base for application of size and shape 
statistical methods developed by the Chief and staff of LSM. 
Several publications resulting from this collaboration are now 
published or in press. 

Many consultations which involve considerable LSM effort do not 
involve new statistical research. . For example, LSM provided many 
computer statistical analyses for studies of atherosclerosis 
conducted by Dr. V.'illlam Roberts, i'iHL3l, during fiscal 78. The 
data base consists of autopsy and other Information from deceased 
atherosclerotic patients, including measurement of major coronary 
vessel narrovnng. Calcium, Iron, hemorrhage, thrombus, 
constituents of plaque, medical history, demographic information, 
and so on. Some studies compared vessel pathology within or 
between certain groups of patients, either comparing diseased 
with control groups or comparing groups with different foi — is of 

31 



I 



heart disease. Other analyses consider patterns of pathology 
within individual patients, such as co-nparlson of Calciun v/ith. 
heart valve danage, comparisons of narrowing of different vessel^ 
or at different points on the same vessel, comparison of vessel 
pathology with vessel size or heart size, and so on. Studies, 
were also made involving demographic factors such as sex and age.!' 
Many analyses were straightforward statistically, but require-?' 
care and attention to detail in problem formulation and computer! 
ana 1 ys i s . | 

Research. BCS research Included projects in b loma thema t I cs,l 
general mathematical methods, and computer science. ResearcHf 
continued on the "symmetric axis" method of shape description,! 
Including substantial Improvements In computer algorithms forji 
general -purpose symmetric analysis for shapes. A cooperatlvJ 
study of human mandible growth and development using symmetric' 
axis methods Is In progress. The principle investigator recently 
visited France, lecturing on visual and geometric theories byl; 
invitation of Professor Rene Thom. 11 



-1 



1 



Cooperative effort continued on a project to develoo a cance 
diagnostic tool using data from a f 1 uorescence-ac t i vate 1 cell 
sorter. A new curve-fitting algorithm for sti^f differential 
equation models was developed, facilitating computer analysis ofij 
larger models. A formal logic appropriate to several varieties ofl 
algebra used In applied mathematics was developed. Research on 
computer storage and retrieval algorithms Is continuing. 

In SMS, research on multivariate statistical methois has been 
active during this fiscal year. Research on size and shape 
methods has drawn Interest from the statistical community, with 
P''0^' ij • i^. Patll of Pennsylvania State University, Dr 
j^an James from C.S.i.R.O., Melbourne, Australia, and Dr. John C. 
Gower, Rothamsted Experimental Station, England, visiting ISr for 
research discussions during the fiscal year. These methois have 
been improved and tested on various biological and me-Jical data 
bases during fiscal 73. A study of procedures for mu 1 t i n 1 e ratio 
estimation was undertaken, with application to a one-factor 
a 1 1 ome t ry mode 1 . 

The Medical Information Science Section system for information! 
storage and retrieval of pathology was reviewed for application' 
to autopsy reports. Further improvements v/ere made in the formal 
rules for automated morphosemant i c segmentation of meJicall 
compound words derived from Greek and Latin, and for paraphrasing! 
them In English and French medical languages. l7ork continued 
towards the development of a formal and practical basis for theli 
construction of computer-oriented medical mi crog loss ar I es . | 



32 



F'jture Plans 

i\o najor shift in laboratory service or research is anticipate:! 
In the coning year. Current levels of support of statistical and 
mathematical program packages, and consultation and user 
assistance v; i 1 1 be maintained or expanded. B ioma thena 1 1 ca 1 and 
b ios tat i s t i ca 1 research projects will be continuations of those 
already initiated and reported here. 



33 



II. LSM PROJECTS AMD ACTIVITIES 



The f ol low 
the year. 
research r 



ing list contains major LS 

Fo 1 1 owi ng th 
e ports are pr 



e list of pub 1 i ca 
esented . 



onsul t I ng ac t I vl t I es 
tions, the LSf individua 



Prenatal bone development in rhesus nonkevs 



DRS. i)urg 
at severa 
with wrist 
a s ta t i s t i 
data, and 



I ca I exami na t 
1 f i xed pos 
-bone OSS i f i c 
cal model usi 
computed nume 



ion of rhesus mon 
t-conception tim 
a t i on exam i ned in 
ng Markov process 
r i ca 1 resu 1 ts bas 



M. Michejda, VR 
key fetuses was conducte 
e intervals an:i at birth 
each case. LSi'l deslgne 
es for ana 1 ys i s of th 
ed on the mode 1 . 



1 



Nurse Job satisfaction e valuation . R. 



study was 
primary vs 
ques t iona i 
response d 
than one y 



developed to 

team nur 

res and in 

ata. LSM has 

ear, and a jo 



descr i be the 1 ev 

sing. LSM assi 

contingency tab 

CO 1 1 aborated on 

int publication i 



Carl sen, NL'RS, CC. This, 
el of job satisfaction o 
sted in the design o 
le analysis of question 

this project for mo 
s i n prepara t i on . 



1 



Behavior of coati 



H. 



Smith, LDN, rJICUHD. Etho.erams 
constructed representing the behavior of coati (raccoon- 
carnivores) in captivity and in the wild. LSM assisted in 
arrangement and editing, and in analysis of data by tv/o-way ' 
and multiple regression. 

Atherosclerosis oatholosv studies . v;. Roberts, IR, PA, m; 
Data obtained from autopsies and medical histories of decs 
atherosclerotic patients was analysed for the effects of 
different factors. LSM provided continuing statistical aJvice 
and computer analyses, primarily using SAS, for evaluation of 
different aspects of the data. 

Hemoglob in kinetics . R. Berger, B. Balko, P. Smith, TD, 
uHLBI. LSM assisted in developing suitable methods for 
curve-fitting and solving differential equations for simulation 
of hemoglob i n-1 igand binding, and experimental control on a 
laboratory minicomputer. 

Redox s tudies of Cytochrome . R. W. Hendler, L3, MHLS I , One 
experiment consists of about 50 spectra. Each spectrum, at a 
distinct voltage, has small sharp -.peaks suoerimposed on broad 
background peaks. LSM assisted in developing methods to separate 
the smal 1 peaks from background and from each other, and to 
explain their growth with voltage in terms of cytochrome 
ox i dat ion . 



Oxygenation o 
prov i ded 
rout I nes 



whole blood . R. M. V/inslow, CL, 

consulting on curve-fitting techniaue, e.; 
for approximating tv/o-d imens iona 1 surfaces 



flHL3l. LSM I 
deve loping 



34 



Funarase kinetics . I. Darvey^ PSl, DCRT, L. Kohn, LPS, N'i\fTD. 

LSM offered advice on possible methods for distinguishing or 

deriving mechanisms to explain the suppression of enzyme activity 
at high substrate concentrations. 

Copper kinetics in Man . E. A. Jones, J. fl . Vierlin?^, DP, 
•JIAMDD and 'W. Rumble, R. ^amodt, 'lf1, CC. \ paper was published 
on uptake of copper by ceru 1 op 1 asm i n and its implications for the 
diagnosis of various copper overload diseases. Discussions have 
been held about possible expanded study. 

Scintillation counting . Ramon Tate, CSL, DCRT, Joe 
Fens termacher, CHPH, NCI. Least generalized variance 
experimental design has been found efficient (in the vector 
minimization sense) in some cases, but there are counter- 
examples. LSf'i is assisting in evaluation of this measure's 
efficiency for scintillation counting experiments. 

Pharmacok i net ics . R. Lutz, R. Dedrick, REI, DRS. Large 
systems of differential equations, simulating drug :! i s tr i but ion, 
are fit to data. LSIl advised on modeling and possible use of the 
Modelaide system, since the formulas are too unwieldy for iMLA3. 

Enzvme kinetics . E. Silbergeli, et.al., LP, 'JIMCns. The 
mechanics of specifying and using Clelands initial rate models in 
MLA3 was discussed and illustrated at length with real data. 

Measurement of kidney size . P. 3rad 1 ey-Moo re, MM, CC. By 
taking lateral X-rays of the kidneys in addition to the posterior 
view, a more accurate measurement of kidney size can be made. \ 
formula was developed for combining the results of the tv/o views 
and adjusting the standard kidney length measurement accordingly. 

Improvement of a standard clustering technique . C. Edelbrock, 
LDP, :jIM(I. Stein's method is being applied to the centroIJ method 
of clustering to produce improved hierarchical clustering 
diagrams (dendrograms). The method uses based estimators of 
multivariate means and provides better predictive value In 
general . 

Study of biochemical polymers . F. Howard, Lf!2, N' I AMP P. 
Polymers of Cytosine and Inosine are generated in vitro under 
controlled conditions, and amount-of product containing specified 
numbers of Cytosine molecules in successive primary sequence 
locations is measured. LSM provided a theoretical formula, 
suitable for computer evaluation, of the amount of product 
expected assuming random grov/th of the polymer In proportion to 
the supply of each base in the preparation. 

Immune function of the soleen in mice . M. Geler, S'.', 'llh'CDS. 
Measurements of labelled antigen persisting In the spleen are 



35 



':1a ys after 
Tia thena t i ca1 
u s i n r " L ^ =* . 



1 ~£n ^'-"' t i on 
I n j 3C t i on 
io :ia 1 i n"; and 



o'^ anti'^o"'y in ■^ 
of the anti 
I n genera t i on 



I ;e -'urin'^ a period o, 
;en . LSf' -Tssiste] I 
of graphical i i s p 1 :i y sj' 



Repetition o~ ori'-ia''y sequence triolets 



LP.C, .-JIAMrc. LS»^ provided advice an 
orobabi 1 i t i es associated with reoetition 
occurring in primary sequences of natural 



I n pro to ins . M. Sai-offJ 
"" analysis on conb i na tor i a 1 

oeot i de cha i n 



of short 
orote i ns 



i 



Water quality data at MjH animal care f^ci 
DRS. Stream water quality parameters 



I ty 



Wl 



son, ES'^i 
'^ .^ i 1 y i 



are inon i tored 
Broad Run, above and below effluent e-nission fron the °ool esv i 1 1 elf 
Mid aninal care facility. LS'^ assisted in locatin- errors i 

MLAB procedures used to ^^tenerate t^raohical displays" of v/ate 
qua 1 i ty da ta . 

SI eao ana 1 vs i s . 



V;. Duncan, ^P, ^^M'^. The analysis focuses o 
the relationshio between Tiental illness and slee 

D i scr i Ti i nan t analysis was enployed to evaluate 
contribution of sleep variables in d i s t i n<;u i sh i n 



oi nor-nal subjects, deoressed (uninolar anr 



exa-n i n i n ^ 

di s orders. 

poss i bl e 

between groups 

bipolar), and Insomniacs patients 

Selection for body weight in mice , 
primary purpose of this project is 
amount of genetic variation remains 



K. Smith, VR, 

to determine if a 
in h i gh 1 y i nbred 



PRS. TH 
s i g n i "^ i c a n t : 
strains of 



it should be oossible 
use to maintain these 



to make changes 
strains wh i ch wou 1 d. 



"^ice. If It can be established that genetic variation does 

exist in these strains, 

management procedures 

reduce the cost of oroduction 

developed In this study are 

objective evaluation of the 

Inbred strains of mice. 



"?l 



The 
d i rec te' 
effects uDon 



statistical 

towa r-^. 
genet i c 



o r o g r a m : 



prov I d I ng 
va r I a 1 1 on 



Kidney measurements . ?. R 
kidney contribution with 
regressions and plots, 
give the true length of angulated 
obtained from r'anera Renography. 



the calculated true length and 
size. 



^ Bradley-'^oore, ^'»', Cr. Comparison of 

hloouran and mercury was done wlt!^ 

Also, an empiric curve was derived tq 

kidneys using projected length 

Correlations were made of both. 



the orojected length with I 



^' I d Library journals studv . J.C. Boggess, L, DRR. A! 
requests for duplication of journal articles obtained in a 
month period by the MIH Library were analysed with re<^^ar-^ to year 
of publication, journal cost, multiple requests of articles. 
Much file manipulation, using SAS, was needed to do va 
analyses. Frequencies, correlations, means were orinted. 



1 

thel 

three! 
I yea r 

etc.jj 
r i ous| 



Organizational oerceotions o^ secretaries 



36 



an 



sup*^ rv I sor s 



W. 



I 
I 



Mowczko, 0AM, MIAMDD. Assistance vias given in the devel op~ien t of 

the co"iD'Jter prograTis used to analyse the data in a co-nparison of 

the organizational perceptions of secretaries and suoervisors in 
selected feJeral agencies. 

Li t i gat i on data . M. P. Lockard, 'i^, OD. \ series of computer 
orogra-ns v/a s written and executed, usin^^ oersonnel data, to 
obtain statistics needed by Justice Department lawyers to defend 
Mi;i personnel practices in a oending litigation. Statistical 
procedures used include regression and breakdown. 

Triglyceride and HPL cholesterol : factors regulating M^L 
cholesterol levels. E. d. Schaefer ! '^ ^•0, ^'dL^.i. ^ database 
containing 1038 normal subjects and 355 hyper 1 I poorote I ne-ni c 
patients was submittei to LS^' for statistical analysis. Several 
analyses of variance and I n tercor re 1 a t i on analyses v/ere 
performed. It was found that ^or both normal and 
hyperl i poprote i nemi c subjects, a negative correlation exists 
between HDL cholesterol ani VLDL cholesterol and between MDL 
cholesterol and plasma triglyceride. MDL cholesterol levels were 
lowest in oatients with fasting chy 1 omi cronemi a and were 
decrease-l In hyper t r I gl ycer I demi c subjects. As has been 
previously reported, females had significantly higher HDI 
cholesterol concentrations than did males. Age was not related 
to MDL chol es terol . 

r'oreol neohr i ne production . C. Lake, 1. Kopin, et.al., L^^S, 
MIMH. Originally this study compared noreo I neohr I ne production 
I -I controls with that In several classes n" hyoer ten s I ves . 
Currently, the effects of drugs (pimozide, llthiu'-', a"^phe tami ne, 
et.al.) on blood pressure, pulse, and nore pi neph r i ne-basel I ne as 
well as response to standing are under study. Subject tyoes 
include normals, sch I zooh ren I cs , cata 1 ep t I cs , manic deoresslves, 
narcoleptics, etc. The Statistical \nalysis System (SAS) has 
been heavily used for data processing and statistical computing. 

OuantitaLlve analysis of interlctal behavior In te-^poral lobe 
eol 1 eosv . p. Pedio, C"I, "iinrDS and P. Bear, '-'arvBri ' Jn i v . 
Patients with right or left temporal eolleotic foci are 
contrasted with normal subjects and subjects v/I th neuro-muscul ar 
disorders. Work continues on this project wif^ oerloilc 
additions to the database and reahalyses. DI scr I ~'I nan t analysis 
and analysis of variance ar° employed. 

>'I crogi nssary for Dermatology . C. Brov/n, College of American 
Dermatologists. ISr asslste-" In evaluation of the S^'OD^R" 
(Systematized "iomeocl a ture for Dermatology) ml crog 1 ossa^y, and 
consulted on its possible use for computer storage and retrieval 
of dermatology data. 

English-German International Data Sank for "ancer . institut fu" 

37 



Doku'^en ta 1 1 on. 



I nf or-na 1 1 on 



Krebsforschun'^szentrum, Heidelberg, V/est 'ler-^iany, 



on develop~ient of 

intended as a tod available for 

International Data Bank for Cancer. 



und Statist!':, Oeutschesl 

LSf' consul te^ 
a conpatible encoier for Pn^rlish and ^ernan 

proDosedl 



creation of the 



','eural i^ipulse ~io deling; 



J. 



R i nze 1 , 



Ml ^ " n '"! 



differentia] equations Tiodeling oroDa.<^a t i on of neuronal 
were obtained. LSM assisted in the praparat 

th ree-d I Tiens i onal graphical displays using '-HAB. 



Partial;! 
i Tipul se: 
i on o' 



'ancer screening analysis 



cance r 
at risk 
i nvol ve J 
integral s . 



d. rhen, FSS, ^"^ I . Designing 4 

screening orogra-n involves selection of classes of peooll 

to be screened. The biased sampling statistical analysis 

reguires numerical evaluation of several associate-! 

LSM assisted in computer evaluation using N"A^. I 



Standard curves for ra -li o i mmunna ssav . ^. Schrecker, lo>ia- 

MCI. Standard curves are required for assay evaluation, wh 
can be obtained by curve-fitting normalization data, 
assisted in curve-fitting a standard empirical formula an' 
model obtained from chemical kinetics, using TL'^B. 

Slood 



Measurements in Jaoan . E. Harris, L'^S, n^^^T. Da 

measurements of red blool counts, albumin, and cholestei 
concentrations were collected from iii+,S25 individuals in Ja 
during the period of December 1, 1370 to lune 5T,1977. '^odels 
forecast future observations were derive^-! ^y using Box-Jenk 
time series techniques. The appearance of the autocorralat 



and partial autocorrelation functions usino- different models 
well as the 3ox-Jenkins goodness-of - f I t significance test 
used as guides in the selection of the simolest model. 



vie re 



T I me 

A f, D 



i^reatinlne Clearance Dq te rm i -i a t i on s . 



'\long with LAS, PCRT, a seri3s 



Kl ipoel ,. 
crea t i n i nej 



Ser i e s o 
Mi AMPH. 
determinations in tv/elve patients were studied to detet — line t'^e 
possible effects of immunosuppressive drugs on renal functions In 
women with lupus nephritis. A cusum technique was used to detect] 
step changes In the time series. ^ 



induction o^ Rat Mammary Tumors . R. Evarts, 



or 

wa s 

gro 

a dm 

sta 

of 

com 

ana 

ave 

bea 

aoo 



hy 

St 

up 

I n I 

I I s 
me 

pa r 
lyz 
ra- 
r i n 
rox 



drox 
ud i e 
of r 
s ter 
1 1 ca 
d I an 
I son 
i ng 
e nu 
^ a 
i mat 



y 1 am i n 
d. HA 
a t s an 
ed to 
1 anal 
1 ate 
of t 
2 x 
nber s 
n i ma 1 s 
i on to 



CMT, ^!r| . The 
e (MA) on the induction of mamma --y tumors 
administration was started at age 3R days 
d at age Sk days for a second group; no 

which served as a contr 
the estimation and com 
biological assay motho 
by FIshers's exact te 



ror 



th I rd group 



con s 1 s t? 



a 
ysi s 

ncy periods by 
umor Inci dences 

2 contingency ta'^les, and the comparison 
of tumors and average weiThts o'^ tumors in 
by either Student's t-test or Sattar 
f'e t-test . 

38 



ef -^e 
In ra t^ 
one 
v/i g 

ol . Th^ 
pari son 
is, the, 
St fo 
o: 



^ 



the' 
tutor 



70 



1 



^ Q - -I V/ 



i l-.O 



I y I • n n 



n o r /^ ^T |-| ^ 



nil -'I 



of JiazGpan b i n i i n 5 './ts deter'^ine"! at various concentrations o^ 



a 



-I 



r'j;^, 



BIolo<^ical assay nethois were 'jsei to ierive a stanJord 



curve froi vyhich dose concentrations 
curves . 



■/e r' 



es t i "la te d 



•or 



unkno'.'/n 



''onkey Diet Fiber Study 



M 



M 



orin, VR3, PRS. Rhesus 



ranJonly divided into four groups of 72 anina 
groups v.'ere fed exper i '^en ta 1 diets or 3";, 5°^, and H 
content respectively while the fourth i^roup was 
Purina diet of 5% crude fiber content. Comparisons 
gains In body weights, the Tiean number of treatmen 
monkeys, the mean number of treatment days for 
diarrhea, and the morbidity of diarrhea were mad 
the four different diets. Statistical methods 

analysis included Student's t-test, Sa t terwa I te ' s 
to the t-test, and Fisher's exact test for anal 



mo n k e y 
1 s each 
% cru ie 

fed a 

of th 
t days 

monkey 
e for r 

used 

a o D r o X 
yz i n5 



s were 
. Three 
fiber 
re?:u 1 ar 
e mean 
for al 1 
s wl th 
ats ''ed 
In the 
i mo t i on 
2x2 



contingency tables. 

''a-K ^umo Ratio in - rv th rocy te s . ^ 

-myotonic dystrophy is a disease characterized by muscle wealcness 



Mo^bs, L''C, MTT'JS 



luman 



and de?^enera t i on as well as 



ten -iency 



"or the muscle 



to 



discharge reoetltlve action potentials. Comparison o^ the Ma-K 
pump ratios In erythrocytes in patients with myotonic dystrophry 
v/ith that of ,a control group was made. Statistical procedures 
used v/ere Student's t-test and the F-test. 

Laboratory Animal Medicine '"omoetency . C. '^c°herson, D'/f, DRR. A 
questionnaire was sent to a random sample of 12t<- veterinarians to 
determine the requirements for labotatory animal medicine 
competency for graduatin-^ veterinary students. Statistical 

procedures used in analyzing the results of this sa-^ole survey 
were analysis of variances and Duncan's multlole range test. 

Numerical Analysis and Statistical Consulti-r^ . Consultation was 
orovidei to R, Marimont, IRP, Ni'-ICDS, on con-iience Intcvals anJ 
significance tests for the parameter estimates of a linear 
combination of exponentials; B. Gladen, L-''., M1"'-!S, on numerical 
Integration; D. W I n terbourne, LCGY, '"'"I, on orobit analysis; M. 
Potter, LC^GY, .*JC I , on biological assay techniques; J. Slack, 
LC3GY, "ICI, on techniques to compare regression lines; and ^. 
Tbakur, ^RR3, rilCMMP, on fittln-"g a oolynomial using orthogonal 
polynomials and unequally weighted data points. 



39 



1 

I 



PU3LICAT!0r'S 



Vierlins, J. M., Shra-^er, R. I., Rjnble, W. F., A.a-iodt, R., 

Beman, M. D., ani Jones, 1. ^. : Inconoration of radiocoppei 
into cerul onl asnin in nor-rial subjects with primary blliar^^ 
cirrhosis an i l/ilson's disease. ^a stro^n t"ro1 op:v 71+ b, SS^-RSo'^ 
1973. ^ ' ■ '■' 

Graitsnn, M., and Dunham, G.: Traitenent Autonatique du] 
Francasis Medical, Cahiers de Lex i col o-^ i e<; . Vol. 30, 'mo 1 
1977. 

Dunham, G. S., Pacak, '1. G., and Pratt, A. W. : AutomaticJ 
Indexing of Pathology Data, Journal of the American "Society foi 
In'^ormation Science . Vol. 29, Mo. 2, 1973. 

°acak, M. G., and °ratt, \. V/. : Identification a^ 

Transformation of Terminal Morphemes in ^'eiical ':nq:lish ''art M; 
f^ethods of Information in ^^edicine . Vol. 17, "o. 2, 1178. 

Knott, Gary D.: "A Mumbering system for 'Binary Tr'^^'s" r/\r?/| I 
Vol. 20, ^!o. 2, pp. 113-115, 1977. ' ' '' 

Shapiro, '' . B., Habbersett, M. c.. Merman, C. J., and Smith,! 
E.: Pattern Classification of nynecolo:;i': Specimens ^n^lyzed 



v/I th a Flowml crof 1 'jorometer . Presentei at 1"^^ '"on^-^rence on 
Pattern Reco.inltion and Image Processing, Chicago, Ml., "ay 33 
June 2, 197S. 



1 



1 



Lleblick, A. K. , Symmes, D., Me\vman, J. ")., and Shapiro, M 

3.: Development of the Isolation Peep in Laboratory ^r^'-^. 
Squirrel Monkeys. The Journal of Animal ^.ehavlor . In press. 

^utchinson, '"- . : Embedding and un so 1 vah i 1 I ty theorems for modular] 
lattices. Al gebra Universal i s 7, kl-Zk, 1977. 

Hutchinson, G.: A duality principle ^or lattices an i ca te '-Tor I :; si 
of modules. J. of °ure and Aonlie-! M?ebra 19, 115-119, 1977. T 

Blum, M. and 'lagel, R.: Shape descriotlon usin-^ v/e I gh tedll 
symmetric axis features. Pattern Recognition , in press. " *} 

i 
Kamel, I. A., Cheever, ^. V.'., Elwi, \. m., Mosl-^ann, J. E.( 

and Danner, R. : Schistosoma Man son I and S. '-Jaema tob i um| 

Infections in Egypt. I. Evaluation of Techniques for Recovery 
of Viorms an i Eggs at Mecroosy. The American Journal o" Tr'-^nlca! 



»<qH 



di cine an ! Mvgi en^ . 2 6:696-701, 1977 



Cheever, A. W., Kame 
and Danner, R.: 



A., ^Iv/i, A. 



M., 



Sch I s tosoma 
40 



'o 5 ; ~!a n n , 



M 



an snn 1 



and 



''aem^ to'^ I urn 



I 



Infections in Egypt 

The 



at fJecropsy. 

Hy.^iene . 2^^:70 2-71'^, 1177 



i. 'Quantitative ^a n s I to 1 o-^ i ca 1 Fin-iln^^s 



\'-^er \ C3n .Journa 1 



"ronic.^i *^eiioinq ani 



Cheever, \. \l . , Ka-ne 1 , I. \., FUvi, ,^- . m., ^^osinann, J. E., 

Danner, R. and Sippel, .1. ^.: Sch i s toso'^in '^ansoni ani S. 

^ae■T^a tob i Li-i Infections in '^gypt. 111. t^xt r.^ohepa t i c Patholncry, 
The ATierican Journal of Tropical ^^edicine ani Mv^i-^ne . 27'55-75 
1977. 



Mosi-nann, J. E. 



an' 



•Ja-ics, F. r. : "lev,- Statistical ^'ethods for 



Allo-ietry v/i th '"^ppl i ci t i on to "I'-ir 
^v^l u t i on . in pross. 



T o -] — v( [ -) T o [ 



^Inc'^Si 



"osi-i^nn, J. E., ^"alley, J. D., Cheever, A. '7. .-in-t ^l-r-k, C. 
5.: Size and Shape Analysis o'P Schistoso-ne E.^^^-Pounts in 

Egyptian "utopsy Data. Si one tries , in press. 



Ko'iel, I. A., El v/i, A. M., Cheever, ^. W., Mosinonn, J. E. 

and Hanner, R.: Schistosona Mansoni an'^ S. Mae-na tob i u^n 

Infections in Egypt. !V. Hepatic lesions. The ^•ner ' can .loiir^al 

of Tropical Medicine and ilygiene , in pi'ess. 



41 



.Ci ',\y'\iiU [:'j uo\ 



>}.. . V\ -'Ma'VM rr I (vr.ji^ri Mj'Mil 



.11 </ 
ii:- ^r-icc) |;i:.Ai.Ti:, ; . '. ' 1 1 ■.•; , /.'.;) ..fUAi,: 

LOT ICf. (.•- 



1 



ZOl CT 00001-07 



LSMi 



r.M\n c:vii;:D 



Oct. 1, 1977 through Sept. 30, 1978 
Automated Data Processing of Medical Language 






TiTLLS Or ri;i:;C;lf'A!. l(,VL:TIGATCr:S A:v[J ail GUitf, 



I 

I 



PI: 


M. 


G. Pacak 




A. 


W. Pratt 


OTHER: 


G. 


Dunham 




S. 


Harper 



Computer Systems Analyst LSM DCRT I 

Director DCRT ■ 

Computer Programmer LSM DCRT 

Computer Programmer LSM DCRT I 



M. DeMeyts-Graitson Guest Worker 



LSM DCRT 



1 



CCC-.-LhAi I!:G UMl. (H any) 



Laboratory of Statistical and Mathematical Methodology (LSM) 



Medical Information Science Section 



'"'Division of Computer Research & Technology, NIH, Bethesda, Md. 20014 



t._- 
I.' 



i-.A:,iLyj; 



'y:^ 



''2. '5^ 



OlIifG 



0.0 



1 



■rcr. ArrKor:^iATf. rox(EG) 

(0 ii'-'J'A'i riT..-CTP 



n (;0 iiG'.:Ai, TIG.UES 



[3;(c) i,;iTi.--R 



(al ) KiNCR- n (u:) IMLRVUVS 

J "A,",V C;" '.iGGfv (::.0 ..erdr oc Jc;.l - ur.derlii.c kcv..ordi) I 

The major objective of the project is to develop a formal and practical basis | 
for a lexicon of the language of medicine. This must include the possibility 
of specialty microglossaries , morphosemantic and syntactic paraphrase rules t 
for automatic recognition of synonymous noun phrases, and internal semantic |i 
structuring . 



The system for information storage and retrieval of pathology data in natural 
language was maintained for future use. 



42 



Project Description: 

Work on the infectious disease nomenclature as a paradigm case for morphosemantic 
analysis of the compound words of medical language was continued. A draft 
publication was prepared describing: the results of distributional analysis, 
the -ITIS alloforms, and a paraphrase algorithm for synonymous expressions in 
this nomenclature. Analogous work is in progress for French. 

Research on the metalanguage for medical lexicons focused on the neglected 
relation between the lexicon and syntax. Areas of study included: the 
representation of medical language information for its actual clinical and 
scientific uses, morphosemantic analysis and paraphrase, parts of speech 
both semantically and syntactically based, convenient rule notations sufficient 
to describe the linguistic processes and reducible to efficient automation, 
the sharing of inference capabilities between grammar and the lexicon. Several 
internal reports were produced. 

Dissemination of computer based medical lexicons was made to 4 extramural 
institutions and 2 users within NIH, 

A data base for morphosemantic and morphosyntactic analysis of medical French 
was extracted from multiple sources. 

An interactive "frame" or "menu selection" language data collection system 
was constructed as a tool for nomenclature building and its simultaneous study. 

Future Efforts: 

1) Completion of work on the infectious disease nomenclature and generalization 
of these findings to morphosemantic analysis of other groups of medical compounds. 

2) Refinements of grammar rules and further development of the metalanguage for 
semantic and syntactic structuring of the computer oriented medical lexicon. 

3) Prepare previously collected French morphosemantic data for automated 
linguistic analysis. 

Publications: 

Graitson, M. , and Dunham, G. : Traitement Automatique du Francasis Medical, 
Cahiers de Lexicologie , Vol. 30, No. 1, 1977. 

Pratt, A. W. : The Use of Categorized Nomenclatures for Representing Medical 
Statements, in Computational Linguistics in Medicine . Schneider, W. , and 
Sagvall-Hein eds., North-Holland Publishing Co., Amsterdam, 1977, pp. 45-53. 

Dunham, G. S., Pacak, M. G., and Pratt, A. W. : Automatic Indexing of 

Pathology Data, Journal of the American Society for Information Science , Vol. 29, 

No. 2, 1978. 

Pacak, M. G. , and Pratt, A. W. : Identification and Transformation of Terminal 
Morphemes in Medical English Part II, Methods of Information in Medicine . Vol. 17, 
No. 2, 1978. 

43 . 



r.iJ-.Cl M: ;■;:; {'-J !.0T u%o l:,i:, -.p'.cc} I,; AIT!', li ';, M I i :, /'.^i v.tir/.:,: f 

)L):'L|-:. ■ ■-•i.iM .M,,-;C._ 

i.o; ic!. 0" 
li<Tf(AH'j;,/'.L f-.L:,LA:,r.ii fTiojECT 



;:uv;.i.i. 



ZOl CT 00012-06 



I'Li^ico cjVl;.:.o 



Oct. 1, 1977 through Sept. 30, 1978 



LS^^' 



TITLL Cr (;,0J Cl (l-O f.l.jrJcU-r^ or lovi) 

Biological and Visual Shape 



.:"::, l;ll;.'i::,v a'.;? p.;iim;u /FFiLi.'.Ti> 



"ITLLi Or riMI.OIf'AL I I.VlCT iGATCriS A:, 3 /■: L CTIiuH 



fi::r[:c:iC:.AL 



PI: H. Blum Res. Gen. Phys. Scientist LSM DCRT 

OTHER: R. L. Webber Chief, Clin. Invest. Branch CIB NIDR 

, R. Nagel Senior Staff Fellow CIB NIDR 



i 



i 

1 



cc:i?iratI::g r:.!U (n any) 

DMG, CIB, NIDR 



L/.tS/SK/,;.C!: 



Laboratory of Statistical and Mathematical Methodology (LSM) 



1 



^^ccTio:; 



Biomathematics and Computer Science Section 



""Dlvisxon of Computer Research & Technology, NIH, Bethesda, Md . 20014 



lOiAL 



0.9 



F;;cr!;c;sio:,AL: 
0.9 



OlIICi;: 



0.0 






!J (?) iiU"A'i t;7.v;;cTS rj (b) i;:;':a!; Tl3Sll^s 



[^(c) UL liHLR 



;/.; m;rv C; 



^:00 -..cni-,: 01 IcDS - u.-idcrl i^.c U 



c)-,crJs; 



This project develops and applies a new geometry of biological shape that give 
a natural and efficient description to a variety of biological objects at 
vastly differing levels: chromosomes, cells, organs, organisms. . 

Applications are to (1) automation of shape analysis for diagnosis and taxonomy 

(2) the psychology and neurophysiology of shape processes in vision and (3) 

the description and understanding of organ and organismic development . 1 



44 



Project Description: 

The overall objective of this project is to develop a formal descriptive 
language natural to biological shapes and apply this language to the variety 
of problems arising in main areas of biology and medicine: taxonomy, 
neurobiology and organismic development. This would permit a better modeling 
and understanding of these processes and also allow for the automation of 
many shape processes now done by humans . 

The methods employed stem primarily from a new geometry based on growth as 
the primitive process, and conceived by the principal investigator. It is 
applied to a variety of problems, both to clarify the biological processes 
taking place and to develop the mathematics in new biologically relevant 
directions. These include cell and tissue description from light microscopy, 
shape descriptions of developing organisms, chromosome description, visual 
psychophysics and visual neurophysiology. 

Computer programs for symmetric axis analysis of arbitrary two-dimensional 
shapes were improved by giving added flexibility to users and addition of 
new output options. Studies of error introduced by quantizing continuous 
shapes are continuing. Investigations of interpolation algorithms, differential 
geometry and minimal energy considerations for curves, and information of 
boundary point sets are continuing. 

A general method for extracting symmetric axis descriptions and for detecting 
edges in gray-scale pictures is still under development. 

Symmetric axis methods were applied to a study of growth and development of 
the human mandible. 

Publications : 

Blum, H. and Nagel, R. : Shape description using weighted symmetric axis 
features. Pattern Recognition , in press. 



45 



'■ If ;■:'/: I..:; '.'Ar\l : I'.i i;kva7 



fh^j- CT ;iii:-'Bi;: (;o iiOi 



I I L / '. .' : 



u:,L- t.'i 1 :. '.. ; jc c , 



I Ui'Llf. ! • /IV: -■ ,,V;C(. 
LOT ICC u' 

n/TnA!'i..';AL HCi,i;A;,cii (liO.if.cT 



ir.o.KC] f;!':';i(.;< 



1. 



ZOl CT 00011-04 LSM 



Trrrrcr" 



Oct. 1, 1977 through Sept. 30, 1978 
Discrete Mathematics and Applications 



I 



•,Av;3, LA;;Gr.AT:;:Y r 



■ i;:STiiUTc Arr iLiATU.s, amj titles cr ri.i;<i.; 



H.VLGTlGATCr.S a:.J ALL CliiCI 



PI: G. A. Hutchinson Research Mathematician LSM DCRT 



I 

I 



t:cop;:RATi!:G ii;, ns (it cny) 
NONE 



LAb/5A.,,C;' 

Laboratory of Statistical and Mathematical Methodology (LSM) 

Biomathemati cs a n d C om puter Sc ience Section 

i:.'-:iT,iL n..Li LCAAiio:; 

Division of Computer Research & Technology, NIH, Bethesda, Md. 20014 



I TOTAL I-A: VlARG: 

0.5 



0.5 



ClHLn: 

0.0 



n (b) iii;"A;: ^^igsulS 



13(c) cLiran; 



(i;cci: ArrRGrRiAT^ i:.C),(i:s) 
M (u) iiu'vA'; iL;;!.'iCTs 

:j (^1 ) !M:;r:v S G i^ j-i^^v'f^s ; 

;;.'.'".A';Y C*^ uGan [Z':0 ..ordi; or Ic.s - ui'dcrlinc kLVr/ordi ) 

A complete formal logic generalizing equational logic as applied to group 
theory , the theory of vector spaces , and related algebraic theories was 
developed. 



46 



Project Description: 

Objectives: 

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. 

Methods Employed and Major Findings: 

Groups and vector spaces are standard mathematical tools used in various fields 
of physics, chemistry, and other exact, theoretical sciences. A system of 
formal logic applying to many questions of group theory, vector space theory, 
ring theory, and related theories was developed. This system is a generalization 
of equational logic for these theories, and proofs are obtained by manipulation of 
algebraic expressions as in equational logic. This is, the formulas in 
question are free of logical connectives and logical quantifiers, and the formal 
proofs in this theory tend to be short and easy to follow. The system developed 
here is broader than equational logic, having the additional capability of 
dealing with questions concerning normal subgroups of groups, subspaces of 
vector spaces, ideals of rings, and so on. The major result proved about this 
formal logic is its completeness: a formula of this type is provable by 
the formal methods if and only if it is true for all algebraic structures 
under consideration. For example, a formula describing a property of groups 
has a formal proof in this logic if and only if it is true for all groups. 

Significance to Biomedical Research and the Program of the Division: 

General purpose mathematical techniques and computer programs implementing 
them are made available to the biomedical research community. 

Proposed Course: 

Work is continuing on theoretical studies to develop new computer methods 
and improve existing methods. 

Publications: 

*Hutchinson, G. : Embedding and unsolvability theorems for modular lattices. 
Algebra Universalis 7, 47-84, 1977. 

*Hutchinson, G.: A duality principle for lattices and categories of modules. 
J. of Pure and Applied Algebra 10, 115-119, 1977. 

(A Reported as in press in previous fiscal years.) 



47 



•.■•iT':'o:;i-:; :,f k;.Cl i:.r i-f'^'Air/. [/cha'.u!. 

in,.jJ:.Cl I.UVjii; (1,0 I1CT u:.c l!.i;. i,pict) 



U.:.. i'l.f'At:!".:'.! ff- 
iai.TI', Ll'i'^Mir : A'.ii wn.CAl 

(U:'LI(. ' : /Dm L,:i,VICl 
I1OI let t.v 

(KlR.'J'L'fiAL H£^i.A,';OII (aOJECT 



I 'i'vircT r;ii'';::.i( 



ZOl CT 00039-01 



LSMi 



rruic^ CCVf.i;^) 



Oct. 1, 1977 through Sept. 30, 1978 

T\'ili~cr ii^('^' n (i-O char.jcUTL or le-.i) 

Linear Methods in Statistics 



;.;"ri, la;;c;:'t::>y am; icstiut^ AFriLiATi.'.:, 



,0 TITLCS Or f'f,i:,CIPAL I I.VuCT I GATOhS A:.u A!.I GTH^K 



PI: J. D. Malley Staff Fellow 



LSM DCRT 



ccc-p:.rati;;g u;.ii:, (if ar.y) 
None 



ij.^/cu/:.:.-\ 



Laboratory of Statistical and Mathematical Methodology (LSM) 
Statistical Methodology Section 



Si iiv'i c a:.c: i.ccamo:; 

Division of Computer Research & Technology, NIH, Bethesda, Md. 20014 



1>TAL I'AltViAr 

0.5 



fSOFtSSIC'.AL: 

0.5 



0.0 



-■rdr r, 



(iO HU-Ai; TISSUES 



D (c) i:ciTi-,LR 



■Afi'i' Cf 



v-^ 



c5 - i/.'idcrl i '.c kc)r,ordi) 



Linear methods in statistics are being studied, with the general linear model 
serving as a point of departure. Famr-My confidence limits for ratios of 
I sample means from multivariate normal distributions were developed. The method 
I have utility in size and shape studies (see the multivariate statistical 
j analysis intramural research project report) as well as in any study where 
I ratios of normal means are important. Theoretical work in linear algebraic 
I groups continued, and a paper on the relation between algebraic structure and 
' automorphism groups is in preparation. 



48 



L__ 



Project Description: 

The overall objective of this project is the study of linear methods in 
statistical analyses. During the past year, in addition to family confidence 
limits for ratios of normal means, and work in automorphism groups and 
algebras, the linear methods were utilized in the understudy of log linear 
models for contingency tables. 

