HANDBOOK OF PHYSIOLOGY section 1: Neurophysiology, volume m HANDBOOK OF PHYSIOLOGY /.- NEUROPHYSIOLOGY \ I .1 1 Ml I Historical development of neurophysiology Neuron physiology Brain potentials and rhythms s, iim,i\ mechanisms Vision \ l higher function we must turn our attention to man himself and to the mind of man. Here there is no full- toned harmony and the physiological voices are few and faltering. I III MOUNTAIN TOP Those who hope to solve the problem of the neuro- physiology of the mind arc like men at the foot ol a mountain. Thev stand in the clearings thev have made on the foothills, looking up at the mountain they hope to scale. Bui the pinnacle is hidden in eternal clouds and many believe il can never be conquered Surely if the day does dawn when man has reached complete understanding of his own brain and mind, it may be his greatest conquest, his final achievement. In the meantime, what are scientists to say of the things unseen, the problems yet to be solved? Are they entitled to a special privilege as prophets? Whal should thev sav in advance about the nature ol the mind, the existence <>! the spirit.' What about monism, dualism, Cod.' In his Rede Lecture at Cambridge in 1933, vn Charles Sherrington said: "I reflect with appre- hension that ,1 great subject can revenge itself shrewdl) for being too hastily touched To the question of the relation between brain and mind the answer given bv a physiologist 60 years ago was Tgnorabimus.' .. . The problem lodav has one virtue a1 least, it will long oiler in those who pursue ii the comfort that to journe) is better than to arrive; but thai comfort assumes arriv al. Some of us perhaps because we are too old or is it, too young think there mav be arrival at last." There is onl) one method that a scientist may use in his scientific work. This is the method of ob- servation of the phenomena of nature followed by comparative analysis and supplemented by experi- mentation in the light of reasoned hypothesis. Neuro- physiologists who follow the rules of the scientific method in all honesty will hardly pretend that their own scientific work entitles them to answer these ques- tions. Ivan Pavlov described the learning process, elabo- rating the mechanisms of conditioned reflexes. Sher- rington explored the basic mechanisms of subcon- scious reflex action. Both were good scientists, and neither claimed that he had solved ultimate truths ouiside the confines of an animal laboratory. Those who assert lodav that the work of Pavlov proved the truth of materialism draw a premature 1 441 '442 HANDBOOK OK PHYSIOLOGY NKTROPHYSIOLOGY III conclusion. Those who hail Pavlov as the great mollis and set Sherrington up against him as the leader of dualism go far beyond the evidence. Sherrington in his laboratory was a cautious scien- tist who worked without preconceived prejudice. His pupils hardly suspected that outside the laboratory he was a poel and a philosopher. His book of poems u.is published when he was 68 years of age, and his philosophical monograph, Man on Hi* Nature, when he was 80! At the age of go, when he wrote the intro- duction to a new edition of his scientific magnum opus, he ventured no farther than to s.i\ : "That our being should consist of two separate elements, offers, I suppose, no greater inherent improbability than that it should rest on one only."1 Although in the laboratory he adhered to critical objectivity, he had faith that "the relation between brain and mind" would be solved the mountain would be scaled in some distant day. Until then, he pointed out, neurophysiologists must cam- on using the "language of separation" without prejudice as to what the final truth may prove to be. To speak of mind and brain is to employ the language of the com- mon man, the language of dualism. But we can do nothing else if we would get on with the job. Outside the laboratory, lew scientists are to be con- red poets or philosophers. But, nevertheless, most reasonable men must wonder about (hose mysteries which are not to be solved by the scientific method in their lifetime; such mysteries as: what lies beyond the e? what is the origin of life, the design of the uni- \ else, the nature of truth3 Scientists are men as well as workers. They enter and leave life like other men. "There is a season and a time to every purpose under the heaven," and all must play, in turn, the role of child, husband, lather .mil cldci Good scientists may well conclude that there is virtue in the concepts of life that have been passed along from father to son within the memory of in. in Christian, Jew, Moslem and Hindu may have an intelligent faith in God .md still be completely objective in scientific thought. The atheist can do the same, il he is also .1 good scientist. I hus eat h man should turn to his workshop in his own 'clearing' without prejudice, and without linger- ing too long on the street corner to debate with Berkleian philosopher, or materialist or dualist. And I hasten now to return to my problem of writing an introduction to the chapters oi this section on "The Neurophysiology ol Higher Function." From the Introduction to the second edition ol / ' / iivr Action "I th London Cambridge, 1947. HIGHER Ft Nl I [i iNS Man is said, by man, to lie higher in the evolu- tionary scale than all other creatures. But how are we to define "higher" as applied to the neurophysiology of the nervous system? Let us say that the neuronal mechanisms that are essential to consciousness are high, higher than certain reflex mechanisms. It is clear that in the central nervous system there are many involuntary mechanisms that have little or nothing to do with conscious states. The manifold activities of the decerebrate animal bear witness to this. Above this there is the neuronal activity asso- ciated with "consciousness" and with the preparation for voluntary action, an activity that is essential to the very existence of these things. Voluntary activ its is produced by a flow of poten- tials along a well-known pathway from cerebral cor- tex to muscle. And the paths of sensory inflow that carry related information inward are well known as far as the cerebral cortex. But the cortex is not the end of the sensory pathway nor is it the beginning of the motor pathway. It is the organizing and integrating activity that comes between this 'CiiMin input and this motor out- put that constitutes the physiological basis of the mind. Here is the "higher function.' Here, the highest levels of integration are to be sought. Physiologists have contributed little .it this level, although the work of MagOUn on the reticular acti- vating system, and such contributions .is those in the preceding chapters by Bremer, Brookhart, French, Jasper, Pribram, Kaada, Green, Pampiglioni and Gloor constitute a promising beginning. However brief it may lie, each sta^e must have .1 sequential time value. I here is a time allowance for a) afferent conduction, h\ centrencephalic organiza- tion, (i volitional efferent conduction and d) motor contraction. A man's awareness oi environment dur- ing any moment ol time cannot be present during the afferent conduction of that moment. It must follow. It must precede the Outgoing volitional conduction. if appropriate action is taken. This sequence must hold although preparation and overlap are, no doubt, important in such processes. In borrow from the thinking of William James, consciousness is continuous during our waking hours, but its content is never the same. Like a mountain brook it tumbles p.ist in iis rocky bed but the watei cannot return nor be held in its place, from the standpoint of the physiologist too, the electrical poten- tials pass in swift and everchanging variation through the higher circuits of organization. NEUROPHYSIOLOGICAL BASIS OF HIGHER FUNCTIONS OF NERVOUS SYSTEM 1443 It might be considered hypothetically that that mechanism is highest which is functionally the closest to the zone of departure of the stream of potentials which, flowing outward, determines voluntary action. In the ganglionic zone of departure the pattern of the plan must become the pattern of the emerging motor stream. On arrival at the muscles, the pattern of that stream is translated into act or word. Abstract thinking, which has no outward expres- sion, differs no doubt from that which produces con- tinuous voluntary action such as talking, or writing or playing football, and yet many of the functional units of integration must be the same. CLINICAL SURMISES There are many inferences that come to a clinician. Perhaps it may serve a useful purpose to sel down a few observations and thoughts which seem to shadow forth the outlines of functional organization in the human brain. Substantiation of most of these state- ments may be found in previous publications else- where. Many patients have been operated upon under local anesthesia to cure them of recurring attacks of focal epilepsy. In the course of such operations, nearly all areas of the cerebral cortex have been excised at one lime or another, all except the most precious, the major speech areas of the dominant hemisphere. Furthermore, all areas, including; those devoted to speech, have been stimulated repeatedly since this is one way of discovering the abnormal zone which must In- excised if the sufferer is 10 Ik- relieved of his fits. These subjects, who help the surgeon with so much steadfast courage, have reported to him movements, sensations and certain psychical phenomena that are produced by application of ,i gentle electric current to the cortex. These constitute- data that no laboratory 'preparation' can provide Electrical stimulation of the cerebral cortex of a fully conscious man produces positive 'physiological' responses, but onK in certain areas. These excitable areas yield, when conditions are favorable and the stimulating current is at a threshold level, three prin- cipal types of physiological responses: a) motor, h) sensory and c) psychical. 