Publications: None 



49 






-ALT!', :l^c^T^.■;, a'.o w:i.rA!i: ! 
rOrLK. !■• /'.Ill -l.:vicl 



1.01 \Ct r,' 
iNTr'Ai'L'iiAi i;i:iCA;;cti fr!0,;EtT 



ZOl CT 00013-04 



ftH:c:j CuVLKti) 



Oct. 1, 1977 through Sept. 30, 1978 



lITLt Cr fnOjrr (:0 ch:;rjcUTs cr l»ii) 

Multivariate Statistical Analysis 



:,;.'■':, la::c;;at;;y a',3 i'::7ni;T^ Arr iliatu:.3, a^.d inicc of fKK.ciFAi. i;,vl:tigatO;;S a;.:: all cti 



.-lA'.'l P, ;-'r//.; ! 



■ Zf.Zi'J C.'.' 7mi. ih'OJrCT 



1 



PI: 


J. 


E. 


Mosimann 


OTHER: 


J. 


D. 


Malley 




R. 


N. 


Banner 




C. 


B. 


Clark 




I. 


R. 


James 




F. 


C. 


James 



Chief LSM DCRT 

Staff Fellow LSM DCRT 

Computer Systems Analyst LSM DCRT 
Computer Systems Analyst LSM DCRT 
CSIRO, Melbourne, Australia 
Florida State University, Tallahassee 



i 




Office of the Chief, LSM, DCRT 



:i.s; nuu A;,i;' i^ccatic; ^ 

Division of Computer Research & Technology, NIH, Bethesda, Md . 20014 



TOTAL h: 



''■■'i::r 



FhOrLSSlO.'.AL: 

1.2 



jj {;) (;i.^'''A'; iL!L-JcCT5 



D (t.) hu"A'; ri^iu.s 



OlhTJi: 



0.0 



■T (c i;lIti.;r 



:j (=0 



[j (.!?) r.TLfiViE'.vr, 



"ArtV i.: 



:.rrJ^ cr Jccs - u;.deriir,c' kcv.crJ 



Multivariate statistical methods ( size - shape methods ) for analyzing ratios 
which follow a lognormal distribution have been developed. Exact statistical ij 
tests have been developed and applied in two biological studies: the |i 
distribution of schistosome eggs in man at autopsy; morphological measurements 
of birds (see publications) . Work on the theoretical meaning of size-shape 
concepts for statistical distributions continued, with a new characterization 
of the Dirichlet distribution being given by Ian James and Mosimann. Work on 
separating error effects from model effects by using information in the sample 
mean vector along with the covariance matrix, was begun with the work on 
schistosome egg counts. This work has been continued by Malley, who has 
analyzed organ weight data in laboratory rats. 



(. 



50 



Project Description: 

The overall objective of this project is the study of multivariate statistical 
methods for the analysis of data which take the form of ratios or proportions. 
During this past year there was continued development of statistical tests 
and study of data. These included data of A. Cheever, NIAID, on the 
distribution of schistosome eggs by organ at autopsy, data on morphological 
measurements of birds of F. James, NSF, (currently with Florida State University) 
as well as data on organ weights of laboratory rats. Lectures on this work 
were given by Mosimann at I'Universite de Montreal (Pierre Robillard Memorial 
Lecture), and the University of Parma, Italy. 

Publications: 

Cheever, A. W. , Kamel, I. A., Elwi, A. M. , Mosimann, J. E. and Danner, R. : 
Schistosoma Mansoni and S. Haematobium Infections in Egypt. II. Quantitative 
Parasitological Findings at Necropsy. The American Journal of Tropical 
Medicine and Hygiene . 26:702-716, 1977. 

Mosimann, J. E. and James, F. C: New Statistical Methods for Allometry 
with Application to Florida Red-winged Blackbirds. Evolution , in press. 

Mosimann, J. E., Malley, J. D., Cheever, A. W. and Clark, C. B.: Size 
and Shape Analysis of Schistosome Egg-Counts in Egyptian Autopsy Data. 
Biometrics , in press. 



51 









u.: 



Ai, !■■:.:. T rf 



■.I- li.i^ -fice} jiw.ALTIi, li''j:M I'":, A:,f) r'Ll.|-Af,£ 



UOTlCf. C' 
IlilTiAKUiiAl. kUCAnCIl PROJECT 



I i;^-..''XT ;;u:'[t! !< 



Z01 CT 00010-02 



TiTurCF'TiTcj'cf (rO chjrcct.-TG or lei-O 



Oct. 1, 1977 through Sept. 30, 197J 



Nonlinear Equations 



1 
1 



FiiCr CCri! 



.:.'T:;iY a'.d p.:Tni:T£ aff iliatu:.3, a;.d titlcj. cf 

•■,L rt.H:.C:,:.LL LI.'iAC^D C:i THE Pi;CJCOT 



^HK.CIrAI. iLVCC-TIGATCiiS AM) ALL OTHF;; 



PI: 
OTHER: 



R. I. Shrager 

G. D. Knott 

E. Hill 

J. E. Fletcher 



Mathematician LSM DCRT 

Computer Specialist LSM DCRT 

Mathematician LAS DCRT 

Research Mathematician LAS DCRT 



I 



I 



Ci.OrLRAl l.'.G iJMU (H eny) 
LAS, DCRT 



la-/e;<a:.C; 






Laboratory of Statistical and Mathematical Methodology (LSM) 



ClCTIC 



Biomathematics and Computer Science Section 



i;.iTiTcic a:.l> LucATio:; 

Division of Computer Research & Technology, NIH, Bethesda, Md. 20014 
t T ;i .v.A.'i V l; a s^ fp.o,^ as iT?y^\LT '~ loi \\iK\ 

1.0 1.0 0.0 



C!.:CK Arr-.^C;- ,HIATF tox(fs) 
J (0 HU'vA-; c-1;b,:ICTS 



D (t.) li'JVAfi TIGSULS 



3(c) CLIThFR 



■lii!^ iiijtl^ JKiZ^ |^URv|fwc 

S.,V'-'-,:\V C'l '..CixJ, (^v'O i.ord:-. or Joo.s - uridr rl i i.c kcjwords) 



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. 



52 



Project Description: 

Objectives : 

To develop methods for solving nonlinear equations frequently encountered 
at NIH. 

Methods: 

A continuing effort is made to create methods or extend existing methods to 
solve problems in a host of NIH applications, and to house those methods in 
accessible computer programs or routines. Modelaide and MLAB are two examples. 

Major Findings: 

The Levenberg-Marquardt method for non-linear least squares, and the variant 
of R. I. Shrager which handles linearly constrainted parameters, have been 

extended to the L^ and L norms. Revision of the manuscript is now in progress. 

1 <=° 

The MLAB curve-fitting routine has been improved, convergence is accelerated 
so that solutions can be found with fewer iterations and function evaluations. 

The MLAB differential equation solver is being revised to take less space, on 
the average, for partial derivative formulas. Among the schemes being considered 
are 

1) including an equation in the stiff set only when accuracy and desired 
step size seem to demand it. 

2) evaluating some partial derivatives numerically. 

Significance to Biomedical Research: 

These methods are now being applied to problems in human metabolism, cell 
growth, chemical kinetics, and spectral analysis (UV, IR, CD, ORD, NMR, ESR) . 

Proposed Course: 

As the methods are proved in test and practice, they will be incorporated into 
easy-to-use systems like MLAB, and as a result, the systems themselves should 
evolve to do more useful work with less human and machine effort. 

Publications: 

Vierling, J. M. , Shrager, R. I., Rumble, W. F., Aamodt, R. , Berman, M. D., 
and Jones, E. A.: Incorporation of radiocopper into ceruloplasmin in normal 
subjects with primary biliary cirrhosis and Wilson's disease. Gastroenterology 
74, 4, 652-660, 1978. 



53 



■f-'.;j'.CT ,'<UVb:i; {vo hOl u-.o thi;- :-^icc) jii.-.ALTH, iivC'.ll;", A'. 'I WCLI/.iiL 

LOT i::- f,r 
imU.':y:c'.iA{. HilCf^iiCH fRCJ£CT 



IKOJLCI ;,U";;Li; 



ZOl CT 00008-04 



n 



LSi 



r'^HiCO CjVL,'..-. 



Oct. 1, 1977 through Sept. 30, 1978 



TITLE Cr ff-.OJ'CT {;0 charge Ivri or leti) 



Pattern Recognition 



:,A';s, LA;;C:;ATo;iY a'.li ilctitdt;: ArriL!AT!:\', a:.o titlc^ cr pri;.c!;7.l i:.v:ctigatc;;s a:,d ail gii-ci 
RCFE^:!c:.Ai. r:Roo;.:iLL ei.gaclO c;; thl icojcct 



PI: 
OTHER : 



M. Shapiro 

C . Herman 
M. Cassidy 

D. Symmes 
J • Newman 
A. Lieblick 



Research Mathematician LSM DCRT 

Senior Surgeon LP NCI 

Biologist LP NCI 

Section Chief BB CH 

Scientist BB CH 

Scientist BB CH 






CCOPLRATI.'.G U.'.llo (it sr,y) 



Laboratory of Pathology, NCI 



.A[:/2nA',Ci 



Laboratory of Statistical and Mathematical Methodology (LSM) 



lactic:, 



Biomathematics and Computer Science Section 



:.'.'^titl'U a;.Ci lccatio:i 

Division of Computer Research and Technology, NIH, Bethesda, Md . 200 

total .'■■<,:•. I L.-.as: 1 fhcf lssici.aL: lcii:Ei7r ~ ~. 



14i 



,.-. I L.-.ad: 

2.0 



FHCFLSSiei.AL: 

2.0 



0.0 



D (!■) HUMAt; TI3SU£S 



S (0 ;-nT!:.R 



ioi-lii.i.- ko),ycrJb) 



ciiCCK Arr:,or;;iAT: £;ox(l;) 
J (a) i:u::a:; si'?j:;cts 

u (:^i) Mit;c:^s n_(f') i"Ti.;;viFws 

SL;.V„V.ARV CF «ORn [200 wcrdr or Jcr.s 

Computer pattern recognition methods have been developed for general use and 
have been applied in two problem areas. 

1. C-LA3, 2nd edition. 

An extensive revision of C-LAB , a system for Cluster Analysis , has been 
undertaken to add some important algorithms, make improvements on existing 
ones, make the manual clearer, and to keep up-to-date with MLAB, with which 
it is compatible. 

2. Pattern recognition of monkey vocalisation records. 

Studies of similarities in vocalization patterns were completed and published, 

3. Detection of cervical cancer using f lowmicrof luorometer data. 

An algorithm for classifying data produced by a f lowmicrof luorometer, an 
instrument which makes measurements on individual cells, was developed based 
on 209 samples collected in phase I of the study. 



54 



< 



Project Description: 

Objectives: 

The main objective is to provide an easy-to-use package of pattern 
recognition programs for the use of NIH researchers and to apply these 
and similar techniques to particular problem areas. 

Methods: 

The standard pattern recognition methods plus some of the most recent 
work has been programmed and applied in a number of problem areas. 

Significance to Biomedical Research: 

Pattern recognition techniques are now being widely used on biomedical 
data for classifying objects, finding relationships between variables, 
and for processing biological images. These applications of artificial 
intelligence has led to both automatic processing and a better under- 
standing of data. 

Proposed Course: 

A wider range of pattern recognition and algorithms will continue to 
be developed and applied. 

Publications: 

Shapiro, M. B., Habbersett, M. C. , Herman, C. J., and Smith, E.: Pattern 
Classification of Gynecologic Specimens Analyzed with a Flowmicrof luorometer . 
Presented at IEEE Conference on Pattern Recognition and Image Processing, 
Chicago, 111., May 30 - June 2, 1978. 

Lieblick, A. K. , Symmes, D., Newman, J. D., and Shapiro, M. B.: Development 
of the Isolation Peep in Laboratory Bred Squirrel Monkeys. Accepted for 
Publication in The Journal of Animal Behavior. 



55 



i^MlTHSONIAfi ^Clo.Ct i ;,f ORVAT 1 0;i t./OHAf:G 
PnCJ^CT ■■■iU.'.i3EH (Oo NOT uzc this opice) 



U.o. DEPARTMu!;! OF 
lEALTH, EOUCATU;., AMj WELFAR 
PUfiLIG HfALlH OcRVICK 
NOTICE Ci- 
IHTRAlIIJIiAL REGCARCH F^RCJECT 



PROJECT fiL!'-!BtR 



ZOl CT 00009-04 



LSM' 



PER I on COVERJO 



Oct. 1, 1977 through Sept. 30, 1975 



TITLE CF PROJECT (iO rhar-ict 



er- or ie; 



Research Topics in Computer Science 



fiA."EO, LABORATORY A'.:. 1 ,'151 i 'UTE AFFI L I -i T 1 0:;S, A'iD TITLES OF Pfti:.'ClPAL INVESTIGATORS A:;D ALL OTHER 
FiVOFtGGIC;;AL FLR:0;;i-iEL EriGAOEO Cu n-iE PI'CJECT 

PI: G. D. Knott Computer Specialist LSM DCRT 



i 



ccope!';ati:;g umto (if c.ny) 
None 



LAB/Br.-AiJCH 



Laboratory of Statistical and Mathematical Methodology (LSM) 



; c r I c;, 



Biomathematics and Computer Science Section 



I:.CT|TUTl Af.D LOCATIO:; 

Division of Computer Research & Technology, NIH, Bethesda, Md . 20014 



ICTAL ;,:a:;yearo: 
0.6 



FR0rEo3l0:.AL 

0.4 



OTHER: 

0.2 



1 



□ (c) riJMAfi T ISSUES 



,^ (r) l.'EITIlER 



CHECK APPROPRIATE pnx(ES) 
"J (e) Hli.'vAf; SUB.>"CTE 

n (3 1) MIk OR.:, n (-^2) i: !TERVIE.v S 

£0,V..'vARY OF f.'ORK (200 wcrdi or Ici.i •- jncerlinc kej.vorJs) I 

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



I 



Currently, an exhaustive survey of storage and retrieval methods is underway. 
I This includes the recently introduced k-d tree method. ' 

Optimal item orderings in split hashing schemes and certain interesting algebraj 
I characterizations of fixed permutation open addressing methods are also being 
studied. Research on trie methods are also being conducted. 



56 



A 



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. 

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

Optimal item orderings in split hashing schemes and certain interesting 
algebraic characterizations of fixed permutation open addressing methods 
are also being studied. 

Research on trie methods, which involves storing items in trees so that the 
path to the item is determined by its key, is underway as well. 

Publications: 

Knott, Gary D.: "A Numbering system for Binary Trees", CACM, Vol. 20, 
No. 2, pp. 113-115, 1977. 



57 



October 1, 1977 through Seotenher 31, lj7r' 

•!ATIO'^\L I'JGTITUTZG 0^ ''^_\LT: 

livision of Conp'.iter f^esearc^ ancl Tec'^nolory 

Junnary of '-)ranc^ \ctiyities ]. .'3"T 

'. ^XTA ''.A'; ^^E.■'E';T Z.'^^IC':' 3. J . Ennet t '..'ard 

■ r a n c '^ 3 h I e f 



''ission and Function 

The Data f'anaFenent "ranch (r''",:) provides advice ^and 
assistance to research I nves t i n tors, prorran officials 
and administrators throuthout MIH in plannin-^ for and 
obtaining conputer data processing; services. in this role 
the branch is a central '; I f- resource for co-'puter systens 
analysis, desif^n and profTann i ng- . 



Scone o'^ \ctivities 

Df'iS staff desig-n and create conpu ter-based data nana'^enent 
systens for specific users and train those users. They 
also teach courses about sone data nanare'-ion t an'' 
pro^fann I n» tools, provide advice on data -i^na,p:enen t 
techniques to '1!!' pro'^ranners and serve as consultants on 
conputer based systens proposed by other '.'!'' frrouns ■f^or 
inplenentat Ion by contractors. Finally, '^''1 creates and 
maintains general purpose, user-oriented progrann i ng tools 
to speed building and inorove operation o-P soeci'^ic 
applications systens. 



P e s i r n ? h i 1 o s o o b y 

f'any conputer apolications are . s tra i g'-' t ^or-/ard . Others 
involve an extended period of ' expl ora t i on to de'^ine the 
users needs, the appropriate infornatlon processin<- 
techniques and the conputer nethods best suite- to the 
users' c i rcuns tances . Clients pay ^or DM'i v/ork under the 
■Mr. service and supply -i^und as './ell as for runnin'^ the 
conpleted systens in the 'III'- Conputer Center. There'f^ore, 
the OiiC aoproach to projects can best he described as 
conservative but progressive using, v,'^^en possible, 
existing progranming tools to nininize develonnent tine 
and costs and to naxinize systen reliability and ease c^ 
na I n tenance . 

58 



The branch develops nost systens 
requesting organization. This 
conpleted computer application 



•f^or oper?itinn ^y the 

places control of the 

in the hands of the user. 



assures each user operational independence 
frees the branch to work on new projects. 



rnn 



U'i 



and 



Experience shows that lar.^-e complex projects, involving 
many users, may flounder for or n;3n i Zri t i ona 1 and 
sociological reasons, as o^ten as technical reasons. 
Fortunately special plannin'^, training the users, and 



care in implementation 
d i f f i cu 1 t i es . 



oi 



the system can avoid many o^ the 



FY7R IM ghl i ^hts 



In FY73 D;1B a^^a 
i nvo 1 V i ng V i r tu 
of the "MH. 
in-patient care 
than 12r^ for 
pro. f^ ram direc 
approximately 
development o'P 
30 ", o^ the p 
"■13 v/or'; and an 
project list i 
the C;iB \/ork. 
of par t i cu 1 ar i 



in v/orked on more than one 
ally every P-ureau, Institu 

Over kS"; "/ere orojects 
, clinical research, or ep 

lahoratory researc^^ p.reP- 



1 1 on. 



manat^emen t 



anr 



3''j ^nr hione-'ical co'-t-i'.i r i C 
data p r o c e s s I n !^ a n ^ a n a 1 y t 
rojects required les"^ t*"an 
other 3n'j less than '"'''^i h 
n Section 11 gives a vie.' 
The follov/ing items h i .^rh 1 I 
n teres t : 



bun'-' 
te '"• 

for 
I d em i 
s ; al 
ad'^ 
p t i o n 
i c tn 

^r' h 
ours 
o^ th 
gh t a 



re^^. p 
nH ^' 

appl i 
o 1 o f y 
mos t 
i n i s <■ 
s ar^. 
ols. 
ours 

e bre 

f G'// 



ro J ects 
i V i s i ^' n 
cat i ons 

; more 

'.= ^: ^or 

ration; 

^;- ■f'or 

\bout 

e a c ^ o ■^ 

^he 

adrh o^ 

areas 



s n ^ 



The tv;o largest efforts in the Data f',ana<Tei""ien t Branch 
fiscal year were the 'Materiel I 
Clinical Information Utility 
three man years respectively. 



^^ 



anage'ient System an 
They represent four 



the 
and 



The fi 
v/ as i m 
best 
wh i ch 
2 50, on 
of th 
FEDSTR 
Admi n i 
Obi iga 
Al thou 
procu r 
which 



rst 
pi em 
be 

enab 
n o 
i s 
I P 

s t ra 
t i on 
gh P 
emen 
to b 



phase of the Materiel '.ana^-ement Svstem 
ented in January, 137". This imple^ien 
characterized as a procurement contr 
les data entry and edit of the apn 
rders generated each year. The prima 
phase are (1) Printed Purchase Or 

order cards for . the Oeneral 
tion, (3) Order ITeports for DP!! 

Transactions for the Central Accounti 
base I does not represent a panacea 
t and payment problems. It does create 
egin addressing them constructively. 



tat 1 

ol s 

rox i 

ry o 

ders 

Se 

an 

nr' S 

fo 

a b 



on can 
ys tern, 
m a t e 1 y 
u tpu ts 
, _(2) 
r v i c e s 
d (i;) 
ys te^ . 
r -i I K 
ase on 



As the first step in what is projected to be ^u 1 1 
base support for all administrative computer syste- 



the 



Phase 



the 



o demonstrates that a 



data 
IS at 
Mj i 1 d i n g 



59 



block approach to database nana!r:enent is aopropriate and 



that adni n 1 s tra t i ve sta^^f Is capable 



c a 5 I I y 



s h i -^ t i nt^ 



from a 
s y s t en . 



clerical system to an on-line conputer data entry 



Clinical Information Utility provided ansv/ers to over 
inquiries this year. The combined databases proved 
both flexible and hip:hly useful as an information source 
and the addition o^ some simple statistical analysis 



The 



softv/are rendered the response to many 
usab 1 e . 



nquiries much more 



The primary prop-rammi nq; effort in the Clinical Information 
Utility v/as directed tov/ard the .Medical Information System 
(■■!!S) pur?e data. l.'hen patients are dlschar-^ed from the 
Clinical Center their data, \-.'hich are collected and 
displayed by the MIS, are du rred from that system, v/rltton 
onto ma.f^netic tapes and mailed to DCRT for archivin<T. 

After some preliminary analysis '.-/as dnne to determine 
v/hich data elements viere to '^e archived, pro<'rams '-/ere 
developed using a very simple pattern reco<^nItinn 
algorithm to select the desired data. ^s the nu-^ber o* 
users increased on the MIS and the older users '-'ecame more 
knov/ 1 edgeab 1 e about the system, ne'.; i'lS data collection 
screens v/ere used to input nev; medical data which ^ar^. not 
been included in our original desl-^^n. Consequently, the 
pattern recognition algorithm failed to select all of the 
des I red da ta . 



A nev/ pattern reco^^nition algorithm has been designed 
developed to process two classes o^ purp-e dat^: 
patterns to be archived and known patterns whlc^ are 



to be processed. This leaves a third class 
which is the unknown. The unknown patterns 
to determine which of the previous tv/o 
belong. Once the class has "^een deter-ilned 
recognition algorithm is modified to process 
data . 



no t 

o" puree data 

^re ana 1 yzed 

classes they 

the pattern 

t'^e nev.' [i^^at^'b 



Another significant improvement of t' 



the pa t i en t prof lie. ^ 
developed to display 
computerized clinical 
patient. The purpose of 
physician In analyzing the 
readml tted pa t i ent . 



CIU 
oro^r ams 



this year 
'■/as 'defined 



system ( 

In an Of-derly format all „ 

in-ormation available -^or 

this system Is to aid 
previously collected data on 



v.' a s 
and 
the 
anv 



the 
a 



60 



A Surgical ?atho1og-v Svsten was inplenente^ last ypar 
the Laboratory of ?atholop:y, 'ICI. "ata collecte'^ by 
ourgical Pathology and Postnorte'^ Section have 
Included in this file since January, 1^77. F 
diagnoses are hand coded usinr the Gystenized 'ionencla 
of .■'edlclne (o^If^MEP) and patient infnr-iation nav 
retrieved fron this file using Boolean selections o-^ 
encoded diagnoses. This year data fron 
Hena topa tho 1 ogy area have been added to the systen. '=> 
call for (1) making these data availa^^le =5 part of 
Clinical Infornation Utility an^ C2) addinp- 
retrospective data in the surgical natholo.o-y 
postnorten area. 



for 

the 

been 

i na 1 

tu re 

Up 

the 
the 
1 ans 
the 
0'= 
and 



Several other PMB projects deserve attention 
of the variety of v;ork perforned. 



as examples 



became 
h i p- h 1 y 



As the Cardiac Valve Pepl a cement System 
operational this year, it became a 
Information source for some fifty inquiries, 
among these inquiries v/ere areas such as morbid 
mortality of similar patient populations and surviv 
comparisons among patient populations. For exam 
physician might v.- i s h to analyze survival and m 
among patients v/ho have had a 
and have pure mitral stenosis 
f^esponse to this request 
working day by a computer 
Sranch, '!11L3!. 



o 



va 1 ve repl aced s i nc 
v/ith no ot'^er valve 1 
s easily provided 



',7 I t 



technician 



the 



ful ly 
useful 
Typical 
i ty and 
a 1 rate 
pie, a 
r b i d i t y 
e 1"72 
es I ons . 
h i n the 
S u r p- e r '.' 



The success of this system in the Surgery branch- 
Impelled the Cardiology Branch to begin a similar 
for its patients. The Cardlolof^v Srancb is a^le "ro 
advantage of our previous exnerience and should'h 
operational system by mid-summer o^ 117^. 



has 



e^'^or t 

t a -' 

a ve an 



I n 



^^^ Reproduction Resear ch ^ranrh o^ -IICIT is involve^ ,, 

the preparation of purified proteins (he-orio^in pl^unin 

hormones) by the nrocess o^ -el electrophoresis. Ms i ng ^ 

scanning isoelectric focusing assembly ( Ca ts impooi I s 

Apparatus) it is possible ."to ohtain, across Mme 

quantitative data on the position, banHw i '' t^, " ' ' 

of the protein as it 

physical/chemical data 

enables researchers to 

behavior of the proteins 

refine their purification procedures 



. pea!cs, etc. 

filters through the gel. This 

can provide information v/hich 

^determine rules controlling the 

in the Fel and tn enable t*-em to 



51 



liorkins closely with the research scientist, D'lE developeH 
an interactive svsten for perfornine- the statistical 
analysis and displaying; peaks, which can inter^'pce './ith ? 
non-conputer oriented user. Current!^', \"?r^< 
pro.f^ress i nt' on a nethod for au tona t i ca 1 1 y c^oosinr 
tolerances ^or peak detection. 



! s 

■he 



In an effort to assist in 
of scintillation coun 
development of a set of s 
prof!:rans v/ill run on the 
the needs of a lar^e perc 
the liquid scintillation 
the DCRT Conputer Systems 
betv/een tv/o extremes, 
programs to perform fi 
production basis. Other 
to perform exploratory d 
statistical testin<^ and 
devel opment . 



the data re'^urtion and ap-''lysis 
ter data, r".Z- has hep-i.in the 
tandard computer programs. Thesp 
?nP-in computer and should ansv/er 
entaf^e of scientists v/ho may use 

counter data letter developed by 
Laboratory. The nee^s ran?;e 

Some scientists want a set o^ 

xed functions on a routine 

s could use an interactive system 

ata analyses and then perhaps 

curve fitting: ^or on-line model 



The liquid scintillation counter applications thouf^ht to 
be in most need of automation are radioimmunoassay, 
chromatography and electrophoresis, interactive ^^raphics 
channel ratios and spillover and quench corrections, and 
adenylate cylase activity calculations. 



Curing FY77-7S we produced an interactive v/orkin^r version 



of a self-teaching system which takes care 



most o^ the 



problems encountered in Chroma to"-raph i c- E 1 ectrop^cres i s 
(except automatic peak detection) and v/hich can rira\-i plots 
on any of a number of devices includin?^ the DEC-3U'^ visual 
display and the teletype. Subsequently these plots can be 

U 
peak 



sent directly to the CalComp for offline plottinf^. 
present, v/ork is being done on a "mini "-Lab" type o^ 



detection module v/hich will allow the user to check and 
modify the results o^^ the program interactively. 



Drug induced heart failur 
an anticancer drug. 3y i 
attach to adriamycin ther 
more effectively and at t 
margin of safety for the 
fiscal year, the adr i amvc 
effort among the invest! 
Branch of the 'IC I and th 
begun. Since that tim 
collected and tv/elve risk 
It Is ant i cl pated that o 
estimate patient risk and 
prospect I vel y . 



e 1 I 
dent 
apy 
he s 
can 
i n t 



mits the use o^ 
i f y i n g the risk 
one can adm I n I s t 
ame t i me ornv i de 
cer pa 1 1 en t 



o X i c I t y s 1 1 



u H V , 



5a ta 



piatlonal Drug i3ranch 
e n 
e d 
fac 
ne c 
to 



Managemen 

ia ta on UD 13 pa 1 1 

:tors have been 

:an use these kno 

adjust adriam 



adriamycin as 

factors V.' h I c h 

er this drug 

a r e a s o n a ^^ 1 e 

the previous 

cooper a 1 1 ve 

and Ciometry 

"ranch, \-i a s 

ents has been 

I dent i "^ i ed . 

wn factors to 

y c i n t h e r a D V 



52 



Fu tu re PI ans 

In FY79 and t^e foreseeable future, 0."3 v/ i 1 1 pursue the 
sane direction that has proven successful over the last 
few years. There is still a clear need for a central "IIF 
resource for advice, consultation, systens analysis, 
design and proj^ranni na: on data nanaeenent aoplications of 
computers. The Df'R staff has accunulate-' a v/ealth o^ 
experience in developin;^ 'i I H. data nana^-enent applications 
on the n I H Computer Center systems. it has a reputation 
for reliable, effective work. it has a pov/er^ul set of 
tools to p;enerate nev/ applications and works with experts 
in other DCPT laboratories v/hen more complex 
Stat i s t i ca 1 /ma themat i cal analyses and/or enj?: i neer i nr 
skills are required to satisfy a specific need. 

Of'iB projects will continue to arise from the data 
processing needs of 'li;', a remar'ca'^ 1 e mix In size and 
substance. The best solution/system for a "-iven 
prob 1 em/ pro j ect will, of course, denend on the technology 
available at the time it is undertaken. The prospect o^ 
"intelligent" terminals, as well as more po-./er^ul 
software, mean only that more options v/ill be availa'^lc, 
not that a given user will set a more complicate-^ or 
expensive solution/system. 'iever the 1 ess, s'-'me data 
processing; needs at Mil' will undoubtedly lead to mrire 
ambitious projects. 

Durinp; the FY78 the Data ."'anaremen t E ranch assume^ 
responsibility for nev/ development o'^ all central 
financial system.s at the 'ii'-. This new development '/ill 
follow the ?:eneral approach use^^ for the ''atari el 
M,anap;ement System, i.e. modular datai^ase desi,o-n and -^u 1 1 
inter;ration with the existin?- database system. '.'o nc' 
development will occur until aprreements can be reach"^^ on 
the maintenance of the existing accountinr system an^ the 
methodology to be used to shi^t fron t^^e current 
accounting application to a database deslj-n concent. 

Phase 11 of the I'lateriel flana^ement System is planned for 
implementation by Cecem.ber, 137r. This phase vn 1 1 
complete the receiving and payment cycle. 

The Clinical Support Section plans to continue improyinT 
and developing the Clinical Information Utility. The 
development v/ork will be done mainly on the archivin'^ 
procedure for low activity data and the additi-r^n o^ new 
clinical information databases. In the improvement area 
the section will continue to reduce t^^e cost o^ uo^^tinp- 
and retrieving data for those systems. 

63 



v;hen the Df'S conpletes dpvelopment of the Patient Recorrl 
Systen for the Carr'iolop:y "ranch, '!''L"1, it is intenrle'-! 
that this file be ner?-ed with that o^ the S'^rzer\/ Branch, 
'iHL3l. This integration should result in a hip-hlv useful 



database for patient care 
branches . 



an 



d clinical research for both 



Work 
sys t 
rece 
f'.ai n 
hand 
re PC 
1 abo 
Upda 
Work 
^or 



has recently be.i'un on a radiation safety control 
en for 'II!'. This will enable (1) Control of order, 
ipt and inventory of all radionuclides at 'M'r', (2) 
tenance of a training record on all persons, v/ho 

material and (3) Gathering and 

contractor survey data relatin<^ to 

handle radioactive Hv-^terial and (U) 

processing and activity balancing. 

the first two sys tens and is scheduled 

In the Fal 1 of 107". 



1 e radloact I ve 
r t i n g of all 
ra tor I es wh i ch 
ting of wastes 
has begun on 
i np 1 enen ta 1 1 on 



PROJECT LI3T 



The list belov.' does not include a nunher o^ snail nev/ 
projects and nodest revisions and adHitinns to existing 
systens. These becone literally to nunerous to nention in 
an annual report alt^^ough each is clearly important to the 
client and requires careful work by the OME sta^f. 



A . Clinical '^esearch, ^atlent Care 
and Ep i den i o 1 o"-v 



Retrospective Studv 
Valve '^eo1ace~ient 



C a r d i a c 



In 

con 

Sur 

Ret 

obj 

per 

and 

of 

nor 

sur 

va 1 

has 

'itlL 



10 75 th 
pu ter i z 
gery 
respect 
ect wa 

1 1 nen t 
ass I s 

this da 
t a 1 i t y 
gery, 
ves . T 
prove 
5 I i nve 



*ata 



M 



an 



ed data 
Branch, 
i ve Study 
s to des 
i nf orna 1 1 o 
t the CI I 
ta I n an a 
of pat I 
and to e 
he systen 
n to be 
s t i gators . 



;3 /rpr-)g 

proce 

of Ca 
i gn 
n on 
n I ca 1 
t tenp 
ents 
va 1 'ja 
becan 
a V 



nt 



rd I 
a s 

hea 
3u 

t t 
wh 

te 

e f 

a 1 u 



':■ r a n c 
n'' s 

The 
ac V 
ys ten 
r t 'va 
r gery 
o de 
o un 

the 
ul Iv 

able 



h b e '^ ? n d e v e 1 o nri e n 
y s t en ''^o r the 

s ys ten was call 
alve ^eplacenents 

v/h I ch cou 1 '^ na I n t 
Ive replace'"! ent p 

5r anch in the pro 
fine t^e nor^idi 



■^ I ! 



a I 
at 
ce 
ty 



derv/ent the onen h 



va r i ous procedu r 
ooerational In IC 



es 

77 



O*- 3 

n i C3 1 
the 
The 
n al 1 
I ents 
5 s I n g 
and 
ear t 
and 
and 



source of info mat ion ^or 



64 



Conputerized Patient ^ecorH 
Svsten for Cardio1o?Tv 

V/ork on this systen ^or the Cardiology -ranch, "!':L!:1 bejran 
durin,<T this year. The systen and database structure are 
the sane as that of the Surgery Branch systen. The systen 
nalntains patient nedicai infornation for both in-patients 
and ou t-pa t i en ts . 

The systen v/ili becone operational in nid-sunner 117". At 
that tine the Surgery Branch systen and the Cardiology 
Branch system v/ill be conbined to produce a sinrle systen 
and database to serve both branches. This v; i 1 1 orovide 
for nore conplete data and reduced naintenance effort. 

The conbined database '.-/ill naintain such infornation as 
operative notes, narrative sunnary of f o 1 1 ovz-un sheets and 
office chart, adnissions, patient relatives re'^erence, 
nuclear an,p;iogran, EFT-, X-ray, echo-card i orran, operations 
not at :!ll!, hospitalizations not at 'il!', catheterization, 
clinical visits, pacenakers, dia?:nosis, and autopsy. 



It is anticipated that 
operational by early n70. 



the conbined systen v; i 1 1 be 



■Analysis of Cardiac Valve Data 

This project for the Surgery Branch, '."'L'"!, consists of 
special statistical studies or the graphic presentation of 
results on the datahase created under the Retrospective 
Cardiac Valve Replacenent Study. f'ost of the stU'^ies this 



year v/ere calculations 
probab i 1 i t i es . 



o^ survival tines and survival 



(1) El 
presen 
var iou 
Valve 
pore 1 n 
probab 
w i t h 
occu rr 
conpu t 
probab 
probab 
su rv i v 
nechan 



even nu 1 t i pi e- 1 i ne CALCOf'.P plots v.-ere generat 
t graphically the survival tine dif^^erenc 
s experinental groupings o~ the 13S .".VR ( 

of 



Replacenent) patients 



The 



V R s '. / ere 



The cunulative su 



e and nechanical types-. 
ilities (classical conputation schene) of 
differing PA systenic pressures, RA nean ores 
ences of nitral stenosis and regi.' r ri ta t i on 
ed and displayed. Also presented '-/ere the su 
ilities for the isolated ilVR group versus 
ilities. A significant binding v/as generally 
al tine of the procine valve users as opposed 
i ca 1 users . 



ed to 
es o^ 
." i t r a 1 

both 
r V i V 3 1 
o-roups 
s u r e s , 
■•.'ere 
r V i V a 1 
those 
longer 
to the 



65 



(2) A nunber of propiran outputs '-/ere prepared for 
33 patients who had undergone at least two 
outputs included the fo 1 1 ov; i n.r^ : 



-/ P, ' s 



set of 
The 



a) A printed formatted dunp of selected clinical 
data fields. 

b) Descriptive and correlational statistics for 
the con pu ted survival tine, interval betv.-een 
operations, and a<re variables. 

c) A printout of nean survival ti'-'es for v^-rious 
age and sex categories. 

d) flu 1 t i pi e- 1 i ne C'^lCir° plots depictint^ t^e near 
survival tine for various categories of are, sex, 
functional class, and status (dep.d, lost). This 
was a graphic presentation of c) with the functional 
class variable ad'^e^. 

e) Printer olots displayed the 33 individual survival 
tines versus the interval between operations. 

f) CALCOf'P plots of functional class population counts 
versus tines of function class assi^^nnents (pre-oo 1, 
post-op 2, pre-op 2, post-op 2, or last visit). 



ruture .eeds 
needs arise 
the valves 
catef^ror i es . 



- !nconin<T requests v/ i 1 1 be evaluated as 
in the Gurgery Branch that would further test 



in relation to survival tine 



o- 



va r I ous 



Study o ^ 



Che-is ti 



Pat : 



ron 



Survey Participants, Colle,?e 
Q-^ Clinical Pathologists 



This collaborative study with the ' 
Applied Studies uses data fro-i t^( 



La^ora tor y 



1 r; -I c r 



Conprehen? i ve 



'^nei 



Chemistry Survey collected by the College 
Patholofti s ts. Traverse City, .".ichip-an. The 
represented are for Calciun, ^lucnse, and Chlori'^e. 
dataset Includes results only ^ron t^e participants 



prov I ded 
method on 



a f u 1 1 set 
all s Dec i nens 



12 values and w^o used the 



' 1 can 
data 

The 

'.■;' h O 

sane 



'rograns were conpleted to 



I t 



and 



connu t e 



sone 

r 



statistical paraneters. Fo 1 lov/ 1 n,"- a t'^orour'^ revie--' o 

the conputed paraneters, a detailed analysis was conoleted 

to test the differences between che'-^ical nethc^s usln"; t^^e 
Kruskal '..'all Is techniaues. 



66 



oerum Pool EvaUiation 

This collaborative study v/ith L\S involves poolinp- blood 
samples of nany patients and reading the oooled chenical 
results. This pooling v;as perfot — led over a t'^ree-nonth 
period in v/eekly Intervals. The analysis o^ this data 
v/ i 1 1 reveal effects of variation of analytic nethods over 
tine and v/ill be used as quality control in cc-iparlnp; and 
evaluatint^ other ?:roups. 

Completed to date are pro.Trans that v/ould compute Time 
Trend Analysis using the J'ean-squa red Successive 
Differences approach. Since some unexpected fluctuations 
have been revealed, evaluation of the slope over time v/as 
ana 1 yzed . 



Analysis of 3L7 Meohritis Patients 



This 

DCRT 

sys t 

on 

pat i 

thro 

cate 

Ther 

etc. 

of 

Info 

pa ro 

bone 

on t r 



is a collaborative research project incluHin,!^ LAS of 

and '\R? (Arthritis and Rheumatism branch), ■IIA.'-'"". A 

em was developed to store chemistry and therapv data 

all 3LE (Systemic Lupus Erythematosus) 



ents participating; in ARl] studies from 
u<Th 'lovember, 197G. Patient Oata 
-ories: i.e. Patin,^ of Clinical 
apy. Objective information, La^ Pata, 
^ Each cate°:ory of information contains 
information. For example, un 'er 
rmation", the subcate<^ories Include: 
t i d f lo'.v rate, 1 i p b icnsy, 
marro'-.', etc. The system a 
i es o^ al 1 i tcms . 



nephr i t i s 



carotid 
so " i 1 ■ 



■iovember, Ijn", 

is s to red by 

"an i f cs ta t ion, 

Comp 1 i ca t i ons, 

su'^ca te^^or I es 

the "Objective 

s '; I n b i ops y, 

scan, ^ra in scan, 

a 1 1 o' / -.11 1 1 1 ol e 



• he clinical chemistry data, urine ^ata, an^ t^^' 
hematolot^y data will be suoplied by the Cliniral 
Section from the Clinical Information 'Jtllity 



lu poor t 



'.'ork-Ahle Canr^ r ^a^Ients Emnlnvnpnf 
Stud I es 



This study for the Division o- Cancer Control an-^ 
Rehabilitation, '.'CI, involves poolin-^ o^ nu-erous su^-^==t^ 
bases prepared by contractors in different loraMons ^or 
tumor_re-istry. Each contractor Data Base includes survey 
questionnaires (responses from work-able cancer patients 
and their employers). Editing- of the responses and the 
^requency of each of the patients' employer respons-s have 
been completed. 



67 



So '^ar our analysis has revealei certain inconsistencies 
anons the '-iata snh'-ii tte ! hy ■'. 
inconsistencies are '^ein'T reGOlved 



"erent contractors. 



r-, e s e 



T'le :iaJor renainin,'^ tas'c 

patient 

aiiong Ji 



t'^ 



conpar e 



an 



versus enployer responses, as 
fferent seocraph. I ca 1 locations. 



.'ell as 



evaluate 
r e s no n s e 5 



Use of Population ristribiition to In prove 
csti^ati^n o' Iniivi lual Means in 
Eni ieiiiolofcical Studies 



This study v/i th LAS, DCRT exanines the nronerties of 
clinical chemistry data which ni^ht •- e 1 n to innrove the 
accuracy of ol^served values as estinates o^ the true neans 



of i nd i V i dua 1 s 



iris pr<~icedure uses t!~e v.-ei 



tr.d 



svp.rn'^.e 



to replace observe J values usin'^ the population nean 
using tlie Empirical Bayes estinates. 



and 



,'^enet ra t i n*"^ ''ea^^ Injury Stud^ 

A connu ter i ze^i lata processin.^^ syste'-i \.ns 'ovelopod for 
the f"hief c*^ Laboratory of Fxperinental ""urnloq;y, ".'I'lCD'^, 
to maintain and query a database cnnsisti'^"; o^ nedical 
data on American military ^len sustaininj^ ponetratin-r head 
injuries in the Viet "an v/ar. T^e data • "a i n tenance 
portion of the system became "^uily operational d.irin^: t'"is 



year 



Evaluation and analysis o"^ t'e 



ta 



I s 



con t 1 nu I n' 



to he handled by t!;e Pl-ysical Sciences Laboratory^ P'^R' 



10. '!|fM: Clinical Pvesearch Pata 
'ana:Te' len t Svs te 'i 



An on- 1 i 


ne data nana.'^ene 


Duclisbau 


n ^or use 


v/i th 


c 


present. 


J i fferent 


resea 


even thou 


gh the sar- 


!e ,";roii 


different researcf' 


pro. 


je 


into his 


ov/n file 


that 


d 


study. 


Ponseq uen 


tly. 




file ma i 


n tenance 


1 s 


!i 


doubtful 


1 t i 


S d.( 


= s 


s tandard 


on- 1 i ne 


data 


s 


IT po S S 1 


ble, ma!-.e 


such 


a 


in other 


c 1 i n i ca 1 


resear 



nt system 
1 i n i ca 1 r 
rchers '^a 
p of oat 
cts. ?.ac 
a ta './hi ch 
':!a ta is f 
"" ^ i c u 1 t , 
i rab 1 e t 
ystem for 
system a 
c h e n V i r o 



csearc 

i r ta ! n 
i en t s 
h rese 

■ s P« 
ra.'^nen 

an i 
he ref o 

use '. / 
d a p t a b 
nnen ts 



pen 



I s 
?rr. 
r ti 
teJ 
•'at 
re 
i f- 
le 



oro 
a ta 
i ^fe 

i nv 
her 
nen t 

and 
a i 

to 
i n t 
for 



posnJ ^: 



at fd 

rent 

o 1 ve i 

i ncor 

to '^ 

!up1 

,-. to-r 

'eve 

I 



VV . At 
" i Ins, 
i n the 
po r a t e s 
i 5 o'. /n 
i c a t G d , 
i t y is 
1 o p a 

' " ' -in' 

T e n e r n 1 use 



M 



68 



The proposed systen vyould be sinl Tar tn systens o'^-^ered Sy 
conmercial tine-sharlne comnanies. It v/ould include 
facilities for data entry, up'^atinT, ref ornn 1 1 i n^^, 
retrieval, and statistics and \.'ould he oriented to'.'ard use 
with small scientific databases. Specifications for the 
systen have been approved, and D::^ has undertaken 
inpl ementa t ion as a lonn'-tern develonnent project. 

11 . Purge Data System 

A pattern recop;nition al^^orithm v/as developed to identify 
two classes of pur^e data from the Clinical Center fllS 
Systen: fCnown patterns to be archived in the C I U and 
those patterns which are not to be processed. There is a 
third class of data which is unkno'/n to the alrorithm and 
it will be displayed durinp: execution. The displayed 
class of pursue data v; i 1 1 be analyzed to determine './hich of 
the above tv/o classes it belonrs. Cnce the class has been 
determined the purse data systen '/ill be chan5^ed to 
process it. 



12 . Pat i ent ^rof i 1 p Gvsfem 

A systen of programs is being devel^ppr' for the Clinical 
information Utility to display a profile hy natient. This 
profile will include all conpu ter I ze^-' data for a specific 
patient presently in the Clinical in'^ormatlnn UtMity. 
The format of the profile will be in patient num^pr or-icr 
by date. 



13. Sur-^ical Patho1o<:rv ^v^tpn 



The Surgical Pathology Systen was begun in l"??? for the 
Pathiological Anatomy Department, CC, as part of the 
ongoing Dfin project in support of Clinical Center Data 
Processing. This year it was modified tn include 
Henatopathology data from that department. 

1^ ■ Sickle Cell Anemia Pafiont'^ 

Molecular Hematology Cranch, 'Ir'tri, is studying ?.^-^S 
sickle cell anemia patients who have entpre--' an '.'ll-i stud" 
over thejast 2-2 1/2 years. They are eatherin^ 
information on the clinical course of the diseasp in or^er 
to predict the course in an individual patient. There is 
great variance among patients and it has ^ppn di^^lcult tn 
evaluate treatments. 

69 



P"[^. Update 
the C 1 i n i ca 1 
source of lah test 
"etrieval Systen as 
the tool for querying the confined database. "e also 
provide a patient history printout for use hy the 
phys i c i an/ i nves t i sator as patients re-enter the hospital. 



\!e have built a systen '/hich intef^rates t'-'e 
'lenerator for data entry by the investigator; 
Infornation Utility systen as 
infornation; and the Synbolic Lofic 



As the usefulness of the database is explored usinr 
facilities provided, statistical analysis reauirenents 
will be formulated and further enhancement 
sys ten. 



the 
nts 
nade to the 



15. MiS Study 

The Office of the ITeputy Director, Clinical Center, is 
doing a study of the iledlcal information System to analyze 
its impact on patient care. 

Pro!?rans have been written to plot daily systen 
performance (number of requests and './altinp; time vs. tine 
of day) to deternine peak periods o*^ usa<^e './ith a vie-/ to 
d i s t r i bu t i n,T usage more evenly durin'^ the day. 
Consul at ion with LS'!"' provided a s cor in."- al'^orit'^n -./blch 
v/as impl emen ted . 



"hen t'^e data is all in and run, < 
as to '..'hef^er more elaborate '^lot 



M 



e c I s I -^' n 



•/O'l 



1 -J 



■ 1 1 
s p r 



'e 



'^rrc' 



Cancer Su rv i va 1 



V5 te^i 



^e Survival Systen v/as oririnally ^eveloopH in tli 



to support the End results 



I n 



:"aintenance 
?;oal of this 



and improvenent 
proj ect . 



'^ a n c e r 
t^^e svs 



tud i es 



o^ 



em is t ^e primary 



Ourinn; FY7S a number of 'IVS- re 1 a ted pro'^lens 
in conjunction with certain apolications o 
v/ere resolved. 1,'e also sent the systen to 
Tumor Re;!:istry. 



'■/hich arose 

f the system 

the i 'a:: a i i 



17 . Chronic D i a 1 '/s i s Complication Study 

The staff of the Artificial .K i dnev-Chron i c 
'lIAflDD, is assessing the frequency 
complications of chronic renal diseases in 
direct research efforts to areas 
improvement of therapy. 



u r em I a 

of 
order 



p ro^r am, 

various 

:o better 



o^ max i mum 



5ene 



I n 



70 



To aid in this they have selected ^ive najor dialysis 

centers v/idely scattered over the U.S. and are rev i ev/ i n"- 

hospitalization records of chronic renal ^ailure patients 
(patients on dialysis) during the year 1976. 

Durin/T the past fiscal year v/e conoleted the database 
creation and produced a clean file for analysis. L3''.'l is 

currently providing the statistical support for this 
proj ect . 



IS . CrvDtococcal flenin^itis 

The Office of the Scientific Director, ■.'l^![:, v/as 
following sone menin<?;itis patients us i n.r the C^C-31 
conputer in the Clinical Center. '.Mth thp f ol 1 nv/-up 
infornation coninp: in and '.-/ith the denise o'^ the C"C-31 
It v/as necessary to convert the pmr^rar.s and data to t^e 
IBil 370. Conversion and r es t ructn r i nf- of nro^r^ns too!; 
place in tine to a 1 1 ov/ updated results to he presented at 
a neetinn; of the Infectious Lisease Gociety an-' the 
International Con/^ress of Antibiotics and Cheno theraov . 

13 . Analysis of Risk Factors ^or t^e 
jjevelonnent of Adrianvcin (A) - 
Induced Coniective Keart Failure 

The Adrianycin Toxicity project './as beKun in FY77 for t*^6 
Cancer Therapy Evaluation Pro^ran, '111. Turin'^ FY7.'l data 
entry and initial analysis for U.o. patients was 
completed. \n additional four hundred patients under the 
age of ten v/ i 1 1 be added to the Adrianycin Toxicity study. 
This will serve to broaden the risk factor analysis base 
for this irroup of patients. Follow-uo 1 nf orna t i on ••/ i 1 l' 
also be added for all patients currently on file. Cr. 
Van ;;off, _ the principal i nves t i ,:^a tor on this project, is 
also gathering data on European patients and these w i 1 1 Kg 
added to the file during this conin<T fiscal year. There 
is also sone discussion on establishing an a-frianvrin 
clearinghouse to provide doctors v/ith risk factor analvsis 
for individual patients. 



71 



'MAI'CD 3t'jdv o^^ t^.e Incidence and 
Prevalence of Kidnev and Urinar y 
Tract diseases in the Arned Forces 

This study was bej:un in 1^7U for the r.idney L'isease 
Progran, ".\MXD, to evaluate the occurrence, nor'^idity and 
nortality o-*^ kidney and urinary tract diseases in an 
effort to deternine research needs. Oata for the stu'^y 
v/as made available by the Air Force, Xrny an'^ ;avy. There 
has been difficulty In obtaining- correct dat? fron the 
Army. '.ie received our fourth set of data fron the Army 
during FY 78 and ran a series of reports. The data 
involved additional diaj^nosis codes and covered the years 
January, 1971 through June, 1^77. The orlpiinal ^ata 
requested was for the period of Januarv, 1^71 throuf^h 
December, 1'373. Present plans are to prepare renorts 
usinp; the Army data, assunInF it is correct. ^^ter t^is 
is completed a decision \.'ill have to ^e ma^e to either 
request the additional data -Prom the \\r Force an.'"^ '.'avy or 
eliminate the additional data received -rrom the Army when 
comparinj^ data for the three services. 



Laboratory investigation 



Guryev o^^ Che'-^icals "'einp- Teste""* 
for 3a r c i no <^en i c i L v 

Early In 1^7^? the Care i noprenes i s Testinr^ ."^ro.^rram, 'IGI, 
asked D.'ID to assist the International Aprency for "esearch 
in Cancer, '..'orlri ::ealth Organization, In the development 
of a computerized Information Bulletin on the Survey of 
Chemicals being tested for Carcinogenicity. 

The system should accept randomly at least nine items oer 
record (i.e., (1) country, (2) city, (3) name and address 
of institute, (k) name of chemical, (5) species (strain), 
(^) exposures, (7) sta<^e of experiment, (8) principal 
investigators, and (0) comments-), sort on any o'^ these 
items, generate indices, and" output the renort in a 
columnar format. The source input •■/ i 1 1 come '^rom 
questionnaires distrib'jted by the ''or]'^ !-e.= lth 
Organization to 1 aborator i es/ i ns t i tu tes throu'^hout thp 
wo rid. 

Currently, basic system design questions are being 
clarified. This project will carry over into the coming 
year. 



72 



Chrona to<^r aohv and fi 1 ec t ronhor es i s 
Peak A. n a 1 V z e r 



The Reproduction Research branch of '!IC!'r Is 
the preparation of puri-^ied proteins (henop-lo'^ 
hornones) by the process of j^el e 1 ec trophores 
scanning isoelectric focusing assenbly ( 
Apparatus) it is possible to obtain, a 
quantitative data on the position, bandwldf-i, 
of the protein as it filters through th 
phys i ca 1 /chemi cal data can provide in^orna 
enables researchers to determine rules con 
behavior of the proteins in the Tel and to enp 
refine their purification procedures. 



i nvo 1 V 
in, a 1 
is. U 
Cats in 
cross 
peaks, 
e o;el . 
t i on 
trr 1 1 i 
hie th 



>^ 



I n 



""luni n, 

s i n.'T a 

poo 1 i s 

t ine, 

etc . 

This 

'.7 h i c h 

n p- the 

e~i to 



'.iorking closely over the last ^ive years v/ith t^^e research 
scientist, Di'iC developed an interactive systen 
perforning statistical analysis and displaying- 
v; h i c h can 



i nter f ace v; i th 



tor 
peaks. 



a non-conpu ter oriented user 



Currently, v/ork is progress i nf' on a nothod for 
automatically choosin,;? the tolerances for per': detection. 



This systen was presented 
International Conference at 
in preparation. 



:t the electrophoresis * 7 "" 
T . \ j o u r n r. 1 a r *: i c 1 e is n c ■.' 



'^a r c 1 no p-pnes i s 



n'^orna t 



on - v - 1 u a t i ^ n 



Rurlnp: FY73 v.-e continued development He^-un 



I n 



analysis phase of the Carcinogenesis Cioassa- 
for the Carcinogenesis 
innediate aim is to 
ca re i nogens . 



" e s t i n g 
detect 



.^rogran, 
prev i ous 1 v 



^V7-^ on the 
Rata System 
'iCI. The 

unldenti'^^ie'-^ 



The interactive system for statistical analysis (C'CR3\3) 
\/as upgraded during the year v/ith the i ns t--' 1 1 a t i on 0^= a 
trend analysis nodule. The C^G\S Terminal User's "anual 
underwent extensive revisi'^n to reelect a^'r^ 
revisions of the system. 



t i o n s to 



As the results of the experiments continue to arr 
accelerating rate, usage of this system ^y 
contractor personnel has been high. 



ve 



a t an 
I and 



73 



Gene napping Program 

At the Laboratory of Viral Car ci nop:enes i s, '.'CI, a nap of 
cat chronosones is bein'' develope-^ ".'ith the inne''iaLe 
purpose of studying cat leukenia, and the Ion?-- tern g-oal 
of naking predictions about enzyne defective diseases. 

Cloned cultures of hybrid cat-nouse cells are assaye'^ for 
about kO proteins by sel el ect rop'^ores i s techniques v;i^ich 
not only resolve t^e di'^ferent proteins but v/ i 11 usually 
a 1 1 ov/ the deterni na t i on of .rrenotyoe. 

A program \*/as v/ritten this year to tai^ulate the ^.e^ree of 
simultaneous activation of various .^rene pairs ^or all 
clones according; to prescribed cate<='ories and to calculate 
the percent of discordancy for each protein pair. 'Jsln,?: 
these pair comparisons it can be determined v/het'-ier or not 
any tv/o °;enes are on the same chromosome. '\nother program 
lists the specific clones v/hich exhibit discordancies for 
any tv;o specified enzymes. 

Since a host cell '.'/ill usually only have one donor 
chromosome, the above information collected ^rom a 
sufficient variety of clones './111 a 1 1 ov/ investigators to 
determine on v/hlch chromosome each j^ene is located in the 
donor genotype and thereby map the chromosome location '^or 
the common protein. 



Cvto tox I c I tv 

The 1 mnunopa tho 1 OFy 3ectIon, '.'CI, performs experiments to 
determine toxicity levels. Calculations of 
radioimmunoassay data v/ere done "/Ith a han"-" calcul-^tor, 
graphs o^ samples plotted out by hand, an^ the toxicity of 
various hlolotlcal samples read of^ the a-raph. '.'e 
automated this process to run on the r7C-l'^. The nro'^ram 
Is general enough that It can be used for subsequent 
experiments. ideas for future enhancements to the program 
have been discussed. 



Henetlc 'barker Tracing Cvstem 

This project for the Laboratory of Immuno^ene t i cs , li^l~, 
Involves the analysis, design and creation of a system to 
provide the capability for tracint^ r^enetic markers ■./It'^in 
a colony of rabbits. The initial analysis, desirn of the 
data collection for~s and design o^ the databases are 
completed. The file maintenance and retrieval routines 

74 



are conpleted an'i tested. The ^^irst update of the n^ster 
file was run In June, 107? and elven to the user. \fter a 
fev/ nore update runs, and connletion of the docunen ta 1 1 on, 
the systen will be turned over to the 'j<^er. 



Chenical I nf or^a t i on Svsten 

This project for the Carci nop:en i c Testinp; Pro,'Tran, !C 1 , 

involved the modification of an I .";5 systen developed hy 

Stanford Research Institute to run on the '!iH/CC"T 

computer systen. This effort v/as connleted during- the 

year and the systen turned over 
runn i nrr . 



Circadian r^hvthn in Rodents 

i^ionedical En.f^i neer i n s and I ns t riji-ien ta t i on "ranch (jEI~) 
of the Division of Research Services (lRS) is huildin?^ a 
systen that v/ 1 1 1 monitor, record and report the activity 
level of rodents in ca^es. A ni cro-processor '/ill recnrd 
such factors as aninal tenperature, roon temperature, room 
light, and level of activity \.'ithln a ^iven tine period, 
from sensors located within each study care. The data 
from the micro-processor v/ i 1 1 he transferred to the 
I3f:-37n and stored as l/YLSUR datasets, v/hich \/ill he used 
as input to an update routine. 

The database has been defined and update and plottin<T 
pro/^rams have been v/ritten but testing has ^oer, limited 
until data can be received from the micro-processor. ^ t 
the present time BEIS is havini^; problems v/ith the 
ni cro-processor . 



In Vitro Information System 

The In Vitro In-^ormatlon System (I VIS) is beinr^ Heveloned 
to process information about In Vitro care i nop-enes i s tests 
conducted under the auspices, of the I'ivision o^ Sancer 
Cause and Prevention of the 'rational Cancer Institute. 
iViS may also be used to process data about tests 
conducted by other agencies. The system provides for the 
collection, maintenance, analysis and reoortin,'^ o^ In 
Vitro data. Data for the system will come from 
contract i n,<^ laboratories on input forms v.-ith both 
descriptive and results data. The input forms v/ 1 1 1 be 
converted into machine readable format and be processed on 



75 



computers at the Division 



;o~ip'j ter 



'es earch 



md 



Technology (nCP.T) at "!' 

systen Installed in February, l^yp 

s y s t en u s i n ?; : ' A R f' i V is in progress. 



usin^r the proprietary "'AnXiV 



P e V e 1 o pn e n t 



m 



■utarenic Testing Svsten 



This s 
3 i onet 
£nv i ro 
nu tage 
nul tl - 
I nter-^ 
con tro 
anal ys 
the s 
con pi e 
curren 
pro toe 
d e V e 1 o 



ysten v/ i 1 1 provide pror;rans for the 
ry 3r?nch of the "lational In 
nnental '-'ealth Sciences to nonitor the 
nesis testing. Design specificatio 
terninal, on-line, conversational pro<^r 
ace v/ith laboratory technicians 
lied data collection, data storat^ 
is. During FY7S a neeting v/as '^eld to 
ysten in a si nul a ted operation. \t 
te systen '.vas release'^ ^or live testing 
t nuta"-enesis research contract. The 



ol 



■or 



suspens ion 



tes t i n-^ 



I s 



Env i ronnen ta 1 

s t i tu tes of 

protocol for 

ns cal 1 ^or 

an packa"-e to 

and provide 

e, an'-' data 

T^enons t ra te 

th^t tine the 

as nart n^ a 

nu t ~ ""ones i s 

p. '■' p r re v I s e '^ 



Tcn t 



Liquid oc i n t i 11 a t i on Counter 



^'.O'.i that 3 data logger is 
scintillation output 



for 



! v a i 1 a b 1 e t " s t r- n ■^ a r d 
easv in out to t'^e co- 



ze liquid 
'puter, ':</" 



is developing, on the 



'?-n. 



conpr ehens I ve set 



o ■ 



routines which can ansv/er the needs o^ a reasonably larsie 



percentage 
Counters . 



scientists 



v/ho 



use Liquid Scintillation 



.70 rcint^ version 



of 



■!OS 



of 
the 



During FY73 v/e produced an interactive 

a s el ■f^- teach I ng systen which takes care 

problens encountered in Chrona togranh I c- E 1 ect rophores i s 

(except autonatic peak detection) and './Hich can ^.r^v: plots 

on any of a nunber o'*^ devices including t^^e DEC-'JU'^ visual 

display and the teletype. Subsequently these nlots can be 

sent directly to the CalConp for offline plottin'^'. 



At present, work is being done "on a "nini ''-La"^" 
peak detection nodule v/hlch './111 alio'-' the user 
and modify the results of the r)rop-rBn interactive! 



:ype 



76 



'^ro'^ran Mana^enent an^ \dni n i s t r ^ t i on 



:'IIH International Activities and 
Personnel '^onitorinp- Svsten 

A fe>.-/ years ago, the Data "lanap-enent Branch i -^dI enen te'-' 



this systen 



to 



orovide the Fo^iarty international Center 



with the capability to maintain, report 
database coverinjr the Visitinr; 
'■.'orkers and Foreign Visitors at '1 



on 



ro^Tran, 



an- 



query 



-or e I gn '^ues t 



The systen has proven valuable to the Fogarty Center. 
This year the systen is beinp; re-evalu?ted an^, v/here 
appropriate, upgraded and expanded vi'\th current technology 
to neet current needs. 



I'Jork is being started to identify needs in several o^ the 
other Fogarty offices, '.nth an eye tov-ar^ developing: 



individualized computerized systens for t^^es 



,^f ; 



systens which v/ i 1 1 be conpatible an- 
other. 



i cos, bu t 



ill cor-iolenent each 



""irants infornation Files 3'^sten 



The computerized Review 
Grants infornation Files 
inplenented for the 
the last three years. 



and Evaluation 
Systen v.' as 



D i V i s 



Branch (RAEL) 
^es i gned 



on 



of Cancer Grants, '.'C 



and 
over 



The systen maintains data, ansv/ers 
variety of reports relative to ' 



portion of the system v/ i 1 1 be implemented next 



queries an-* produces a 
grants. A contracts 



'omputerized l' i s t r i bu t i on Svsfen 



A computerized mailing list system was desip-ne-' an- 
implemented for the '11 H Printing an^ Reproduct fon' "ranch . 



The system of 
comprised of 
databases . 



computer pror^rams and 
two subsystems and 



^ne subsystem mainta 



nrocedures 
' their res oec t 
ns a base 



I s 

! ve 

of 



approximately fourteen thousand -MM employees. f'ailinc^ 
labels can be selectively produced usin^ suc^^ information 
as work category, organization, o-ra^e, title, etc. as 
well as any of several special mailing list Indicators. 



77 



The other subsystem maintains a database of about seventy 
five thousand ou ts i de-of -'1 M' Individuals. ,",ai]in<^ labels 
can be selectively produced for any of the nearly seven 
hundred different indicators. 'ioth su'-systens provide for 
selection usin^^r any of the nailinp- list in'^icators in any 
logical combination. 

This system makes it possible for the Printing- and 
Reproduction 3ranch to fill requests for labels usually 
within tv/enty-four hours. 



Library Circulation :Wstem 



This system provides the HT'^T Librarv -.'it'- ='n improved 
computerized method of Inventory control, circulation 
control and to monitor usace on an item (book, journal, 
etc.) basis. The system pays particular attention to ease 
of input, accurate an^ immediate status o''^ circulation 
items, usa.^e statistics, and automatic .p-eneration of 
overdue notices. The system controls overdu^s, and 
provides waltln?- lists and listings o^ loans by bor rov/er . 
It will be used to v/eed out less useful and outdated 
materials, to aid In clearlnr employees leavinr "'I'-i, an'' 
to monitor the v/hereabou ts of all materials. The programs 
are written and av/altlns data input. 



Consultant -lie "ana?e^ent Svstem 

The Consultant File 'lana,n-emen t System for the Revie-/ 
"ranch of Extramural Affairs division, '.'''LCI, "/as 
completed in 1078. 

The system v/ i 1 1 help the Reviev.' Cranc^ nersonnel orp:ani2e 
and maintain a master file of 'consultants' from 
throughout the United States '.i^n participate in ".'i-lLCI site 
visits to various research projects in prof^ress which are 
funded by "IIM .(grants or contracts from ''''L''l. 

The system maintains all I den 1 1 ■^ I cat i on i p-'^orma t i on, 
fields of expertise and exoerience, anH information an^' 
history of all site visits for each consultant. It will 
serve as an aid in ■f^ormln?' site-visit teams, provide 
correspondence assistance and make possi'^le basic 
statistics on t^e consultant and site-visit activity. 

The Review Branch Is currently initializing t^e database. 
'/hen that Is completed, the system •.-.'ill be operational. 



78 



later i el '.'am.'^e'-^erit Svsten 



This 
Inpl 
f'.ana 
oper 
on- 1 
tech 
(IMS 
prov 
oper 
requ 



project is concerned '.-/ith the desi.'^n, develoo-ient ?nd 
enentation of a conputer based 'liH-v.'ide "aterlel 
genent System *or the purpose of inprovin" overall 
atinj: efficiency. The systen operates prinarlly in an 
ine environnent utilizinr database nanap;enent 



no logy, nanel y I B 



D ■ ; I 



infornation .".anaf e'nen t 3yste~i 



). In addition to the on-line functions, the systen 
ides for periodic batch-process i n?- functions to neet 
ational, accounting and nanar:enent control 
i renen ts . 



Phas 
con t 
Janu 
edit 
per 
data 
,o;ood 
na i n 
^urc 
orde 
\dni 
.Obi i 

^cco 



e I '.■/ 
rol 
ary . 
i n^; o 
year 
base 
s a 
t a i n e 
base 
r pu n 
n i s tr 
<;s t i o 
u n t i n 



hie 
svs 

in 
f a 

con 

nd 

Or 

Ch 

ati 
n 



h c 
ten 
vo 1 
1 1 

As 

s i s 

s 

Th 
^.er 
cp.r 
on, 
tra 
yst 



an best be characterized as a 
successfully v/ent into product 
in this phase is the data 
procurencnt orders, sone 25'^ 



ve' 

Ml : 



part of the 
ting of sone 
er vices to 
e prinary output 
s to ^e 
-"s to be 

3) '^rr'.cr ^ 
ns actions that 
en. 



data validatio 
qn^ nnn vendors \i 
the ^e'^eral tov 
.^ h a s 1 p. re 

t^ ^ 

:i 



sent to 
sent to 
eoo r ts ■Po r 



e n e r a 
n s '~' e v 



3 n 



pa sse^' to t^e 



proc 
i on 
en t 

"TO 

n, a 

ho 
ernn 

1 ) 

2 } - 
1 

i e'/, 

u 



LI renen t 

in late 

r y an d 

orders 

ven'^or 

prov i de 

en t is 

""r i nte-^ 

: : 3T ^ I ^ 

e r v i c e =; 

-n^ ^) 

Central 



Phase 11, 
deve 1 opnen t . 



Aeceiving and Payient, is oresentl 



under 



\ninal M location .~'rop:ran 

The Extranural and Collaborative Prograns Croup, 



maintained a colony of aging rats and nice 



1^, has 
the Charles 

River Breeding Laboratories in order to supply rodents o^ 
various ages to investigators stu'^yin?- tbe a<^in'^ process. 
^t present there are three aninal colonies v/ith a fo'.irth 
to be added and plans to expand to eleven strains. The 
job of balancing future supply -.-/ith present and future 
demand has already become alnost unnanan^ea^ 1 o by han'^. 

'!!A wanted a conputerized nethod of: (1) projectin'^ 3^ 
months ahead of time the effects on the colonies o^ actual 
drones, (2) using projected r'ra'\' and colony size 'or 
planning purposes. 



79 



There is nov; a running protrran ',7^1 ch aoplies survival 
rates and projected drav/ to the v.'eekly colony cens'js 
figures and projects colony size for rodents up to H^ 
months of ase as desired. This procran also al1o'./s the 
■•lA to project the ef-^ects of unexpected drav.'s, 
non-routine drav/s, and the addition of aninals to the 
colonies at any ase ?-roup. 



ORP. integrated I nf or'-'a t i on 
''ana<^enent Gvste^ 

The Office of Progran Analysis, D^P., has a compu ter-hased 
systen for handling orottran data. D"R novj v/ou H like nore 
capahillties than the original des i p:n allov/s (e.g., 
interactive queryin'^, plots, etc.) dIus a .~;ore nana^-ea'i 1 e 
uDdate schene. Difficulties in correlating- scientific and 
administrative data must be resolved and a \.'ay found to 
provide shorter response time for filling requests. 

lie have been v/orking '/ith CRR to determine a file desltrn 
and retrieval scheme v/hich can orovide the'i '/ith the 
retrieval logic flexibility they require './ithout 
sacrificing quick response time. Some decisions have been 
made regarding methods of implementation and '.-/ork \7 i 1 ] 
continue during FYTT. 



Primate Reoortin?: Gystem 

This system has been developed to control data on primates 
processed through the '/eterinary "Quarantine unit. Office 
of the Director, DRS. During the fiscal year DCRT support 
of the Optical Character ^ea'ier (OCR) v/as discontinued. 
This necessitated designing a new form ^or keyinr anr! 
making mc^^ 1 f i ca 1 1 ons to a number o^ programs In the 
system. The system has been designed to a 1 1 ov/ for input 
of data either throuf'h batch keyin^- or on-line 
interactively at the user's terminal. \]] pro'-^rams for 
collecting, updating, editing, and reports have ^oen 
completed. ■^ Symbolic Logi c -Re t r i eva 1 (3LR) prop-ran has 
also been provided for the user to ans'.'er ar* hoc queries 
on-line. The system v/as turned over to the user in '^oril, 
1378. 



80 



m 



iase Reports 



As a result of the 
A cad en i c 3c i ence 
have been produced 
prepared for the D 
?ro?:r?.r] Plannin? an 
awards to I ns t i tu 
Professional 5 choc 
Research i ns t i tut 
Research and Develo 
a 1 1 been rece i ved, 
the '.'.at I ona 1 5c i 
developed durinz th 
years data with 
this year and prove 
number of report! 
to ,<^ross discrepanc 
records and ei the 
reports should be c 



reconnenda t 
and Enn-ine 
annually si 
i V i s i on o^ 
d Evaluatio 
t i ons o^ 
Is, lon-pr 
GS and 
pnent Cente 
ed i ted and 
ence 



"ou nd 
e last f i sc 
the i neon i 
d to be V 
no- connonen 
i e s and ',■/ e 
r resubn i t 
onpleted du 



I ons 
Grin 
nee 
Pes e 
n C^ 
I' i -h 
of i t 
Oper 
rs . 
bala 
a t io 
al y 
nr d 
ery 
tor 
re 

n e'./ 
r i n<T 



(C 

■^ n r- r 

X - ) 5 . 
a rc^ 

er 

ros 
a t i nr 

The 
need, 

n (•; 

G? r 

r' t a \'l 

s 'J c c 
T 3 n i z 
a^le 

or 3 

thi s 



th 

A ^ rr 
\ ^j - 

T 

Ser 

an 

Edu 

Di t 

f- 
I" 

det 
an 

3F) 
to 

ess 

?t i 

to 

nen 



ttee on 

repo r ts 

r ts are 



e oonni 

) , these 

he repoi 

vices, 'jffice of 



5 u nn a r 
ca t ion, 
a Is, Ao 
ounda t i o 
a for F 
d con ! es 
The 
m tc^ 
us e'^ on 
^u 1 i n 
ons '.■/ G r e 

rechec 
de': data 
s c = 1 v p a 



I ze Jt':i\. 

i-'eal th 

n-pro-f^ i t 

ns, and 

Y7 7 has 

sent to 

prorr an 

pr ev i ous 

all data 

1 1^ a t a 

alerted 

k t ^ G i r 

Final 

r . 



1 1 



Svsten for Controlling- and 
■"onitorin?: Con plaints n^ 
riscrinination at 'II'' 



1 V I s I on 



Equal 



!ppor tun i ty 



containin--^ infor-iation on fornal 



Th is sys ten for the 'Ml! 
establishes a database 

and Infernal complaints of discrimination at thp Alii to: 1) 
provide statistics associated with processing conplaints 
In a nore timely fashion; and '>.) enable the user to nore 
closely monitor status of conplaints. Desi.'^n 
specifications and codine fornats have ^een dGvelope^'. \n 
update^ proj^ran has been written anH transaction records 
are being collected as input to the naster ^ i 1 <^ . A 
Synbollc Lof^ic Retrieval (3LP) pro.^ran has ^^ 



query the mas ter file, 
system durin.q; this fiscal 
be i ng col 1 ec ted. 



eon './r 1 t ten to 
It is honed to inplenent this 
year althovir^h data is still 



12 . ''ateriel Aana.°:enent System (:"'S) 
'^uerv and Reports 

This project will determine t^^e f'eas I >- i 1 i tv of lis 
'lAR.AIV systen to provide an economical method 



•or 



t '-^e 

the 



selection and reporting; of data fron the ''IH "•'C data'-ase 
\°^'' daily and two of the -/eekly reports produced ^v t^^ 



inquiry routine have been 



v/r I 



ten 



-ARKIV outputs are beinp; evaluated, throu?:h parallel 



'i:iV. Th; 



runs. 



81 



v/Ith the reports fron the T'S Inq'jire. Future plans for 
the use of f'A^KlV, in conjunction '-.'ith the ""'o "^iles, 

auHit tr-iil file 



an 



include the developnent o' 
facility to handle ad hoc requests 



and 



L:> 



.; I V I s I on o' 



es e^ r c 



:jerv I res 



1 



o p- 



I V I ties 



The Of-f^ice o^ the Director, ?•""■.$, requested thnt analysis, 
design and pro.Pirannl ni: services he provided by the n:"i2' to: 



1) Establish a job control systen ^or the systens 
naintenance section (Sf'S) o^ the Bionedicpl 
Engineering; and I ns t runent i on 3 ranch ( ^: E I ~ ) . 
The systen involved '•/ritintT collection, un-^r-rp 
and edit, and daily and v/eekly renort proo-rans. 
The systen '-/as conpleted and turned over to f^e 
user in 'larch, l*^?"; 



?. ) Develop a rental infornation 
, The analysis and desi'^n hav 
this sys ten 



sy 



of forms 



and des i '^n have 
'..'or'-, in pro-^res s 
for data collection, 
programs, procedures ^or 

f an i ns t runen t 
Access !;ethod (IS*-'-'!) file, a 

to s 
;i th 



p 
Ind 



crea t i on o 

fiethod (IS 
reports. Present plans are 

the 



sys ten, in par a 1 1 e 1 
during this fiscal year 



s ten 
en CO 
i nvo 1 

unda 
r ror 
exed 
n'^ or 
tart 

nanu 



fo 

nnl 

ves 

te 

h 

^^e 
o ^ u 
run 
.^1 



r :; 
e ted 
de 
and 
an^l 
q u e n 
c t i n 
n I n 
s vs 



for 
=; i 7n 
ed i t 
i n", 
t i a 1 
n o-^ 
^ the 
ten. 



Iti . Survey of P. i ohaz ardous "aterials 

The initial ohase o^ this autonated 



S'/s 



-^r 



■-; a s 



des i r:ncd to print a questionnaire t^at '..'ill l-trr '--o s-'^r*: 
to all o^ f-e ■'!■• l^hs. ^'ata '/ill '^e colleCred -rr G?ch 
employee \.'ho Is v/orkInT '/It*^ any '-' i ol o"' c? 1 a'^'=-^':'=, tiss-!" 
cultures, etc. liste'^ in the ques t ionna I r •? . 'n nn-i'pr 



<-<^-i 



data collection orofr am has hppp './ritten an^^ teste 
-.eport p-eneratin^ pro-^rams av/ait spec i ■^ i cat i ^ns . 



15 . '"adiation Safety Control S'/ste' 



The "^.adiation Sa'^ety Control Syste'" 
conposed of the "^o 1 1 ov." i nc subsystems: 



^or ^^S 



ill ^e 



1) Inventory and Cioassay 

2 ) T r a i n i n f 

3) Laboratory Survey and Airborne '^e lease 

U) '..'aste i-rocessed and. -\ctivi^y '■a lance 

82 



The total systen should be an interrrate^ svsten canoble o^ 
supplying: data fron one subsysten to another in order to 
produce desired reports. 



'.iaste Processed and ^ctivity 3a lance 
undefined as yet. 



subsysten has been 



The analysis and pro,";ranni n,<' ■'^or the Inventory and 
3ioassay subsysten has been connleted. '..'e Pire presently 
avyaitlnp- the data necessary to build the control file of 
individuals using radionuclide naterlals before './e can 
proceed. Upon conpletion of this file, a parallpl o^^ the 
nev/ and old systens will be run to isolate any 
deficiencies or progrannin?; bu^s. it is estimated that 
the nev/ syston will be suoplenented in Septenber a^ l"^?". 

A neeting v/as held June 1, 1073 to discuss the developnent 
of the Training subsysten. \ fol 1 o.-.'-up neetint v; i 1 1 be. 
held June 20, 1973. The progranninc of the su'^systen '-/ill 
not proceed until the Inventory and '"•ioassay subsysten has 
becone operational. The Laboratory Survey and Airborne 
P.el ease/'.'as te Processed and Activity balance subsystens 
have not yet been defined. It is presune'- 
after acceptance of the Training subsysten, 



^a 1 ance su? 
■■ d this '.-.' i 



1 occur 



IG . ,'! I EMS Grants Mana^-enent 
'!e tv/or k 



P'^o mat i on 



The purpose of this project is to develop a 
nanagenent systen for the Progran Analysis ^rancb. 
This involves the creation o''^ a ""-rants "anarp^ent 
and creating a Retrieval and "eportint^ Systen. 
Ilanagcnent File v/ i 1 1 be created fron a subset r>^ 



Grants data available fron uPG \"?\ 

administrative data to be input by P'^o, 

Petrieval and Reporting Systen involves 
searches of Administrative Codes. 



■he 
t'^e ' 
^i le 



rran ts 

•■IE: S. 
^i le, 
"•rants 



h i e r r . r c ^ 



The 

CPl 



17 . Provide Svstens \nalvsis and i^ro '^r ann i nr: 



Su npo r t f o r 



ro^rans 



This project will involve a connlete analvs 
'lICP's present and anticipated future 



requ i renents . I t wi 1 1 
short range projects. 



I nc 



the 



n" 



' s y 

s nee 



the 

iten 

^ic 



83 



1. Financial i-anapenent 3y?;ten. 

2. Contract Tata Systen. 

3. Interactive On-line Project, Personnel 
and Financial Systens. 

k. Biblios:rapHic Citation Cyst en. 



18 . ".\'^ Central Li'irarv Circulation 
Svs ten 

This project for CRS involved the installation of a 
library control systen purchased fron the '.'niversity of 
South Carolina on the ■;i:V37n systen. 

All pron;ranning including the nodi Election r^ the '.'CC 
ProScans is conplete. The f^.U'l procedures have been 
established, the entire systen testnd ,^nd t^e systen 
documentation written. 

l;e are currently revie'./inp; a request for an 'In "^rocess 
i'.aterials On Order' systen. It Is requested t^^at this 
systen run interactively on the library's TP-ll/^o 
computer, but due to limitations of the current ??P-ll/h'^ 
configurations It may be necessary to develop this, 'In 
^rocess', system on the l;."-37n. 

in . The lational C 1 ea r i n "•'"qms e ^cr Can^cltv 
Qui Id I n^ and "unan Services Inte'^ration 
Prorrran 

The c 1 ear i np-hous e service is beinr nrovi'^e'^ by t^e ^spen 
Systems Corporation under a Dl-'i',' contract. In support o'' 
this effort the HM'^ developed systems for 1) the 
subscriber ^ile; and 2) the literature ^Ile. -or both 
systems, the necessary prosrrams ■^or collectin-^, updating, 
editinf^, printinr- and queryinr: of the files v.'ere provide^' 
and the entire system alon?r with documentation v/as turned 
over to the user during the ^iscal year. 



84 



The total systen should he an Interrate^ systen canohle of 
supplying data fron one subsystem to another in order to 
produce desired reports. 



'alance suhsysten has been 



in' 



Waste Processed and ^ctivity 
undef I ned as ye t . 

The analysis and pro,'Tranni n,<^ ^or the Inventory 
Bloassay subsysten has been connleted. '.,'e ^re presently 
av/aitlnf the data necessary to build the control file of 
individuals usini^ radionuclide materials be^^ore v.'e can 
proceed. Upon completion of this file, a parallel of the 
nev/ and old systems vy i 1 1 be run to isolate any 
deficiencies or programming bu-rs. it is estimated that 
the nev/ system v/ i 1 1 be supplemented in Sentember o^ I"'?''. 