2 Over the years, these data have been recorded carefully in the Montreal Neurological Institute with the invaluable as- sistance of Herbert Jasper, Theodore Rasmussen, Theodore Erickson, Edwin Boldrey, Lamar Roberts, Anne Dawson and a succession of brilliant assistants. SENSORY AND MOTOR RESPONSES There are 'sensory' and 'motor' areas that may be considered secondary or supplementary. These areas yield responses, when stimulated, that seem to the patient just as compelling as those obtained from the better known primary areas. But I shall refer here only to the latter: the excitable motor strip on the Rolandic or precentral gyrus, the somatic sensory area on the postcentral gyrus, the visual sensory area in the calcarine fissure and the auditory sensory area that carpets Heschl's transverse gyrus deep in the fissure of Sylvius. One conclusion is clear: these sensorv areas, that occupy homologous positions in each hemisphere, are way stations in the afferent streams of impulses that lead from the periphery to some deep subcortical target. The precentral gyrus, on the other hand, is a way station in the efferent pathway that arises in sub- cortical gray matter and passes outward to the muscles of voluntary action. Circumexcision of conical areas, sparing these sensor) or the motor areas, does not stop functional conduction through these wa\ stations, for example, when the parietal cortex of the right hemisphere is excised the subject is still able to guide the left hand in voluntary movement according to the information received from the left visual field that entered the brain through the right visual cortical area. The guidance is obviousl) not provided by direct cortical interconnection. It seems lair to conclude that corticocortical 'association1 connections between one functional of cortex and another are of comparatively minor importance. This contradicts a long-cherished hy- pothesis that the cortex was somehow invested with miraculous mechanisms of integration. Electrical stimulation of die cortex, which produces sensor) or motor responses, does so l>\ producing neuronal conduction in the direction of the normal flow through the particular way station to which the electrode is applied. This neuronal conduction ma\ therefore be considered dromic, and the effect is pro- duced by activation of nerve cell stations farther along in the stream of normal flow The electrode applied to the precentral gyrus thus produces simple movement of the opposite fingers by conduction to the anterior horn cell clusters in the cervical spinal cord; or, if applied elsewhere, it pro- duces vocalization, by conduction to the vocalization and respiratory mechanisms in the lower brain stem. Thus, the electrode provides no direct information as to what functional contribution the cortex may '444 II VNDIHMlK nl I'HVSIOI.OCV NEUROPHYSIOLOGY III make at these junction points. It does prove something • ihciut i!h- dromic connections of each area. The electrode, applied to a sensory area, produces seeing or hearing or feeling, depending upon which area is stimulated. The sensation, if somatic, is ting- ling, Dumbness or a sense of movement in some part; if visual, it is moving lights or colors or stars; if auditory, it is a buzzing, ringing, whispering or thumping sound. These are the elements of the sen- sory input.3 The exploratory electrode, delivering, for example, 60 pulses per sec. lasting a msec, with an intensity of 3 v., produces these three different categories of sensation because the cortical areas are connected with three different receiving stations in the higher brain stem. That same electrode, when it applies the same current to the precentral gyrus, produces move- ment because of the corticofugal connections with motor mechanisms in the cord or lower brain stem. However, under normal functional conditions, the How of impulses that activates the precentral gyrus for voluntary movement must arise in subcortical hi, iv matter since excision of cortex in front of the gyrus or behind it does not prevent a man from carry- ing out voluntary action based on sensory informa- tion.' CENTR1 \< I PHALIC ORGANIZA1 H in Therefore, the hypothesis presents itself that much of the organizing neuronal activity between sensory input ,ind motor oulpul must depend upon circuits and ganglionic structures within the higher brain stem, if the brain stem as defined by Herrick is under- stood to include the thalamic nuclei ot the two hemi- spheres. This old centra] structure, including the diencephalon, mesencephalon and at leasl part of the metencephalon, is connected by symmetrical projection tracts to the various functional areas of the i erebral 1 01 tea on either side. Tims bv means of these connections each area oi ould, it seems, be employed in common integrated action. There is no oilier obvious set of essential inlei < oiinecl ions for one cortex with the othei 1 ortex and "in- cortical area with iis neighboring In Chapter l.\. William Nefl points out .ilso thai the "input "I the tensor) systems may be controlled at .1 numba ol levels . . 1 ntrifugal pathways " In I haptei I \\ II [acques Paillard h.is summarized the cortical origin ol the voluntai \ ind ih- stud) hi the patterning ol skilled movements is stimulating. cortical areas. Direct transcortical connections can be interrupted with relative impunity. Even com- plete section of the corpus callosum by a number of intrepid neurosurgeons has had astonishingly little effect on the intellectual activity of man. One-sided removal of, or injury to, any area of cerebral cortex does not abolish conscious thinking. It may change the content of awareness, interfere with voluntary acts, render less effective planned ac- tion, deprive the patient of word symbols — but he still thinks and weeps, perhaps, at his own pitiful incapacity. On the other hand, interference with the centren- cephalic svstem of the higher brain stem produces loss of consciousness. In the presence of deep coma due to a small critically placed local lesion in the higher brain stem, the motor mechanisms may seem to remain intact. The patient lies in bed and move- occasionally as in deep sleep, like the enchanted 'sleeping beauty' of the French nursery tale. It is accepted by those familiar with epilepsv that epileptic discharge originates in gray matter, never in white matter. When epileptic discharge occurs in the gray matter of the centrenccphalic system of the higher brain stem, the patient is initially unconscious because the discharge interferes with local ganglionic function. On the other hand, a partial seizure, due to restricted epileptic discharge in any area of gray matter of the cerebral cortex, is no( associated with initial unconsciousness. The system of nerve fibers and ganglionic centers within the brain stein may be called centrenccphalic since, because of iis central position, it provides symmetrical connections with the whole brain Through it, one may suppose that this part of the cortex, or that part, could be used simultaneous!) or in sequence, depending upon the pattern and the requirements of the existing stale ol consciousness. 11 VSH-BACK VIIVHiKV v\l) COMPARATIV] IN I I RPR] I V I n IN From the temporal cortex of either side, electrical stimulation or local epileptic discharge mav produce phenomena ol an entirely different order from the motor and sensory responses previously produced in animals and man These responses are 'psychical,' to borrow an adjective from Hughlings Jackson. 1 1<- described such phenomena as they were experienced l>v his epileptic patients during small his. "Dreamy states," be called them, "as ii I went back to all thai mi lined in mv childhood." NEUROPHYSIOLOGICAL BASIS OF HIGHER FUNCTIONS OF NERVOUS SYSTEM ■445 Psychical responses may be produced occasionally in conscious man by gentle stimulation of the cortex of the temporal lobe. The response is ordinarily physio- logical and does not outlast the application of the electrode to the gray matter. The responses are of two types: illusion of comparative interpretation of the present, and hallucination of re-enactment of previous experience. Thus, the first type of response, if interpretive, gives the patient a sudden sense of alteration in the meaning of things being seen or heard or experienced at the moment. He finds them suddenly familiar, or strange, or more distant, nearer, fearful, etc. In nor- mal life we assume that such feelings must result from comparison of the present experience with similar past experiences, although this comparison of present with past is carried out subconsciously. If the response to stimulation is hallucinatory, it proves to be a re-enactment, or 'flash back," of previ- ous experience. The patient is aware of all those things to which he paid attention in some previous period of time. The experience unfolds for him at the former rate of speed as long as the electrode is held in place. He is aware of the present and of the past .is well. If he hears music, it is ,1 specific rendition as he heard it years ago, perhaps. He ma) sec the orchestra or the piano or the singer, and he may feel again the emotion roused in him by the music during that distant strip of time. Thus it is clear that the temporal cortex has some sort of selective connection with a detailed flash-back record of the past, most of which has been forgotten as far as the individual's ability for voluntary recall is concerned. HIGHER FUNCTIONAL MECHANISMS It would seem that, under normal circumstances, this flash-back record is reopened when similar experi- ence presents itself. For example, the reappearance of a long-forgotten friend makes available his previous records so that he seems familiar at once, and minor changes in him are obvious. Take another example: the set of present circumstances is compared with similar experiences from the past which may have been dangerous or painful and the subject feels a fear that calls for action. By inference, therefore, one may surmise that the temporal cortex normally plays some role in a sub- conscious scanning mechanism that opens the flash- back memory file and provides a signal of compara- tive interpretation (familiar, fearful, etc.) which rises without warning into consciousness. This scanning of the flash-back memory forms one partially separable mechanism in the higher functions of the brain. The use of words in symbolic thinking, so well described in Chapter LXYIII by Zangwill, is another such mechanism. Some degree of localization of the cortical portion of these mechanisms is possible now, but not the centrencephalic portions. Neverthe- less this is a beginning of the delineation of functional units closely related to conscious thinking. Hughlings Jackson pointed out that there were levels of integration in the nervous system and at- tempted to assign an ascending "representation" and 're-representation' to spinal cord, cerebral cortex and frontal lobes. But it is clear now that, although