A meeting v/as held June 1, 1073 to discuss the develonment 
of the Training subsystem. \ f ol lov.'-up meeting v.- i 1 1 be 
held June 20, 1078. The programming of the su'-^system '-/ill 




in . .'! I EMS Grants r'ana'^ement 
'!etv/ork 



no rma t i on 



The purpose of this project is to develop =■ rrants 
management system for the ?roF.ran \nalysis '"'^a^c^", "IE! 3, 
This involves the creation o'*^ a ''-rants "anare-^ent 



and creating a Retrieval and Reportinf^ System. 
Management File v/ i 1 1 be created from a subset o^ 



Grants data available from, 
administrative data to be input 
Retrieval and Reporting System 
searches of Administrative Codes. 



DRG 
by ?^::, 
i nvo 1 ves 



-ile, 

he 'rants 

the 'ii'ZrG 

RA: File an^ 

ME'^G.' The 

hierarchical 



17 . Provide Systems \nalvsi5 and Pro •^ ram" 
Sun port for 'ilPR Prc^rams 



This project will 
'ilDR's present 
requ i remen ts . I t v; i 1 1 
short range projects. 



involve a comnlete analvsis o^ the 
and anticipated future ^TP syste-i 



include the ^oll o'.,' 



n' 



s ncc I 1 c 



83 



Financial Tanafrenent Systen. 

Contract Data Systen. 

Interactive On-line I'roject, Personnel 
and Financial Systens. 

Bibliop:raphic Citation Cyst en. 



IR . "i'-^ Central Library Circulation 
Svs ten 

This project for CR5 involved the installation of 
library control systen purchasef^ fron the 'Jniversity of 
South Carolina on the ■;ilV37n systen. 

All pror^ranning includinfr the nod i -^ i cc t i on c^ the '.'CC 

pro^rans is conplete. The r-;U'i procedures have been 

established, the entire systen tested and t!~e systen 
documentation v/ritten. 

'./e are currently rev i e'..' i nr^ a request for an 'In ^rocess 
liaterlals On Crder' systen. It is requested t^^at this 
systen run interactively on the library's TP-ll/^O 
computer, but due to linitations of the current PCP-ll/'tl 
configurations it nay be necessary to develop this, 'in 
Process', systen on the IB."-37G. 



n . The iatinnal C 1 ea r i n'-^hous e '^or Can^'citv 
C u i 1 d i n ^ and ' ' u n a n Services I n t e '^ r a t i '"^ n 



The clearinp-house service is hein?- provide'' by t^e ^spen 
Systens Corporation under a CI'"',' contract. in support o-^ 
this effort the Ofl'^- developed syste^-^s ^or 1) t^e 
subscriber file; and 2) the literat^'re ""lie. -or both 
systens, the necessary programs ■f^or collectin'^, updating, 
edltin<^, printinr and queryinr: of the ■f^iles v/ere provide^' 
and the entire systen alono- with documentation v/as turned 
over to the user during the ^iscal year. 



84 



C . Sionedica] Connun i ca t i ons Applications 



Gelective :?is'^pnination o^ 
1 nfornat ion ( SP I ) 

Di'13 continued support of the current av/areness searcH for 

Chenical Biological Activities (C:\C). This service is 

still offered free of char?:e to all researchers at IIH and 

is run biweekly as tapes 3re received fron Chenical 

Abstracts Service in Colunbus, Ohio. Retrospective 

requests are beinr^ referred to the on-line service. 



TOXL l'!E, aval lable at 



:i 



CilD continued support of the current a-./areness search of 
Dlosciences infornation Systen ("lOGIS). Three tines a 
nonth tapes are received ^ron the "ioloric^^l abstracts 
Service and infornation is d i sseni na tc* to the "!!- 
connunity thru the sane vehicle as C""S. 

I'r . C. Gillespie of the ''-\'.' Library has been the orinary 
contact v/ith the "!'' researcher 'vishin" to search this 
data base; he subnits their profiles to riiB for current 
av/areness searching, and to 'L" for retrospective 
searching;. 



Conputer "esearc'^ and Tec^. nicue T eve 1 o f^'^rr t 



Incuirv and "^epor tin-" Sv^ ten (1'^. S) 

! P.S is a proprietarv pacl'.are developed '^y Sir^na "ata 
Corporation. Curin'^ FY7S D"'." continue"^ its user support 
of this systen. This includes na i n ta i n i n.'T the integrity 
of I Ro by testin?r nev; IRS releases ^n^ notil^yin-^ users of 
chancres to IRS, keeping- in'f^orned about lC'"'T hard'.-/are/ 
softv/are changes that nay affect the operation o^ IRS, and 
providin.q; IRS classes. 



Recursive "aero Actuated ''"enerator 
(RflAG Pro ; get) 

"■'AG is a prop;rann i n.f^ lan°:uafe use'^ ■^or the renerntion o^ 

other source pro;^r?nninr lann;uares and .'ata/Text strin-^^. 
The ori,Tinal R'i^G was inplenented in 1G7^ as a conoiler 

i.e., a progran that functioned in thr- nain'-^atcH Job 

strean of the '!!'' Conputer Center, readin,'=' pro'^rans 



85 



written in RM^G lan^ruafre and o^eneretlnp- spcci-^ic prorrans 



for users. Since its inpl enen tat i on, several very 
powerful additional arithnetic and lor^Ical 



bu i 1 t i nto it, but its 
compiler renain the sane. 



node o' 



f e-"' tu res 
access and use 



'.•; ere 
as a 



In FY73 a previously developed Symbolic Lo.'^ic Retrieval 
(SLR) systen was substantially completed. Extensive 
documentation v/as produced for this powerful. Interactive 
re t r i eva 1 sys ten. 



An RMAG22 encoded generator, reading 
specifications of the user's file structure, 
customized COBOL prop;ram which interactiv 



selected records from the user s file. 



f ree 

r^ene r 

e 1 y re 

interactively 



T! 



COBOL program permits its user to 
sentential calculus selection ^^ ci.olo, 
existential quantifiers, soph i s t i ca te^^ ari 
specification, the Rf!AG22 macro^enerator, and e 
macro libraries. Because n:',\022 has been made a p 
it, user's generated SLR is extensible at any time 
future without the need to then modi'^y t^e source p 



^orm 
a t e s a 
t r i eves 
ner a te^ 
utilize 
i m p 1 led 
t'^met I c 
X t e r n a 1 
art o-^ 
in the 
rorrram. 



Also In Fvy? the previously developed P.EPO'^TG'E 
extended to v;ork in conjunction with SLP. data ■^Iles. 



■■•/ a s 



This p 
high-1 
data f 
and u 
ref o ( — I 
f i les 
senera 
,?^enera 
f i 1 es 
be twee 
extend 



roduc 
evel 

lies. 

pdate 
at da 

(f I 1 
1, m 
ted 

i t 
n f I 
I ng P. 



t involved development of a soec 

prof^ranm i ng lant^ua.Te ^or routine o 

Pro.^rams can nov/ be <7;enerated (1) 



files usIn.T i<ey oriented transact 
ta files, and (3) to index textual 
e inversion). Files are assumed 
ultiple logical records per su^^ 
pro.trram can validate or error-chec!( 
reads, using conditions invo 
elds. 
E ^OnTG- ."i 



Th i s year ' s efforts 



to deal wi th SLP 



1 V i n 

cent 

event re 



ial 
pera t 

to 

I ons, 

'^at 

to h 

j e c t . 

data 

p- re 

ere^ 
cords 



pu r Dose 
Ions on 

create 

(2) to 

a ^rom 

ave, I n 

■^ny 

In t"^e 
1 at i ons 

around 



Autotab I I 



pec\i^?-e '^esigne^ to create 
Even non-pro "rammers can 



This Is an easy-to-use softv/are 

and update tables of numbers. 

use the system, which requires only basic infoi — lation 

(about rov/s, titles, column headings, initial values, and 

calculations needed) to produce attractive final reports. 

Related tables can be consolidated so that cbr.n.'-es in one 

table can be 'rippled' through all r~.<-^c.ra lo ^^ 



hers in the 



•roup 



86 



July 1, 1977 through September 30, 1978 

NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 

Summary of Branch Activities 1. PORT 

2. COMPUTER SYSTEMS LABORATORY 3. Alan M. Demmerle 



Chief 



I . SUMMARY 



Function 



The Computer Systems Laboratory (CSL) identifies and solves problems in areas 
of biomedical research and clinical care where real-time data collection, 
analysis, display and experiment control are required, where economic consid- 
erations favor a small computer or where proximity of the computing equipment 
to the work site is important to successful solution. 

The activities of CSL's electronic engineers and computer scientists center 
in these principal areas: computer applications in a clinical environment, 
computer applications in laboratory research, and consultation with research- 
ers in need of computer expertise. 

CSL's method of approaching each project varies as each problem presents 
unique challenges. Sometimes the objectives of a project are clearly defined. 
The investigator knows exactly what he wants automation to achieve. In this 
case the path to solution of the problem is conceptually straightforward, 
even though it is often time-consuming and technically difficult. In projects 
of this kind, CSL staff analyze the user's specific requirements, evaluate 
various alternative solutions with regard to technical merit, time to com- 
pletion, and cost. The hardware and software aspects of the system which best 
meet the user's needs are then specified. 

A different approach is required when an investigator believes his research 
might benefit from automation but does not know exactly how to take advantage 
of the possibilities. In projects of this kind, members of CSL become more 
involved with the research in order to discover how automation can be achieved 
and how research methods need to be altered to utilize this technology. 

When components that meet the user's needs are not commercially available, they 
are developed by the CSL engineering and programming staff. The system is 
refined to the point where it becomes a functioning, integrated part of the 
user's research. Since user's requirements often change as their research pro- 
gresses, collaboration between the users and CSL staff continues beyond the 
initial implementation of the system so that necessary adaptations can be made. 



87 



Scope of Work 

Over the past ten years, CSL has worked with nearly all of the Institutes on 
a wide variety of requirements. Major effort has been expended upon a few 
large projects that have required an investment of ten to twenty-five man- 
years. These projects, in addition to including initial analysis, design, 
development and installation phases, have frequently involved continuing 
collaboration with users in order to adapt systems to evolving research 
requirements. In fact, some of these projects have no distinct ending because 
of continuing modifications to accommodate new research protocols. 

Each year, CSL also undertakes a number of relatively small jobs, ranging in 
size from one man-week to a half man-year. These involve consulting with the 
intramural and extramural programs in the areas of automated data collection, 
display and analysis, data transmission, certain aspects of biomedical sensor 
design and all aspects of real-time computer system design. 

Intramural consultation often includes development of hardware and software 
systems; extramural consultation activities are primarily advisory. The con- 
sulting activities are valuable to us in that they help keep us aware of other 
activities in biomedical engineering, but require only a minimal commitment of 
manpower resources. 

Highlights of the Year's Activities 

During FY78, CSL has continued to utilize minicomputer and microcomputer based 
technology in support of clinical care and laboratory research programs of ^ 
widely varying scope and duration. Three projects which reflect this diversity 
of effort and suggest the impact of CSL contributions to the biomedical com- 
munity are the Computerized Radiotherapy Project, the NIAMMD Distributed Data 
Acquisition and Control Project and the Scintillation Counter Data Recorder 
Project. 

The Computerized Radiotherapy Project is primarily concerned with the treatment' 
planning aspects of radiation therapy. Treatment planning involves selection 
of the type, energy, size and orientation of therapy beams so as to insure 
adequate irradiation of the tumor volume, and requires accurate identification 
of the shape, size and depth of the tumor and surrounding normal radiation- 
sensitive tissue. 

In the past, the methods available for acquiring such information have been 
relatively primitive as evidenced by the use of solder wire techniques for 
external contour determination, the estimation of internal cross-section 
contours from conventional linear radiographs, and the assignment of densities 
based upon representative values obtainable from atlases of cross-sectional 
anatomy. 

Recent advances in ultrasound and computed tomography (CT) technology, however, 
have made it possible to differentiate between tissue that varies in attenua- 
tion characteristics by only 0.5 percent relative to water, thus providing 
a capability for attainment of accurate individual-specific internal and 
external contours. Moreover, the potential use of multiple slices suggests 



88 



that greater conformance of radiation dose to tumor volume can be achieved 
through the use of three-dimensional computation and display techniques. 

The goal of early phases of the project has generally been to achieve improved 
patient care through the integration of ultrasound and CT developments into 
existing computer supported treatment planning technology. Emphasis is now 
directed toward the development of innovative border detection and display 
algorithms, and toward the exploration of sophisticated new treatment planning 
techniques (three-dimensional planning, optimization, dynamic control, etc.). 
This will facilitate the assimilation of evolving ultrasound and CT technology, 
and also provide a basis for in-depth assessment of the impact of this tech- 
nology on radiation therapy. 

The two and a half years that have elapsed since project inception have wit- 
nessed the development of a sophisticated stand-alone system that combines a 
powerful display facility with state-of-the-art treatment planning capabilities, 
Configured around a Digital Equipment Corporation POP-11/70 computer, with 128K 
words of core memory and 40 million words of disc memory, the system maintains 
compatibility with EMI scans via 9-track magnetic tape, provides for scan dis- 
play and manipulation on a 32-level of gray Princeton 801 display terminal, 
facilitates digitization of contours via an Atomic Energy of Canada Ltd. (AECL) 
Grafpen, and supports life-size reproduction of scans and selected contours on 
a Versatec printer/plotter. Treatment planning is accomplished using TP-11, 
a radiation treatment software package developed by J. R. Cunningham and 
marketed by AECL. 

Within the context of a menu executive, programs have been developed that 
facilitate flexible interactive translation of EMI attenuation numbers to dis- 
play intensity levels, permit manipulation of scan displays (sector selection, 
magnification, zooming, etc.), computer designated areas and distances, and 
provide automatic contouring for limited anatomic regions. Substantial cur- 
rent activity is concentrated on developing edge detection techniques and 
criteria adequate for organ and tumor identification. Convolution, probabilis- 
tic texture analysis, and line following techniques predicated upon "a priori" 
anatomic knowledge, are being evaluated both individually and collectively. 

Additional on-going effort includes the development of a treatment planning 
optimization package utilizing linear programming techniques, and the integra- 
tion of ultrasound images into the PDP-11/70 treatment planning system. 
Scheduled near-future activities include a number of modifications to TP-11 
such as expanding the number of permissible contours so that inhomogeneity 
correction computations can fully utilize CT derived density data, software 
development for a color replacement to the Princeton 801 display terminal, 
and implementation of a data management system for tumor registry and sched- 
uling purposes. In the more distant future it is planned to apply three 
dimensional display techniques to tumor volume determination and to resume 
study of dynamically controlled radiation therapy. 



89 



The Computerized Radiotherapy Project is a collaborative project involving 
radiotherapists and physicists of the NCI, and engineers, system programmers 
and application programmers of CSL. To date, CSL has invested 12 man-years 
of effort in the project. In as much as the project is comprised of a series 
of research tasks of unpredictable outcome and consequently uncertain med- 
cal value, neither project duration nor the size of eventual manpower involve- 
ment can be confidently predicted at this time. 

The Computerized Radiotherapy Project illustrates a minicomputer solution to 
a number of diverse but nevertheless centralized requirements. In contrast, 
the Distributed Data Acquisition and Control Project utilizes minicomputer 
and microprocessor technology to handle the data acquisition and processing 
requirements of several instruments distributed among various laboratories 
of the. NIAMDD. Specifically, the Distributed Data Acquisition and Control 
Project is aimed at replacing an earlier Honeywell 515 computer system that 
has, for the past eight years, provided NIAMDD scientists with the ability to 
automatically collect data, display both raw data and computed values, imple- 
ment feedback loops and perform a range of computations including area deter- 
mination, smoothing, curve fitting, normalization of spectra, catenation of 
spectra, etc. 

The existing and replacement systems enjoy similar goals but feature quite 
different technical approaches. An important difference that distinguishes 
the system under development is the use of a microcomputer at each instrument 
to perform all instrument-dependent functions and handle the real-time require- 
ments. Data transfer from the distributed microcomputer to a centeral pro- 
cessor is accomplished through a communication microprocessor using standard 
communication protocols. A significant advantage of this system architecture 
is that the power of a shared central processor is retained while expensive 
specialized multi-user real-time system software development is exchanged 
for development of relatively simpler and lower cost microcomputer programs. 
Yet another benefit is that data acquisition, data storage, data transmission, 
and data processing capabilities can be readily upgraded, as required, through 
memory and peripheral expansion of the microcomputers. Finally, much of the 
software will be duplicated on each microcomputer thus providing for substan- 
tial standardization. 

As in the case of the Computerized Radiotherapy Project, the Distributed Data 
Acquisition and Control Project is a large long-term project that consequently 
imposes heavy manpower demands. The staff assembled to perform the engineer- 
ing design, equipment fabrication, system software development and applica- 
tions programming required, although drawn primarily from CSL receives sup- 
plemental NIAMDD support. To date, CSL has expended approximately five man- 
years of effort on development of this system. Extrapolating from experience 
gained with the existing Honeywell 516 system, NIAMDD research programs can 
be expected to impose a continuing need for CSL to modify and expand the sys- 
tem thus precluding establishment of any firm date for project completion. 



90 



Expectations of system performance are high and it is anticipated that this 
distributed processing system will provide a model for automation of other 
NIH laboratories. 

The Scintillation Counter Data Recorder Project is an excellent example of 
the significant role microprocessors and microcomputers have assumed in CSL 
development of laboratory instrumentation and data acquisition systems for 
the support of biomedical and clinical research. The relatively low-cost 
of microprocessor components and lower development effort which results from 
the use of microcomputers as compared to special -purpose logic enables us to 
undertake projects which previously would have exceeded our resources. More 
importantly, we have been able to significantly improve our turnaround time 
in responding to the requirements of investigators to the point that we can 
often react in periods measured in weeks rather than months. Specifically, 
the Scintillation Counter Data Recorder project involves the development of 
a microprocessor controlled instrument for logging data from liquid scintil- 
lation counters onto digital cassette tape. Scintillation counters are among 
the most widely used analytical instruments in biomedical research; there are 
approximately 300 of them at NIH. They are commonly used to detect and quantify 
radioactive tracers in biological specimens. The output device often supplied 
with a liquid scintillation counter is a teletypewriter or similar unit. 

A typical experiment may require the analysis of hundreds of samples. This 
will produce large volumes of data which must be reduced and analyzed before 
the experimental results can be interpreted. Reduction of the data requires 
extremely tedious calculations, and is often done on a computer. In the past, 
this required that the listing from the teletype be manually transcribed or 
that a paper tape be read into a computer in some way. Either of these is 
almost as tedious as doing the calculations by hand. 

The data recorder was developed with the thought of providing more convenient 
computer access to scintillation counter users, and improving reliability as 
compared to paper tape or manual transcription. The recorder produces an ANSI 
compatible digital cassette tape which can subsequently be replayed on any of 
a variety of commercially available terminals with cassette tape capability. 
Through use of a microprocessor as a control element, it has been possible to 
produce an instrument which can be easily adapted to a wide variety of scintil- 
lation counters with virtually no hardware changes. An additional benefit is 
the ability to perform some data editing and formatting functions within the 
data recorder so as to reduce the cost and time associated with data trans- 
mission. Finally, the microprocessor has allowed the per unit costs of these 
devices to be kept to a minimum due to reduction in parts count and develop- 
ment time. 

In addition to system development, consultation to NIH intramural and extra- 
mural programs has recently been gaining significant prominence among CSL 
activities. The nature of this consultation ranges from brief encounters in 



91 



which engineering advice related to the selection of a particular piece of 
equipment or a computer technique is given to in-depth programmatic assess- 
ments. A typical example is our association with the Carcinogenesis Testing 
(CGT) Program of the National Cancer Institute. NCI has entered into an 
interagency agreement with the National Center for lexicological Research 
(NCTR) of the Food and Drug Administration which will lead to the implementa- 
tion of various aspects of NCTR's Research Support System as a computer-based 
information and data management system for CGT. CSL's role is to assist NCI 
in the evaluation of various specialized hardware and software components of 
this system. In addition, we are advising NCTR on the development of micro- 
computer-based, programmable data acquisition terminals for use both by NCI 
contractors and NCTR. A request for proposals for these terminals has 
resulted from our collaboration with NCTR; proposals are expected to be 
received from prospective vendors in late suirmer, 1978. At that time, CSL 
staff will participate with NCTR and NCI personnel in the evaluation of pro- 
posals and the selection of a contractor. 



92 



II. ANNOTATED PROJECT AND ACTIVITIES LIST 

1 . Clinical Research, Patient Care Projects 

Computerized Radiation Therapy, NCI, RO : CSL, in collaboration with the 
Radiation Oncology Branch, NCI, has developed a computer system to utilize 
the detailed contour and density information available from computer assisted 
tomography to improve radiation treatment planning. Our system allows the 
radiotherapist to review scans of a tumor area at a video terminal, vary the 
gray-level spectrum to emphasize a particular structure, or zoom on an inter- 
esting feature. Contours, which may be computer derived or manually defined, 
are then processed by software algorithms procured from the Atomic Energy 
Commission Ltd. of Canada, to provide sophisticated individualized treatment 
plans. Current emphasis is on the development of suitable image enhancement 
and edge detection techniques necessary for accurate organ and tumor identi- 
fication. A data management capability, which would include development of 
a tumor registry and management of scheduling activities, is also planned. 

Cardiac Intensive Care Unit Patient Monitoring Computer System, NHLBI , SU : 
A real-time, on-line patient monitoring system was developed to serve the 
four bed postoperative recovery room of the NHLBI Surgery Branch. The system 
provides beat-to-beat analysis of the electrocardiogram and arterial pressure 
waveform, in addition to continuous monitoring of venous pressure, body 
temperature, urine output, and blood loss. Derived parameters are stored for 
72 hours in a disc memory and real-time data retrieval is performed from 
either bedside or nurses station display terminals. Additional features 
include a five-minute electrocardiogram memory, the computation of cardiac 
output by the dye dilution technique, and the automated detection of common 
ventricular arrhythmias. Currently, the arrhythmia analysis programs are 
being modified to allow the monitoring of patients with cardiac pacemakers, 
and programs for the computation of cardiac output by the thermal dilution 
technique are being completed. Future plans call for the automated generation 
of alarms based on variations in monitored parameters, the the connection of 
the operating rooms to the computer system in order to allow vital sign 
monitoring during the operative procedure. 

Medical Intensive Care Unit Patient Monitoring Computer System, CC, DCCM : A 
commercially developed computerized patient monitoring system is being pro- 
cured for the newly authorized Medical Intensive Care Unit. This system will 
provide for the periodic data logging of vital signs from the nine bed care 
unit. Vascular catheterization facilities within the unit will also be auto- 
mated to provide on-line collection of hemodynamic data as needed. Long-term 
data storage is provided by disc memories and on-line data retrieval is 
accomplished from numerous display terminals within the care unit. A mass 
spectrometer subsystem will be integrated into the commercial monitoring sys- 
tem in order to provide for the analysis of respiratory gases. In addition, 
urine scales will be added to each bed to allow the automated monitoring of 



93 



urine output and automated infusion pumps will eventually provide the capa- * 
bility for closed-loop therapy under computer control. Patient monitoring 
algorithms will be developed to allow the correlation of respiratory gas | 
composition data and urine output values with vital signs obtained by the i 
commercial monitoring system. New techniques for the automated evaluation 
and treatment of critically ill patients will be investigated. Programming f« 
modifications to the existing commercial patient monitoring system will 1 
ultimately customize the system to NIH applications. 

Cardiac Scintillation Probe, NHLBI, IR C8, and CC, NM : CSL is continuing 
development of the microprocessor based cardiac scintillation probe. The 
purpose of this instrument is to allow continuous, noninvasive monitoring 
of left ventricular function. The microcomputer and electronics are mounted 
on a cart which can be easily positioned at the bedside. The system is pro- 
grammed to automatically (1) acquire scintillation data from the probe, (2) 
organize the data into a composite left ventricular time-activity curve, (3) 
permit the curve to be constructed only from heartbeats havinq a user-selected 
range of lengths, (4) correct the curve for left ventricular background, (5) 
calculate ejection fraction, and (6) display the curve and all calculated 
values of ejection fraction as a function of time from the onset of monitoring 
Clinical results, reported to the American Heart Association in November 1977, i 
have demonstrated the utility of this system. We are currently working toward J 
extending the capabilities of the system for use in the Cardiac Catheteriza- I 
tion Laboratory to produce real-time pressure-volume curves in addition to 
monitoring LV volume. The system will be used there for drug intervention 
studies. 



Cardiovascular Signal Analysis, NHLBI, IR SU : This project was terminated 
during the FY78. 

S-T Segment Mapping, NHLBI, IR CB : Work on this project was suspended during 
FY78 and the status is therefore unchanged from the previous reporting period. 

Automated ECG Processing, CC and NHLBI, IR CB : The Clinical Center's ECG Ser- 
vice requested assistance in specifying and acquiring an on-line ECG system to 
assist staff physicians in analyzing and reporting current ECG's and comparing 
serial ECG's. CSL has reviewed available literature, commercial systems, and 
DOD's Tri-Service Medical Information System procurement effort. A draft 
request for proposals containing functional specifications has been prepared 
for a minicomputer-based ECG system. The system is to perform the ECG anal- 
ysis functions mentioned above and also have provision for further development 
particularly with regard to integrating it with the Clinical Center's Medical 
Information System. The RFP is currently;awaiting action by the appropriate 
Contracting Office; funding of the system is expected to be accomplished with 
the FY79 money. 



i 



94 



2. Laboratory Investigation Projects 

Flow Microfluorometry, Cell Sorter (FMF) NCI, I, LP and NIAID, LMI : The 
demand for FMF instruments has increased here at the NIH and in the research 
community at large during this reporting period. Here at the NIH, the demand 
for data collection and analysis exceeded the capability of the PDP-n/40 
which was shared by both the FACS II FMF (Dr. Wunderlicht I, NCI) and the LASL 
FMF (Dr. Herman LP, NCI). A computer dedicated to each FMF was proposed by 
CSL, DCRT, and a major portion of the effort expended during this reporting 
period was aimed at providing an essentially identical dedicated system for 
each FMF. A POP-11/34 was purchased and the existing dual system software 
was altered and optimized to support a single FMF instrument. Both FMF instru- 
ments now have dedicated computers and the emphasis has returned to improving 
existing and implementing new support software. Primarily, improvements to the 
software will considerably reduce the time spent by researchers at the computer 
console while using data display and analysis programs. This will be accom- 
plished by allowing preselection of parameters that will automatically be 
applied to one or many data files without operator intervention. The capabil- 
ities provided under a new release of the operating system (RT-11 3-B) will be 
required to support the larger application programs and therefore existing 
programs will be adapted to run under it. 




A requirement for the separation of cells based on more sophisticated combina- 
tions of the present measured cell parameters began a new aspect of CSL sup- 
port for the FMF instruments. The crucial factor in any implementation of a 
more sophisticated separation scheme is the time required to make the sort 
decision after the FMF cell measurements are made. The direction in which to 
sort one cell must be determined before the next cell's arrival into the deci- 
sion process. The temporal spacing of the cells in the FMF instrument affect 
the maximum decision time that can be taken for sorting. This time interval 
spacing between cells was thought to be some random distribution, but no pub- 
lished data was available to confirm this. A hardware, cell to cell, time 
interval measurement system was designed and fabricated by CSL and added to 
the FACS-II - PDP-11/40 system. With this addition, it was possible to quan- 
tify the distribution of the time intervals between cells for various average 
cell flow rates and then, project the number of cells that would be lost as a 
function of the average cell flow rate and a fixed decision time. Another use 
for this measurement is being explored within DCRT. Through statistical anal- 
ysis of the time interval distributions of various types of cells and various 
types of cell preparations it may be possible to quantify some biological char- 
acteristics of the cells such as the cells' affinity for one another. 



95 



is- 
emented 



Distributed Data Acquisition and Control System, NIAMDD : A prototype D 
tributed Laboratory Data Acquisition and Control System is being 1mplemcMi,cu 
for NIAMDD, in Building 2, as an eventual replacement to the Laboratory Com- [j 
puter System developed here in CSL several years ago and which has now become*' 
overloaded. The new system will consist of a network of remote microprocessor; 
connected in a star configuration through a comnunicatlons microprocessor to * 
the existing Honeywell 516 (H-516) computer. Laboratory data will be col- 
lected locally in real-time and transmitted at a convenient time to the com- 
munications processor using DOCMP, a standard communications protocol. 

Significant progress occurred during the past reporting period, particularly 
in procuring and fabricating those links essential for the first laboratory 
node and link to the existing Honeywell 516. A prototype computer (LSI-11) 
chassis was fabricated with a larger back plane (6X9 slots) than available 
in commercial units. Five additional units based on the prototype were sub- 
sequently built under contract. Sufficient computer modules (central process- 
ing unit, memory, input/out) have been purchased for three functional com- 
puters and depot maintenance (1 spare of each module). The communication pro- 
cessor and its link to the H-516 are in place and software implementing the 
protocol for the transfer of files is being developed and tested for this linl 



Display and menu routines have been prograrrmed for the laboratory (LSI-ll) 
computer. The effort involved in developing the hardware and software for th 
communication processor and much of the laboratory software such as display 
and menu routines are shared and will require minimal changes as other labora 
tory computers are added. An interface between the C-118 spectrophotometer 
and the laboratory computer has been designed and is being fabricated. Real- 
time data acquisition and control software is being developed to support the 
C-118. This Interface and support software are dependent on the requirements 
of the C-llS and each new laboratory installation will have similar instrumen 
dependent requirements. The technique to be used to implement the serial link 
between the conmuni cation processor and laboratory computer is being evaluate 
and a decision on the most time expedient route to implementing a subset of 
DDCMP (Digital Data Communications Message Protocol) will soon be made. 



k 

1 



I 



Mass Spectrometer, NIMH, LCS : This is a collaborative project with Dr. Marke 
NIMH, that is designed to automate ion focusing and data acquisition from an 
LKB 9000 mass spectrometer. Ion focusing is accomplished using a PDP-11/10 ^ 
computer to control the accelerating voltage of the spectrometer. Although II 
data acquisition and control are essential, the compelling reason for develop 
ing this system was to allow a user to create and use his own data manipula- 
tion routines, a feature not present on commercially available mass spectro- |: 
meter data systems. During this reporting period, all hardware and software •" 
required for trained personnel to begin using the Selected Ion Recording Sys- 
tem (SIRS) for chemical analysis in the laboratory environment was completed. 
An operations manual has been prepared to explain SIRS to laboratory personnel 

In April 78 the SIRS hardware was moved from the development area in Building 
12A to Dr. Markey's lab in Building 10, Room 3N325. The system is in daily 
use, and we are adding Improvements based on their laboratory experience. 



96 



Improvements already completed include; 1) a hardware addition to the Phoenix 
ADC/PDP-ll interface to reformat the ADC data directly into the computer's 
double precision integer format, freeing the CPU for more efficient real-time 
operation, 2) software enhancements to allow the automatic acquisition and 
processing of data from a series of samples without reinitialization between 
samples, 3) provision for a real-time display of acquired data during the 
acquisition process. 

Considerable production time has been lost because of multiple failures in the 
LDK-9000 mass spectrometer. Preliminary versions of software and hardware 
developed for SIRS have been requested by and supplied to FMC Corp., Princeton, 
N.J., and Karolinska Institute, Huddinge, Sweden. 

Scanning Electron Microscope (SEM), NIAMDD, LCP and NCI, LP : CSL participated 
with Dr. William Hagins (LCP NIAMDD) and Dr. William Banfield (LP, NCI) in 
specifying additions to an ETEC Autoscan SEM to permit its use for quantita- 
tive x-ray analysis of biological samples. The instrumentation requirement 
was to retain the SEM as a microscope with 200A resolution while providing 
the capability to analyze x-rays generated by the electron beam in the sample 
by both Energy-Dispersive (EDS) and Wavelength-Dispersive (WDS) Spectrometry 
simultaneously. Both EDS and WDS require a dedicated computer for acquisition 
and data processing and display. 

There were several commercial systems available or under development that would 
satisfy the instrumentation requirement; the task was to identify the system 
that offered a cost-effective solution with good expectations for a successful 
technical integration with the existing SEM. CSL is preparing a recommendation 
for the EDS and WDS instruments, after several visits to sites using SEM's with 
these features and numerous inquiries to potential suppliers. 

Microanalysis Facility, DRS, BEIB : BEIB is undertaking the establishment of a 
microanalysis facility which will conduct electron energy-loss spectrometry 
studies with investigators of several institutes at NIH and NIMH. These 
studies will involve computer acquisition and processing of spectral data, 
investigation of techniques for computer control of the electron microscope, 
and investigation of image processing and pattern recognition algorithms. 

In February 1978, CSL's assistance in the computer system design, procurement, 
installation, and hardware and software development was requested. After dis- 
cussions with the researchers involved with this project, CSL recommended the 
purchase of a PDP-n/5Q computer system in May 1978, and initiated the purchase 
of such a system at a cost of $99,000, $40,000 to be contributed by NINCDS and 
the balance by CSL. 

Delivery of this computer system and a Kevex Model 7000 Analytical Spectro- 
meter is expected by October 1978. One of the first goals will be to inter- 
face this instrument to the computer system and adapt software to operate it. 

Microcomputer-Based Satellite Controller, NIMH, LBEB : This year, we began 
the task of extending the capabilities of a laboratory computer system^ 
developed several years ago by the addition of a microcomputer satellite. 
The DEC LSI-11 microcomputer is being configured as a versatile programmable 

97 



multichannel pulse generator. The microcomputer will replace an existing ^ 
system of manually controlled timers, wave-shapers , and pulse generators 
which control various aspects of evoked response experiments. Software 
running in the host PDP-11/40 computer will use information supplied by the 
user to compute parameters which determine the various timing relationships 
of the output pulses. At the time of execution, these parameters are trans- 
ferred to the satellite where the microcomputer will generate the specified 
pulses on the chosen channels. In contrast to some of our other LSI-11 pro- 
jects, the satellite requires no console, no operating system, and no auxil- fi 
iary storage. Through this experiment we are expanding our host-satellite |[ 
cormiunication capabilities and are developing techniques for configuring and 
programming PROM-based DEC LSI-11 systems which should prove valuable in futur 
appl ications. 

NIEHS Computer Facility, NIEHS : The Biometry Branch of NIEHS is responsible 
for providing computer facilities for the Institute. The data processing 
needs include scientific and statistical computation, simulation, and labora- 
tory data processing. Since 1978, these needs have been met with a DEC GT-44 
Computer System which was purchased and installed on the advice of CSL. CSL 
has since that time provided support in the form of technical advice and con- 
sultation services. i 

Faced with an anticipated further increase in demand placed upon their sat- |] 
urated system, the Biometry Branch requested CSL to make recommendations for 
expansion of their computer facility. In March 1978, CSL reviewed the cur- ' 
rent and anticipated future data processing requirements of the NIEHS and 
recormiended 1) the purchase of a DEC PDP-11/70 computer system to augment the 
present GT-44 laboratory data processing system, and 2) investigation of 
methods to upgrade the telecommunications service to the DCRT Central Computeil 
Facility. *'' 



I 



Based on these recommendations, NIEHS is purchasing the suggested computer sy| 
tern. CSL has begun a survey of available telecommunication multiplexors and | 
concentrators which might be used to improve telecormiunications service, and 
plans to present the findings to NIEHS and CCB for possible implementation. 
CSL will also continue to provide support for the Biometry Branch computer 
system. 

Pulsed NMR Spectroscope, NCI, DCT : The development of a microprocessor Ij 
based acquisition and control system for use in conjunction with a Bruker 
Pulsed Nuclear Magnetic Resonance Spectrometer was started in early 1978. 
This unit, when completed, will replace the manual timing unit normally used 1 
with the spectrometer. The user will be -able to pre-program a series of *' 
complex pulse sequences and have them generated under microprocessor control 
while experimental data is simultaneously acquired via an analog-to-digital | 
converter. This method will not only relieve the experimenter of the tedium | 
of setting a series of switches for each run but will also allow a new 
flexibility in the types of sequences that can be generated. * 

Bioassay Information System, NCI, DCCP, and FDA, NCTR : The National Cancer 
Institute has entered into an interagency agreement with the National Center 
for Toxicological Research which will lead to the implementation of various I 

I 
98 f 



aspects of NCTR's Research Support System as a computer-based information 
and data management system for NCI's Bioassay Program. CSL is serving as a 
consultant to NCI in evaluating various specialized hardware and software com- 
ponents of this system, and is providing consultation to NCTR on specifica- 
tions for the development of microcomputer-based, programmable data acquisi- 
tion terminals. A request for proposals for these terminals is complete; 
proposals are expected from prospective vendors in late summer, 1978. CSL 
will participate in the evaluation of proposals. 

Laboratory of Immunology, NIAID : CSL is expanding the capability of a labora- 
tory data acquisition system which we developed several years ago. The pur- 
pose is to accommodate increased use of the system by providing for the simul- 
taneous acquisition, storage, and processing of instrument data from multiple 
instruments, as well as supporting interactive terminal processing. By 
September of 1978, the system will be completed, documented, and operational 
for use as described above for a Beckman amino acid analyzer, a Guilford 
spectrophotometer, and three fraction collectors. Provision is made to record 
the data on disk, annotate it with appropriate identifying information, anal- 
yze data from previous runs, and simultaneously and independently execute 
unrelated computer programs such that none of these functions will in any way 
affect or endanger the others. Provision has also been made for transmission 
of data to the DEC 10 for more sophisticated analysis. The transmission 
technique is described separately in this report as the Computer Communi- 
cations (CLINK) project. 

CSL Data Recorder (Previously reported as Scintillation Counter Data Logger) : 
The CSL Data Recorder was designed for the unattended recording of data from 
liquid scintillation counters onto digital magnetic cassette tape. The 
recorder is connected between a scintillation counter and the counter's tele- 
type, and in no way affects the operation of either the counter or its tele- 
type. At the completion of an experiment, the tape may be removed from the 
recorder and played back to a computer for analysis using a commercially 
available cassette tape data terminal. Three Data Recorders, one for NIAID 
in Phoenix, Arizona, and two for NICHD in Building 10, have been successfully 
in use since mid-1977. Eighteen additional units are now being fabricated. 
These will be used in laboratories in NCI, NIDR, NICHD, and NIA, and should be 
installed and operational by late summer 1978. 

Potentiometric Titration Controller, NHLBI, IR LC : A microprocessor-based 
system for measuring and controlling the electrical potential of an aqueous 
solution while monitoring optical density has been developed in conjunction 
with the Lab of Cell Biology (LCB), NHLBI. The degree of oxidation or 
reduction can be determined spectrophotometrical ly and controlled by pulses 
of electric current. The ability to control solution potential also allows 
complete optical spectra to be taken at fixed redox potentials. This past 
year, the system was used to study the respiratory chains of bacterial and 
mammalian systems. In particular, it was found that there are more redox 
components in E . Co 1 i respiratory membranes than have been previously 
recognized. A new multi-channel spectrometer used in conjunction with the 
existing microcomputer will allow the investigator to take complete optical 
spectra in 2.5 milliseconds compared to the present time of 30 to 40 seconds, 
a factor considered of great significance due to the extreme volatility of 
the experimental conditions. 

99 



Intramural Laboratory Support : Support for intramu 
through the development of specialized computer sys 
and control systems has long occupied a significant 
The current year represents a continuation of this 
this category require the investment of substantial 
are therefore reported separately. Those which can 
ranging in duration from a few hours to a few weeks 
separate projects. A number of the tasks are illus 
applications in which microcomputers may be used, 
number of these projects are included. 



ral research laboratories 
tems and data acquisition 
portion of CSL resources 
trend. Many projects in 
manpower and resources and 
be accommodated by effort 
have not been reported as 
trative of the variety of 
Brief descriptions of a 



A. Dr. C. B. Anfinson of NIAMD has requested that we develop a peak detectinc 
device for use with a high pressure liquid chromatograph. The unit is to 
detect peaks, and, on the basis of that detection, control the advance of a 
fraction collector so that separated material from the chromatographic column 
may be collected. The problem is complicated by the fact that the peaks are 
incompletely separated and therefore often strongly overlap rather than returr 
ing to baseline. A microcomputer will be used for performing the arithmetic 
operations required by the slope detection algorithms used to solve this pro- 
blem. Work on this project was begun in mid-June, 1978 with the expectation 
of completing it by late summer, 1978. 

B. An interface designed for NIDR provides the investigator the capability tc 
transmit data from a Nicolet VA-500B spectrum analyzer to a PDP-11/70 compute) 
The data is simultaneously recorded on cassette tape via a Texas Instruments 
733 ASR terminal as a backup should the PDP-11/70 not be available. In addi- 
tion to the data transfer capabilities, the interface provides high-speed 
(1200 baud) communications between the PDP-11/70 and the TI terminal. The 
CSL unit incorporates a commercially available single board microcomputer as 
well as circuitry and software developed by CSL. The unit has been in use 
since January 1978. 