there are advancing stages in the progressive organization that make conscious thinking possible, these stages cannot lie assigned 10 separable major areas of the brain. A neurophysiologist might well attempt to separate the levels . > t organization invoked in normal volun- tary activity on the basis oi the progressive lapse of time, the period of afferent conduction, the period of organization and integration, the period of efferent conduction. From a functional point of view the mechanisms of that middle period of time are the highest and seem the most complicated. They form the physical basis of the mind. Doubtless, many of the mechanisms that are usable in this middle period .ire .ilso v.iiiuuslv employed during aimless conscious states and in constructive abstract thinking. Ihe neuron circuits of these functional mechanisms are to be found in the higher brain stem and in the cerebral cortex, joined together in action patterns lli.it form themselves and vanish and form anew, making combinations never twice the same. Genera- tions of workers must employ the scientific method to Study man in health and in disease Thus the shadowy outlines we perceive today will take clear form in a neurophysiology of higher func- tions. And if man comes to understand himself in mind and body he will have made his greatest con- quest. Perhaps then he will recall the prediction that "the truth shall make vou free." CHAPTER LX Sensory discrimination WILLIAM D. N E F F Laboratory of Physiological Psychology,* University of Chicago, Chicago, Illinois, CHAPTER CONTENTS Historical Summary The Beginning of Scientific Inquiry The Modern Period Neurophysiological Basis of Sensory Discrimination The Primary Sense Modalities Discriminate Dimensions of Sensation Intensity discrimination Quality discrimination Space discrimination Pattern discrimination Relation of the Primary Sensory Systems to Other Systems to report verbally when a stimulus change occurs; often he is asked to respond by pushing a button or by making some other nonverbal response. When a lower animal is used as an experimental subject, it is usually trained to indicate the stimulus change by in. iking a movement, such as Hexing a leg, pressing .1 lever with its foot or moving from one position to another in the test apparatus. In each case, the dis- crimination consists simply of a motor response to a change in some aspect of a physical stimulus which acts upon the sensory end organs. in other chapters, the function of each of the sensory systems has been discussed in detail. It is the purpose of this chapter to consider the informa- tion we have for all the senses in a less specific fashion and to look for common principles which may have emerged in the search for neurophysiological ex- planations of sensory discrimination. To the sensory physiologist or psychologist, the meaning of the phrase 'sensory discrimination" needs no explanation. To others, its meaning may not be clear immediately. Therefore, a brief ex- planation is in order. The living organism makes a sensory discrimination when it shows by its behavior that it has made a response to a change in a physical stimulus applied to one of its sense organs. It may re- spond to the presence or absence of a particular stimulus or may make a choice between two or more stimuli. When man is used as a subject in an experi- ment on sensory discrimination, he may be instructed 1 The Laboratory of Physiological Psychology is supported by the Office of Naval Research and the Air Force Office of Scientific Research. HISTORICAL SUMMARY2 The nature of the questions asked and of the ap- proaches taken in current research on sensory dis- crimination can best be understood after a brief consideration of the history of scientific investigation of the sensory systems. The Beginning of Scientific Inquiry There is no evidence to indicate that learned men of pre-Hellenic times attempted to account for how man perceives the world by making careful observa- tions and logical speculations; they were inclined to account for natural phenomena in terms of the super- stitions and religious dogmas of their particular civilizations. Greek scholars did observe carefully and specu- lated wisely. Handicapped by lack of knowledge — 2 This historical summary is based for the most part on secondary sources. For more complete accounts of the history of the neurophysiology of sensation, the reader is referred to the works of Boring and others (26, 28, 30, 46, 50, 63, 64 '447 '448 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III limited and inaccurate notions about the physical environment, incomplete information about gross anatomic. il structures, almost no evidence, and mis- taken ideas about the relation of sense organs to the brain or other parts of the body— they nevertheless proposed theories of perception which, although wrong, were reasonable. For example, Empedocles (ca. 490-430 B.C.I mi induced a theory of sensation which can be stated approximately in these words. The organ of vision contains a lantern and the organ of hearing, a bell or gong. The lantern in the eye is lighted and the gong in the ear rung by outer light and sound, re- spectively. The light in the eye and the ringing in the ear are then conveyed in some fashion to the 'point of sense' and thus the light and sound are perceived. Theophrastus (ca. 372-287 B.C.) criticized Empe- docles' theory of hearing as follows: "Empedocles explains hearing by stating that it is due to intra- aural sounds. But it is strange of him to suppose that he has made it self-evident how we hear by merely Stating this theory of a sound, as of a gong, within the ear. For suppose that we hear the outer sounds by means of this gong, by what do we hear the gong itself when it rings?"3 Translating this argument into modern terms, we might think of Empedocles as representing those investigators and theorists who have emphasized peripheral analysis and who have proposed so- called 'theories' of sensation which are actually theories of end organ function. And Theophrastus stated quite aptly, almost 2500 years ago, the criti- 1 1 m which has sometimes been made in recent years, namely that having accounted for how light, sound, odors or other sensory stimuli are received and analvzcd In the sense organs, there still remains the formidable task of discovering how this informa- tion is coded and transmitted into the central nerv- ous system and how it is utilized there to bring about discriminatory responses Nbl only Empedocles and Theophrastus bul man) others of the Greek philosophers contributed to the l ih<- contrasting viewpoint! <>l Empedocli and Theophrastus is taken from Beare 117). the muscle sense, was recognized and added at the start of the nineteenth century. Despite this early show of interest, some 2000 years elapsed with little further advance in knowledge of sensory mechanisms. There were, of course, im- portant discoveries in the fields of physics and anat- omy, discoveries which prepared the way for later scientific investigations. The physics of light and sound were at least partially understood and, equally important for later psychological and physiological investigations, instruments were developed for con- trolling production of sound and light. In anatomy, it became an acceptable practice (in some places and in some times) to dissect animals or human cadavers, and the development of the compound microscope and of techniques lor preparing tissues for microscopic examination led to the discoverv ol many details of the structures of the end organs and of their connections with the central nervous system. The philosophers of the seventeenth and eighteenth centuries also made a contribution by emphasizing the importance of the senses as the essential channels by which the living organism is able to have an awareness of the external environment. The field was set then at the beginning of the nineteenth century for the launching of a scientific attack upon the problems of sensory discrimination or, as it might have been put by the scholars of ihat period, the problem of how man knows the external world. In a brief historical summary, it is impossible to give due credit to all who made contributions in any given field. As a mnemonic device, it is excus- able, perhaps, lo select the names of some men who have traditionally been accepted .is leaders in doing research and in inlluencing the research and thinking of their colleagues and students. The Modern Period It is fitting lo begin a discussion ot modern re- search in season discrimination with .1 reference to Johannes Miiller and his doctrine of specific nerve energies. The doctrine or theory that Miiller ex- pounded, first in 1826 and in greatei detail in 1838 and [840, was not new with him. The same basic ideas can be traced back at least to John Hunter in 178(1 and were expressed indirectly by Thomas Young in i8oj and (putc explicitly hv Charles Bell in 1H11. It was, nevertheless, Miiller who gave the 'theory1 a name, even though a misleading one, and who formulated in detail 10 laws or propositions and SENSORY DISCRIMINATION 1449 supported them by evidence based upon careful observations. The main points of Miiller's doctrine were the following. a) Sensation is an awareness of the states of the sensory neural pathways and not of the environment directly. Miiller gave strong emphasis to this part of his theory, stating it in slightly different forms in several of his 10 laws. He undoubtedly felt that this emphasis was necessary in order to refute the theory which had been accepted with little change by most scholars and scientists from the time of Empedocles, namely that an image of the stimulus entered the sense organ and in some manner was transmitted to the 'sensorium' or sensing center in the brain. b) When a given sensory nerve is excited, the same kind of sensory experience results no matter what the nature of the stimulus. For example, when the optic nerve is excited by light waxes striking the retinal receptors, by mechanical pressure (ex- ternally or internally produced) directly affecting the optic nerve or by electric shock applied to the nerve, the result is a visual sensation. For the audi- tory nerve, the result is always a sensation of sound, and so on for the other sensory 1 1 h k [ . 1 1 i t ic> c) The same physical stimulus applied to dif- ferent sense organs or sensory nerves gives rise to different sensations, in each ease to the sensation appropriate to the organ or nerve in question. For example, electric shock applied to the optic nerve produces a visual sensation; to a tactual nerve, the sensation of touch. Miiller accepted the concepts of 'adequate stim- ulus' and 'specific irritability' of sense organs which had been expressed by other biologists before him. He gave a straightforward answer to die very im- portant question: where does the 'specific energy' reside, is it a characteristic of the peripheral nerve or of centers in the central nervous svsiem.' His answer was that he did not know. He explicitly stated: "The peculiar mode of reaction of each sense, after the excitement of its nerve, may be due to either of two conditions. Either the nerves them- selves may communicate impressions different in quality to the sensorium, which in every instance remains the same; or the vibrations of the nervous principle may in every nerve be the same and vet give rise to the perception of different sensations in the sensorium owing to the parts of the latter with which the nerves are connected having different properties" (50, p. 166). The doctrine that activity in separate peripheral nerves or in their central connections forms the basis for different sensations was soon extended not only to account for vision, hearing, touch, smell and taste but for the sense qualities within each of these modalities.4 As we have noted above, Thomas Young is often credited with having suggested a specific nerve energy theory a number of years before Miiller's first publications on the subject. Young's suggestion was not made in an attempt to account for differences among the major senses but to account for the different colors perceived visually. He proposed the notion that there might be three kinds of receptors and connecting nerve fibers sub- serving t he three primary colors. Helmholtz made use of the principle of specific nerve energies to explain different qualities of sensa- tion in both vision and hearing. He elaborated upon Young's earlier hypothesis and, giving due credit to Young, postulated specific nerve energies for the primary colors. The theory which he put forward is still referred to as the Young-Hclmholtz theory of color vision. In hearing, Helmholtz sought to explain pitch perception in terms of the specific nerve energv principle. The Helmholtz resonance theory of hear- ing stated that resonators in the cochlea analyze sounds so thai sensory receptors and their connecting nerve libers are selectively stimulated, and that each discriminablc pitch is represented I >v a specific receptor or receptors and corresponding neural connections. In the search for specific receptors for individual sense qualities, Blix in ifWj and Goldscheider in 1884 explored the surface of the skin for spots ha\ ing specific sensitivit) to the qualities pressure, warmth, cold and pain. They, and later von Ftev , were able to identifv spots for pressure, warmth and cold, with the question of spots for pain being a matter of some controversy. Von Frey also attempted to identifv particular receptor endings in the skin for each of tin- separate qualities. For the other senses, a direct attempt to identify separate receptors for different sensory qualities was less easy. That the sensory receptors for taste were associated with the papillae on the tongue had been 4 Here and in the discussion which follows, modality is used in referring to one of the principal senses, namely to vision, hearing, somesthetic sense, kinesthetic sense, vestibular sense, taste or smell. The term 'attribute' will be used to refer to any discriminable dimension within a given sense, e.g. differences in intensity, in space, in duration or in quality. For a definition of 'sense quality,' see p. 1457. 1450 MAMJI'.i II >K 111- 1*1 IS SH il I n.s NEUROPHYSIOLOGY III deduced bs Haller in 1765; Bell and Miiller agreed with this opinion. Morn in [825 showed experi- mental]) that different regions of the tongue were mure or less sensitise to various taste stimulating substances. Tin- taste buds were identified and de- scribed by both Schsvalbe and Losen in 1807. But c\l other sensors organs does, ni> attempt was nude in classif) qualities of vestibular sensation. Nevertheless, Fluorens .md later investigators did seek in discover separate linn linns ini the major subdivisions of the vestibular end ni',;. in, name!) the semicircular canals, the mi tde .md the " < m!i I oincident with the development ol these new hypotheses .is in die anatomical and physiological bases ui sensor) discrimination was another line oi attack upon the problem of how the physical world is perceived by the living organism. E. 11. Weber in 1834 began, and G. T. Fechncr during the period 1 85 1 -1 879 and W. Wundt from 1858 to 1920 ex- tended a series of experimental investigations in which the aim was to establish systematic relation- ships between the physical attributes of external stimuli and the psychological attributes of sensation. Their svork in this new field, later labeled psycho- physics, established the foundation upon which was built a new branch of science, experimental psychol- ogy. Although the procedures which they and their followers used did not call for direct observation of physiological events, the theories of sensation to which the psychophysicists contributed have usually been expressed in terms of physiological events. In many instances, the results of psychophysical studies base set problems for physiological investigations. Psychophysics has also made important contributions to physiology of sensation by its desclopment of methods of measurement, particularly methods of measuring absolute and differential thresholds, and by creating and assisting in satisfying a need for instruments which make possible the production and exact control of physical stimuli. The last half of the nineteenth century was marked by a number of other developments which were important to the advance in scientific knowledge of sensory discrimination. Increased interest in the sense organs and the central nervous system led to their careful examination by microscopic techniques. The techniques, at first crude the dissection and examination of fresh tissues under the microscope — were rapidl) improved. Sense organs and neural tissues were treated with fixatives or hardening solutions, thus making dissection easier. A next and more important step forward was taken when analy- tical staining methods were discovered lis Golgi in 1873 and Ehrlich in 1886. Their methods and tlio-e developed b) others such as Ramon \ Cajal, Nissl, Weigert and Marchi began to reveal for the Inst time details of the structures w hich mediate sensation. The neural pathways were traced from sense or- gans into the central nervous system and, at least parti) because of the influence of specific nerve en- erg) theory, localization of separate regions sub- seising iln different senses was sought in the highest centers, particular!) in the cerebral cortex. 1 In- methods of architectonics were applied, and the cortex was subdivided into man) regions of sup- posedl) special function, the sensor) projection SENSORY DISCRIMINATION '45' regions being prominent in most maps so pro- duced. In experiments on animals, regions of the cerebral cortex which were thought to be sensory' projection areas were ablated and the effects on sensory- dis- crimination were observed or tested crudely. In the clinic the effects of brain damage on sensory per- ception of patients were observed. During the last part of the nineteenth century, considerable progress was also made in the use of electrical stimulating and recording techniques to explore the central nervous system. As Brazier (30) has pointed out, it was known by the end of the nineteenth century that changes in electrical activit\ can be recorded from the cerebral cortex in response to sensory stimulation. Better instrumentation was needed, however, before the full value of electro- physiological methods could be realized. There have been a number of significant methodo- logical advances which have marked the study of the senses during the twentieth century. First, and most important, has been the rapid development of new techniques in electrophysiology. Second has been the use of the behavioral testing methods of the com- parative psychologist in studies of sensory function in animals. Third has been tin- refinement of tech niques of experimental surgery. Fourth has been the improvement of instruments for production of sensory stimuli; in many instances, science lias profiled lii 1111 the expansion of electronics in industry, fin- ally, a most important trend which can be seen today is the application of many techniques in a single experiment. Methods of electrophysiology, of neuroanatomy, of psychophysics and of psycho- physiology are often used in a single investigation. NKI ROPHYSIOLOGICAL BASIS OF SENSORY DISCRIMINATION Tin Primary Sense Modalities In his doctrine of specific nerve energies, Johannes Miiller offered an explanation, in anatomical and physiological terms, for discrimination of the pri- mary sensory modalities: vision, hearing, touch, taste and smell. Others extended the theory to account for differences in sense qualities. In examining the present status of neurological theories of sensory discrimination, let us also start with the broader problem of how the sense modalities are differ- entiated and then go on to the more difficult task of evaluating the evidence for the neural mechanisms underlying discrimination of attributes of sensation.4 The specific nerve energy doctrine assumed that each primary sense modality had its separate nerve supply and separate representation in the central nervous system. An accumulation of evidence from experiments using a variety of anatomical and neurophysiological techniques has shown the es- sential correctness of this basic assumption. Thus, nerve fibers from the retina can be traced to the lateral geniculate bodies in the thalamus and thence to the striate cortex of the occipital lobes. In the auditory- system, similarly, a direct restricted pathway can be traced from the cochlea to a limited region of the cortex of the temporal lobe via the primary cochlear nucleus, superior olivary complex, inferior colliculus and medial geniculate body. Receptors for the skin senses project 10 the parietal lobe, again having p.issed upwards through definite channels in the spinal cord and