This microcomputer data recorder for NIDR will log data from a spectrophoto- 
meter onto cassette tape. The recorder will record from one to thirteen 
readings per specimen upon comnand from a front panel switch of the recorder 
The recorder will also record control and status information which will be 
used when the cassette tape is played back to a PDP-11/70 computer. This 
project should be completed by September of 1978. 

C. A microcomputer controlled data logger was designed and built for NICHD 
to replace an existing punched paper tape data acquisition unit. The recorder 
accepts an analog signal from a gel scanner, converts this signal to digital 
form, displays the digital data and records it on digital cassette tape. The 
analog signal can be sampled at ten different rates from 1 sample/second up 
to 10 samples/second. These and other characteristics of the device can be 
altered by simply revising the microprocessor program. This system has been 
in use since May 1978. 

3. Program Management and Administration : I^lo Projects 



« 



I 



100 



i 

I 



4. Biomedical Communications Project 

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 helped coordinate the 1977 
Conference at the Thoraxcentrum, Rotterdam, The Netherlands, and edit the 
Conference Proceedings. We are also involved in the planning of the 1978 Con- 
ference which is to take place at Stanford University. 

5. Computer Research and Technique Development 

Computer Communications, DCRT, CCB : It is often necessary to be able to trans- 
fer information between laboratory minicomputers and a large computer system 
such as the DECsystem-10 in order to exploit the large system's data analysis 
and software development facilities. To meet this need, a software package 
called "CLINK" (Communications LINK) has been developed jointly by CSL and CCB. 
CLINK enables communications between DEC PDP-11 's (the most commonly used 
minicomputer at NIH) and the DECsystem-10. 

CLINK provides the following capabilities: 

1. Terminals physically connected to a PDP-11 which is, in turn, con- 
nected to the DECsystem-10 can operate as "virtual terminals"; i.e., the 
terminal appears to be a normal terminal connected directly to the DEC- 
system-10 and can be used to perform all the usual time-sharing functions. 

2. Data files can be transferred from the PDP-11 to the DECsystem-10. 

3. Data files can be transferred from the DECsystem-10 to the PDP-ll . 

An experimental version of CLINK was implemented and extensively tested in FY77 
The first production version (CLINK Version 2) was stabilized and made available 
for distribution in November 1977. CLINK was extended to run under RT-11 V3 
S/J and F/B operating systems on the PDP-11 in January, 1978. The RSX-llM 
operating system version was extended to run under RSX-llD V5.2 by Lilly 
Research Laboratories; this software was made available to DCRT for distri- 
bution with the original CLINK software in May, 1978. Currently, CLINK is in 
operation on over 15 PDP-11 systems on campus and has been distributed to at 
least 40 sites outside NIH. 



101 



CO 



-o 

O) 

> 
t— c 

o 

Q r- 
•r- 

c 



i«i i.i i<i i^ i«i 

O LO LD C O 



CM CNJ V£> • r— I — 



V 

.o 
E 

•4-> 

o. 

0) 
00 

I 





^—^ 




l/l 




S_ 


+-> 


fO 


s- 


<D 


o 


>- 


M- 




M- 


c 


UJ 


fO 




s: 




— ' 



CM CM C\J CVl 

r- r- .— r- O O 

•tj- n I— I— 



. tn <:s- m UD CM 1,0 






. m CM 1^ 



3 



o 

UJ 

•-3 

o 

q: 
o 



>- 
CD 



■u 




u 


^ 


OJ 


<D 


•"-5T3I 





ro 


s_ 


0) 


Q. 


— 1 



N ^4 O O M 

■M +-) a o •(-> 

a)cuajj=x:a)'D.c 

>,>,>,U Ur— I— o 
OOOOOO^CyOQ-COO 



M^4NOOOOOOOOO 

r— ^~ t— "f"" T" 'f^ '^^ *^ '^ '^ T"" *^" O 

333XXXXXXXXXS- 

^ ^ g- qt ni (1) (11 f|i fll fl) f|l flJ4-> 
(J U CJi — I — 1 — I — I — I — I — I — I — (/) 

t/)ooooi:i.a.Q-Q.CL.Q-Q.a_Q_o 



UJ 
Q. 
X 



Q. 

•a: 
tj 

o 
z 



a: 

o 



Ll- 
O 

_I 
UJ 















Q 






































t— 1 


Q 




































< 

































2: 






>— » 


■z. 




a. 


























2: 






Z 


< 




1 








a: 

1— 



















»» 






oa 


■Z. 




** 















— 1 

























H-H 








z 























ca 


_J 


*» 














f« 


















03 ^ 









5 








i. 




00 


















00 






<» 


-M 




oa 




-M 




Q 






*• 












•» 




Ci 


1-^ 


(/I 








C 




U. 






t-H 












ca a: 




H-l 




>, 




a. 















CD 












1— 1 






#1 


00 









C_) 




oa 






_1 

















«^ 


»_i 






_i 














"^ 








Qi ZD 




cc: - 




1— 1 





r— 








oB 




a. 






2: 








t/1 


s 


»-i^ )— • 




ca 


Z 







A 























* 





ca 




! 




s_ 









c 











•» 








t— 1 *« 





- _i 




^ 


#* 


■i-> 




CIl 









Q 






S_ 








C_J — 


Q 


>— ' zz 




2: 


S_ 


c 




2: 














a; 




0) 




z ca 




C2 S 


'■O 




OJ 







<c 


ca 


+J 




1— •> 






t— 




(J 




1 


#* 


_J 


CJ 


■o 


-l-> 


(_) 




t— 


^M 


• ^- 




C_5 « 






r— 




3 




•» "■" 


t-J 


3Z " 




c 


s_ 






^ 


UJ 


I/l 




ca c_) 












OJ 




>,2: 


CJ 


^ 1/1 


a; 


« 


c 


c<5 






ca 


• ^ 




■z. 


t— 1 




U 




s_ 




Q. 






1— 1 




t/1 






#« 




3_ 




»* 


< 




4-J 


-0 


0) 




fO » 


M 


«• l/l 









c 


00 


^ 


A 


^r 




t»i #* 


p— « 




C 


s_ 


M- 




S_ +J 


-*-> 


0) >i 


^ 





^^ 










00 


CJ 




= 


■z. 










C 




0) -r- 


•r- 


j2 .— 


li-H 




r^ 




_l 


0" 


Oi < 




Z Ol 









CL 







^ c 


C 


fO 


CQ 


«t 


Ol 


4-> 







Q 






■u 


f^ 






Q.C_J 




1— r3 


ID 


s_ c 


1 


cno 


•r- 


<• 


i/i 




03 




- (/) 


>- 




c 


3 










Q- < 


"T" 


C 




l/l 


^~ 





#> 


■I-) 


>T >^ >- 


o- 







I/O 


>~ 




C 0) 


OJ 




■Z. 


•r— 


•* 


•^ 


s: 


s- 


>> 


03 


4J 


a. 00 







•^ 




CT 




s- 


s- 


C r- 




t/1 


>i 


3 







•M 


Q 


•r- 


"0 


^— 




+-> 


>- 





(U 


•r- (0 


« 


« 


«« 


un 


S- 


^-' 


z 


•r- 


• ^ 




^— 


u c 







t: 


u' 


r— 




-t-J 


C_) 


•r- C 


cn 


Ol 


■l-> 


CJ 




2: 


r— 


■0 


•r— 


on 


c 




u 








■4-) 


(d 




4J CD 


c 


<_! 


gj 


< 


*t 




•^ 


0) 


u 


•-- 


3 




4-1 


4-> 


■T— 


•r- 


•I- OJ 


OJ 


rT3 -1- 


•1— 









;_ 


^ 


U 


(/I 


03 


S_ 4-! 


^ 


u 


•^ 


03 


-o 


1— 


■0 > 


> 


1— OO 


Q. 


S_ 





(T3 


OJ 


Q 


03 


03 


U. 


j-> 03 


^ 


OJ 


1— 


U 


u 




fO •<- 


.^ 


^~ 


Q.Q. 


^ 


4-J 


4-1 


C 


U. 


ca 




CJ ~ 


1-^ 


"O 







03 


4J 


Q; 1/1 


tn 


•1- i. 


03 







rO 


3^ 


4-J 






s- 


OJ u 




u 


a 


ja 





(J 


c: 


c 


+J n3 


2: 


e: 


3 


Q 





t/1 


u 


0) 


Q. 


14- 





•^ 


03 




OJ 


-a OJ 


Ol 


C -— 














OJ 


• ^- 


0) 


4-> 


00 M- 








u 


_1 


C 


•<-: 


0) 4-i 


-kJ 


•>- 3 


■!-> 


UJ 


4I 


•3 


s- 




l/l 


-i-j 


3 


C 




OJ 


^ 




•^ 





Ni C 


c 





r~ 







OJ 


4J 


UJ 


>> 


3 


^ 


■"^ 1— t 


5-c: 


OJ 


^— 




i- 


•r- ^^ 


».-t 


00 1/1 


l! 


-3 


s- 


-l-J 


U 






3_ 


— 


2: 


^ 




*— 


03 


c/1 


Cl. 


s_ 




na 


g 


OJ 


CJ 


3 


OJ 


CT 


03 


= 





2: > 





rO 





U 


U 




OJ u 


r^ 


u > 


CD 


^ 




ja 




X 


3 





c_; 


03 


♦J 


4-i 


'r— 


3 


OJ 




■M ra 


n3 


"3 


OJ 


ro 


£ 




Jo' 


■^ 


« 







-C l/l 


03 


fO 


4-> 


^ 


4-> 




3 •<- 


U 




oo 


E= 




S 'il 




C 








00 


OJ 1/1 


U 





C 


03 


3 




CL-O 


•^ 


■5 TD 







3 


_l -u 


i/i 


c 


U 


U 


^^ 


l/l 03 







OJ 


U 


Ci. 




= s- 


-a 


^ ^ 


I— 


4-> 





l/l 


o) 


ra 


U 


'J 


UJ 


r— 


ja 


^^ 


-.-> 


4-> 


S 




iT3 


O) 


rO rO 


1 


3 




•^ 


rt3 


u 


• ^- 


•^ 


»— 1 


3 T- 


fO 


00 





C 







c_J 


s: 


c_) 


OO 


< 


u. 


Q 


S 


00 


2: 


2: 


z: 


a_ ca 


1 


(_) 


a. 


1— ' 






102 



III. Publications and Presentations List 



Plexico, P., Songco, D., and Ostrow, H.: Microcomputers in Biomedical and 
Clinical Applications, National Bureau of Standards--IEEE Conference on 
Microcomputer Based Instrumentation, June 1978. 

Miller, M. , Powell, J., Sharrow, S., and Schultz, A., Jr.: Rapid Data 
Collection, Analysis and Graphics for Flow Microfluorometry Instrumentation. 
Review of Scientific Instruments , Bookhaven, National Laboratory, Upton, 
Long Island, N. Y., August 1978. 

Ader, R. , Lepley, A., and Songco, D. : Utilization of a Microprocessor in a 
Pulsed NMR Spectrometer. Journal of Magnetic Resonance 29 , January 1978, 
pp. 105-111. 

Hendler, R. , Songco, D. , and Clem, T. : Automated Electrodic Potentiometry 
System. Analytical Chemistry , Vol. 49, No. 13, November 1977, pp. 1908-1913. 

Bacharach, S., Green M. , Ostrow, H. , Redwood, D., and Johnston, G.: ECG- 
Gated Scintillation Probe Measurement of Left Ventricular Function, Journal 
of Nuclear Medicine , Vol. 18, January 1977, pp. 79-84. 

Bacharach, S., Green, M. , Borer, J., Douglas, M. , Ostrow, H., and Johnston, 
G. : Real-Time Scintigraphic Cineangiography. Presented: Computers in 
Cardiology Annual Meeting, October 1976, St. Louis, MO. Published in Pro- 
ceedings. Available: IEEE Computer Society , Long Beach, CA 98403, IEEE 
Catalog No. 76CH1150-1C. 

Bacharach, S., Green, M. , Borer, J., Douglas, M. , Ostrow, H., and Johnston, 
G.: A Real-Time System for Multi-image Cardiac Studies. Computer Methods , 
David E. Lieberman, Editor, Book Chapter, December 1977, C. V. Mosby Co., 
St. Louis MO. 

Bacharach, S., Green, M. , Borer, J., Douglas, M. , Ostrow, H., and Johnston, 
G. : A Real-Time System for Multi -Image Gated Cardiac Studies. Journal of 
Nuclear Medicine, Vol. 18, January 1977, pp. 79-84. 



103 



October 



1, 1977 through September 30, 1978 



NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



Summary of Branch Activities 
2 . Computer Center Branch 

I. SUMMARY 
Function 



1 . DCRT 

3 . Joseph D. Naughton 

Chief 



The Computer Center Branch was established to provid* 
efficient and effective central computational services foi 
the NIH on a f ee-f or-service basis under the NIH Service and 
Supply Fund. The objective of the Computer Center is t< 
incorporate proven advances in computer technology an( 
methodology into a reliable, accessible, economic Computer" 
Utility to support NIH scientific and managerial program.s. Thi 
need to preserve the integrity of data and insure th( 
uninterrupted availability of computer services is its' 
primary concern. Its challenge is to achieve new improvements; 
while maintaining the quality of service to science and itsl 
management at NIH. 



The main concern 
I use the comput 
the system, "How 
or improvements a 
of the Utility 
rapidly changing 



Activities 



The N 

a p 

commu 

mult i 

DECsy 

f acil 

telep 

in 

throu 

(syst 

comp 1 

( syst 

or 

compl 



IH Compu 
owerf ul 
nication 
-compute 
stem-1 , 
ities 1 
hone lin 
research 
ghout N 
ems h 
emented 
ems sof 
acquired 
ex requi 



ter 



r s 
ea 

ink 
es t 

IH. 
ardw 
by 
twar 

jr- 
1 

reme 



of user 
er to b 

fast 
re ess 
and t 
s tate-o 



Center 
network 
cili tie 
u b s y s t e 
ch hav 

thes e 
o over 
laborat 

The 
are) 
a 
e) des 
rom 
n t s of 



s of the Computer Utility i 
etter accomplish my job?" a 
can I get results?" Continu 
ential to maintain the r 
o provide services consis 
f-the-art of com.puting. 



s "How can] 
nd once usingi 
al additions 
espons iveness| 
tent with the! 



I 



des ii 



o 
s. Th 

ms , 
ing un 

two s 
1 ,500 
ories 
comput 

are 
very 
igned 
t h e r 
the NI 



f 

e nu 
the 

ique 

ubs-y 

term 

a 

ing 

c 

com 

and 

sour 

H re 



imp 1 
mode 
cleu s 
IBM 
capab 
stems 
inals 
nd a 
and c 
ontrol 
plex 

impl 
ces a 
search 



emen 
rn 

is c 
Syst 
ilit 

and 
of V 
dmi n 
ommu 
led, 
set 
emen 
nd 

pro 



ts , a 

comp 

cmposed 

em" 37 

ies . C 

conne 

arious 

istrati 

nicatio 

bal 

of comp 

ted by 

adapted 

gram. 



nd 

uter 
of 



ommu 

ct 

type 

ve 

ns 

ance 

uter 
th 
to 



opera 
s 

two la 
and 
nicati 
them 
3 loca 
offi 
equipm 
d 

progr 
e Cen 

meet 



tes 
and 
rge 
the 
ons 
by 
ted 
ces 
ent 
and 
ams 
ter 
the 



1 
I 
I 
1 



104 



Services and facilities 
standard programming Ian 
SAIL, Assembly Language), 
(IMS), and a large and va 
users with terminals there 
WYLBUR, ISO, and t 
DECsystem-10 , which fa 
output of jobs and pe 
(using FORTRAN, BASIC, 
The Center also provide 
the output of printed 
paper and microfiche 
dimensional pictures on 
"sketch pads" for ad 
those examining macro mole 



provided by the 
guages (e.g., FO 

a data base/data 
ried library of ut 

are also interact 
he time-sharing 
cilitate creation 
rmit direct int 
CPS, APL, MLAB an 
s a variety 
material or grap 
Programs to 

cathode ray t 
vanced graphic 
cular structures a 



Center include 

RTRAN, COBOL, PL/I, 

management system 

ility programs. For 

ive systems such as 

services of the 
, submission and 
eractive computing 
d other languages), 
of facilities for 
hies information on 
show two or three 
ube displays and 

projects such as 
re also available. 



In addition, the Center provides extensive training, 
technical documentation, information services and assistance 
with problem diagnosis for its users throughout NIH. It also 
offers limited sources for data entry and recurring program 
maintenace and recurring production application. 

Scope of Work 

The Computer Center operates the NIH Computer Utility 24 hours 
a day providing services to over 6,000 users, including research 
scientists and program manager from all NIH Bureaus, Institutes, 
Divisions and Offices. 

IBM System 370 Facility: During FY78, the workload handled 
by this facility grew 18% over the previous year. ( Paqe 113 
is a chart of workload growth since 1967.) At year's end, the 

This number 

the utili ty , 

Interactive 

risen to over 



Center was handling over 400,000 tasks per month, 
includes 9,600 mainstream jobs a day processed by 
with 80% of them completed within one hour, 
timesharing sessions via remote terminals have 
4 ,500 a day. 



DECsystem-10 Facility: The level of activity on the 
DECsystem-10 has shown steady growth over the past year. The 
number of user sessions now exceeds 12,000 per month, while 
the number of registered users has climbed to over 
1,200. The number of connect hours has grown to over 12,500 
per month while total runtime of the processors has 
remained steady at approximately 320 hours per month. 
Unfortunately, system response time has degraded 
noticeably due to the lack of sufficient CPU capacity 
for handling the workload. 

Manpower: Only a small part (about 12%) of the 

operating expenses of the Computer Center are personnel 

costs, but those 142 people are the absolutely essential 

ingredient in the Center's success in providing cost effective 



105 



computing at NIH. About two thirds of them have responsibi 
at a daily operating level to insure that all work is compl 
in a timely and effective fashion within the Computer Util 
The remaining third are responsible for maintaining 
integrity of the complex hardware and software syst 
educating and assisting users in the use of the Utility and 
design and implementation of new computer based services 
facilities. While the overall workload of the Comp 
Center increased 18% in FY 1978, the number of 
available to do the work remained essentially constant. 

Highlights of this years activities 



pe( 



FY78 
Cent 
serv 
ma jo 
made 
f aci 
that 



was 
er p 
ices ; 
r ope 
in 
litie 
tot 



the 
rovi 

add 
rati 

gr 
s, 
aled 



1 It 
ded 
ed ne 
ng s 
aphic 
and 

ove 



h c 

MIH 
w f ac 
ystem 
se 
users 
r 355& 



onsecutive year that the Compute 
with planned improvements to existin 
ilities and reduced rates to users. Tw 
s were upgraded, improvements wer 
rvices, printing and communication's 
received the benefit of rate reduction 



Upgrading of Systems 
Communications Upgrade . 



mode 

and 

tele 

the 

betw 

user 

are 

comb 

the 

spee 

port 

can 

Cent 

prev 

to 

mode 

inte 

six 

July 

avai 

the 



ms f 

the 
proce 
new m 
een 
s. T 

the 
ined 

same 
d an 
s on 
handl 
er . 
ious 

serv 
ms w 
rrupt 
week 
and 
labil 
total 



or 

DEC 
ssin 
odem 
the 
he n 

lat 

wit 

nu 

d t 

any 
e th 

Thi 
syst 
ice 
ere 
ion 

pe 
is n 
ity, 

sys 



all 
10 r 
g ca 
s inv 

Comp 
ew mo 
est 
h new 
mber 
ype o 

inte 

e ful 

s imp 

em wh 

eac 

repl 
in se 
riod . 
ow in 
hi 
tern. 



the 
epr es 
pabil 
olved 
uter 
dems 
in c 
Auto 
for 
f the 
ract i 
1 ran 
rovem 
ich r 
h sp 
aced 
r V i c e 
The 
full 
gher 



The 
intera 
ented 
ities 
eigh 

Cente 
operat 
ommuni 
speed 

a te 

parti 
ve sy 
ge of 
ent e 
equire 
ecif ic 

at t 
. Use 

ent ir 
-time 
reli 



comp lete 
ctive ter 

a ma jo 
for the U 
t months 
r , the 
ing at sp 
cations 
software, 
rminal s 
cular ter 
stem ope 
termi nals 
liminated 
d segmen 

speed an 
he Compu 
rs modems 
e convers 
prodxict i v 
ability 



change 
minal s 
r imp 
tility. 

of 
telepho 
eeds of 
technol 

allows 
ys tem 
minal t 
rates 

suppor 
the 
tation 
d termi 
ter Ut 

were i 
ion was 
e use p 
and gre 



to new 
y stems 
r oveme 

The 
clos e 
ne CO 

up t 

ogy. 

all u 
regard 
hey ar 
in the 
ted by 
inef f i 

of 
nal ty 
ility 
nstall 

compl 
r ovidi 
at er e 



high 
, WYLBU 
nt in 
convers 
coope 
moany a 
o" 1200 

The 
sers t 
les s o 
e using 

same w 

the Co 
ciency 
line 
pe. Ov 

wi thou 
ed duri 
eted in 
ng inc 
f f ic ien 



J 

speed 
R, TSa 
thd 
ion to 
)n] 

\i 

IS 

liall 
the| 



ration 
nd th 
bau 
modems 
o d 
f 

. All 
ay and 
mpute 
of the 
groups, 
er 60 
t any 
ng a 
early] 
r e a s e d 
cy for 
( 



Computer to Computer Communication link 

Computer to computer communications became a reality during 

the year as CLINK, a program that allows many 

laboratory minicomputers to interchange information with 

the Central Computer Utility became a production service,' 



106 



Files can be sent to or from the Central Facility to 
specialized laboratory computers throughout NIH. A powerful 

communications protocol insures that all of the transmitted 

data is received completely and accurately. CLINK also 

allows terminals connected to these laboratory computer to be 

used as ordinary terminals with either the DECsystem-10 or 

with WYLBUR and TSO, on the IBM System 370. 

The CLINK system is a significant milestone in the support 
of scientific and laboratory computing. It recognizes the fact 
that specialized minicomputers and microprocessors can 
profitably team with the Central Utility's processor power 
and data storage facility to exploit the advantage of both. 

New Surface Display System 

The hardware for a high level display system was installed 
and special graphic system software was designed and 
implemented. Beginning October 1 this new facility will go 
into regular service providing graphic support for the 
modelling of macroraolecular structure. The system provides 
highly interactive manipulation of molecular structure using a 
graphic tablet and pen. One form of modelling involves 
the manipulation of the structure to fit clouds of electron 
density which are derived from c ry stal logr aphic data. The 
data bank developed for the AMSOM atlas can also be used as 
a source of macromo lecular structure. An algorithm was 
developed for the representation of the Van der Waals surfaces 
of atoms as shaded and colored spheres. This algorithm 
makes it possible to display whole macromo lecules or the 
active site for extremely large ma cromo lecules. The system has 
been used to provide new insights into the structure of the 
active sites of serine proteases and the structure of 
collegen. The vector portion of the display system gives 
information about distances and angles between atoms as well 
as structural patterns while the surface representation 
display gives information about the packing of atoms. 
The results of modelling can be obtained from the system 
using a camera station which provides the possibility of 
instant filming, high quality plate filming as well as movie 
filming. Several movies on different macromo lecular subjects 
are in production and planning. The system is being used by a 
broad population of biochemists- and protein cry stallogr aphers 
who come to the system for working visits which last from a long 
day to a week. The system will be used in the next year 
to model the interaction between proteins and their substrates 
as well as to study the structure of the environment in 
which the proteins function. The possibility of doing these 
studies as well as the probability of success is largely 
determined by the hardware, algorithm and program development 
just finished. 



107 



11 



Services and Facilities 

DECsystem-1 Improvements . DECsystem-10 users benefitted fro 
an increase in the speed of the mercury communicatio 
link, installed last year, which connects the DECsystem-10 
and the System 370. The transmission speed of the link wa 
increased from 4800 baud to 9600 baud. This link has prove 
so popular that an additional 9600 baud line was installed 
between the two sytems to avoid queueing problems at either; 
end of the link. 



APL users with Tektronix CRTs were able to util 
graphic capability, the APL-GRAPH-11 package. The 
interactive graphics language with English-like 
was designed to make APL graphics functions readable 
to use. APL's concise and easy manipulatoin of 
matrix data makes it an ideal language for graphic ap 

Computer conferencing was introduced to the NIH commu 
the year. This experiment in teleprocessing a 
participants to record their views and feelings 
topic. Conferees can also read the views of 
comment or argue for or against their points. The 
of computer conferencing is that the participants 
to gather at one place, at the same time, 
participate to whatever degree they wish and at the 
is most convenient for the individual. 



A new digitizing system for inputing graphic data was offere^ 
to DECsystem-10 users. The system uses a CRT termina^ 
with a graphics tablet to transform graphic data to a list of x! 
and y coordinate values. This data can then be easily analyze( 
using MLAB or any other convenient data analysis program. 




370 Improvements . A new Computer Out 
device was installed at the Compute 
DatagraphiX 4561 is 75% faster than the 
it replaced, and produces microfiche 
and with improved readability. The 
of microfiche over conventional print 
speed, rapid information retrieval, low 
of copies, and ease of handling, stora 

The installation of a second hig 
subsystem enabled the Computer Center 
smaller standard size paper for pr 
improving operational efficiency. The 
(14-7/8 X 11) was replaced with a new 
8-1/2. The smaller print size permit 
information (132 columns) to be prin 
paper while maintaining readibility. 
provides an effective 20" increase in 



put M 
r Cen 

COM 
mor 

prim 
ed out 
cost , 
ge and 

h s 
to o 

inted 
old 

stand 
s the 
ted 
The us 

the p 



icrof i 
ter . 

equip 
e e 
ary 

put ar 
unlimi 

distr 

peed 
f fer u 
outp 
stand 
ard si 
same 
on th 
e of s 
r int 



Im ( 

The 
ment 
f f ici e 
advant 
e thr 
ted nu 
ibutio 

prin 
sers a 
ut w 
and p 
ze of 
amount 
e sma 
mall p 
speed 



COM), 
ne 

tha 

ntly 

ag 

up 

mber 

n 






! 

ting' 

nevi 

hiloi 

aper 

ller» 



aper 
o 



1 



108 



a 3800. 
SPOUT and 
recognized 
increasing 
bytes and 



Other paper size are still available to users through 

JES facilities. The improved system performance 

from the transition to MVS last year permitted 

the region size limit for all job classes to 1000K 

^j ^^^ the elimination of class H. This change allows large 

programs to run in all job classes without resorting to the use 
of overlay structures or segmentation. The use of new large 
region size did not have any adverse impact on either 
turnaround time or system performance. 

Terminal Improvements 

After a three year interuption for procurement the 
Computer Center is again offering terminals to users. A 
new CRT, the NIH7000, has been selected and is being installed 
in laboratories and offices throughout NIH. The terminal 
features a customized keyboard, a larger buffer memory in 
which up to 200 lines may be stored, and numerous special 
features for use with NIH developed software systems. 
Twenty-eight lines of 80 characters are displayed with 
extremely sharp characters. The new terminal operates at 
1200 baud and has a microprocessor controlling all the 
regular and special functions. 

A procurement to select two types of printing terminals 
to replace the IBM model 27^1 terminal which has served as the 
NIH standard for many years was also completed. New 
hardcopy terminals that operate at 1200 and 300 baud have been 
selected and will be available to users during the coming year. 



Rate Reductions . 
370 offered 
DECsystem-10 reduc 
of 25^ became 
DECsystera-10 use 
for online disk st 
System 370 bat 
reduction in May 
CPU utilization, 
has provided c 
year. As a result 
received more comp 



Both the DECS 
rate reduc 
tion of 6% an 

effective i 

rs also real 

orage was r 

ch service 

which rang 

The constant 
ost reducti 
, users of t 
uting than th 



ystem- 1 
tions 
d an IBM 
n Move 
i z e d s a V 
educed 
users b 
ed from 
tuning 
ons f 
he NIH 
ey budge 



and the 
during the 
System 37 
mber for 
ings of 28% 

to $ .1 
enefited fr 
87o to 287o 
of the Comp 
r the 1 It 

Computer U 
ted for. 



IBM System 

year. A 

reduction 

all users. 
when the rate 

per block, 
om a second 
depending on 
uter Utility 
h consecutive 
tility again 



User Education and Assistance 



A crucial factor in the effectiveness of any Computer Center 
is the assistance it provides its users in learning how to use 
its facilities and services effectively and in overcoming 
any problems which they may encounter. For the tenth successive 
year the Computer Center offered a full program 
Training Courses and Seminars. Four, three month terms were 
offered for the first time this year. 



of Co mp uter 
'ere 
Audio visual and 



109 



computer 
year and 
training at 



assisted courses also 
wel 
all levels 



made their debut during the 
were well received. Seventy-three courses provided 

over 1500 students this year 
These courses covered the entire range of general purpos 
programming languages, special services and facilities, 
programming aides, job control languages, use o' 
statistical and mathematical packages (BMD, SAS, SPSS, PSTAT 
MLAB, MODELAIDE), query and reporting facilities' (IRS) 
and programming tools (RMAG 21, PDP-11 tools). Seven seminar 
taught by DCRT mathematicians and scientists covered sucH 
diverse topics as computer netowrk synchronization, data 
structures and file organization, design of assemblers, 
curve fitting, simulation using GASP II, linear programmin' 
and Fourier and Laplace transforms. Unfortunately, some 65 



applicants 
because of 



had to 
the lack 



be denied 
of personnel 



admission to training courses 



Liaison (PAL) unit received 
or telephone requests fo 
Programmer T 



■ s fort 
'roublej 

I its: 
andj 

inicaJ 

i 

li 



The Programming Assistance and 
more than 17,000 personal 
assistance and analyzed some 3,400 written 

Reports (PTRs). Equally important, the Center helped its 
users avoid many problems through the publication 
distribution of technical documentation. Its Tech 
Information Office used automated files to insure that all 
users received new or revised copies of technical manuals 
relative to their work. The Center celebrated the lOt 
anniversary of the publication of INTERFACE, its technical' 
notes covering new services, facilities, procedures, 
diagnostics and programming hints in June. Seven 
issues containing some 400 pages of technical information 
for NIH computer users were published during the year. 

Overall the Technical Information Office stocked and 
distributed some 70,000 pieces of documentation to over 3,600 
active computer users, including technical reference manuals 
and guides and the Computer Center Users Guide. Both the 
Users Guide and the Timesharing Guide as well as many 
technical reference manuals were revised extensively throughout 
the year to keep abreast of changing technology and 
procedures . 

Future Plans 

Unfortunately, most of the activities which were identified 
as "future plans" in last year's report appear under the samel 
title again this year. This lack of progress can be attributed) 
to two primary causes; 1. Delays and burdens introduced by 
procurement procedures; and 2. Inadequate manpower to 
keep abreast of increasing workloads. As a result, the 
majority of the significant accomplishments planned for the past 
year appear again as "future olans." 



no 



J 



The Center is still faced with severe shortages in 
several critical areas, online data storage space, physical 
space, data security, manpower, air conditioning and 
electrical power. Although contracts have been let to improve 
the physical plant to provide the necessary utility services and 
to improve security of the Computer Utility and the data it 
handles no real progress has been made. 



As 

of 

move 

the 

spac 

This 

DECS 

all 

firs 

to 

syst 

comp 

requ 



nothe 

Build 

of s 

seco 

e if 

wil 
ystem 
user 
t flo 
post 
em wi 
uter 
ired 



r yea 
ing 
om.e o 
nd f 
adequ 

I pr 
-10 e 
-or ie 
or of 

of fi 

II be 
area 
under 



r came 
12B sti 
f the C 
loor o 
ate hou 
ovide 
quipmen 
nted f 

Buildi 
ce box 

instal 
from pu 

the Pr 



to an 
11 re 
omput 
f Bu 
sing 
addi t 
t cur 
acili 
ng 12 
es ) 
led t 
blic 
ivacy 



end 
ma in 
er C 
ildi 
can 
iona 
rent 
ties 
A. ■ 
and 
o pr 
area 

Act 



, the 
s a dr 
enter 
ng 12 
be f ou 
1 com 
ly in 

to 
Lock ed 

a 
ovide 
s and 
of 19 



long 
earn . 
staff 
to be 
nd f o 
puter 
Build 
be 1 

u ser 
compu 

bett 
to pr 
74. 



await 

If ever 

into i 

convert 

r the re 

space t 

ing 12 a 

ocated t 

ou tput 
terized 
er sepa 
otect pe 



ed completion 
completed, the 
t will enable 
ed into machine 
maining staff, 
o relocate the 
nd will enable 
ogether on the 
boxes (similar 
access control 
ration of the 
rsonal data as 



New WYLBU 
most po 
completio 
delayed 
possibi li 
the maj 
undergoin 
service 
users . 
the desig 
user doc 
when comp 
coming 
commands , 
and many 
years . 



R, a 
pula 
n th 

for 
ty t 
or 

g e 
bef o 
Cons 
n of 
umen 
lete 
year 

mac 
ot 



comp 
r 

is ye 

th 

hat i 

part 
xtens 
re i 
idera 

trai 
tatio 
d. T 
pr 
ro ca 
her n 



lete 

inte 

ar. 

e 

t wi 

s 

ive 

t is 

ble 

ning 

n t 

he s 

ovid 

pabi 

ew f 



ly 

ract 

Con 

past 

11 a 

have 

int 

off 

ef 

cou 

o in 

y ste 

ing 

liti 

eatu 



rede 

ive 

side 

tw 
ctua 

be 
er na 
ered 
fort 
rses 
sure 
m wi 

th 
es , 
r es 



sign 

sy 

ring 

y 

iiy 

en 

1 t 
as 

is 
and 
a s 

11 b 

em 
doc 

whic 



ed ver 
stem, 

that 
ears t 
happen 
complet 
esting 

a sta 
cu rre 

the pr 
mooth t 
e avail 

with 
ument 
h have 



s ion 
is a 

it 
here 

this 
ed a 

to i 
ndard 
ntly 
epara 
rans i 
able 
more 
prep a 
been 



of 

gain 
has 

is 
ye 
nd 

nsur 
se 

bei 
tion 
tion 
to 

pow 
rati 
requ 



the 

Pl 
air 

the 
ar . 
are 
e a 
rvi c 
ng i 

of 

to 
user 
erf u 
on 
este 



Cent 
anned 
eady 

dist 

Most 
cur re 

reli 
e to 
nveste 
ext en 
New WY 
s in 
1 edi 
f acili 
d for 



er 's 

for 
been 
inct 
of 
ntly 
able 

NIH 
d in 
si ve 
LBUR 

the 
ting 
ties 
many 



As NIH research programs expand to include more and more 
research and patient care data and management information, 
data storage capacity continues to be the single more serious 
problem facing the Computer Center. 

The tape library has overflowed its space many times over 
and online disk space is a constant problem. The future 
should see the introduction of a Mass Storage facility. Such a 
unit could store extemely large amounts of data in a relatively 
small space while reducing the manpower burden for computer 
operators and reducing overall storage costs. Data stored 
on a mass storage system is transf errable in a matter of 
seconds to online disk units where it is directly 



111 



acces 

sole 

comp 1 

FY79. 

the 

perso 

avail 

suf f e 

space 

Plans 

past 

delay 

incre 

highe 

no i 

procu 

temoo 



sible by the main processors. A system study an 

source justification document for such a system was 
eted this year and will be forwarded to the GSA in earl-. 
The only thing which can be assured at this time is thai 
request will be delayed and will require additional 
nnel effort to defend. Unless a mass storage facility ia 
able in the immediate future, many NIH programs will 
r severely due to the unavailability of online storagj 



to upgrade online disks have been well defined for thJ 
two years but procurement regulations continue to 
installation. New units would enable the Center t 
ase its online storage capacity almost four-fold, provide 
r data transfer rate and provide improved reliability at 
ncrease in operating cost. Because of more critica 
rement issues, this activity has been suspende 
rarily . 



Recent industry improvements in computer systenj 
technology, together with new NIH program requirements 
for increased computational capacity /power and new functions 
logically require that major components of the NIH Comp 
Utility be replaced by new, more cost effective units 
addition to new faster processors and disk drives with 
larger storage capacity for both the System 370 and the| 

and 
long 
and 



'Uteri 
Inl 



DECsystem-10 
improvements 
comimunication 



facilities , 

are already 

controllers 



terminals 



all types, 



a mass storage 
overdue for tape 
batch remote job 



system,! 
drives , 
entryi 



Unfor 

every 

years 

abrea 

progr 

under 

total 

DEC) 

activ 

staff 

p r e V e 

of th 

and 

procu 

outco 

The 

compu 

requi 

strai 

right 

it w 

neces 

resea 



tunat 

imp 
has 
St o 
am . 
take 
com 
of 
ity 

of 
nting 
e NIH 
oppo 
remen 
me. 



ely , 
rove 
pr ev 

f 

The 

the 

peti 

hard 

will 

th 

mu 

pro 

rtun 

t(s) 



c 

ment 
ente 
the 
Comp 
lar 
tive 
ware 

CO 

e C 
ch s 
gram 
ity 
can 



onstant 
that 
d upgrad 

needs 

uter Cen 

gest si 

reproc 

, softw 

nsume t 

omputer 

ignif ica 

The c 

of p 

not pos 



content 
NIH has 
ing the 

of the 
ter is n 
ngle task 
urement of 
are and 
he major 

Center d 
nt accom.pl 
ost in man 
r e p a r i ng 
sibly be r 



ion 

ide 

prese 

NIH 

ow f 

it 

i ts 
maint 

port 
ur ing 
i shme 
power 

and 
ecove 



wi 
ntif 
nt 

b 

orce 
has 
en ti 
enan 
ion 

th 
nts 
, do 

CO 

red 



th 
ied 
equi 
iome 
d b 
ever 
re c 
ce 

of t 
e n 

in 
liar 
nduc 
r ega 



GSA 
in the 
pment 
di cal 
y the 

attem 
omplex 
servi c 
he pro 
ext y 

di rec 
s and 
ting 
rdless 



over 

past few! 

to keep' 

research 

GSA to 

pted--the 

(IBM and 

e. This 

f essional 

ear thus 

t support 

lost time 

such a 

of the 



providing responsive, cost effective 

in response to the complex and dynamic 

the NIH research program depends on fast 



success of 

ter support 

rements of 

ght forward procurement of the proper equipment at the 

time. If 
ill be impos 
sary to prov 
rch and mana 



the discussions with GSA of the past continue, 
isible to predict if or when the equipment 
'ide proper support to the NIH biomedical 
igement orogram will be available. 
112 




113 



ISMITHSOfJU'. CCi;'.C£ I'.f OR'^ATIOri EXCHA.VGt 
PHOJcCT NUMEjcR (Do NOT use this space) 



U.S. DlPartmei.t of 

HEALTH, EOUCATIOfi, AhJ WCLFARE 
PUtiLIC HEM TH CERVICE 
NOTICE Of 
INTRAKURAL RESEARCH PROJECT 



PERIOD COVERED 

October 1, 1977 - September 30, 1978 



PROJECT NUMBER 



-t 



Z01 CT 00031-03 CCf 



TITLE Cf PROJECT (50 characters or less) 



Atlas of Macromoleoular Structure on Microfiche (AMSOM) 



f.-AMES, LAECRAT:f,Y a;.C institute AFFILIATIO^o, a;.0 titles of PR|;,CIPAL investigators AfO ALL OTtxR 

PROFtSSio.'.AL PEftsc.';;;EL E.';sAGED cr, the project 

P.I. Richard J. Feldmann, CCB, DCRT, Computer Specialist 
Other: R. L. Tate, Ph.D., CSL, DCRT, Senior Staff Fellow 



cco?erati:;c U:.IT3 (i: a 



Department of Chemistry, Brookhaven National Laboratory 



LAS/ SRA.'.CH 

Computer Center Branch 
SECTic;; ~ ~~~ 



ir.STiTUTE a;;o lccatic. ' ' ' — - 
DCRT, NIH, Bldg., 12A, Room 3009, Bethesda, MD 2001 a 

TOTAL MA.'.^'EA.T:: [ .-nOFESS 1 Oi.AL: 

0.5 .05 



C.iECK APPROPRIATE 50x(ES) 
U [i) HU.v.a;. SJ5JECTS 



OTHER: 







n (b) HU.VA.'; TISSUES 



n (c) NEITHER 



n (at) .Vi^ 



'2 (a2) K.TERVIrAE 



SJ.'"'ARY Ci 

The 
was 

cry s 

surv 

for 

repr 

shad 

carr 

this 

colo 

well 

diss 

way 

with 



p-HS-cOiC 



- "Ct.n (200 words or less - underline kcy-oras) 

Atlas of Macromoleoular Structure 
published in March of 1977 is bein 
tallographers for teaching a 
eilance and structure validation i 
a supplement to the atlas. -A d 
esenting the Van 6er Waals su 
ed colors has been implemented. 
ied out to determine the best medi 

type of information. The curre 
r microfiche using the already de 

as side by side stereo presenta 
emination using color xerography i 
of avoiding the cost of color sep 
printing. 