brain stem. Pathways lor nerve libers from taste and smell receptors have been carefully traced into the central nervous system, but here the evidence for final projection centers is not as clear .is for the visual, hearing and somesthetic systems. Pathways for kinesthetic receptors are parallel and in close relation to those lor the skin senses and ap- parent!) end in the same region of the cerebral cortex. Although it was at one time thought that the vestibular system differed from the other af- ferent s\ stems in that it did not project to the cere- brum but had onk reflex connections via lower brain-stem centers, evidence has been obtained in recent years which indicates an afferent pathway paralleling that of the auditory system and ending in an area of the temporal lulu- which is usually con- sidered 10 be part of the auditory projection region. On the afferent side, at least, the structural organiza- tion for discrimination based upon a place or topo- graphical principle appears to be present in most mammalian species for all of the major senses. First scientific evidence for topographical organi- zation of the sensory systems came from clinical studies of the effects of brain lesions produced by disease, accident or surgery, and from experimental ablation studies in lower animals.5 For the visual and tactual systems, in particular, it was clear that limited lesions of ascending pathways or of relatively small cortical areas result in sensory deficit. For the 5 Brief historical summaries of this topic are available (14, 46, 63, 64, 66, 95, 137, 157, 189, 213). 11 \M)K()(IK <>r l'HYSIdl.c H.Y NEUROPHYSIOLOGY III other systems, the evidence was, and to some extent still is, controversial. Studies of cortical cytoarchitecture and, more im- portant, tracing of anatomical pathways by selective staining techniques provided further evidence of the basic structural organization of the sensory systems.5 The most important advances, however, were made when electrophysiological methods were used to trace the sensory pathways in the central nervous system. The use of the evoked potential method has been of special value in mapping the areas of the cerebral cortex to which the primary sensory systems project. (Sec particularly Chapters XYII through XXI, XXIY and XXX of this Handbook for studies of individual projection areas.) The increased knowledge gained through electro- physiology has not led to a sudden clarification of the significance of topographical organization. Rather, it has brought to light a complexity of organization, the significance of which remains to be explained. For example, at the cortical level, the primary pro- jection areas as defined by cytoarchitecture, myelo- architecture or tracing of degenerating pathways to or from thalamic nuclei were once thought to be unitary regions confined to relative!) small portions of the main cerebral lobes. Evoked potential maps have shown thai for the visual, somesthetic and auditor) systems, at least, and for most mammals ili. 11 have been siudied, the regions of primary pro- jc< linn are more extensive than was formerly thought and the) arc not unitary, that is there may be two or more projection areas or two or more subdivisions ol the projection region for each system, [n all mam- mals (man included) thai have been examined, two or more somatic areas ha\i- been found, evidence foi dualit) of projection being that the different parts oi the bod) project in an orderly fashion in two separate but adjacent areas. Likewise, in most mam- mals two visual areas and two auditor) areas can be defined (see < hapters XXIV and XXX). Evidence for dual projection of visual and auditor) systems in pi mi. itcs is not conclusive, but this lack of evidence ma) be due to difficulties encountered in exposing and in adequate!) exploring the relevant regions. The dual projection areas discovered b\ the evoked potential method in man) instances have also been shown to be different with respect to their thalamic connections. Foi the visual, auditor) and somatii ■ in- there i- some evidence thai there ma) be three or mon projei areas which must, in terms ol in ii organization, be considered as distinct, aaming o) multiple projection anas has no1 been simply the subdividing of the regions classified as projection cortex by the older anatomical methods; parts of the cortex formerly called 'association' areas have now been shown to have direct afferent con- nections with the thalamic nuclei of the main sensor) systems. Consequently, the sensory projection areas, as now defined, occupy a larger portion of the total cortex than was once thought to be the case. In mammals below the primates, there are only small strips of 'silent' cortex lying between the visual, auditory and somatic projection regions. Even in primates, the extent of so-called association cortex is reduced in cortical maps based upon the latest electrophysiological studies. While it was recognized that many reflex re- sponses to sensory stimulation can be carried out v ia connections in the spinal cord or in brain-stem centers, it has been the common view, until relatively recent times, that conscious discriminations or learned discriminations depend upon the intactness of the appropriate regions of the cerebral cortex. Despite convincing evidence from experiments on lower animals and the lack of critical evidence from clinical investigations of man, this view is still ex- pressed or implied bv man) writers of textbooks and even by some authors of current research publica- tions. Evidence bearing on the role of the cerebral cortex in sensory discrimination will be dealt with in more detail below when the neural mechanisms of different kinds of discrimination are discussed. In conjunction with the present discussion of the topo- graphical organization of the primary sensor) modalities, it is appropriate to note that, with the possible exception of vision in man, evidence from both clinical and experimental investigations in- dicates that for each ol the sensor) systems that has been studied carefully, some kinds of learned dis- crimination can be made after complete destruction of the cortical projection anas of that system.6 In Subprimate mammals, it h.is been shown that a learned discriminatory response to a sensory cue such as the flashing ol a light, onset of a fairl) loud ionc or tactual stimulation of the skin can be estab- lished after complete or near!) complete ablation of all cerebral cortex ( 32, 75). As we shall see, the capacit) foi sensor) dis- crimination may be altered bv more limited abla- ' It iniisi be kept in mind, however, thai we do not have clinical cases with lesions suitable foi comparison with lower .mini. ils in which the sensory projection areas of a single system have been complete!) ablated bilaterallj SENSORY DISCRIMINATION •453 tions confined to cortical projection areas, but the ability to make certain discriminations remains.7 The effects of interrupting sensory pathways at thalamic, tectal, bulbar or lower levels cannot be summed up so readily. Evidence to date is quite limited and will have to be considered in relation to particular kinds of discrimination. Discriminable Dimensions of Sensation A primary sensory modality may be defined as a sense organ or system of sense organs which has its adequate stimulus and which is connected to higher centers by its own separate nerve pathway.8 Diffi- culty arises in deciding upon the total number of primary modalities. Vision, hearing, smell, taste, kinesthetic and vestibular systems can be differ- entiated readily. A question remains as to whether or not the somesthetic sense should be divided into a number of modalities, at least into three (touch, temperature and pain) and perhaps into more (light touch, deep pressure, warmth, cold and pain). This problem may be resolved when sufficient evidence is accumulated from experiments or, as seems more likely, it may cease to be considered an important problem. For the purposes of the present chapter, we shall classify as separate sensory modalities the following: a) vision, b) hearing, c) somesthesis (touch, tempera- lure and pain), d) kincsthesis, e) vestibular sense, f) taste and g) smell. Discovering the neurophysiological basis for dis- crimination between primary sense modalities i^ only the beginning of an understanding of sensory discrimination. The next step is to look for mecha- nisms to account for discrimination within each sense since man and lower animals can make dis- criminations based upon differences in the charac- teristics of the stimuli affecting a single sensory modality. On the basis of psychophysical experiments in which the essential method was that of controlled introspection, Wundt, Titchener and other psy- 7 In discussing the effects of central nervous system lesions, we shall be concerned in this chapter with sensory capacity which can be demonstrated after the lesions, in many instances there is a loss of a learned discriminatory habit following cen- tral nervous system damage, but relearning occurs. See Chap- ter LXI for review of studies on the neural basis of learning. 