114 



On Micr 
g used 
nd re 
s being 
i splay 
rface o 
Invest 
a for t 
nt poss 
V e 1 p e d 
tion on 
s being 
aration 



of iche 

by bio 

search. 

done a 

system 

f a mac 

igation 

he diss 

ibiliti 

AMSOM 

paper . 

i nvest 

norma 1 



lid 



(AMSOM) wn 
chemists anc 

Literatur 

s preparatiorf 

capable of 

romolecule ir 

s are bein^ 

emination of 

es are stereq 

vehicle ast 

Incremental 

igated as a 

ly associate 



.'^MITHSOMa:. GCU'.Ct i:.FOR'.'AtlON E;CHAf!G£ 
PHCJcCT NaVSEfi (Oo NOT uie this space) 



U.S. DEPARTMEtiT OF 
HEALTH, EDUCATION, A;iO WELFARE 
PUBLIC HEALTH CERVICE 
HOT ICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00036-02 



CCB 



PERIOD COVERED 



October 1 1Q77 tn September ^0. \<17& 

TITLE Of PROJECT (60 characters or less) 

Graphic System for the Display of Biochemical and Biomedical 
Objects 



fiAS'ES, LABORATGF.Y A.'.C INSTITUTE AFFILIATIONS, ANO TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTMER 
PROFESSIONAL PERSONNEL ENCAGED ON THE PROJECT 

P.I. RichardJ. Feldmann, CCB, DCRT, Computer Specialist 

Others: Thomas K. Porter, CCB, DCRT, Computer Specialist 

Charles R. T. Bacon, CCB, DCRT, Computer Specialist 



CCOPERATI.'.G UNITS (if ar.y) 

None 



LAS; BRANCH 

Computer Center Branch 



SE 



..omp 
cTicn 



NSTITUTE A.ND LOGATICN 

_ DCRT. NIH. Bldg.. 1 2A . Room 3009. Bethesda. MP 20014 



TOTAL MANYfAn:-. 

2.5 



FSCFESSIONAL: 



2.5 



OTHER: 







^- '- ) HUMAN SUBJECTS 



Ci-ECK APPROrRlATE BOX(ES) 



□ (b) HU.VAN TISSUES 



□ (c) NEITHER 



□ (al) I.MNO'<S [] (a2) INTER Vhj>S 



SUMMARY OF WCR.-- (200 ,.ortc or 

Implementation of 
manipulation of 
done at various 
software. The c 
mainframe compute 
focus of hardwa 
operating charac 
achieve maximum 
which was obtai 
and modified. Sy 
as general sys 
algorithm for t 
spheres served 



lei3 - underline kcy-ordi) 

graphics system for 
biochemical and biome 
levels of hardware, 
ommunications link be 
r and the PDP-11/70 
re and systems pr 
teristics of the lin 
data transfer. The g 
ned from the system ve 
stem software for the 
tem infrastructure ha 
he representation of at 
as the starting point 



the representat 
dical objects i 

system software 
tween the DECs 
graphics computer 
ogramming work . 
k are being tu 
raphics system 



ndor has 

f r ame 
ve been 
oms as 



been imp 
buffer 

develop 
shaded , 



for the develo 



ion and 
s being 
and user 
ystem- 1 

was the 
The 
ned to 
software 
lemented 
as well 
ed. An 

CO lor ed 
pment of 



115 



high level user programs for manipulation of macromo lecular 

structure. High level program development has centered about 

system supplied FORTRAN as no other language seemed to 
supply the needed operating characteristics. 

A one micron densitometer is being added to the system to enhance 
the range of data which can be processed by the system. 



The system 
scintograms , 



has been used to manipulate data from nuclear 



T16 



October 1, 1977 through September 30, 1978 

NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



Summary of Branch Activities 1. PORT 

2. LABORATORY OF APPLIED STUDIES 3. Eugene K. Harris 

Chief 



I. SUMMARY 



Function 

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

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 dynamic flow processes, transport of 
substrate to tissues and macromolecular interactions; 

in collaboration with clinicians, to develop and apply 
mathematical or statistical theory and computer programming to 
improve diagnosis of disease and assessment of treatment; 

. to engage in independent research in applied mathematics, 
statistics and computer systems necessary to provide a sound 
theoretical basis for collaborative studies. 

LAS consists of two sections in addition to the Chief's office: 
1) Applied Mathematics Section (J. E. Fletcher, Ph.D., Chief), 
including specialists in applied mathematics, numerical analysis, 

computer science, mathematical physiology; 2) Medical Applications 
Section (J. J. Bailey, M.D., Chief), including physicians, electronic 

(biomedical) engineers, computer systems analysts. The Chief, LAS, 
is a biostatistician. 

Scope of Work 

The Laboratory of Applied Studies works on projects in basic and 
clinical biomedical science. In large part these are carried 
out in collaboration with other groups at NIH and elsewhere in the 
U.S. and abroad. The collaborating investigators this year included: 
. biochemists and pharmacologists - at U. of Iowa, Northwestern 

University, and other universities, working on models for receptors 
of drugs or other ligands, and the kinetics of enzymes in membranes, 



117 



physiologists and chemical engineers - in the U.S., U.K., 
Scandinavia, and Germany studying the transport of substrate 
within the microcirculation, 

clinicians - at NIH in the cardiology and pulmonary branches 
of MKLBI, 

clinical chemists and pathologists - at NIH (Clin. Path. Dept., 
CC) , in California, elsewhere in the U.S., and in the U.K., 
engaged in the collection of and study of reference values in 
laboratory medicine, 

electrocardiologists, biomedical engineers, and health adminis- 
trators - in the U.S., Canada and Europe concerned with the 
evaluation of computer-based ECG interpretation. 

During FY 78 LAS staff members participated in various teaching 
and consulting (or advisory) activities: 

J. Fletcher continues to serve as Chairman of the Mathematics 
Department, FAES, and taught advanced calculus in the NIK graduate 
s chool ; 

J. Bailey was a member of several site-visiting teams for NKLBI 
concerned with computer analysis of exercise ECG's, biophysical 
s tudies in pediatric cardiology and studies in cardiovascular 
pathology; 

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 has also served as an invited 
speaker at meetings of American and European pathologists concerned 
with applications of statistical theory to problems facing clinical 
laboratories. 

Highlights of LAS Program Activities during FY 78 

1 . Hardware analog model implemented on MAC-16 to simulate cone 
cell responses : E. Pottala, R. Covacci (LAS), S. Vallerga, R. Momann 
(Lab. of Neurophysiology, NINCDS) : The most satisfactory/ mathematical 
model currently available to describe retinal cone cell activity is 
the Ray lor-Hodg kin-Lamb model involving several nonlinear differential 
equations to describe sodium-potassium fluxes. Such models are too 
time-consuming and expensive for routine use on a digital computer; 
however, the analog hardware representa-tion enables extremely rapid, 
inexpensive simulation of cone cell response to a wide range of 
graded light stimuli, including estimation of parameters and graphic 
display of response cur/e. It is hoped that this model system will 
be expanded to include a small network of cone, rod and other cells 
involved in the retina's processing of light stimuli. 



118 



2. Extracting accurate, precise estimates of ven tilation 
parameters from radioxenon scintigrams of the lung ; B. Bunow, 
B. Line, M. Horton (LAS), G. Weiss (PSL) : Extensive computer 
simulation, paralleled by mathematical analysis, has developed 
previously unavailable information on the precision and accuracy 

of various methods of estimating washin-washout rates of radionuclide 
exchange from scintigraphic studies of lung ventilation using 
radioxenon. These results have not only quantified the cumulative 
effects of random variation in counting statistics but have also 
tested quantitatively the effects of different isotopes of xenon 
and the benefits of increasing dosage over shorter time periods, 
and have allowed a rational decision on the most suitable 
scintigraphic area to represent a lung compartment. 

3, Role of myoglobin in facilitating the diffusion of oxygen 
in muscle : J. Fletcher, M. Bieterman (LAS) : The Krogh capillary- 

t issue model and associated nonlinear oxygen-hemoglobin dissociation 
have been used to study the role of myoglobin in facilitating diffusion 
of oxygen in muscle. Taking blood flow rate into account, the 
resulting mathematical model has required development of new numerical 
methods of solution, particularly at boundary layer regions near the 
supplying capillaries. Preliminary results indicate that facilitation 
of diffusion through binding proteins such as myoglobin can be 
important at low flow or high tissue metabolic rates. 



II. Annotated list of LAS projects 

Mathematical modeling of biological processes : J. Fletcher (LAS) , 
A. Spector (University of Iowa): Development and application of 
mathematical models and numerical methods in studies of substrate 
transport in the microcirculation, diffusion processes in physiology, 
and macromolecule-ligand binding equilibria. 

Simulation of physiological systems ; E. Hill, J. Fletcher, E. Harris 
(LAS); B. McLees (CC) : Exploration and testing of computer programs 
simulating responses of physiological systems to determine the 
usefulness of such programs as consultants or quality control 
mechanisms in patient care. 

Mechanisms of active transport; biochemical kinetics : B . Bunow 
(LAS); J. Rinzel (NIAMDD) ; J. De Simone et al. (Medical College of 
Virginia) : experimental and mathematical studies of the energy 
mechanisms for active transport, and of multi-state biochemical 
kinetics in cells and membranes. 

General mathematical and computational methods : E. Hill (LAS) ; 
R. Shrager (LSMM) : Study of methods of fitting non-linear models 
utilizing other than least squares criteria. Evaluating methods 
of organizing large data files for rapid storage and retrieval. 



119 



Computer aided analysis of electrocardiograms ; J. Bailey, M. Horton 
(LAS): Separate studies, conducted in collaboration with the 
Cardiology Branch, NHLBI, and a cardiologist in Glasgow, Scotland, 
to evaluate the utility of leading computer programs for ECG 
interpretation. Both studies include clinical documentation. 

Computer-based modeling of pulmonary gas exchange : Many LAS staff 
members; staff of Diagnostic Imaging, (CC) , and Pulnonary laboratory 
(MIX3I). Utilizing scintigraphic and other clinical data on lung 
function to construct sound mathematical and computer-based mocels 
of ventilation and perfusion in the lung. 

Computer systems for diagnostic imaging : J. Bailey, M. Douglas, 
B. Line (LAS); H. Ostrow (CSL) ; M. Green and others (Diagnostic 
Imaging, CC) : Development and application of computer systems 
to such diagnostic imaging activities as ECG-gated radionuclide 
angiocardiography, functional mapping and other scintigraphic studies 
of kidney, brain, heart and lung. 

Hybrid computing to analyze physiologic signals and construct 
simulation modelsl E. Pottala, R. Covacci (LAS), S. Vallerga 
(Lab. Neurophys., NINCDS) : Using LAS minicomputer system (MAC-16) 
for hardware simulation of physiologic functions (e.g., retinal 
cell activity) and analysis of analog signals (myogram, 
ultrasonogram, etc.) 

Statistical research in clinical pathology ; E. Harris (LAS) , 
G. Shakarji (DMB) ; clinical chemists in California, U.K.: 
Application of variance component and time series analysis to 
description of reference distributions of clinical laboratory 
tests, to serial studies of normal biochemistries, and to the 
design of criteria for recommended precision and accuracy of 
analytic methods. 



III. Publications 



Bacharach, S.L., Green, M.V., Borer, J.S., Line, C.R., 
Bradley-Moore, P.R., Ostrow, H.G., and Johnston, G.S.; Real-time 
collection, analysis and display of nuclear medicine data. Proc. 
7th Symposium Comp. Prog. Tech. Nucl. M6d. , 1977 (in press). 

Bailey, J.J., and Horton, M.R.; Reproducibility of Version 2 of 
the IBM program with and without the serial comparison option. The 
Fifth International Congress of Electrocardiology , 1973 (in press) . 



120 



Bradley-Moore, P.R., Klickna, J., Line, B.R., Jones, A.E., and 
Johnston, G.S.: Nuclear medicine reporting system — renal. 
In Howard, B.Y. (Ed.): Proceedings of Seventh Symposium on Sharing 
of Computer Programs and Technology in Nuclear Medicine, Computer 
Assisted Data Processing . ERDA COKF-770101, 1977, 7: pp. 92-105. 

Brereton, H.D., Line, B.R. , Londer, H.N., O'Donnell, J.F., Kent, 
C.H., and Johnson, R.E.: Gallium scans for staging small cell 
lung cancer. JAMA , 1978 (in press). 

Bunow, B.: Chemical reactions and membranes: A macroscopic basis 
for facilitated transport, chemiosmosis , and active transport. 
Part I. J. Theor. Biol. , 1978 (in press). 

Bunow, B.: Chemical reactions and membranes: A macroscopic basis 
for active transport. Part II. J. Theor. Biol. , 1978 (in press). 

Douglas, M.A. , and Green, M.V.: A System for Computer Generation of 
Left Ventricular Masks for use in Computerized ECG-Gated Radionuclide 
Angiocardiography. Nuclear Cardiology: Selected Computer Aspects . 
Society of Nuclear Medicine, New York, May 1978, pp. 119-123. 

Douglas, M.A., Green, M.V., and Ostrow, H.G.: Computer generation of 
left ventricular masks for use in computerized ECG-gated radionuclide 
angiocardiography. Computers in Cardiology , 1978 (in press). 

Dunham, R.G., Line, B.R., and Johnston, G.S.: A comprehensive 
software system for producing functional maps. In Howard, B.Y. (Ed.): 
Proceedings of Seventh Symposium on Sharing of Computer Programs and 
Technology in Nuclear Medicine, Computer Assisted Data Processing . 
ERDA CONF-770101, 1977, 7: pp. 337-345. 

Fletcher, J.E.: A generalized approach to equilibrium models. 
J. of Physical Chemistry , 81: 2374, 1977. 

Fletcher, J.E.: Mathematical modeling of the microcirculation. 
Math. Biosciences 38: 159-202, 1978. 

Fletcher, J.E.: The mathematical sciences in research at NIH. 
Newsletter, Conference Board of the Mathematical Sciences , 
Vol. 13, No. 2, 17-20, 1978. 

Fletcher, J.E. and Spector, A. A.: Alternative models for the analysis 
of drug-protein binding. Molecular Pharmacology 13: 337-399, 1977. 

Green, M.V., Brody, W.R., Douglas, M.A., Borer, J.S., Ostrow, H.G., 
Line, B.R., Bacharach, S.L., and Johnston, G.S.: Ejection fraction 
by count rate from gated images. Jour. Nucl. Med. , 1978 (in press). 



121 



Green, M.V., Bacharach. S.L., Douglas, M.A., Borer, J.S., and Johnston, 
G.S.: Sources of Virtual Background in Multi-image Cardiac Studies. Nuc leaj 
Cardiology: Selected Computer Aspects . Society of Nuclear Medicine, 
New York, May 1978, pp , 97-106. 

Harris, E.K. and Shakarji, G.: Use of the population distribution to improvi 
estimation of individual means in epidemiological studies. J. Chronic ' 
Diseases , 1978 (in press) . 

Harris. E.K.: Statistical principles underlying analytic goal setting in | 
clinical chemistry. Proceeding of the College of .American Pathologists 
Second .Annual Aspen Conference (1 9 76) on Analytical Goals in Clinical 
Chemistry , pp. 115-135, 1977; to be reprinted as a supplement to the .Amer . 
J. of Clinical Pathology , Summer, 1979. 

Hill, E.: A comparative study of very large data bases. Lecture Notes in 
Computer Science , Vol. 59, January 1978. 

Hill, E.: .Analysis of an inverted data base structure. In Dattola, R.T. (EdJ 
Proceedings of the International Conference on Information Storage and ' 
Retrieval . .AC.M-SIGIR, New York, Vol. 13, No. 1, May 1978, pp. 37-64. 

Kushner, T.R., Line, B.R., Barharach, S.L. and Johnston, G.S.: A spirometri 
method for gating .xenon ventilation studies. In Howard, B.Y. (Ed.): Pro- 
ceedings of Seventh Symposium on Sharing of Computer Programs and Technology , 



I 



in Nuclear Medicine, Computer Assisted Data Processing . ERD.A CONF-770101. 
1977, 7: pp. 207-215. 

Line, B.R.: .A command processing system for the analysis of scintigraphic 
data. In Brill, .A.G. (Ed.): Proc . 5th International Conference on Infor- 
mation Processing in Medical Imaging. .Nat. Tech. Info. Service, ORNL/BCTIC-. 
June 1977, 5: pp. 436-467. 

Line, B.R., Dayhoff, R.E., and Bailey, J.J.: An algorithm for the produc- 
tion of regional gas partial pressures and blood contents from scintigraphic 
and physiologic data using an alveolar gas e.xchange mode. In Howard, B.Y. 
(Ed.): Proceedings of Seventh Symposium on Sharing of Computer Programs and 



Technology in Nuclear Medicine, Computer Assisted Data Processing . ERDA 
CONF-7-Olbl, 1977, 7: pp. 196-206. 

Line, B.R., Fulmer, J.D., Re}-nolds, H.Y., Roberts, W.C, Jones, A.W. , Harris 
E.K., and Crystal, R.G.: Gallium-67 citrate scanning in the staging of idio 
pathic pulmonary fibrosis: Correlation', with physiology, morphology and 
bronchoalveolar lavage. .Amer. Rev. Resp. Dis ., 1973 (in press). 

Pottala, E.W., Covacci, R. , Colburn, T.R., and Vallerga, S.: Hardware Model 
of a tiger salamander cone cell. Proceedings of the 3Qth ACEMB, .Alliance 
for Engineering in .Medicine and Biology, Bethesda, MD., 1977, pp. 369. 



122 



Rautaharju, P.M., Ariet, M., Pryor, T.A., Arzbaecher, R.C., 
Bailey, J.J., Bonner, R., Goetowski, C.R., Hooper, J.K., F-lein, V., 
Millar, C.K. , Milliken, J. A., Mortara, D.W., Pipberger, K.V., Pordy, 
L., Sandberg, R.L., Simmons, R.L., and Wolf, H.K.: Report of Task 
Force III: Computers in diagnostic electrocardiography. In 
Proceedings of the 10th Bethesda Conference: Optimal Electro- 
cardiography. Am. J. Cardiol. 41; 158-169, January 197S. 

Shrager, R. and Hill, E.: Curve-fitting in the L^ and L„ norms. 
SIAM Reviews , (Chronicle), April 1978. 

Shrager, R. and Hill, E.: Some properties of the Levenberg method 
in the L, and Loo norms. Mathematics of Computations , 197 8 (in 
press) . 

Spector, A. A., and Fletcher, J.E.: Nutritional Effects on Drug-Protein 
Binding. In Hathcock, J.N., and Coon, J. (Eds.): Nutrition and Drug 
Interrelations . New York, Academic Press, 1978, pp. 447-473. 

Spector, A. A. and Fletcher, J.E.: Transport of Fatty Acid in the 
Circulation. In Dietschy, J.M. , Dietschy, J. A., Ontko, J. A., and 
Gotto, A.M. (Eds.): The Physiology of Lipids and Lipoproteins in 
Health and in Disease . Bethesda, MD., American Physiological Society, 
1978, pp. 229-249. 

Spector, A. A. and Fletcher J.E.: Fatty acid binding by serum albumin. 
In Peters, T., and Sjoholm, I. (Eds.): FEES Federation of European 
Biochemical Societies, 11th Meeting Copenhagen 197 7 . New Yorl:, 
Pergamon Press, 1978, Vol. 50, Coll. B9, pp. 51-60. 

Vallieres, B.J., Johnston, G.S,, Jones, A.E., and Line, B.R.: 
Scintigraphy with gallium-67 citrate. I. Metabolism. 
Tome 106: 1275-1281, 1977. 

Vallieres, B.J., Johnston, G.S., Jones, A.E., and Line, B.R.: 
Scintigraphy with gallium-67 citrate. II. Clinical aspects. 
Tome 106: 1382-1391, 1977. 

van Rijk, P.P., de Graaf , C.N., Jambroes, G., Bacharach, S.L., 
Green, M.V., Line, B.R., and Bailey, J.J.: Functional imaging in 
nuclear cardiac studies. Computers in Cardiology . IEEE Catalog No. 
77cn2154-2C, IEEE Computer Society, Long Beach, Ca. , 90803, 1973, 
pp. 9-17. 



123 



SMITHSONUrj SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00033-02 



LAS 



PERIOD COVERED 

October 1 1977 to September 30 1978 
TITLE OF PROJECT (80 characters or less) 



Analysis of Coupled Transport anr" Piochenical Kinetics 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



CTITRS: 



B, EunoT.' 



Senior Staff FelloT-r 



LAS DCRT 



J. rernovez Prof. Univ. of Tech. Conpie?ne, France 
J. DeSimone Prof. Univ. of Va. (?Iedical Collep.e) 
n. Mikulecky Prof. Univ. of Va. (Medical College) 
G. Weiss Chief PSL PCRT 

J. Rinzel Mathenatician NIA>n^D 

R. Miller Staff Fellow NIAMDD 



COOPERATING UNITS (if any) 



None 



lab/branch 

Laboratory of Applied Studies 



SECTION 



Applied Mathematics Section 



INSTITUTE and LOCATION 

DCRT, NIH, Bethesda, Md. 2001 A 



TOTAL MANYEARS: 
1.0 



PROFESSIONAL: 
1.0 



OTHER: 



CHECK APPROPRIATE 80X(ES) 
n (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIE.S 



n (b) HUMAN TISSUES 



H (c) NEITHER 



SUMMARY OF WORK (200 *ords or less - underline keywords) 

This project investigates tvo fundamental problems in biology: 
(1) role of dynamic patterns in enbryological and evolutionary biology 
and (2) kinetics of enzymes located in cell membranes. The first area 
involves a demonstration of the role which simultaneous reaction and 
diffusion might play in the formation of biological patterns such as 
those which determine the shape of organs. The second area involves 
investigation of the effect of diffusion limitation by cell mem.hranes 
on the kinetics of enzymes contained in them. Digital computer 
sLmulation and bifurcation theory are the principal tools of these 
investigation. 



124 



PHS-6040 
(Rev. 10-76) 



(1) D>'namic Patterns 

Background and Objectives: 

The objectives of this investigation are twofold. The first is a 
study of dynamic patterns in reaction-diffusion systems obeying real- 
istic biochemical kinetics. The objective here is to demonstrate the 
structure of solutions to the descriptive equations of these systems 
which develop spontaneously when unstructured solutions are perturbed. 
The solutions obtained with the mathematical model are to be compared 
with results obtained in the laboratory with a working model. The 
adequacy of reaction diffusion models for explaining certain aspects 
of embryological pattern formation needs further clarification. For 
example, it is important to know the modification to be expected in 
the patterns when the size of the system is altered or the rate of 
enzymatic activity, as might occur due to mutations or change of tem- 
perature. 

Significance to Biomedical Research: 

Patterns generated by reaction-diffusion processes have been im- 
plicated in embryological organogenesis. There are a number of un- 
answered questions in this area which we address. 1) Are the patterns 
generated by reaction-diffusion sufficiently varied as to be capable of 
producing the many shapes seen in embryonic development? 2) Is the 
relation between the pattern obtained and the underlying reaction scheme 
sufficiently close that we might infer the latter from the former? Do 
the patterns succeed one another in a consistent way as parameters are 
changed, as would occur during growth? 

Progress in FY 78: 

It has been shown that a large number of different patterns can 
be obtained with a single reaction mechanism. Most of this variety 
can be explained without any detailed knowledge of the details of the 
reaction mechanism. Many different patterns can exist for the same 
values of all parameters in the mathematical model. In order for a 
reaction-diffusion mechanism to be operative in embryological pro- 
cesses, it is necessary that an additional mechanism be present to 
select among the possible patterns. The patterns show no unique 
succession with growth. These results suggest that the role of 
reaction-diffusion patterns in embryology is likely to be more com- 
plicated than has been recently speculated. A manuscript summarizing 
accomplishments during FY 77 has been completed and will be submitted 
shortly. Results from FY 78 have been presented internally and a manu- 
script is approaching completion. 

(2) Kinetics of Enzymes 
Background and Objectives: 

A second direction of investigation is the effect of cell membranes 



125 



and surface structures on the kinetics of membrane associated enzymes. ' 

These enzymes are important in membrane transport, metabolic energy J 

transduction, and intercellular communication. The goal here is to dis- I 

cover means of analyzing kinetic data from membrane-associated enzymes ' 
which permit correct conclusions about kinetic constants and mechanism. 

Significance to Biomedical Research: |i 

The design of pharmacological agents with which to modify the ' 

function of membrane-associated enzymes requires correct estimates ^ 

of interaction between the enzyme and the agent. Only bv taking 1 

the effect of the membrane into account can such estimates be obtained ^ 
for enzymes in membranes. A kinetic analysis appropriate only to en- 
zymes in solution has been widely misused in studies of enzymes in 
membranes. It is likely that many conclusions regarding the mechanism 
of such enzymes as well as the strength of their interactions with 
substrates have been incorrectly estimated in this way. j 

Progress in FY 78: .----. 

An analysis has been developed which permits correct determination 
of the affinity constants between membrane-associated enzymes and their 
substrates, products and specific inhibitors. Distinction between 
different reaction mechanisms is possible for these enzvmes, but the 
patterns which they produce in conventional studies of mechanism are 
quite different from what is expected for enzymes in solution. In 
particular, membrane-associated enzymes typically appear to be co- ' 

operative due to the effect of the membrane. Application has been 
to the pharmacology of a sympatholytic inhibitor of acetylcholinesterase 
and to the kinetics of tripeptide transport in E. coli. A manuscript 
reporting on this work has been submitted for publication. 

Proposed Future Course of Investigation: 

Future efforts in the dynamic patterns area will involve consolidation 
of theoretical and experimental procedures. Work in the enzyme liinetics area, 
will be elaborated to consider more complicated kinetic schemes speci- 
fically appropriate to membrane transport and Tietabolic processe'^". 

Publications : j 

1. r,unow, B.: Chemical reactions and m.embranes: a macroscopic basis 

for facilitated transport, cheniosmosis, and active transport, i 

part I. J. Theor. Biol .. 1978 (in press). ! 

2. Bunow, B.: Chemical reactions and membranes: a macroscopic basis 
for Active Transport, Part II. J. Theor. Biol ., 1973 (in press). 

3. Bunox;, B.: Kinetics of membrane-associated enzymes: effect of 
transport limitation by boundary layers and membrane structure. 
(Submitted to Biochemistry, 1978) . 



126 



SMITHSONIAN SCI 
PROJECT NUMBER 



NCE INFORMATION EXCHANGE 
Do NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE OF 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00002-08 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978. 



TITLE OF PROJECT (80 characters or less) 

Computer Aided Analysis of Electrocardiograms 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: J.J. Bailey Chief, Med. Appl. Section LAS DCRT 
M.R. Horton Computer Systems Analyst LAS DCRT 
J. Gottdiener Chief, ECG Laboratory CB NHLBI 



COOPERATING UNITS (if anj-) 

P. Lawrence, Cardiology Dept., Walter Reed Army Medical Center 

P. MacFarlane, Medical Cardiology, Glasgow Royal Infirmary, Scotland 



lab/branch 

Laboratory of Applied Studies 

SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NTH, Bethesda, im 20014 



TOTAL MANYEARS: 

1.3 



PROFESSIONAL: 
1.2 



OTHER: 



0.1 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

□ (al) MINORS (i2) INTERVIEWS 



□ (b) HUMAN TISSUES 



[f[c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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). 



127 



PHS-60*0 
(Rev. 10-76) 



Background and Objectives: 

In the past fifteen years numerous computer programs implemented 
upon a variety of computer systems have been developed for analysis of 
routine ECG's. Computer processing of ECG's has become a sizeable 
enterprise in many parts of the country, including both commercial 
bureaus offering service for a fee and non-conmercial (academic or 
government) centers establishing the capability for themselves. 

Since 1970, LAS in collaboration with the Cardiology Branch (CB) , 
NHLBI has studied various programs to determine which, if any, would be 
useful to implement on the NIfi campus. As a result of this work, an 
evaluation methodology was evolved which was published in 1974 and has 
become one of the standard references in the field. Although a program 
was selected and implemented in 1974 for daily use at Nil!, additional 
programs and other computer systems continue to be evaluated as nossible 
improvements to the Mill system v;ith regard to cost and accuracy. Past 
evaluations have included the ECAM-D program (1964) , the 'fayo-IBM program 
(1968), the experimental IIM program (1971), and the AVA 3.4 (Pipberger) 
program (1975) . 

Progress during FY 78: 

A series of 300 ECG's were collected on patients in the Glasgow 
Pvoyal Infirmary, (GRI) , Scotland. Clinical documentation of the 
patient's cardiovascular status by non ECG means in the form of enzyme 
studies, cardiac catheter laboratory data, etc. was obtained whenever 
such investigations were warranted, i.e. in most cases. The standard 12 
lead ECG data was analyzed by the IBM program at NIK; modified McFee 
lead (XYZ) ECG data was analyzed by a program developed at GRI. The 
results of these two programs are being compared with respect to 
accuracy where clinical documentation exists and with respect to 
cardiologist agreement in all other cases. 

A series of 110 ECG's were collected by CB, MHLBI on patients with 
a variety of conditions including hypertension, valvular disease, and 
coronary disease. In all cases, the size of the left ventricle has been 
estimated from ultrasound measurements of the septal wall thickness, 
posterior wall thickness, and internal end-diastolic diameter. The IBM 
program uses ECG features to compute points toward a diagnosis of left 
ventricular hypertrophy (LVH) in a manner somewhat similar to the 
Romhilt-Estes point score. The AVA 3.4 (Pipberger) program uses Frank 
lead (XYZ) ECG features to compute probabilities for LW.. The results 
of these two computerized LVE algorithm.s' are being correlated with 
ultrasound measurements of left ventricular mass. 

A method, published in 1974, for using interleaved digital samplings 
from the same analog tracing to test reproducibility of programs has been 
extended and applied to the ECAJJ-E program (1975) and Version 2 of the 
IBM program (1976). Guidelines for evaluation of ECG programs which were 
developed by LAS have been adopted by the American College of Cardiology 
(ref. 1). 

128 



Several manuscripts describing the above work are in preparation 
and one will be the subject of an invited paper at the 5th International 
Congress of Electrocardiology in Scotland. 

Members of LAS and CSL developed specifications for a minicomputer 
system to be dedicated to ECG analysis which the Clinical Center plans 
to procure. 

Significance: 

The estimated number of computer-processed ECG's in North America 
in 1971 was 600,000; in 1972, it was one million; and in 1974 it was A 
million. In view of this exponential increase in computer usage, it 
will become ever more important to have methodologies and guidelines by 
which ECG computer systems can be evaluated. 

These studies seek to establish the diagnostic limits of ECG itself 
and the degree to which computerized algorithms may achieve these 
limits. Important evaluation methodology continues to be developed, 
which may have a significant impact on the further diffusion of computer 
technology in electrocardiography. 

Proposed Course: 

The testing of these programs with clinically documented data, 
i.e., cases with correlative data from cardiac catheter laboratory 
studies, echocardiography, or scintigraphic studies, is expected to be 
completed within FY 79. 

Publications and Abstracts: 

1. Rautaharju, P.M., Ariet, M., Pryor, T.A., Arzbaecher, R.C., 
Bailey, J.J., Bonner, R. , Goetowski, C.R., Hooper, J.K., Klein, V., 
Millar, C.K., Milliken, J. A., Mortara, D.W., Pipberger, K.V., Pordy, 
L. , Sandberg, R.L., Simmons, R.L., and Wolf, U.K.: Report of Task 
Force III: Computers in diagnostic electrocardiography. In 
Proceedings of the 10th Bethesda Conference: Optimal Electro- 
cardiography. Am. J. Cardiol . Al: 158-169, January 1973. 

2. Bailey, J.J., Norton, M.R.: Reproducibility of Version 2 of the 

IBM Program with and without the Serial Comparison Option. The Fifth 
International Congress of Electrocardiology , 1978 (in press) . 



129 



SMITHSONIAM GCIEi'lCE I ;iFCRMAT I Oli EXCHAriGEl U.S. DEPARTMENT OF 

PROJECT fJUMBER (Oo NOT use this ssace) HEALTH, EDUCATIC:, A.'iO WELFARE 

PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl Ct 00034-02 



LAS 



PERIOD COVERED 

Oct. 1, 1977 to September 30, 1978 



TITLE OF PROJECT (SO characters or less) 

Computer-based Modeling of Pulmonary Gas Exchange 
and Respiratory Mechanics 



NAMES, LABORATORY AfJO INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: J.J. Bailev Chief Medical Appl. Section 

Senior Staff Fellow 
Chief, Applied Math Section 
Senior Assistant Surgeon 
Chief, Pulnonarv Branch 
OTHERS: E.K. Harris Chief, LAS 

M.Pv.. r.orton Computer Systems Analyst 

T.E. Stibolt, Jr. Senior Staff Fellc- 

A.E. Jones Asst. Chief 

J.D. Fulmer Chief 

B.D. McLees Ch, Medical Intensive Care Unit 

A.R. Mitz Engineer 



J.J. Bailey 
B.J. Bunow 
J.E. Fletcher 
B.R. Line 
R.G. Crystal 
E.K. Harris 
r.orton 



LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


PB 


M^LRI 


LAS 


DCRT 


LAS 


DCRT 


LAS 


T^pprp 


DI 


CC 


PB 


>1I1LBI 




CC 


LAS 


DCRT 



COOPERATING UNITS (if any) 

Diagnostic Imaging Branch 
Pulmonary Branch 



CC 

NHLBI 



lab/branch 

Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section, Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIK, Bethesda. MD 20014 



TOTAL MANYEARS: 

2.0 



PROFESSIONAL: 
2.0 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
D (a) HUMAN SUBJECTS 



n (b) HUMAN TISSUES 



B (c) NEITHER 



□ (a1 ) MINORS 



INTERVIEWS 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves a collaborative effort of LAS with the 
Diagnostic Imaging Department, CC, the Pulmonary Branch, MHL3I, and 
the Medical Intensive Care Unit, CC. It 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. 



130 



PHS-6040 
(Rev. 10-76) 



Background 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 which tended 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 this has been changed by the advent of the 
digital computer and also by newer diagnostic tools, viz: pulmonary 
scintillation, cardiac catheterization, multiple inert tracer gas 
analysis and dynamic compliance studies. 

This program involves the Pulmonary Branch, NHLBI; the Diagnostic 
Imaging Department, CC; the Medical Intensive Care Unit, CC; and the 
Laboratory of Applied Studies, DCRT. The objective is to construct 
computer-based models of pulmonary gas exchange and repiratory mechanics 
founded upon sound physiological considerations and descriptive 
mathematical formulations. 

Progress during FY 77: 

A computer simulation model was developed to test the accuracy and 
precision of various methods for extracting ventilation parameters from 
radioxenon studies of the lung. The simulation model showed how the 
results are affected by variations due to counting statistics and 
revealed how much improvement could be obtained through the use of a 
better isotope (xenon 127 instead of xenon 133), by increased dosage over 
a shorter time, and by using more pixels to represent a lung compartment 
(ref. 7). 

The Kelman procedure which relates gas tensions and contents in 
blood for given values of hemoglobin, hematocrit, pH, etc. has been 
revised to incorporate Adair binding constants for hemoglobin, to remove 
unnecessary corrections in the formulae and also sources of numerical 
imprecision, and to utilize more efficient computer algorithms. A 
modification of the Kelman routine was used to determine regional gas 
contents from ventilation-perfusion scans of the lung (ref. 1) . 

A method for studying the distribution of inhaled radioxenon at 
different lung volumes has been devised, (ref. 2). With this method the 
effect of gross movement upon the extraction of respiration parameters 
can be studied. A method using radiogallium citrate to tag white cells 
was used to study active inflammatory processes in the lung and to 
correlate regions of inflammation with regions of ventilation-perfusion 
abnormalities (refs. 3-5). 

Significance: 

These computer-based models when combined with data from 
scintillation, cardiac catheter, and pulmonary laboratory studies should 

131 



allow a quantitative description of pulmonary pathophysiology on a 
regional basis. They should allow separation of diseases (e.g. 
bronchitis from emphysema), separation of disease components 
(obstructive vs. restrictive vs. vascular), assessment of severity of 
disease component, and prediction of the degree to v/hich each component 
compromises overall pulmonary function. 

Proposed Course: 

The sources of greatest variation and potential error in estimating 
ventilatory parameters for xenon scintigraphy relate to counting 
statistics (Poisson noise), masking errors, gross respiratory movement, 
and variable background. These sources can be minimized through the 
implementation of: controlled patient breathing; reduced re-breathing 
reservoirs for radioxenon; hardware markers for end of washin/beginning 
of washout; and new software algorithms for masking and background 
subtraction. 

The Kelman procedure (NIH version) will be validated with real data 
from the Clinical Center. So far as is known such validation has not 
been done before. 

Publications and Abstracts: 

1. Line, 3.R., Dayhoff, R.E., and Bailey, J.J.: An algorithm for the 
production of regional gas partial pressures and blood contents from 
scintigraphic and physiologic data using an alveolar gas exchange 
model. In Howard, B.Y. (Ed.): Proceedings of Seventh Symposium on 
Sharing of Computer Programs and Technology in .luclear Medicine , 
Computer Assisted Data Processing . ERPA COMF-770101, 1977, 7: pp. 
196-206. 

2. Kushner, T.R., Line, 3.R., Bacharach, S.L., and Johnston, G.S.: A 
spirometric method for gating xenon ventilation studies. In Howard, 
B.Y. (Ed.): Proceedings of Seventh Symposiun on Sharing of Computer 
Programs and Technology in ^luclear .[edicine, Cor.puter Assisted Data 
Processing . ERDA COMF-770101, 1977, 7: pp. 207-215. 

3. Vallieres , B.J., Johnston, G.S., Jones, A.F., and Line, 3.R.: 
Scintigraphy with gallium-67 citrate. I. metabolism. Tone 
106: 1275-1281, 1977. 

4. Vallieres, B.J., Johnston, G.S., Jones, A.E., and Line, B.R.: 
Scintigraphv with gallium-67 citrate.- II. clinical aspects. Tome 
106: 1382-1391, 1977. 

5. Brereton, H.D., Line, B.R., Londer, K.M., O'Donnell, J.F., Kent, 
C.K., and Johnson, R.E.: Gallium scans for staging small cell 
lung cancer. JAMA . 1978 (in press) . 



132 



6. Line, B.R., Fulraer, J.D., Reynolds. H.Y., Roverts, W.C., Jones, 

A.E., Harris, E.K. , and Crystal, R.G.: Gallium-67 citrate scanning 
in the staging of idiopathic pulmonary fibrosis: correlation with 
physiology, morphology and bronchoalveolar lavage. Amer. Rev . 
Resp. Pis . , 1978 (in press). 



133 



SMITHSONIAN SCIEf 
PROJECT NUMBER (C 



JCE i::formation 
Oo NOT use this 



EXCHANGE 
space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATICN, AMO WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00003-07 



LAS 



PERIOD COVERED 

October 1, 1978 to September 30, 1978 



TITLE OF PROJECT (30 characters or less) 

Computer Systems for Diagnostic Imaging 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: J.J. Bailey Chief Med. Appl. Sec. 
M.V. Green Ch, Appl. Physics Sec. 
B.R. Line Senior Assistant Surgeon 



OTHERS :M. A. Douglas 

M.R. Horton 

E.W. Pottala 

S.L. Bacharach 

T.B. Stibolt, Jr, 



Computer Systems Analyst 

Computer Systems Analyst 

Engineer 

Physicist 

Senior Staff Fellow 



P.R. Bradley-Moore Staff Fellow 

R.G. Dunham Computer Specialist 

G.S. Johnston Chief 

II. G. Ostrow Engineer 



LAS 


DCRT 


DI 


CC 


LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


DI 


CC 


LAS 


DCRT 


DI 


CC 


DI 


CC 


DI 


CC 


CSL 


DCRT 



COOPERATING UNITS ^if any) 

Diagnostic Imaging Department, CC, NIH 
Computer Systems Laboratory, DCRT, NIH 



lab/branch 

Laboratory of Applied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MP 2001A 



TOTAL MANYEARS: 

3.1 



PROFESSIONAL: 

3.0 



OTHER: 



0.1 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

(al) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



g (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project involves computer-based mathematical analysis, 
pattern recognition, and image processing in support of diagnostic 
activities in the Diagnostic Imaging Department of the Clinical Center 
and collaborating Institutes. Applications include computerized 
ECG-gated radionuclide angiocardiography and scintigraphic studies of 
regional cerebral blood flow , renal dynamics , gastric fluid dynamics, 
and pulmonary ventilation-perf usion relationships as well as computer 
analysis of ultrasonograms . 



134 



PHS-6040 
(Rev. 10-76) 



Background and Objectives: 

Since FY 72 LAS with engineering support from CSL, DCRT and in 
collaboration with the Diagnostic Imaging Department, CC, has 
accomplished acquisition of minicomputer hardware and the development of 
software necessary to process data from three scintillation cameras in 
the Diagnostic Imaging Department. 

The objective of this program is continuing 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. Application 
to other noninvasive imaging systems such as ultrasonograms, and the 
incorporation of other physiologic signals with the imaging studies 
(e.g. ECG-gated heart studies) are also under investigation. 

Progress during FY 78: 

During FY 76-77 the extended nuclear utility (EMU) system was 
developed (ref. I), which allowed the rapid modular construction of 
complex processing programs using newly developed program modules from a 
library of already developed algorithms. This system was used 
extensively in the development of radionuclide angiography, pulmonary 
scintigraphy, and radionuclide renography projects (ref. 2). During FY 
78, the algorithms developed under EOT have been largely upgraded and 
re- implemented under the real time executive (RTE) . RTE is a new vendor 
supported system which allows the nuclear medicine minicomputers to 
become multi-user, multi-task systems. RTE allows rapid development of 
modules which formerly could only be achieved under ENU. 

Previous work in FY 76 showed a significant enhancement of 
radionuclide renography by the use of functional maps and since FY 77 
functional maps have come into routine clinical use (ref. 3). Currently 
several dogs are being studied, before and after surgically induced 
segmental renovascular occlusions. These canine models will be used to 
document the sensitivity and specificity of functional maps in detecting 
and localizing renovascular disease. 

Since FY 72, DI and LAS have collaborated upon development and 
refinement of algorithms to produce radionuclide angiograms and 
ventricular volume curves simultaneously while scintillation data is 
being collected from the patient (refs. 4-5). Usefulness of functional 
maps in this area also has been studied '.(ref . 6). In FY 76 a package of 
image processing routines (IHAGE) was developed on the PDP-10 System. 
This package is capable of several functions: e.g. edge detection, 
spatial derivatives, iterative deconvolution, noise detection and 
application of selective filters. 



135 



Progress during FY 78: 

In FY 78 this image processing package was used to design an 
algorithm for automatic detection of the left ventricle in radionuclide 
angiocardiography (refs. 7-8). Currently the algorithm is being 
implemented on the nuclear medicine minicomputer system in order to 
facilitate the real time system. IMAGE was also used to design an 
automatic detection algorithm for kidneys. During FY 78 a Biomation 
hardware buffer was interfaced with a Picker ultrasound system and an 
HP minicomputer. A few abdominal ultrasonograms have been captured in 
digital form and displayed on the Evans and Sutherland device with color 
coding. 

There has been continued development in pulmonary scintigraphy 
methods (see project on Computer-based Modeling of Pulmonary Gas 
Exchange and Respiratory Mechanics) . 

Significance: ^"~ 

Scintigraphy is a noninvasive tool which sho^TS considerable detail 
concerning the dynamic function of an organ on a regional basis. 
Computer processing not only enhances scintillation images but allows 
quantification of the dynamic function. Real time implementation of 
algorithms on the minicomputers allows the clinical investigator to 
perform repeated studies on patients with exercise or therapeutic 
manipulation, thus better elucidating the nature of the patient's 
pathophysiology. 

Ultrasonography is a diagnostic imaging tool with wide application. 
It offers an excellent opportunity to obtain serial studies of certain 
disease processes at frequent intervals non-invasively and without the 
risk of radiation exposure. Image enhancement and automatic pattern 
recognition should greatly improve its clinical utility. 

Proposed Course: 

The canine models will be extended to include obstructive 
nephropathy, toxic nephritis, and mass lesions (e.g. hematoma). When 
renal functional maps are validated by canine models, this noninvasive 
diagnostic method should greatly improve the evaluation of disease state 
and response to therapy in patients with lupus nephritis (NIAMDD) , renal 
hypertension (NHLBI) , and neoplasia affecting the kidneys and urinary 
tract (NCI) . 

NHLBI has solicited the collaboration of LAS and DI in studying 
Thallium 201 images of the myocardium as an adjunct to radionuclide 
angiography in detection of coronary disease (ref. 9). 

Clinicians in the Diagnostic Imaging Department propose to study 
techniques for improving ultrasound Lmaging of abdominal structures. 
These include grey scale expansion and windowing, use of color, area 
measurement with possible extension to volume measurement, and reduction 
of spatial distortion. 

136 



Publications and Abstracts: 

1. Line, B.R.: A command processing system for the analysis of 
scintigraphic data. In Brill, A.G. (Fd.): Proc. 5th TntrpmaM'nn^l 
Conference on Information Processing in Medir.nl TTnaftnp . Nat. Tech. 
Info. Service, ORNL/BCTIC-2, June 1977, 5: pp. ^456-467. 

2. Dunham, R.G., Line, B.R., and Johnston, G.S.: A comprehensive 
software system for producing functional maps. In Howard, B.Y. 
(Ed.): Proceedings of Seventh Symposium on Sha ring nf rnnpnrpr 
Programs and Technology in Nuclear Medicine^ Compiirpr A^gicirpH riara 
Processing . EPJ3A CONF-770101, 1977, 7: pp. 337-345. 

3. Bradley-Moore, P.R., Klickna, J., Line, B.R., Jones, A.E., and 
Johnston, G.S.: Nuclear medicine reporting system — renal. In 
Howard, B.Y. (Ed.): Proceedings of Seventh Svmpn?;1uTn on Shnring nf 
Computer Programs and Technology in Nuclear Medir-inp, rnrnpnrpr 
Assisted Data Processing . EPDA CONF-770101, 1977, 7: pp. 92-105. 

4. Bacharach, S.L., Green, M.V., Borer, J.S., Line, B.R., 
Bradley-Moore, P.R., Ostrow, H.G., and Johnston, G.S.: Real-time 
collection, analysis and display of nuclear medicine data. Pror. 7t-'-: 
Symposium Conp . Prog. Tech. Nucl . Med .. 1977 (in press). 

5. Green, M.V., Brody, W.R., Douglas, M.A., Borer, J.?., Ostrow, E.G., 
Line, B.R., Bacharach, S.L., and Johnston, G.S,: Ejection fraction 
by count rate from gated images. Jour. Nucl. Med . . 1978 (in press). 

6. van Rijk, P.P., de Graaf, C.N., Jambroes, G., Bacharach, S.L., 
Green, M.V., Line, B.R., and Bailey, J.J.: Functional imaging in 
nuclear cardiac studies. Computers in Cardiology . IEEE Catalog No. 
77CH2154-2C, IEEE Computer Society, Long Beach, Ca., 90803, 1978, 
pp. 9-17. 

7. Douglas, M.A., and Green, M.V.: A System for Computer Generation of 
Left Ventricular Masks for use in Computerized ECG-Gated Radionuclide 
Angiocardiography. Nuclear Cardiology: Selected Computer Aspects . 
Society of Nuclear Medicine, New York, May 1978, pp. 119-123. 

8. Green, M.V., Bacharach, S.L., Douglas, M.A., Borer, J.S., and Johnston, 
G.S.: Sources of Virtual Background in Multi-image Cardiac Studies. 
Nuclear Cardiology: Selected Computer Aspects . Society of Nuclear 
Hedicine, New York, May 1978, pp. 97-106. 

9. Douglas, M.A., Green, M.V., and Ostrow, H.G.: Computer Generation of 
left ventricular masks for use in computerized ECG-gated radionuclide 
angiocardiography. Computers in Cardiology . 1978 (in press). 



137 



PROJECT 



SMITHSONIAN SCIENCE I TiFORMAT I C.'i EXCHAflGE 
NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00006-08 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

General Mathematical and Computational Collaborative Efforts 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: E. Hill 
OTHER: R. Shrager 



Computer Systems Analyst LAS DCRT 
Mathematician LSMM DCRT 



COOPERATING UNITS (if any) 



None 



lab/branch 

• Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 2001A 



TOTAL MANYEARS: 

1.5 



PROFESSIONAL: 

1.5 



OTHER: 



CHECK APPROPRIATE BOX(£S) 
n (a) HUMAN SUBJECTS 

□ (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



S (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

A curve fitting methodology utilizing the Levenberg-Marquardt 
method for fitting non-linear models was extended to norms other than 
L (least squares). Heuristics are proposed and programs have been 
developed which offer these methods as practical tools. Manuscripts 
describing these methods have been submitted for publication. 

A second project area involved a comparison of methods for 
organizing very large amounts of stored data which facilitate rapid 
storage and retrieval . Existing hashing methods were evaluated and 
ne\7 algorithms are proposed for very large data bases . A methodology 
for evaluation and comparison of schemes is described. 



138 



PHS-6040 
(Rev. 10-76) 



Curve Fitting 

Background and Objectives: 

This project is concerned with the development of two areas of 
computer applications and mathematics in the biosciences. The first 
area is concerned with Nonlinear Model Fitting and Parameter 
Estimation. The "standard" methods of least squares have been 
extensively developed for fitting nonlinear models to experimentally 
obtained data with estimation of unknown parameters in these models as 
a basic objective. For many applications, the least squares 
criteron is not appropriate and may introduce unnecessary computational 
difficulties. For example, fitting data taken over a wide range of 
values of the independent variables, and fitted without weights, tends 
to bias the results toward the data associated with larger values of 
the independent variables. The problems of ill-conditioning with 
systems having several parameters are also well known. Until recently, 
methods for fitting models with nonlinear parameters were not available 
with other than the least squares criteria. The first part of this 
project is concerned with such criteria. 

Progress in FY 78: 

In FY 78 the Levenberg-Marquardt Method for nonlinear curve-fitting 
in the L Norm, (L, = Min Z ly - y| ,•) and L norm, [L = Min 

Max I y - y |]- has been implemented in Fortran IV. Articles describing 
these methods and some applications have been submitted for publication. 

Proposed Course: 

The data fitting routines are available in DCRT computer library 
form. Applications will be made on an as required or requested 
basis. No further analytical development is anticipated in this 
area. 

Large Data Bases 

Background and Objectives: 

A second project area involves a comparison of methods for 
organizing large amounts of stored data on direct access storage 
devices to facilitate fast retrieval of- the desired information. 
This project was undertaken in order to classify the performance of 
the various retrieval methods and their associated data structures. 

In a very large data base involving retrieval and updating, the 
major factor of immediate concern is the average number of accesses 
to the direct access storage device to complete a request. The 
average number of accesses to store and retrieve items on a direct 
access storage device was considered for hashing methods and other 
techniques . 

139 



Significance to Biomedical Research: 

Such methodology has application to the design of storage and 
retrieval schemes for medical records, computer-aided diagnoses, 
medical linguistics, and other technical areas requiring search and 
retrieval in very large data or decision files. 

Progress in FY 78: 

The algorithms and performance measures described in FY 77 have 
been edited and published. 

Proposed Course: 

The algorithms developed from the described studies are 
available in DCRT computer libraries, and applications will be made as 
the need arises. No further analytical development is anticipated in 
this area. 

Publications: 

1. Shrager, R. and Hill, E.: Curve-fitting in the Li and L norms. 
SIAiM Reviews , (Chronicle), April 1978. 

2. Shrager, R. and Hill, E.: Some properties of the Levenberg method 
in the Li and L^, norms. Mathematics of Computations , 1978 (in 
press) . 

3. Hill, E.: A Comparative Study of Very Large Data Bases. Lecture 
Notes in Computer Science 59: January 1978, 

4. Hill, E.: Analysis of an inverted data base structure. First 
International SIGIR Conference, May 1978. 



140 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo MOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
NTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00004-07 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE Of PROJECT (80 characters or less) 

Investigation of Hybrid Computing for the Construction of Simulation 
Models and for the Analysis of Physiologic Signals 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



E.U. Pottala 



OTIIEP^: J.J. Bailey 

M.R. Ho r ton 
R. Covacci 
S. Vallerga 
A.R. Mitz 
R. Norman 
T.B. Stibolt 
M.A. Douglas 



Elec. Engineer 

Chief Medical Appl. 

Section 

Computer Systems Analyst 

Visiting Scientist 

Visiting Scientist 

Elec. Engineer 

Staff Fellow 

Senior Staff Fellow 

Computer Systems Analyst 



LAS DCRT 



LAS 


DCRT 


LAS 


DCRT 


LAS 


DCRT 


LMP 


NINCDS 


LAS 


DCRT 


UTP 


NINCDS 


LAS 


DCRT 


LAS 


DCRT 



COOPERATING UNITS (if any) 

Laboratory of Neurophysiology, NINCDS 



lab/branch 

Laboratory of Applied Studies 



SECTION 

Medical Applications Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20014 



TOTAL MANYEARS: 



2.1 



PROFESSIONAL: 

2.0 



OTHER: 



0.1 



CHECK APPROPRIATE B0x(£3) 
□ (a) HUMAN SUBJECTS 

n (al) MINORS n (^2) INTERVIEWS 



□ (b) HUMAN TISSUES 



3 (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

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 , ultrasonogram , and -scintigram . 



141 



PHS-6040 
(Rev. 10-76) 



Background and Objectives: 

In some simulation models, certain pieces or functions can be split 
off and implemented in hardware circuitry or a set of microprocessors. 
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. And third, some models are so complicated and 
extensive, that their implementation on a large scale digital computer 
is not feasible; whereas 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 FY 76) . 

Since FY 72 the LAS Laboratory minicomputer system (MAC 16) has 
been developed and utilized for various projects including (1) the 
construction of physiologic simulation models and (2) the processing of 
physiologic signals. Since FY 72 the system has been interfaced with 
the Marquette tape drive (for routine ECG's from the Clinical Center); 
with the Honeywell 7600 analog tape transport; with a general purpose 
switch-filter network; with a real time spectral analyzer and ensemble 
averager; and with a neural control panel for simulation of neural 
networVis (FY 75) and central nervous system subsvstems (cerebellum, FY 
76). 

A general advantage of this system is that an investigator can 
automatically pre-process (edit, filter, and digitize) dynamic 
physiologic data so that optimal use of a large scale digital computer 
can be obtained. This was demonstrated in electromyograms from subjects 
on a muscle fatigue protocol studied by the National Institute of 
Occupational Safety and Health (FY 75) . This facility was also used for 
electrocardiogram (ECG) and ventricular pressure data in monkeys (F^' 76) 
and also ECG data from Glasgow Royal Infirmary (see ECG section) . 

Progress During FY 78 : 

The Laboratory of Neurophysiology, NINCDS, is studying the 
electrophysiology of the retina of the larvae tiger salmander. The 
Baylor-Hodgkin-Lamb model of a cone cell requires six compartments 
connected by six linear differential equations and three compartments 
involving sodium and potassium fluxes related by three nonlinear 
differential equations. This model has been implemented in hardware by 
LAS (ref. 1-2). It was also implemented in software on the PDP-10 so 
that the hardware model could be thoroughly checked out (ref. 3). The 
model parameters were adjusted to best fit the actual photoresponses 
(membrane potential) recorded intracellularly from cones in the retina 
of the larva tiger salamander. Stimuli consisted of flashes and steps 
of light. The responses of the model and the real cone cell were in 
good agreement for 10 millisecond flashes and 0.7 second steps of light 
whose intensities ranged over 2.5 log units with the intensity of the 
unattenuated light at 1.5 X 10**17 photons per second per square 
centimeter. 



142 



The LAS minicomputer system is being upgraded by the incorporation 
of a Micro-Arithmetic Processor (MAP) . This hardware device can process 
large arrays very rapidly; for example, a Fast Fourier Transformation of 
1024 data points can be produced in 20 milliseconds. The array 
processing capability will enable interactive design of filtering 
algorithms for many different physiologic signals and also the 
implementation of such filtering algorithms in real time. 

Proposed Course: 

Work will continue on the hybrid modeling system for retinal 
physiology consisting of hardware models of cone cells, rod cells, 
horizontal cells, and bipolar cells connected together in a network and 
driven by the LAS minicomputer system. This hybrid modeling system 
should permit a rapid test of possible synaptic configurations to 
determine the one which best fits observed photoresponses and eventually 
the designing of the proper electro-physiological experiments to verify 
it. 

An operating system for the MAC-16 system will be developed which 
will provide the user with easy access to compilers and major programs, 
will handle I/O for all peripheral devices, and will automatically 
allocate system facilities and storage during execution of user 
programs. 

The installation of the MAP will allow extensive studies into use 
of orthogonal transforms and other tools for filtering of physiologic 
signals and image enhancement algorithms for ultrasonographic and 
scintigraphic data. 

Publications and Abstracts: 

1. Pottala, E.W., Covacci, R. , Colburn, T.R., and Vallerga, S.: Hardware 
model of a tiger salamander cone cell. Proceedings of the 30th ACE>t3 . 
Alliance for Engineering in Medicine and Biology . Bethesda, MD, 1977, 
pp. 369. 

2. (ABSTRACT) Vallerga, S., Covacci, R. , and Pottala, E.W.: Hardware 
simulation of cone cell photoresponse. Seventh Annual Meeting , 
Society of Neurosciences , Bethesda, MD, 1977, pp. 579. 

3. (ABSTRACT) Covacci, R. , Pottala, E.W., and Vallerga, S.: Cone 
responses: a computer-driven hardware model. 50th Ann iversary 
Spring Meeting of the Association for Research in Vision and 
Ophthalmology , 1978, pp. 163. 



143 



SMiTHSOfiur; scie'jce iriFORMATiot; exchange 

PROJECT NUMBER (Oo NOT use this space] 



U.S. DEPARTMENT OF 
HEALTH, EDUCATICIi, AfiD WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PfiOJECT 



PROJECT fJUMBER 



ZOl CT 00005-08 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

Mathematical Modeling of Biological Processes 



NAMES, LABORATORY AfJD I.NSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI; 



J.E. Fletcher 



Chief, Applied Mathematics 

Section LAS DCRT 



OTTTER: M. Bietenaan 

A. Spector 

B . Bunow 



Mathematician, AMS 
Professor, Univ. of Iowa 
Med. School 
Biomathematician, AMS 



LAS DCRT 



LAS DCRT 



COOPERATING UNITS (if any) 
None 



lab/brai.ch 

Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MD 20014 



TOTAL MANYEARS: 
1.2 



PROFESSIONAL: 

1.2 



OTHER: 



CHECK APPROPRIATE BOx(ES) 
n (a) HUMAN SUBJECTS 

□ (al) MINORS □ (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



S (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project consists of three areas of investigation: (1) 
Mathematical Modeling of Substrate Transport in the Microcirculation , 
(2) Approximation Techniques and Numerical Methods for the Solution of 
Transport and Diffusion Processes in Biomedicine, and (3) Mathematical 
>todels of Binding Equilibria . All areas have in common the development 
of conceptual models as mathematical formulations from basic physical, 
biochemical, or biomedical principles. Methods for solution via 
computer are studied; those which appear theoretically sound and 
efficient in practice are used to produce unknovm parameters from 
experimental data, simulate laboratory experiments, and/or to validate 
experimentally determined ranges of variation in biomedical phenomena. 



144 



Pk'^.Anin 



Objectives: 

The objectives of this project are to develop mathematical models 
which can be used to simulate laboratory or clinical experiments and to 
explain, interpret and/or predict physiologic behavior and limits. Such 
models may lead to better understanding of basic biological processes 
and suggest new experimental approaches to problems of biomedical and 
biochemical importance. Because of manpower limitations, only a few 
such models can be considered in depth. These are detailed in the 
following paragraphs. 

(1) Mathematical Modeling of Substrate Transport in the Microcirculation 
Background : 

The mathematical modeling of substrate supply to tissue from the 
microcirculation has been under study since FY 69. The tissue substrate 
of primary concern is tissue oxygen. Such 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, hemoglobin 
characteristics, tissue metabloic rate and other physiologic parameters 
have been examined. The complex interaction of microcirculatory 
geometry, nonlinear oxygen dissociation properties, intracellular 
binding proteins, and substrate dependent metabolic rates requires such 
a detailed description to achieve physiologic validity. These models 
require the solution of coupled distributed parameter models which are 
of a nonlinear type. 

Significance to Biomedical Research: 

Such modeling is necessary to reconstruct local tissue 
microcirculatory dynamics since direct measurements are generally not 
possible and microcirculatory function must be inferred from boundary 
observations. Studies of this type are used to predict tissue oxygena- 
tion and reoxygenation in ischemia, hypoxia, anemia, coronar\' 
obstructions, sickle cell anemia and other conditions of substrate 
pathophysiology. Recent studies have considered the contribution of 
diffusion facilitation through oxygen binding to tissue myoglobin. 

Progress in FY 78: 

The contribution of binding proteins- to substrate movement in 
tissue was considered in FY 73. A new model was developed which 
included the previously neglected effects of flov;, capillary length and 
radius, and dissociation Icinetics in capillary blood. The resulting 
mathematical model required nonclassical solution techniques; a 
methodology for the complete system of equations is currently being 
explored. Preliminary results with simplified versions indicate that 
diffusion facilitation can be quite important at low flou or high 
metabolic rate tissue conditions. 



145 



A review of existing modeling literature in this area in FY 78 
resulted in a published review paper. This article is expected to 
attract and stimulate new modeling efforts on some of the more difficult 
and unresolved questions relating to tissue substrate supply. 

Proposed Course: 

Mathematical techniques for the solution of model equations will be 
explored for accuracy and applicability. New nur.erical techniques 
currently under development will be finalized and applied to the models 
describing the complex interactions of facilitation, metabolism, and 
blood-oxygen extractions. As the approximation technology is developed, 
efforts will be made to consider rhythmic or periodic conditions which 
relate more closely to cardiac and skeletal muscle function. 

(2) Approximation Techniques and Numerical Methods for the Solution of 
Transport and Diffusion Processes in Biomedicine 

Background: 

The mathematical modeling of flow, transport, and chemical reaction 
in the microcirculation has generated mathematical models of a 
nonclassical type. Specific mathematical and numerical techniques to 
treat those equations were not available in the existing mathematical 
literature. The development of such techniques, and the continuing 
search for new efficient and economical techniques is the objective of 
this research area. 

Significance to Biomedical Research: 

The effective modeling of dynamic physiologic processes requires 
efficient methods for the numerical or approximate solution of systems 
of lumped or distributed parameter models. Such models are required to 
describe the complexity of dynamic interacting systems such as flow, 
diffusion, and chemical reaction basic to normal and pathophysiology. A 
key to parametric exploration and simulation by such models is clearly 
the ability to efficiently solve the model equations. The collection 
and characterization of such techniques also makes them available to the 
NIH biomedical community for its general use. 

Progress in FY 78: 

Numerical methods used to solve linear and nonlinear ordinan/ and 
partial differential equations describing physiologic flow and 
diffusion processes were investigated and implemented in computer 
programs. These programs are designed to be useable by those not 
expert in the relevant mathematical theory. The methods include finite 
difference techniques which are appropriate for biological transport 
problems and finite element methods, an approximation procedure studied 



146 



and used extensively in mathematics and structural engineering. This 
method has just recently been applied to nodels describing biological 
problems. Specifically, this method is incorporated in a novel 
iterative continuation scheme used to solve coupled nonlinear 
reaction-diffusion equations with flux boundary conditions. With 
respect to reliability, accuracy and cost, preliminary results compare 
favorably with published solutions of siniilar problens obtained by other 
techniques. 

Proposed Course: 

These programs can be used in -.ore conprehensive sti:dies than 
described in the above applications, and will fom the basic tools for 
computing solutions to such models. Efficient and easy to use numerical 
methods will reduce the cost and decrease the time necessary to simulate 
a laboratory test or wet lab experiment. This effort will continue at a 
level of effort determined by section priorities and manpower. 

(3) Mathematical Models of Binding Equilibria 

Background: 

Mathematical models of macromolecule-ligand binding equilibria, 
have been investigated since 1956. This continuing effort has revised 
many of the concepts related to the binding of ligands to macromolecules 
and has produced an interactive methodology for the fitting of nodels to 
data and other computer oriented tools for the analysis of data from 
laboratory experiments. 

Significance to Biomedical Research: 

The fitting of models to experimentally obtained data is a 
procedure used to determine unknovm parameters in mathematical models. 
The proper choice of a model and the ability to determine the unknown 
parameters is a basic tool of biomedical research. Such procedures 
broaden biomedical laiowledge and add to basic scientific knowledge only 
if the models represent the underlying biological process, and the 
unknown parameters can be readily and accurately determined. A thorough 
and continuing critique of such models and their validity for the 
interpretation of current laboratory and clinical experiments is 
therefore essential to the progress of science. 

Progress in FY 78: 

A generalized analysis of equilibrium models was completed and 
published in FY 78. Examples were constructed demonstrating the non- 
uniqueness of equilibrium models in predicting the microscopic binding 
process. Generalized computer programs to treat the alter-cative models 
were also developed. It is not anticipated that analytical work in this 
area will continue, although applications will continue, to be considered, 



147 



Proposed Course: 

Applications of existing methodology to data analysis will continue 
to be raade as they are requested by collaborating laboratories. Soi?.e 
new concepts in menbrane-receptor studies are being considered and 
applications such as the analysis of cholesterol exchange in lipid 
bilayers are under study. The direction of this project will continue 
to be in the examination of the transport and exchange of lipids, fatty 
acids, and cholesterol in laboratory experiments. 

Publications and Abstracts: 

1. Spector, A. A., and Fletcher, J.E.: Nutitional Effects on Drug- 
Protein Binding. In Hathcock, J.N., and Coon, J. (Eds.): Nutrition 
and Drug Interrelations . New York, Academic Press, 1978 pp. 447-473. 

2. Fletcher, J.E.: A generalized approach to equilibrium models. 
J. of Physical Chemistry 81: 2374, 1977. 

3. Spector, A. A. and Fletcher, J.E.: Transport of Fatty Acid in the 
Circulation. In Dietschy, J.M., Dietschy, J. A., Ontko, J. A., and 
Gotto, A.M. (Eds.): The Physiolotry of Lipids and Lipoproteins 

in Health and in Disease . Bethesda, MD., American Physiological 
Society, 1978, pp. 229-249. 

4. Fletcher, J.E. and Spector, A. A.: Alternative models for the analysis 
of drug-protein binding. Molecular Pharmacology 13: 387-399, 1977. 

5. Spector, A. A. and Fletcher, J.E.: Fatty acid binding by serum albumin. 
In Peters, T., and Sjoholm, I. (Eds.): FFB5 Federation of European 
Biochemical Societies, 11th Meeting Copenhagen 1977 . New York, 
Pergamon Press, 1978, Vol. 50, Coll. B9 , pp. 51-60. 

6. Fletcher, J.E.: Mathematical modeling of the microcirculation. 
Math. Biosciences 38: 159-202, 1978. 

7. Fletcher, J.E.: A singular perturbation solution for a model of 
facilitated diffusion in striated muscle tissues. SIA.2-^ National 
Meeting. May 25, 1978. 



148 



SMITHSONIAN SCIENCE INFORMATION EXCHANGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00035-02 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1973 



TITLE OF PROJECT (80 characters or less) 

Mathematical Models and Simulation Programs in Physiology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 


E. 


Hill Jr. 


OTHER: 


J. 


Fletcher 




E. 


Harris 




B. 


MacLees 




A. 


Robertson 




R. 


El in 



Computer Scientist 

Mathematician, Chief 
Statistician, Chief 
Staff Physician 
Pathologist 
Pathologist, Chief 



LAS DCRT 

LAS DCRT 

LAS DCRT 

ICF CC 

CP CC 



CP 



CC 



COOPERATING UNITS (if any) 
None 



lab/branch 

Laboratory of Applied Studies 



SECTION 

Applied Mathematics Section 



'^'M¥,*^itf.'^^'°'^ 



ethesda, MD 20014 



TOTAL MANYEARS: 



0.5 



PROFESSIONAL: 

0.5 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

□(al) MINORS D (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



[5 (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Tliis project has as an objective the identification and 
classification of the various simulation programs used in physiology . 
These programs are being examined for utility in a clinical environment, 
accuracy of representation of normal and abnormal physiologic conditions, 
and the identification of areas of defective or missing physiologic 
relationships. Selected programs will be tested and evaluated in a 
teaching environment (Clin. Path. Dept., NIK) and an intensive care unit 
(Medical Intensive are facility, NIH) . 



149 



PHS-6040 
(Rev. 10-76) 



Background and Objectives: 

The objective of this project is to investigate the use of 
simulation programs in physiology as diagnostic and patient 
management aids to physicians and other clinical staff. A previous 
contractual effort, conducted in FY 77 by the George Washington 
University, has revealed that many such programs now in limited use 
are deficient in their physiologic description of interdependent 
metabolic pathways. One such deficiency, concerning acid-base 
balance, is now being considered in detail by this laboratory. 

Significance to Biomedical Research: 

Simulations offer the clinician the opportunity to try a 
proposed course of medical treatment on an ideal (computer) patient, 
without the attendant risk of injury to the actual patient. 

Progress in FY 78: -~- 

The computer simulation of Electrolyte and Acid-Base Disorders 
is the first simulation program to be tried in a clinical setting at 
the NIH. This model was chosen because electrolyte and acid-base 
equilibrium data can be measured. 

This program was developed by Dr. K. Bleich. 
Bleich's program is written in MIIS, a dialect of Ml,T!PS. To make 
this program available at NIK on the IBM S/370 computers, it had to 
be translated to Standard MU>CPS. 

Considerable time was spent trying to convert Dr. Bleich's 
Acid-Base Program to Standard MUMPS. Two major software problems 
prevented the completion of this task. The first difficulty resulted 
from errors in the translator and the second problem arose from 
errors in the Standard MUMPS implemented on the S/370 computers. 

After the translation efforts failed, a Fortran IV copy of the 
Bleich program was obtained from Dr. A. Forrey. 

The I/O for this program was redesigned for the IBM S/370 
computers at NIH. This program is now running on the NIH IBM S/370 
computers and efforts are being made to evaluate the model's 
performance in a clinical setting. 

Proposed Course: 

Pending the outcome of feasibility studies, current trial 
applications, and manpower availability, future efforts will consider 
the development of physiologic simulators with accurate descriptions 
of interdependent pathways. Implementation of these programs in a 
variety of clinical settings will follow the developmental stage. 
Continuing efforts, requiring the cooperation of the Clinical 
Pathology Department (CC) , will involve the testing of an acid-base 

150 



balance program with existing data on electrolyte, blood gas, and pH 
measurements from patients before and after specific therapeutic 
intervention. 

Publications and reports: 

None 



151 



SMITHSONIAfJ SCI 
PROJECT NUMBER 



NGE INFORMATION EXCHANGE 
Oo NOT use this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00007-10 



LAS 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

Statistical Research in Clinical Pathology 



NAMES, LABORATORY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI; 



E.K. Harris 



Chief, Lab. of Applied Studies LA.S DCRT 



OTHERS: G. Shakarji 

G.Z. Williams 

S.S. Brown 



E.A. Robertson 



Supv. Systems Analyst 
Director, Institute for Health 
Research, San Francisco, Ca. 
Clinical Chemistry Service 
Clinical Research Centre 
Harrow, England 
Clinical Pathology 
Department 



mm DCRT 



CC 



COOPERATING UNITS (if any) 
None 



LAB/BRAf.CH 

Laboratory of Applied Studies 



SECTION 



INSTITUTE AND LOCATION 

DCRT, NIH, Bethesda, MP 2001A 



TOTAL MANYEARS: 

0.5 



PROFESSIONAL: 

0.5 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
B (a) HUMAN SUBJECTS 

n (al) MINORS □ (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



□ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

In cooperation with Dr. G.Z. Williams and staff of the Health 
Research Institute (San Francisco) , records of some 30 different 
biochemical and hematological tests performed annually over a 4-7 year 
period on several hundred healthy volunteers have been analyzed to test 
the adequacy and usefulness of 3 statistical forecasting models described 
in previous reports and publications. An initial report of this work is 
in preparation. The theory of Empirical Bayesian estimation has been 
explored to determine its usefulness as a way of improving estim.ates of 
individual mean values of biochemical constituents. Initially, applications 
to epidemiological studies have been investigated (Ref. 1). Work 
currently in progress tests the application of this theory to patient 
data from hospital laboratories. Another cooperative study examines the 
distribution of within-person variance in healthy volunteers and its 
effects on reference ranges and other medical diagnostic criteria. 
Finally, joint studies with cooperating pathologists have begun on the 
power of various statistical models and tests to detect short-term trends 
in biochemical quantities. 



PHS-6040 
(Rev. 10-76] 



37" 



Objectives: 

To investigate applications of statistical theory, particularly 
the use of variance components and the theory of discrete time series, 
to the interpretation of clinical laboratory measurements and the 
evaluation of analytic methods. 

Background and Progress during FY 78: 

The database gathered through the health monitoring program of the 
Health Research Institute (HRI) , San Francisco, represents one of the 
largest extant collections of serial biochemistries on normal volunteers. 
Hence it offers an unusual opportunity to study the suitability of 
various recently published statistical models and analytic methods aimed 
at detecting step changes and trends in short series of biochemical data 
from healthy persons. During the past year, in cooperation with 
statistical and computing staff at HRI, data have been collated and 
analysis begun on over 30 biochemical and hematological analytes in 
several hundred individuals who have undergone at least 4 annual 
examinations, with several weekly retests around each time of 
examination. Preliminary results show that a nonstationary , random walk 
model of within-person variation is appropriate for a substantial 
proportion of individual series, particularly hematological variables. The 
practical effects of using a "critical forecasting range" based on both 
stationary and nonstationary models are being explored in this database. 
Initial studies are underway on the power of these models and related 
statistical tests to detect trends in short series of biochemistries in 
the presence of analytic and biological background variation. 

Utilizing a smaller database of weekly measurements on 37 healthy 
male volunteers over a 5-month period, collected in a cooperative study 
with the Clinical Research Centre, Harrow, England, the second in a new 
series of reports on normal variation is being prepared. This analysis 
focuses on the heterogeneity of intrapersonal variation in each of 10 
common analytes, and the effects of such heterogeneity on (a) width of 
reference ("normal") ranges and (b) the applicaMlity of coranonlv 
accepted criteria concerning the amount of change bet^'een succesivp. 
observations which should trigger investipative medical activity. 

A common problem in epidemiological studies and in tlie 
interpretation of hospital laboratory data is the inaccuracy of a single 
observation as an estimate of an individual's underlying true value 
(e.g., blood pressure, alkaline phosphatase, etc, etc.). During this 
past year, an intensive study, both theoretical and applied to real data, 
was undertaken of Empirical Bayesian estimation procedures to improve the 
accuracy of single measurements. These methods involve replacing the 
individual observation by a weighted average of the observation and the 
mean of a reference distribution. The expected gain in accuracy depends 
partly on whether the single measurement is typical or extreme, partly on 
the ratio of average within-person to between-person variation. An 
initial publication of results useful for epidemiological studies will 
appear in late 1978 or early 1979 (Ref. 1). The process of collecting a 



153 



suitable database to test the usefulness of Empirical Bayes estimates in 
assessing clinical laboratory reports is just beginning in cooperation 
with the Clinical Pathology Department, CC, Nil:. 

It was previously reported that statistical criteria to determine 
suitable goals for the accuracy and precision of clinical laboratory test 
methods had been developed in response to a request from the College of 
American Pathologists (CAP) . These criteria v^ere based on stochastic 
models of within-person variation which, together with earlier estimates 
of physiological and analytic variance, were applied to several realistic 
contexts of use of laboratory tests. The invited paper part of a 
conference on the subject of analytic goals sponsored by the CAP was 
published during FY 78 (Ref. 2). In April, 1973, in concert with 
representatives of the CAP, these statistical principles v/ere presented 
to European pathologists at a conference in London under the aegis of the 
World Association of Societies of Pathology. Recommendations of this 
meeting are expected to be transmitted to the World Health Organization 
for distribution to member nations who nay wish to apply them in 
accordance with specific national needs. 

Significance to Biomedical Research: i 

Tl;e definition and estimation of analytic and biological variance 
components provides an essential basis for the objective interpretation 
of clinical laboratory tests in patients and healthy persons alike. The 
development, testing and eventual routine use of stochastic models to 
describe and forecast sequential results of laboratory tests in 
individual cases has already proven useful when applied to periodic 
monitoring of healthy individuals as part of a general program of 
preventive medicine. These methods of data analysis require the use of 
standard computer program packages as well as construction and 
imolementation of special algorithms for computer-based laboratory 
reporting systems. A spin-off from these research efforts has been the 
development of versatile computer systems for storing, updating and 
retrieving serial information on multiple laboratory results for 
individual patients. These systems are currently being employed in the 
Hypertension-Endocrine Branch (::HL3I) and the Arthritis and PJieumatism 
Branch (NIA^TDn) . These developmental and associated consulting 
activities have expanded the services which Dr?,"- offers the Mil' clinical 
community. 

Future Course: 

Collaborative analysis of data collected by t'\e Healt'n Research 
Institute to determine the relative usefulness of stationary v=!. non- 
stationar;/ models for detecting step changes in serial biochemistries^ 
of normal' subjects will probably be concluded during FY 79, as will the 
analvsis of variance heterogeneity based on the more intensive study 
undertaken at the Clinical Research Centre near London. Studies with 
the Clinical Pathology Department at Nil: on application of Empirical 
Bayesian methods to laboratory data are expected to deepen during the 
coming year, while investigations into the power of statistical 

154 



models to detect trends in short series of biochemical measurements will 
expand further, utilizing data from various clinical laboratories. 

Publications: 

1. Harris, E.K. and Shakarji, G.: Use of the population distribution 

to improve estimation of individual means in epidemiological studies. 
J. Chronic Diseases , 1978 (in press). 

2. Harris, E.K.: Statistical principles underlying analytic goal-setting 
in clinical chemistry. Proceedings of the College of American 
Pathologists Second Annual Aspen Conference (1976) on Analytical 
Goals in Clinical Chemistry , pp. 115-135, 1977; to be reprinted as a 
supplement to the Amer. J. of Clinical Pathology , Summer, 1979. 



155 



October 1, 1977 through September 30, 1978 



NATIONAL INSTITUTES OF HEALTH 
DIVISION OF COMPUTER RESEARCH AND TECHNOLOGY 



Summary of Branch Activities 1. DCRT 

2. PHYSICAL SCIENCES LABORATORY 3. Geo rge H. Weiss 



Chief 



SUMMARY 



I 
1 

I 



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 the 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 six 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 Wor k 

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. These collaborative projects 
are done jointly with approximately ten other investigators including 
two major projects with data being generated by off-campus scientists. 

Highlights of the Year's Activities 



During the past year the members of the Physical Sciences Laboratory 
have made significant progress in projects started in past years, 
as well as initiating a few new ones. 

Dr. Ferretti continued his studies of the application of NMR techniques 
to determining conformational properties of small proteins. A particularly 
interesting investigation recently begun is the determination of the 



156 



metal ion binding sites of bleomycin, an important agent in cancer 
chemotherapy. Con|^derable progress has been made in assigning the 
resonance of the C NMR spectra of bleomycin. 

Drs. Parsegian and Tinker conducted an extensive survey of electro- 
static interactions in proteins, using the computer programs developed 
by Richard Feldmann of DCRT. They discovered several features of such 
interactions that help to maintain contact within proteins by an intricate 
matching of positive and negative charges. 

Drs. Nossal and Brenner continued work on the application of laser 
spectroscopy to the measurement of elastic properties of biological 
gels. Experiments are in progress to apply Doppler laser techniques 
to measure tissue blood flow. 

Drs. Darvey and Weiss developed a theory allowing one to specify 
optimal designs and analyses of biochemical kinetic measurements. They 
have examined the effects of different weighting schemes for least-squares 
analysis of the data, as well as the effects of choosing different 
substrate concentrations on the accuracy of estimates of the kinetic 
parameters. Dr. Weiss has developed a general theory of interpolation 
error in computerized tomography with parallel ray machines. The results 
enable one to choose the number of views required to achieve given resolution. 

Dr. Nossal has been appointed to the Board of Editors of the Biophysical 
Journal and Dr. Weiss has been appointed to that of Separation Science 
and Technology. Dr. Parsegian has been presented with the NIH Director's 
Award . 