8 This is, of course, an oversimplified definition in view of the known interconnections among the many sub-systems within the central nervous system: it is, nevertheless, a convenient working definition for the present discussion. chologists of their time concluded that sensory experience can be described in terms of basic at- tributes such as intensity, quality, duration, ex- tensity and clearness (attensity).9 In the sections which follow, we have chosen to discuss the neuro- physiology of sensory discrimination under headings, many of which are reminiscent of the categories set up by psychologists of the introspection school. In attempting to bring together data from phys- iological, psychophysical and psychophysiological ex- periments, a complication arises which can lead to confusion unless it is recognized in advance and kept in mind during the subsequent discussion. In phys- iological experiments, parameters of the physical stimulus, such as intensity and frequency (or wave length), are varied and physiological changes, such as rate of impulses in a nerve fiber and number and position of fibers excited, are measured. In psycho- physical experiments, the same physical parameters are varied and discriminations of the experimental subject recorded. The results of psychophysical studies have shown, however, that changes in a single dimension of discrimination are not neces- sarily a function of changes in a single dimension of the physical stimulus. Thus, brightness discrimina- tion in vision is a function not only of light intensity but also ol wavelength, and pitch discrimination in hearing is a function of both frequency and intensity ol sound. The physiological events which are the basis of brightness or of pitch discrimination must then, in each case, be a function of at least two parameters of the respective physical stimulus. In psychophysiological experiments, physiological events are varied and changes in sensory discrimina- tion measured under different conditions of periph- eral stimulation, the usual parameters of physical stimuli being varied, but in this case measurements are made of changes in discrimination before and after manipulation of a physiological condition. Difficulty arises when the results of physiological experiments, e.g. the variation in flow of nerve impulses with change in intensity of peripheral stimulation, are used in hypothesizing a neural mechanism to account for a change observed in a psychophysiological or psychophysical experiment. If, for example, pitch discrimination is a function of both intensity and frequency of sound, then the neural correlates of both have to be considered in formulating a neural mechanism to account for pitch discrimination. 9 For a review of the problem of defining dimensions of sen- sation, see Boring t. - 7 ) • '454 II \M)HIM 'k Ml PHYSIOLOGY NEUROPHYSIOLOGY III From future experiments new kinds of information are needed to fill in the gaps which now exist. We need in particular to know more about: a) patterns of neural activity elicited In peripheral stimulation and b) effects of systematically varying patterns of nerve pathway or nerve center activity on dis- criminatory behavior. INTENSm DISCRIMINATION. In the very first experi- ments in which electrical signs of activity in sensory nerves were recorded successfully during stimulation of the sense organs supplied by the nerves, it was apparent that increase in intensity of stimulation leads to increase in rate of firing of nerve impulses. It was also noted that as intensity is increased, more nerve fibers are activated (3, 5). Through the use of techniques designed to isolate single sensory nerve fibers and through the use of better methods of electrical recording, analysis of the responses in single first-order fibers of the visual (87), auditory (203), tactual (5, 6), kinesthetic (140- 142), vestibular (133, 134) and taste (166) systems has now been accomplished. In all instances, a systematic relationship between strength of stimulus and frequency of nerve impulses has been found. The picture has been complicated, however, by the further finding that in all of these systems there are many fibers in the first order neurons in which a 'spontaneous1 discharge occurs, i.e. a regular firing of nerve impulses occurs even when the end organ is not being stimulated. This phenomenon was noted by Adrian in his 1926 report (3) and again by Adrian & Zotterman (5); but they could not rule out the possibility that in the receptors of resting muscle or in tactual receptors not being intention. ilk stimulated, a mild degree of stimulation was taking place due to such factors as muscle tonus or skin tension. In other systems, e.g. vestibular, visual and auditory, it appears less likely that the spontaneous firim; which has been observed (88, 133, 134, 203) can be accounted for in terms of mild external stimulation ol receptor organs. A more likerj ex- planation is that metabolic processes within the end organ give rise to the spontaneous discharges. In addition to spontaneous discharge and the in- e in frequency ol discharge with increase in intensity ol stimulation, another phenomenon has been ui sei ved in first ok lei neurons ol the \ estibular and taste systems. Lowenstein & Sand (133, 1 34) have isolated single libers from the ampullae ol the semi, in nl. 11 canals ol the ra\ and have found that these fibers fire spontaneous!) when the end organ is at rest, fire at an increased rate when the head is rotated towards the side of the semicircular canal from which a recording is being made, and fire at a rate below that of spontaneous discharge when the head is rotated in a direction away from the side of the canal being tested. Pfaffman (167) has observed inhibition of spon- taneous discharge in single fibers of the chorda tympani of the rabbit. The inhibition occurs when taste receptors are stimulated by weak solutions of sodium chloride; other substances such as potassium chloride and hydrochloric acid produce increased discharge. While similar findings to those reported by Lowen- stein & Sand for the vestibular system and by Pfaff- man for taste have not been observed for first-order fibers of the other sensory systems, observations ob- tained from microelectrode recording in peripheral neurons one or more synapses removed from the end- organ receptors reveal a number of phenomena which must be taken into consideration in relating neural discharge rate to peripheral stimulus in- tensity. In the optic nerve of the frog, Hartline (88) has described three types of nerve fibers, classified in terms of their responses to peripheral stimulation. Fibers of one type respond to onset of stimulations, after an initial burst of rapid activity, these fibers maintain a somewhat slower steady rate of response during the period of stimulation. A second type responds with a burst of impulses at onset and at cessation of stimulation but remains silent in between. Fibers of a third type fire only at cessation of stimu- lation. Recording from cell bodies of second order neurons of the auditory nerve of the cat, Galambos & Davis (69, 70) have discovered single units which fire spontaneously in the absence of end-organ stimula- tion, at an increased rate during stimulation by tones of some frequencies, and at a decreased rate during stimulation bv tones in a different part of the fre- quency range. Other fibers of the auditor) system show only increased rate of discharge when the end organ is stimulated. No 'off responses were reported by Galambos & Davis [asaki & Davis (205), re- cording with microelectrodes from second order auditor) libers in the guinea pig, failed to find fibers in which spontaneous discharge was inhibited b) tonal stimulation. Tas.iki (204) suggests that the difference between the results obtained bv him and Davis and those reported earlier bv Galambos & I ).iv is 1 1, 1 the cat ma) be due in part to difference in SENSORY DISCRIMINATION '455 size of electrodes used and to their placement in the cochlear nucleus. In the Tasaki & Davis study, the microelectrodes were sufficiently small to record intracellular potentials; those used by Galambos & Davis were larger and could only record extracellular potentials. Furthermore, Tasaki points out that he and Davis probably recorded from the ventral cochlear nucleus where there is an absence of cell bodies with long dendrites and consequently an absence of large external potential fields generated by the cell bodies when their surfaces are uniformly activated. Galambos & Davis, on the other hand, were more likely recording from the dorsal cochlear nucleus where there are many cells with long dendrites which can generate large potential fields around them. In unpublished experiments, Galambos & Tasaki report having observed "suppression of spontaneous discharges with submicroscopic elec- trodes inserted into the dorsal cochlear nucleus of the cat" (204). With electrodes in the mitral cell layer of the olfactory bulb of the rabbit, Adrian (4) recorded the activity of single neural units and observed 'spon- taneous' discharge in the absence of intentional stimulation and increase in discharge rate when the concentration of substance stimulating the olfactory receptors was increased. Results for all the senses agree that, in addition to an increase in rate of impulse discharge in some units, increase in intensity of end-organ stimulation results in the excitation of more nerve fibers. There is also evidence that some sensory fibers have higher thresholds than others and are therefore activated only by high intensities of stimulation (fig, 142, 1 * > 7- , 203). Although the simple relationship between stimulus intensity and rate of neural discharge is complicated by such phenomena as spontaneous neural firing and increased rate of discharge in some neurons at the same time that rate is decreased in others, it seems a safe conclusion that if both frequency of dis- charge and increased number of fibers are con- sidered, increase in intensity of end-organ stimulation is represented in the peripheral pathway of each sensory system by increase in total flow of nerve impulses. For sensory centers in the brain stem and cere- bral cortex, there is only limited and, for the most part, indirect evidence relating frequency of nerve impulses in individual nerve fibers and spread of response or number of nerve fibers activated as func- tions of peripheral stimulus intensity. It has been shown that increase in intensity of peripheral stimula- tion may result in increase in neural discharge in midbrain centers, but, as would be expected, the relation is not always a simple one. As in more peripheral centers and pathways, factors such as spontaneous discharge and its suppression by stimu- lation in some instances and different thresholds of excitation complicate the picture (65, 71, 73, 84, 106, 107, 178, 206, 208). Studies of responses evoked at the cerebral cortex support the hypothesis that the total number of impulses arriving at higher neural centers is corre- lated with stimulus intensity. At threshold intensities of peripheral stimulation, a small evoked response can be recorded tvpically from a relativelv limited region of the relevant cortical projection area of the deeply anesthetized preparation. As intensitv is in- creased, the amplitude of the evoked response is increased and responses can be obtained from a wider expanse of cortex. Insofar as these changes can be taken as indicators of increased frequency of nerve impulses arriving al the cortex and of increase in number of projection fibers conveying response, 1 Inn the evidence from examination of evoked cortical activity supports the hvpothesis stated above (94, 123, 210). Investigations using microelectrodes to record responses of individual cortical units also provide evidence that with increase in intensity of peripheral stimulation more units are fired and in many units more impulses per second are elicited (".47. 56)- Establishing a relationship between intensitv of stimuli applied to sense organs and the total flow of nerve impulses in the neural pathways and centers connected with these end-organs is an essential first step in discovering the neural basis of intensitv dis- crimination, but i( is not sufficient as an explanation. It is also necessary to relate neural activity to be- havioral discrimination. As one example of an at- tempt to do this, Granit (821 has cited the studies of Enroth as evidence for the proposition that bright- ness discrimination in vision is directly related to total flow of nerve impulses in the visual system. Enroth recorded the response of single fibers of the optic nerve when the eye was stimulated by light at dif- ferent flicker rates and different intensities. She counted the number of nerve impulses elicited by the last four flashes of light preceding and immedi- ately following the point at which fusion was reached. When frequency of nerve impulses was plotted against flicker fusion frequency for a large sample of nerve fibers ('on' and 'off' type fibers both being ' IV' II WDHdllK OF I'HYSIUl ( ii.Y M UROPHYSIOLOGY III included), a lineal relationship was found. Since the frequency at which flicker Fusion occurs is propor- iKin. il in the logarithm <>f the brightness of a flicker- in- light, Enroth (and also Granit) make the in- ference i hat brightness is directly related to nerve impulse flow. Experiments arc yel to be done in which processes in sensory neural pathways and centers are con- trolled and manipulated directly while behavioral dis- criminations are observed. First steps in this direction have been taken by the utilization of implanted electrodes to record neural activity from or to stimu- late neural centers in unanesihetized animals. Another method, one almosl as old as the history of scientific investigation of sensation, has gisen Us valuable data about the pans of sensors systems which are necessar\ for different kinds of sensory dis- crimination and has, moreover, enabled us to make some inferences about the aeurophysiological proc- esses which underlie sensors discrimination. The method is that of testing discriminatory capacity in animals after selective ablation of neural centers or (he transection of neural pathways. The validity of the results obtained b\ use of this method depends on a number of factors, particularly: a) the definition of appropriate neural 'units' to be rendered non- functional by ablation, //) the exactness of surgical procedures, , i the adequacy of pre- and postopera- tive methods of measuring discriminatory capacity and Bg). Unfor- tunately, similar experiments have not been done at the primate level, but the peripheral mechanisms are probably not essentially different from those of the forms studied. ( )n the basis ol anatomical observations, LeGros Claris (39—41) proposed that color discrimination might be accounted lor in pari by a place or topo- graphical principle. The lateral geniculate nucleus ol primates is divided into six layers, three receiving libers from the contralateral nasal hemiretina and three from (he ipsilateral temporal hemiretina. Clark suggested that each Layer transmits impulses from die modulators related to one of the three primary c omponents of the trichromatic theory of color vision, six layers being needed to eonvev (he impulses from (he two e\es. I he e\ idence presented by ( 'lark to SUp- poi I his h\ pothesis has been criticized b\ Walls U 1 j I The proposal must be regarded as a tentative one, needing further experimental confirmation. In other recenl studies, a search has been made loi electro- physiological correlates ol color vision. Some e\ idence has been reported indicating that amplitude, latenc) and shape oi responses evoked hum die visual cortex varv with (he color Used in Stimulating (he e\ e 1 ;-, 1 j-', 1 _•■•;, 135). The results reported to date are only jestivc II these characteristics of central nervous system response do have .1 relation to coloi vision, it would appe.11 advisable 10 explore them in the primate or in another mammal known to have good color vision rather than in the cat which has been shown to be deficient in color discrimination (44, 51, 143)- Studies of behavioral discriminations of color be- fore and after the production of lesions in neural pathways and centers of the visual system provide almost no evidence that will help in formulating a hypothesis as to a neural basis of color vision. In the monkey, there is evidence that discrimination on the basis of color cannot be made after bilateral ablation of the visual cortex (1 15). In other animals such as the dog and cat, which have been used most often in be- havioral studies of vision, experiments indicate that at best the intact animal has a very low order of color vision. As pointed out above in discussing; electro- physiological investigations, such animals are ques- tionable subjects for color discrimination studies. Reports from the clinic, likewise, furnish very little additional information on the role of the visual cortex or of lower centers in color discrimination. Electrical stimulation of the striate areas in man snves rise to reports of visual sensations which include that of color (163, p. 143). Ablation of parts of the visual areas in man produces scotomas in which complete absence of color vision is usually reported. The evidence, though it is limited, docs not lend much support to an explanation in which color dis- crimination is accounted for on the basis of separate sense-organ and neural units for the separately dis- criminable colors. On the other hand, it still remains a possibility that there are a number of separate color receptors in the retina and that they maintain to some extent independent central nervous system connec- tions. It is more probable that different colors are represented in the central nervous system by patterns of activitv which vary with respect to temporal and Spatial arrangement of neural activitv in groups of neural units. I Ielmholt/'s extension of the specific nerve energy theorv to account for pitch discrimination in hearing lias met with greater success than the similar Sugges- tion lor color vision, lbs resonance theorv of hearing proposed that sounds are analyzed bv the cochlea in such 111. inner that for eacdi discriminablc pitch, there is a separate receptor unit which is excited bv a given frequency or narrow band ol frequencies. (Cochlear processes are considered bv Davis in Chapter XXIII and central auditor) mechanisms bv Ades in Chapter XXIV ol this Handbook. > Although the original simple teson.mce principle ol analvsis has been proved inade- quate, it is generall) accepted that peripheral analvsis does lake place, that high tones excite receptors in the SENSORY DISCRIMINATION '459 base of the cochlea and that as frequency of the stim- ulus is lowered, the region of maximal excitation moves towards the apex. It is also generally agreed that the degree of cochlear analysis, particularly for low tones, is not sufficient to account for the fineness of pitch discrimination. If pitch discrimination de- pends upon a place principle, then neural processes must result in some 'sharpening' or 'channeling' so that frequencies that can be discriminated activate separate neural units. To demonstrate such sharp- ening in a sensory system, Bekesy (20) has used a di- mensional mechanical model of the cochlea with the tactual receptors of the forearm of a human subject serving as the receptor cells of the basilar membrane; the traveling waxes which under stroboscopic light may be seen moving along the surface membrane ol the model are felt by the subject as a vibratory stimu- lus applied to a fairly sharply localized region on the forearm. The phenomena demonstrated by BeJcesy's model imply that 'inhibitory' or 'suppressor1 processes in the afferent nerve pathways from the skin result in a channeling of neural activit) . The fact that the impulses in peripheral and central neural pathways are synchronous in frequency with stimulating tones throughout a wide range of fre- quencies m.iv also be used in accounting for pitch discrimination. Wever (220) has suggested that fre- quency of nerve impulses may provide the principal cue for discrimination of tones in the lower range, perhaps up to 400 cps, for intermediate frequencies, a combined place and frequency principle may operate (400 to 5000 cps) with place being of primary importance for the highest audible frequencies (above 5000 cps). Several investigators have recently called attention to the long-known but often disregarded fact that there are at least two kinds ol subjective experience which we tend to include under the single heading of pitch perception. For example, we perceive tones such as those produced by an oscillator or organ pipe and we say that one tone is higher or lower in pitch than another. We can also make similar judgements about sounds which have an intermittent, rougher, noisier characteristic; these sounds can be matched in pitch with pure tones but they are by no means identical to the observer (49, 125, 126, 146, 180). As Licklider (125) notes, not only are there two attributes of pitch sensation, there are also two characteristics of the physical stimulus for pitch, namely frequency and periodicity. He postulates two neural mechanisms to account for the two attributes of pitch sensation. One is the classical frequency analysis according to place; the second is a neuronal autocorrelation analysis based on periodicity. The latter operates only for frequencies in the lower range, perhaps up to 1000 cps. There is not only conclusive evidence for frequency analysis in the cochlea but also for topographic pro- jection of the cochlea in neural pathways and centers up to and including the auditory areas of the cortex (69, 71, 84, 94, 106, 107, 109, 124, 203, 205, 208, 210, 211, 227, 228). The fact of such projection does not necessarily mean that it provides the basis for pitch discrimination. As Lashley ( 120) has pointed out, the maintenance of systematic spatial organization lrom peripheral end organ to cortex max be the result of the mechanics of embryonic development. Wires may be strung side by side in a telephone cable, but mes- sages