Theory and Application of Nuclear Magnetic Resonance Spectroscopy 

A continuing subject of investigation is the optimal design of NMR 
experiments to insure specified accuracy in the determination of relaxation 
times^in the minimum time. We have found that time savings of the order 
of 30% can be achieved by proper design, over the standard inversion- 
recovery method. Different approximation techniques are being applied 
to the Bloch equation to extend its use to situations of experimental 
use. In particular, perturbation theory has been used to demonstrate 
that correlation spectroscopy is practical even when the spin system 
is driven into a nonlinear response region. 

Consulting Services 

A considerable amount of time has been spent in developing techniques 
to correct and query a large data base on head injured veterans of 
Vietnam. An important conclusion allowed by our work to date is that 
the use of anticonvulsant drugs has not led to a decrease in the indicence 
of post traumatic epilepsy. A further problem in which progress has 
been made is that of interpolation error in computerized tomography. 
We have developed an error analysis applicable to asymmetric objects 
that allows one to relate the errors to the number of views available 
and to the interpolation method that is used. 

157 



II 

I 



II. PSL PROJECTS AND ACTIVITIES FY 78 



Theory of Biochemical Separation Techniques. George H. Weiss, PSL, 
in collaboration with Dr. D. A. Yphantis, University of Connecticut. 



I 

This project develops mathematical theory for the planning and interpretatiori 
of experiments with such techniques as chromatography, electrophoresis, " 
and ultracentrifugation. J 

The Role of Electrostatic Forces in the Organization and Properties I 
of Macromolecular Systems . Stephen L. Brenner, PSL, V. A. Parsegian, 
PSL. This project studies the role of electrostatic forces in determining J 
the mutual arrangement and interaction of macromolecules in aqueous ( 
salt solutions. (This project is inactive) 

Theory and .Application of Nuclear Magnetic Resonance Spectroscopy . J 

James A. Ferretti, PSL, George H. Weiss, PSL. This project includes 

the development of new methodology in nuclear magnetic resonance and 

the application of NMR to elucidate chemical properties of molecules ] 

of biological interest. ' 

Correlation Function Spectroscopy/Laser Light Scattering. Ralph J. 

Nossal, PSL, Stephen Brenner, PSL. The laser inelastic light scattering I 

spectrometer is being used to measure elastic parameters of gels as 

well as parameters of cell motility. Several collaborative experiments 

are being planned using optical techniques developed by this project. 

Cell Motility and Chemotaxis. Ralph J. Nossal, PSL, Stephen L. Brenner, 

PSL, and George H. Weiss, PSL. This project develops methodology for 

the interpretation of experiments related to cell locomotion and chemotaxis. 

Theor;^' and Measurement of Intermolecular Forces. V. A. Parsegian, 
PSL, George H. Weiss, PSL, James E. Kiefer, PSL. The object of these 
studies is to develop the theory of electrodynamic forces in biological 
media, and to develop experimental methods for measuring these forces. 

Studies in Mathematics and Statistics . George H. Weiss, PSL, James 
E. Kiefer, PSL. Several disparate studies are included in the project. 
The present set of studies include the optimization of enzyme kinetic 
experiments, and an analysis of the dimensions of contrained polymer 
chains. 

Consulting Services. George H. Weiss, PSL, Ralph J. Nossal, PSL, and 
James E. Kiefer, PSL. Members of the PSL give consulting assistance 
to other scientists at NIH and elsewhere, in the areas of the physical 
sciences and applied mathematics. 

III. PUBLICATIONS 

Blumenfeld, D. E., Weiss, G. H. : Sampling errors in the measurement 
of traffic noise. J. Sound and Vib . 53, 111-116 (1977). 



158 



Blumenfeld, D. E., Weiss, G. H. : Curve fitting the probability 
distribution of acoustic noise from freely flowing traffic. Transp 
Res. 12, 111-114 (1978). 

Blumenfeld, D. E., Weiss, G. H. : Can nonlocal gap acceptance functions 
be distinguished from local ones? Transp. Res, (to appear). 

Brenner, S. L., Nossal, R. J.: Correlation functions for light 
scattering from soft gels. Macromol. 11, 207-212 (1978). 

Brenner, S. L., Gelman, R. A., Nossal, R. J.: Laser light scattering 
from soft gels. Macromol. 11, 202-207 (1978). 

Brenner, S. L. , Nossal, R. J., Weiss, G. H. : Number fluctuation 
analysis of random locomotion. J. Stat. Phvs. 18 , 1-18 (1978). 

Brenner, S. L. , Parsegian, V. A., Gingell, D. : The effects of 
image forces on double-layer interactions. J. Phys. Chem. (to appear). 

Brooks, R. A., Weiss, G. H. , Talbert, A.: A new approach to interpolation 
in computed tomography. J. Computerized Tomog. (to appear). 

Chew, D., Weiss, G. H. , Brooks, R. A., Di Chiro, G .: Effect 
of noise on the detectability of test objects. Am. J. Roentgen, (to 
appear) . 

Correia, J. J., Weiss, G. H. , Yphantis, D. A.: An extrapolation 
method for reducing equilibration times in sedimentation equilibrium 
experiments. Biophys. J. 20. 153-168 (1977). 

Cowley, S., Fuller, N., Rand, R. P., Parsegian, V . A. : Measurement 
of repulsion between charged phospholipid bilayers. Biochem. (to appear). 

Dishon, M. , Weiss, G . H. : When do transient double peaks occur 
in pH gradient electrophoresis? Analyt. Biochem. 81. 1-9 (1977). 

Dishon, M. , Weiss, G. H. : Numerical inversion of Mellin and two- 
sided Laplace transforms. J. Comp. Phys. (to appear). 

Gail, M. H., Weiss, G. H. , Mantel, M. , O'Brien, S. J.: A solution 
to the generalized birthday problem with application to allozyme screening. 
J- Appl. Prob. (to appear). 

Gingell, D. , Parsegian, V. A., Todd, I.: Experimental evidence 
for long-range attraction between a red cell and a hydrocarbon surface. 
Nature 268, 767-768 (1977). 

Kiefer, J. E., Parsegian, V. A., Weiss, G. H. : An easily calculated 
approximation for the many-body van der Waals attraction between sphere 
and wall. J. Coll. Interf.Sci. 63, 161-162 (1978). 



159 



Kiefer, J. E., Parsegian, V. A., Weiss, G. H. : Some convenient 
bounds and approximations for the many body van der Walls attraction 
between spheres. J. Coll. Interf. Sci. (to appear). 

Lis, L. J., Rand, R. P., Parsegian, V. A.: Measurement of electro- 
static forces between lecithin bilayers charged by divalent cations. 
Biophys. J. 21, 213a (1978). 

McAlister, M. , Fuller, N. , Rand, R. P., Parsegian, V. A.: Measurement 
of surface pressure in and repulsion between approaching phospholipid 
membranes. Biophys. J. 21, 213a (1978). 

McNeil, D. R. , Weiss, G. H. : A large population approach to the 
estimation of parameters in Markov population models. Biometrika 64, 
553-558 (1977). 

Nossal, R. J.: Quasielastic laser light scattering by flexible 
polymer networks. J_^ Appl . Phys . (to appear). 

Nossal, R. J.: Factors affecting the reliability of capillar MIF 
(Migration Inhibition Factor) assays, in Theoretical Immunology , Marcel 
Decker, 1978, 121-142. 

Oppenheim, I. 0., Shuler, K. E., Weiss, G. H. : Stochastic theory 
of nonlinear rate processes with multiple stationary states. Physica 
88A , 191-214 (1977). 

Parsegian, V. A., Weiss, G. H. , Schrader, M. E.: Macroscopic 
continuum model of influence of hydrocarbon contaminant on forces causing 
wetting of gold by water. J_^ Coll. Interf. Sci. 61 , 356-361 (1977). 

Parsegian, V. A. : Considerations in determining the mode of influence 
of calcium on vesicle membrane interaction. Soc . for Neurosci ■ Symp . 
11. 161-171 (177). 

Rubin, R . J., Weiss, G. H. : Boundaries of constrained random 
flight polymer chains. Macromol. (to appear). 

Rubin, R. J., Weiss, G. H. : Span of a random flight model of a star- 
branched polymer chain. Macromol. 10 , 332-334 (1977). 

Simon, R. , Hoel, D. G., Weiss, G. H.-: The use of covariate information 
in dichotomous response experiments. Comm. in Stat . . A6 , 777-788 (1978). 

Tinker, D. 0. , Parsegian, V. A. : Interactions between subunits 
of protein oligomers. Biophys. J. 145a , 142 (1978). 



160 



>( 



Weiss, G. H. : Comments of a model of polymer growth. J Coll 
Interf. Sci. 61. 199200 (1977). ~^ - 

Weiss, G. H. , Brooks, R. A.: Integration errors in image reconstruction 
of circularly symmetric objects. Theory and Application of Statistical 
Mechanics , Plenum Press, 1977, 669-684. 



Sc 



Weiss, G. H. : Transport equations with quadratic nonlinearities . Sep. 
i. and Technol. (to appear). — 



161 



SWITHCC.'IIAf; :CI£NCc i:,FCRWATIC,'( EXCHANGE 
PROJECT NUMEcR (Oo NOT use this space) 



U.S. OEPARTMc.'lT OF 
HEALTH, EO'JCATICfi, A:;j WELFARE 
PUBLIC HEALTH lEHVlCE 
NOTICE Of 
IKTRANUfiAL RESEARCH PROJECT 



PROJECT ,'JU,M3ER 



ZOl CT 00017-06 PSL 



PERIOD COVERED 

October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (30 characters or less) 

Cell Motility and Chemotaxis 



NAMES, LABORATORY A^.O INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PEHSCNNEL ENGAGED ON THE PROJECT 



PI: R. J. Nossal, Research Physicist, PSL, DCRT | 
Other: G. H. Weiss, Chief, PSL, DCRT | 
S. L. Brenner, Staff Fellow, PSL, DCRT 

1 

1 
1 


COOPERATING UfMTS (if any) 

L. Lipkin, M.D., Image Processing Unit, DCBD, NCI 


lab/branch 

Physical Sciences Laboratory 


SECTION 


INSTITUTE AND LOCATION 

Div. of Computer Research & Technoloev. NTH. Bprhpsdfl. Mprvl^nH 


TOTAL MANYEARS: 
0.7 


PROFESSIONAL: 
0.5 


OTHEJ?: 

0.2 ( 


CHECK APPROPRIATE B0X(E3) 
D (a) HUMAN SUBJECTS R (b) HUMA/I T 1 SSUES ^(0) NEITHER 

n(al) MINORS n (a-) INTERVIEWS 


SUMMARY OF WORK (200 words or less - underline keywords) 

This project has been undertaken to study various aspects of 
cell locomotion and chemota'xis. Analytical expressions to quantitate capillary 
migration (MIF) assays have been derived. New procedures for measuring 


macroscopic coefficien 
computer assisted trac 
lymphokines affect the 


ts of cell migration are being developed, including 
king techniques. Studies of the manner in which 
migration of individual leukocytes are in progress. 

162 


PHS-60-J0 
(Rev. 10-76) 







Ce ll Motility and Chemotaxis 

This study concerns phenomena relating to cell locomotion and chemotaxis. 
Recent emphasis has been on examining certain immunologic aspects of leukocyte 
migration. 

Recently, as part of this project, a general mathematical theory for 
interpreting results of capillary migration assays for cellular immune 
sensitivity (MIF tests) was derived. However, various basic parameters of 
leukocyte movement which are necessary for quantitating the assay are not well- 
known. Consequently, collaborative experiments now are being performed with 
Dr. Lewis Lipkin (DCBD/NCI) which involve studying the response of neutrophils 
to various chemical factors ( "lymphokines") produced by stimulated lymphocytes. 
Specialized measurement techniques have been devised, an example of which 
is a scheme where occupation number fluctuations are analyzed to determine 
mobility coefficients of migrating cells. Also, algorithms and computer 
programs have been developed in order to adapt an automated microscope 
system for cell tracking experiments. In addition, this instrument will be 
used to examine the behavior of neutrophils when responding to chemoattractants 

Publications : 

Nossal, R. : Factors affecting the reliability of capillar MIF 
(Migration Inhibition Factor) assays. Chapter 5 in Theoretical Immunology , 
ed. G. I. Bell, A. S. Perelson and G . H. Pimbley, Marcel Dekker, N.Y., 
pp. 121-142 (1978). 

Brenner, S. L., Nossal, R. and Weiss, G. H. : Number fluctuation analysis 
of random locomotion. Statistics of a Smoluchowski process. J. Stat Phys 
18, 1-18 (1978). '-^ 



163 



feMITHGOriUr; SCI£f.C£ !.'.FCRMAT1CM EXCHAtlGE 
PROJECT fJUMBEfi (Do WOT a:e this space) 



U.S. DEPARTMENT OF 
HEALTH, EDUCATICfi, AHO WELFARE 
PUBLIC HiALT-i :£RVICE 
NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



PROJECT f.UM3ER 



ZOl CT 00022-11 



PSL 



PERIOD COVERED 



October 1, 1977 to September 30. 1978 



TITLE OF PROJECT (80 characters or less) 
Consulting Services 



NAMES, LASCRATORY Af.D INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 



G. -H. Weiss, Chief, Physical Sciences Laboratory , PSL, DCRT i 



COOPERATING UNITS (if any) 

W. F. Caveness, M.D., Chief, LEN, NINCDS 

R. A. Brooks, Ph.D., SN, NINCDS 

3 — H — GPiU , M n,, Ph.D., Riometry, NCI 



LAB/BRANCH 



Ph 



ly.sirfll ^ripnrpc! T.ahn-rafnry 



SECTION 



NSTITUTE AND LOCATION 

Division of Computer Resparrh F. Tprhnnlnoy^ NTH^ Rp|-hpsHa^ Mfl 7C)n^L 



TOTAL MANYEARS: 

0^6 



PROFESSIONAL: 



0.6 



OTHER: 



Ji^ 



CHECK APPROPRIATE BOx(ES) 
H (a) HbMA.N SUBJECTS 

n (al) MINORS G (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



□ (c) NEITHER 



SUMMARY OF WORK (200 «orts or less - underline keywords) 

Members of the Physical Sciences Laboratory provide consulting 
services to the scientists and physicians at NIH in the areas 
of applied mathematics, statistics, and the physical sciences. 
A continuing project is the construction and querying of a large 
data base on head injured veterans of Vietnam. We have also 
developed an error theory for interpolation in computerized tomography 
In a third project we have developed an accurate approximate 
solution to the generalized birthday problem , a classical problem 
in probability, with application to a genetic assay . 



164 



PHS-6G<0 
(Rev. 10-76) 



Consulting Services 

A large effort has been put into the correction and querying of a large 
data base on head injured Vietnam veterans. Most of the effort has been 
in the correction of coding errors and errors that have been made in mis- 
interpretation of battlefield and hospital records. So far approximately 
13,000 records on 1,030 veterans are included in the data base. Several 
studies are either being planned or are presently in progress. Among these 
are the effects of anticonvulsant drugs on the supression of post traumatic 
epilepsy, and the utility of different operations on veterans with injuries 
to both lobes. A preliminary conclusion of the analysis is that treatment 
with anticonvulsant drugs does not lessen the incidence of past traumatic 
epilepsy. This conclusion follows by a comparison of data from the Vietnam 
war with data from the two World wars and the Korean war. This work was 
done as a collaborative with Dr. W. F. Caveness, NINCDS, and several other 
neurosurgeons outside of NIH. 

We have continued our study of interpolation errors in computerized 
axial tomography with Dr. Rodney Brooks of NINCDS. Specifically we have 
derived a formula for the interpolation error due to an offset delta function 
image. This will enable us to study the interpolation error for asymmetric 
images. Another study on computerized axial tomography relates to the 
effects of noise on the identification of output images. Together with 
E. Chew of the University of Maryland and R. A. Brooks we designed and analyzed 
an experiment to test whether smoothing helps in the detection of different 
images and whether there is any difference in detectability when images 
are presented on a console or as polaroid pictures. We found, using 50 
untrained observers and 8 radiologists that smoothing improves detectability 
in some, but not all, instances, and that there are no significant differences 
in presentation by console or by photograph. 

Together with Dr. Mitchell Gail of NCI we have derived an accurate 
approximate solution to the generalized birthday problem. This problem 
arises in the context of a genetic assay. The solution to the problem 
involves a novel form of perturbation theory and may be applicable to 
many other combinatorial problems. The theory was tested on assay data 
furnished by S. J. O'Brien, NCI, on the contamination of cell cultures. 
The theory was found to be accurate over the entire range of probabilities. 

Keyword Descriptors: Head injuries, post traumatic, epilepsy, anticonvulsant 
drugs, computerized axial tomography, interpolation errors, smoothing, 
birthday problem. 



165 



n 



Publications 



Weiss, G. H. and Brooks, R. A.: Integration errors in image reconstructj 
of circularly synmietric objects in Theory and Application of Statistical f 
Mechanics , Plenum Press, 1977, 669-684. «l 

Chew, E., Weiss, G. H. , Brooks, R. A., Di Chiro, G.: Effect of i 
noise on the detectability of test objects. Am. J. Roentgenology (to appear)! 

Gail, M. H., Weiss, G. H. , Mantel, N. , O'Brien, S. J.: A solution | 
to the generalized birthday problem with application to allozyme screening. | 
J. Appl . Prob. ( to appear) . 

Brooks, R. A., Weiss, G. H. , Talbert, A. J.: A new approach to I 
interpolation in computed tomography. J. Computed. Tomog. (to appear). 



166 



3M1THS0N1AI1 SCIENCE I ;;FOR,'-'AT 1 Oil £XCHA?iGE 
PROJECT NUMBER (Oo NOT use this space) 



U.S. DEPARTMENT CF 
HEALTH, EDUCATION, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAHURAL RESEARCH PROJECT 



PROJECT NUMBER 



ZOl CT 00021-07 PSL 



PERIOD COVERED 

October 1. 1977 to September 30. 1978 



TITLE CF PROJECT (80 characters or less) 

Correlation Function Spectroscopy/Laser Light Scattering 



NAHES, LASORATCRY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 
Other: 



R. J. Nossal, Ph.D., Research Physicist, PSL, DCRT 
S. L. Brenner, Ph.D., Staff Fellow, PSL, DCRT 



COOPERATING UNITS (if any) 

H. Saroff, Ph.D., Laboratory of Biophysical Chemistry, NIAMDD 
J. Gladner, Ph.D., Laboratory of Biophysical Chemistry, NIAMDD 
R. Bonner, Biomedical Engineering & Instrumentation Branch, DRS 

LA3/ branch"' R. Geiman, Ph.D". , Laboratory of Biochemistry, NlTTR 
Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 



Di'v- of Cnmp 



^pr Rpsparrh ft Tpchnnlngy, NIH, Bpthpsda. Md , 



TOTAL MAIiYEAhG: 



n. 9 



PROFESSIONAL: 



0.7 



OTHER: 



n.2 



CHECK APPROPRIATE BOX(ES) 
G (a) HUMAN SUBJECTS 

[J (al) MINORS □ (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



□ (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

Experimental and theoretical studies have been performed to develop 
laser inelastic light scattering methods for studying biological gels . 
The technique is being used to examine the strength of fibrin clots. Other 
studies are being performed in order to 'understand how laser Doppler technique 
can be used to measure capillary blood flow in tissues. 



167 



FHS-6040 
(Rev. 10-76) 



Correlation Function Spectroscopv/Laser Light Scattering 



(1 
I 



1 



The primary objective of this project is development of laser inelastic 
light scattering techniques for performing rapid and precise measurements on 
biological systems and materials. In principle, any process giving rise to 
refractive index fluctuations can be monitored. The instrument which we hav 
constructed has been used to measure diffusion coefficients of macromolecules 
swimming speed distributions of motile microorganisms, elastic moduli of gelai 
and blood flow in capillaries. j 

During the past year we perfected a scheme for measuring elastic J 

coefficients of dilute polymer networks and soft biological gels. Several I 

collaborative studies utilizing this new technique are in progress; one, " 
for example, involves relating the mechanical strength of fibrin clots to the 

nature and extent of interchain cross-linking (with H. Gladner, NIAMDD) . I 

Techniques for studying stiff gels also have been devised, and experiments I 
with polyacrylamide gel models have been performed (with R. Gelman, NIDR) . 

Necessary mathematical theory has been developed. I 

Also, experiments have been undertaken in regard to laser Doppler 
measurements of tissue blood flow. Dr. R. Bonner (BEIB) is developing a | 
laser Doppler flowmeter for clinical use, and our spectrometer is being used J 
to examine related questions concerning light scattering from surface micro- 
vasculature. Synthetic flow models have been devised to explore relation- 
ships between measured spectra and such variables as blood density, flow rate! 
and back-scatter illumination from surrounding tissue. ' 

Keyword Descriptors: 

Laser light scattering, macromolecules, diffusion coefficients, correla-' 
tion functions, gels, blood flow. 

Publications : 



Brenner, S. L. and Nossal, R. : Correlation Functions for light scatter-] 
ing from soft gels. Macromolecules 11, 207-212 (1978). 

Brenner, S. L., Gelman, R. A., and Nossal, R. J.: Laser light scatteril 
from soft gels. Macromolecules 11 , 202-207 (1978). 

Nossal, R. : Quasielastic laser light scattering by flexible polymer 
networks. J. Appl. Phvs. (to appear). - 



168 



IsMlTnSCtilAr; ZC'.'JXi I .'.FCSVATI ON iXCHA.NGE 
PROJECT MUMBEfi (Oo KOT use this space) 



U.S. D£PARTV£.'(T OF 
HEALTH, E^iUCATIC;, A:,: .ELFARt 
PUBLIC HEALTm :£F,VICE 
NOTICE OF 
INTRAMUSAL ReSEARCH PfiOJECT 



PROJECT NUMEEH 



ZOl CT 00024-03 



PSL 



PERIOD COVERED 



October 1, 1977 to September 30, 1978 



TITLE OF PROJECT (80 characters or less) 

Studies in Mathematics and Statistics 



NAMES, LASCRATORY A:,3 INSTITUTE AFFILIATIONS, ANO TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSICNAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 
Other: 



George H. Weiss, Chief, Phys. Sci. Lab, PSL,DCRT 

D. G. Hoel, Chief, Biometry Branch, EBB, NIEHS 

E. Blumenfeld, Lecturer, University College, London 

J. Rubin, Senior Scientist, NBS 

G. Darvey, Visiting Scientist, PSL, DCRT 

0. Tinker, Visiting Scientist, PSL, DCRT 

E. Kiefer, Research Mathematician, PSL, DCRT 



COOPERATING UNITS (if any) 
None 



LAB/BRANCH 

Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATIC;. 



Division of Computer Research & Technology, NIH, Bethesda, MP. 7C)C)^L 



TOTAL MANYEARS: 

0.5 



PROFESSIONAL: 



0.5 



OTHER: 



CHECK APPROPRIATE 60x(£S) 
n (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



_Q.J^ 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

This project includes several unrelated studies in applied mathematics 
Most of the time spent on this project was in the are of optimization 
of enzyme kinetics experiments. Methodology was developed for enzymes 
that follow Michaelis-Menten kinetics. Further work was done on the 
theory of the dimensions of constrained polymer chains . 



169 



PHS-6040 
(Rev. lC-76) 



Studies in Mathematics and Statistics 



I 

( 
t 



We have examined several aspects of the problem of optimizing 
kinetic experiments for enzymes that follow Michaelis-Menten kinetics. 
The first is a specification of the types of error that can occur in 
kinetic experiments, i.e., instrumental and pipetting errors. These 
have different effects on the behavior of observed variances of measured 
initial velocities as a function of the variable parameters. In particular, m 
we have been able to explain some experimental measurements in the biochemic^ 
literature on the basis of the distinction between absolute and relative 
errors suggested by an analysis of error sources. A second problem 
relates to the specification of optimal sets of substrate concentration I 
for single stage design of Michaelis-Menten experiments. A third problem ' 
in this area involves the comparison of different forms of weighting 
in least squares analysis of Michaelis-Menten experiments. This project I 
involved simulation of the experiments and data analysis on both the { 
nonlinear and linearized form of the equations. Surprisingly, a properly 
weighted form of the linearized equation led to the most accurate determinatil 
of parameters. Our results suggested the use of hybrid weighting for j 
improved accuracy. This does not seem to have been discussed before 
in the statistical literature. A subject for further investigation 
is that of two stage optimization in which data collected in the first 
stage is used for optimization in the second stage of an experiment. 

We have, in collaboration with Dr. R. J. Rubin, analyzed the maximum i 
extension of polymer chains that are constrained in various ways by I 

using random walk models. 

i 
Keyword Descriptors : experimental design, Michaelis-Menten reactions 
least squares analysis, simulation, polymer chains. 

Publications : 

Oppenheim, I. 0., Shuler, K. E., Weiss, G. H.: Stochastic theory 
of nonlinear rate processes with multiple stationary states. Physica 
88A , 191-214 (1977). 

Weiss, G. H. : Comments on a model of polymer growth. J_^ Coll . 
Interf. Sci. , S_l, 199-200 (1977) 

Simon, R. , Hoel, D. G., Weiss, G. H. : The use of covariate information 
in the sequential analysis of dichotomous response experiments. Comm. 
in Stat. A6, 777-788 (1977). 

Blumenfeld, D. E., Weiss, G.H.: Sampling errors in the measurement 
of traffic noise. J. Sound and Vib. 53, 111-116 (1977). 



170 



McNeil, D. R ., Weiss, G. H. : A large population approach to 
estimation of parameters in Markov population models. Biometrika 64 
553-558 (1977). ^ ' 

Blumenfeld, D. E., Weiss, G. H. : Curve fitting the probability 
distribution of acoustic noise from freely flowing traffic. Transo 
Res. 12, 111-114 (1978). ^ 

Dishon, M. , Weiss, G. H. : Numerical inversion of MeUin and two- 
sided Laplace transforms. J. Comp. Phys . (to appear). 

Rubin, R. J., Weiss, G. H. : Span of random flight model of a 
star-branched polymer chain. Macromol. 10, 332-334 (1977). 

Blumenfeld, D. E., Weiss, G. H. : Can nonlocal gap acceptance 
functions be distinguised from local ones? Transp. Res, (to appear). 

Rubin, R. J., Weiss, G. H. : Boundaries of constrained random flight 
polymer chains. Macromol (to appear). 



171 



SMITHSCNIAfJ GCI^fiCE i:;FCRMATI ON £XCHA,'1GE| 'J.S. DEPARTMENT OF | PROJECT NUMBER 

PROJECT ,'iU.MSER (Oo NOT use this space) HEALTH, EOUCATICN, A;;D WELFARE 

' PUBLIC HEALTH CEflVlCE 

NOTICE Of 
INTRAMURAL RESEARCH PROJECT 



ZOl CT 00025-03 



PSL 



PERIOD COVERED 
October 1, 1977 to September 30, 1975 



TITLE Of PROJECT (80 characters or less) 

Theory and Application of Nuclear Magnetic Resonance Spectroscopy 



NA.MES, LABCRATCRY AND INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 

PI: James A. Ferretti Ph.D., Research Chemist, PSL,DCRT 

Other: E. D. Becker, Chief, Laboratory of Chemical Physics, LCP.NIAMDD 

G. R. Marshall Professor of Physiology, Dept. of Physiology, 

and Biophysics, Washington University School of Medicine, 

St. Louis, Mo. 

G. H. Weiss, Chief, Physical Sciences Laboratory, PSL,DCRT 



COOPERATING UNITS (if any) 

Laboratory of Chemical Physics, NIAMDD 



lab/branch 

Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 
Div. of Computer Research & Technology, NIH, Bethesda, Maryland 



TOTAL MANYEARS: 
1.5 



PROFESSIONAL: 



1.5 



OTHER: 



CHECK APPROPRIATE BOX(ES) 
n (a) HUMAN SUBJECTS 

n (al) MINORS G (a2) INTERVIEWS 



□ (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF WORK (200 words or less - underline keywords) 

The purpose of this project is to develop new methods in nuclear 
magnetic resonance spectroscopy and also to apply NMR to the study of small 
proteins. In particular, development of the correlation method of 
obtaining NMS spectra is of special interest. An experimental and 
theoretical study of interference effects in correlation spectroscopy 
has been undertaken. Saturation effects in correlation NMR are currently being 
studied. Investigations of the solution conformation of derivatives of 
Angiotensin , Bradykinin and Bleomycin are in progress. In these systems 
we have demonstrated the importance of the contribution of internal 
motion to the relaxation behavior. 



172 



PHS-6040 
(Rev. 10-76) 



We are continuing to study spin-lattice relaxation times, T^, spin-spin 
relaxation times, T„ and nuclear Overhauser enhancement (NOE) factors in 
small peptide systems. These systems include Angiotensin II, Bradykinin, 
Tetragastrin, and various derivatives. The magnetic field dependence of 
T , T and the NOE has been carried out at 25 MHz, 45 MHz, and 67.9 MHz, 
for the aliphatic carbon atoms of these various peptide systems. From these 
data we have been able to evaluate the effective correlation times for both 
overall and internal rotation in these molecules. The studies have allowed 
us to specify the nature of the molecular association in solution as well 
as to study the effects of constraints at the receptor-bound conformation 
of these hormones. Another important conclusion demonstrated by these studies 
is that the neglect of internal flexibility of these molecules leads to 
erroneous conclusions concerning the magnitudes of the overall correlation 
times. Measurements at a minimum of two magnetic field strengths are 
necessary to evaluate these molecular parameters. 

We have initiated studies on Bleomycin (m.w. •-■ 1500). It is a glyco- 
peptide antibiotic of low toxicity and is effective against a variety of 
human neoplasms. Its complexes with radioactive metals are useful as tumor 
scanning agents for diagnostic purposes. While the Cu(II), Zn(II), and Co(II) 
derivatives of Bleomycin are inactive, the Fe(II) complex exhibits enhanced 
DNA degrading activity. Initially we studied the carbon-13 NMR spectra of 
Bleomycin and completely assigned all the 55 resonances. We have also 
carried out T and NOE studies at 45 MHz and 67.9 MHz as the basis for carry- 
ing out metal ion binding studies. Preliminary results on Cu(II) and Mn(II) 
have permitted us to locate tentatively the metal ion binding site in 
Bleomycin. 

Non-linear effect including saturation in NMR correlation spectroscopy 
of uncoupled spin systems has been investigated. The basis of this study 
has been to carry out a perturbation expansion of the Bloch equation in 
order to obtain the first correction term to the response of the spin system. 
We find that for slow sweep rates, the rules of correlation spectroscopy 
hold only for very low power levels. In the case of fast passage the rules 
hold for power levels which correspond to effective flip angles of at least 
60°. These results demonstrate the practicality of correlation spectroscopy 
even when the spin system is driven into a non-linear response region. 

We have continued to analyze various aspects of NMR spin-lattice relaxa- 
tion time T , studies determined by the inversion-recovery, fast inversion- 
recovery and saturation recovery methods. We have found under which 
conditions fast inversion-recovery or saturation recovery should be the 
selected technique. We find that under optimum conditions, typical time 
savings of about 30 percent may be realized relative to the standard 
inversion-recovery method. 



173 



SMITHGOMAN SCIENCE IJ.FCRMATIC! 
PROJECT NUMBER ;Do NOT use ihi: 


1 EXCHANGE 
. space) 


U.S. DEPARTMENT CF 
HEALTH, EIJCATICN, AND WELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 
INTRAMURAL RESEARCH PSOJECT 


PROJECT NUMBER i ' 

ZOl CT 00026-03 PSL^_ 


PERIOD COVERED 1 
October 1, 1977 to September 30, 1978 * 


TITLE Of PROJECT (80 characters or less) 1 


Theory and Measurement of Intermolecular Forces 1 


NAMES, LABORATORY AND INSTITUTE Af F IL 1 AT 1 CfiS, AND TITLES Of PRINCIPAL INVESTIGATORS AND ALL OTHER 1 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT | 


PI: V. A. Parsegian, PSL.DCRT 


G. H. Weiss, PSL,DCRT f 


D. 0. Tinker, University of Toronto | 


J. E. Kiefer, PSL, DCRT 


Others: R. P. Rand, Brock University | 


L. Lis, Brock University 1 


S. Cowley, Brock University 


M. McAlister, Brock University - 


N. Fuller, Brock University 1 


D. Gingell, Middlesex Medical School ' 

1 


COOPERATING UMTS (if any) 


None 


lab/branch 


Physical Sciences Laboratory 


SECTION 


INSTITUTE AND LOCATION 


Division of Computer Research & Technology, NIH, Bethesda, MD 


TOTAL MAMYcARS: 


PROFESSIONAL: 


OTHER: 


2 


2 


0.3 1 


CHECK APPROPRIATE =Ox(ES) ' 


D (a) HUMAN SUBJECTS Q (b) HUMAN TISSUES Q (=) NEITHER 


□(al) MINORS n (a2) INTERVIEWS 


SUMMARY OF WORK (200 -ords or less - underline key^iords) 


This project aims to understand the role of long range forces in 


biological phenomena. A major topic has been the measurement of forces betweer 


phospholipid bilayer membranes immersed in water. We are also' measuring inter- 


molecular forces between lipids in the same membrane. A new technique has ■; 


been devised for detecting the reaction of charged particles with membrane 


surfaces . 


Using the DCRT electronic display we have found two classes of forces 


between protein molecules forming dimers and tetramers (such as hemoglobin). 


These are intricate fitting of electrostatic charges and contact between 


hydrocarbon surfaces which are repelled by water. 


174 



PHS-604Q 
(Rev. 10-76) 



Theory and Measurement of Intermolecular Forces 

We continue to progress in developing methods for computing and detecting 
intermolecular forces. This has been accomplished in model cell membranes 
immersed in water, cells adhering to surfaces, protein particles forming 
functioning dimers and oligomers and in the reaction of charged particles 
with membrane surfaces. 

With Professor Peter Rand of Brock University we have successfully made 
measurements of forces between phospholipid bilayer membranes. This includes 
several estimates of the van der Waals force between bodies in water. The 
results published so far are the beginning of a series of systematic studies 
on the physical properties of cell membrane lipids. We hope to extend these 
Chen to direct observations of forces between natural cell membranes. So 
far we have worked successfully with several model preparations. We have 
measured the important electrostatic repulsion between charged membranes. 
We have been able to measure also forces between molecules within the same 
membrane. This finding has opened up new means to determine the mechanical 
properties of artificial and natural membranes. 

In one particular instance we have observed the adsorption of charged 
particles (ions) to membrane surfaces by measuring the effect of that 
adsorption on electrical forces between membranes. The force measurement 
then becomes a measure of the reactivity of ion with surface as it depends 
on electrical potential of the reacting surfaces, distance between membranes, 
and the ionic conditions of the bathing medium. 

We have analyzed the contact of chemical delivering vesicles with nerve 
membranes to distinguish the role of vesicle membrane nerve membrane forces. 
It seems that the enormous forces required to remove water from membrane 
surfaces act as a serious barrier to contact and to the fusion of vescicle 
and membrane necessary for delivery of chemicals across the nerve wall. From 
our new knowledge of these forces it appears that chemical modification of 
membrane must occur to allow chemical release from nerve. 

With Dr. David Gingell of London, England, we have been studying the 
interactions of red cells with each other and with artificial materials. 
Guided by the theory of forces as developed in this laboratory, we have 
devised experiments demonstrating that cells can be held to surfaces by 
long-range electromagnetic forces, which have now been shown to act at 
thousands of Angstroms separation. 

During the sabbatical visit of Professor David Tinker of Toronto we made 
a survey of the contact forces between protein monomers. We believe 
we have discerned an unrecognized feature of protein contact: particles that 
remain in permanent contact touch over surfaces of hydrocarbon. Surfaces 
between monomers that can separate (dissolve) and rejoin hold together 
by an intricate matching of positive and negative charges that stud a 
non-polar surface. 



175 



Publications 

Cowley, S., Fuller, N. , Rand, R. P. and Parsegian, V. A.: MeasuremenJ 
of repulsion between charged phospholipid bilayers. Biochemistry (in press) 
(1978). — — ^-__^__^_ 

I 

Gmgell, D., Parsegian, V. A., Todd, I.: Experimental evidence for 
long-range attraction between a red cell and a hydrocarbon surface Nature i 
268 , 767-768 (1977). ' ' ^^^ j 

Parsegian, V. A., Weiss, G. H. and Schrader, M. E.: Macroscopic 
continuum model of influence of hydrocarbon contaminant on forces causing I 
wetting of gold by water. J. Coll. Int. Sci. 6J. 356-361 (1977). • 

Kiefer, J. E., Parsegian, V. A. and Weiss, G . H. : An easily calculabl 
approximation for the many-body van der Waals attraction between sphere and 
wall. J. Coll. Int. Sci. 63 161-171 (1978). 

Parsegian, V. A. : Considerations in determining the mode of influence 
of calcium on vesicle membrane interaction. Society for Neuroscience 
Symposia II. p. 161-171, ed. W . M. Cowan and J. A. Ferrendelli. 

Tinker, D. 0. and Parsegian, V. A.: Interactions between subunits of 
protein oligomers. Biophys. J. 2J. 145a (1978). 



I 



McAlister, M. , Fuller, N., Rand, R. P., and Parsegian, V. A.: MeasuramenI 
or surface pressure in and repulsion between approaching phospholipid 
membranes. Biophvs. J. 21 213a (1978). j 

Lis, L. J., Rand, R. P., and Parsegian, V. A .: Measurement of electro- 
static forces between lecithin bilayers charged by divalent cations. Bioohysl 
J^ n 145a (1978). ' ' ' j 

Kiefer, J. E., Parsegian, V. A . and Weiss, G. H.: Some convenient boun 
and approximations for the many body van der Waals attraction between spheres 
J ■ Coll . Int . Sci . (in press). 

Brenner, S. L., Parsegian, V . A. , and Gingell, D.: The effects of image' 
forces on double-layer interactions. J. Phys. Chem. (in press). ° : 



176 



oMITKGC.'MAIJ CCIdNCE Ir.rCRM.MI C.'i E<CHA,'JGE| 
PROJECT uUWSEH ^C- NOT u-.e this space) 



U.3. D£PART:.'£MT CF 
HEALTH, £DL1CAT10:,, a;;0 -.'ELFARE 
PUBLIC HEALTH SERVICE 
NOTICE Of 



PROJECT rUMEER 



INTRAMURAL RESEARCH PSOJECT ^01 CT 00014-11 PSL 



PERIOO COVERED 



October 1, 1977 to September 30, 1978 



TITLE Of PROJECT (60 characters or less) 

Theory of Biochemical Separation Techniques 



NAMES, LASORATCRY A.'iO INSTITUTE AFFILIATIONS, AND TITLES OF PRINCIPAL INVESTIGATORS AND ALL OTHER 
PROFESSIONAL PERSONNEL ENGAGED ON THE PROJECT 



PI: 
Other: 



G. H. Weiss, Chief, Physical Sciences Laboratory 
D. A. Yphantis, Professor of Biology 
University of Connecticut 



COOPERATING UNITS (if any) 
None 



lab/ BRANCH 

Physical Sciences Laboratory 



SECTION 



INSTITUTE AND LOCATION 

Div. of Computer Research & Technology. NIH. Bethesda, MflryLgnd 



TOTAL MANYEARS: 



0.1 



PROFESSIONAL: 



0.1 



OTHER: 



0.0 



CHECK APPROPRIATE SOX(£S) 
□ (a) HUMAN SUBJECTS 

n (al) MINORS n (a2) INTERVIEWS 



n (b) HUMAN TISSUES 



n (c) NEITHER 



SUMMARY OF sORK (200 words or less - underline keywords) 

This project explores applications of mathematical techniques 
to biochemical separation techniques such as electrophoresis . 
Little was done on this project in the past year except to start to look at 
practical ways of processing data from solutions of heterogeneous proteins. 



177 



PHS-OQ40 
(Rev. 10-76) 



Only a small amount of effort was expended on this project in the past 
year pending further experimental work on the theory developed for the 
acceleration of equilibrium sedimentation experiments. We have started 
looking at possible ways to process data from equilibrium sedimentation 
measurements on inhomogeneous proteins. 

Keyword Descriptors : Ultracentrifugation, equilibrium, sedimentation, in- 
homogeneous proteins. 

Publications: 



Correia, J. J. Weiss, G. H. . Yphantis, D. A : An extrapolation method 
for reducing equilibration times in sedimentation equilibrium experiments. 
Biophysical Journal 20, 153-168 (1977) 

Dishon, M. , Weiss, G. H. : When do transient double peaks occur in pH 
gradient electrophoresis? Analytical Biochemistry 81 . 1-9 (1977). 



178 

^ J' rj o 



NUH 

l|^ Amazing Research. 
Amazing Help, 

http://nihlibrary.nih.gov 



10 Center Drive 

Bethesda, MD 20892-1150 

301-496-1080 




' iJ'iJ' \'^,\'A^l^ -. 



3 1496 00821 8680