sent over these wires ma) still be in a frequenc) code. Nevertheless, there remains the fact that in a system such as that formed by the auditory neural pathways, spatial arrangement of axons in tracts and of the cells upon which the) end must inevitably play a role in determining the manner in which afferent events produce the elletent activity which exenlu.ilh results iii the final acts of sensor) discrimination. To the inxcstigatoi searching for an explanation in neurophysiological terms of sense-quality discrimina- tion, the rather beautiful pit lure ol' lonotopie organi- zation in the auditor) system is, at first, both reassur- ing and exciting, reassuring in that it suggests that a simple principle, spatial organization, max be funda- mental for discrimination of a sensory quality and exciting because of the possibilities lor experimental test. When the experimental lest is made, the results are disappointing, at least at liist consideration. Complete bilateral ablation of the tonotopicallx organized audi- tory cortex produces little or no elleet upon the ca- pacity of experimental animals to discriminate changes in frequency of tones (36, 144, 156, 179). Of course, the tonotopic organization exists in sub- cortical centers and, as we have alread) seen, learned responses to sound cues can be made in the absence of all cortex, nexei theless, the results may seem some- what surprising in \ iew of the fact that the auditory cortex provides a better structural possibility for dis- crete spatial differentiation than lower centers which have a lesser number of nerve cells and consequently lesser possibilities of discrete connections. At least, it might be expected (if topographic organization is important in sense-quality discrimination) that the fineness of frequency discrimination might be altered by ablation of the auditory cortex. Since the initial studies of Blix, Goldscheider and 1460 HANDBOOK Ol I'insKlLOGY m 1 Roi'HVsioi.iKJV hi von Frey, there has been a continuous search for special receptors subserving touch, deep pressure, cold, warmth and pain. (These arc considered in detail in Chapters XVII through XIX of ihis Hand- book.) The early discovery of "spots' of special sensi- tivity mi the skin encouraged the belief that each of these different qualities" has its own kind of re- ceptor. In fact, .1 little evidence and considerable imagination led to the assignment of a particular type of receptor for each quality: Mcissner's corpuscles for touch, Pacinian corpuscles for deep pressure, Krause end bulbs for cold, Rullini cylinders for warmth and free nerve endings for pain. The results of experiments in which 'dissociation' of these separate miim' qualities was produced by peripheral nerve section and regeneration (24, 29, 43, 58, 89, 1 1 7, 154, 177, 190, 209), or by procedures designed to produce temporary alteration of skin sensitivity, and the dis- covery that the qualities were selectively disturbed by spinal cord lesions (57, 86, 105, 197, 214, 221, 222), all added up to a quite convincing argument in favor of a place (or specific nerve energy) explanation of somesthetic qualities. Despite the Imlk of evidence ac- cumulated by researchers holding this point of view and despite many thousands of words written in support of the evidence, there were a few dissenters w ho cited phenomena difficult to account for by a place theory and who produced alternative theories and some evidence to support these theories (for details, sec Chaptei XVII). Particular credit must be given to Nafe (i", 1 153); to Bishop (23); to Jenkins (10 ;>, and to Wcddell, Sinclair, Lelc and their collaborators (K)i) for questioning the adequacy of the evidence produced to support the explanation of discrimination of pain, touch, warmth and cold, simply in terms of a specific receptor and specific neural unit theory. It now seems clear that an adequate theory of dis- crimination of somesthetic qualities must recognize thai the same receptor endings are stimulated l>\ differenl kinds of stimuli and that discrimination of quality cannot be based upon events in separate sensory-neural units but only upon differences in temporal and spatial pa 11 cms of events in the same oi in similar units (85, 91, 121, 192,217 219,229 231). The possibility remains that al the spinal cord level there is some differentiation of the paths subserving the qualities ol somesthesis At the thalamii and corti- 11 1 )i modalities Since tic same principle "I specific nerve energies has been used to account foi the difference in quality and the difference in modality, it has made little difference to research whether touch, temperature and pain were classified parate modalities "i as separate qualities "I 9 esthesis cal levels there is almost no evidence to suggest that the separate qualities arc represented in spatially separate areas (187). Studies of the effects ol ablation of the cortical projection areas of the somesthetic sy stem or of lesions in subcortical pathways and centers are, in com- parison to similar studies for vision and hearing, rela- tive!) scarce and the quality of the evidence is less satisfactory because of the difliculty in adequately controlling the parameters of stimulation. In experi- mental investigations only the tactual and kinesthetic senses have been examined with any care. The kinds of deficits which occur arc described in other sections of this chapter. Early psychophysical studies of taste indicated that different spots or regions of the tongue were par- ticularly sensitive to stimuli which produced the sensations of salty, sweet, bitter and sour. Whether these are the basic taste qualities, and the only ones, may be questioned. (See the discussion of taste by Pfaffmann in Chapter XX of this Handbook, 1 At least they are the ones most readily differentiated and named by the human subject not only under experi- mental conditions but in common experience. At- tempts to identify receptors having structural charac- teristics which might indicate differences in function have met with no success. Nor has it been possible to explain satisfactorily on the basis of the chemical composition of a given substance the taste that it will produce. When substances known to arouse the sensations of salty, sweet, sour and bitter are applied to the tongue of experimental animals, records of activity in single fibers of the nerves supplying taste receptors show that some fibers are activated l>v acid alone, others hv acid or salt, and still others bv acid or quinine (ibfi, 167). Experiments bv Kimura & Beidler (11 1) provide evidence thai the same taste bud may be excited by chemical substances which in man arc known to produce more- than one quality of sensation. This kind of experimental analysis in peripheral units leads to the inference that in response to substances which produce separate taste qualities, patterns of neural impulses, which differ both spatially (different nerve fibers) and temporally (frequency in individual fibers) arc transmitted to the central ner- vous system Electrical stimulation of the glossopharyngeal and chorda tympani have made possible the mapping ol cortical areas for taste in the rat and cat (21, 159) although the possibility thai these a teas may be tactual areas for the tongue cannot be completely ruled oui SENSORY DISCRIMINATION I 46 I Bilateral ablation of the cortical areas for taste in the rat has been shown to result in a deficit in discrimina- tory ability (21). A similar deficiency in taste dis- crimination has been found in the monkey after bi- lateral ablation of the cortex of the anterior part of the island of Reil, the operculum and the anterior supratemporal plane (13). The cortical areas for taste in the monkey have not been defined by electro- physiological methods. Of all the sensory systems the investigation of the olfactory system has been particularly unrewarding. (Present knowledge is summarized by Adey in Chap- ter XXI of this Handbook.) Schemes to classify stimuli which arouse sensations of odor have not been satis- factory even from a subjective standpoint. While man can discriminate different odors on a qualitative basis, there are no particular qualities which appear to be distinctive in the same sense that sweet, sour, salty and bitter are for taste. As in taste, substances which arouse similar sensations are not necessarily similar in chemical composition. It is difficult to expose the sensory receptors for olfaction and also difficult to isolate and record from single neural uniis of the peripheral olfactory system. Adrian (4) has been successful in recording activity of units of the olfactory bulb and the results of his experiments are of special interest in thai it appears that different patterns of response arc elicited by different stimulating substances. Lack of accurate knowledge of the central connec- tions of the olfactory system, as well as inadequate control of the parameters of stimulation, has made impossible anything but the crudest kind of experi- mentation through the use of the ablation method. About all that can be said at present is that the only- lesion which has been shown to have an effect upon capacity' for olfactory discrimination is bilateral sec- tion of the olfactory tracts (7, 8, 200, 201). space discrimination.1'- For in. in and for many of the higher mammals, the visual and somesthetic systems are usually thought of as being the senses of primary importance in discriminations involving the localiza- tion of objects in space. This is true, vision is the sense used in most instances for localization of objects at a distance; touch as well as vision, for objects near the body. The auditory system is of less importance. Sources of sound can be localized as to angular position with reasonable accuracy, although some 12 For an interesting discussion of spatial and temporal events in the central nervous system as related to the space and time dimensions in the external world, see Davis (48). Space discrimination is discussed by Teuber, Chap. LXV, this volume. confusions arise; only crude estimates can be made of distances. Considered not as a system providing accurate localization of objects in space but as a system which provides a background of information about environment, it may be argued that the auditory system is not of secondary importance in space dis- crimination. This function of auditory cues has been stressed by Ramsdell (172) and by Myklebust (150) in discussing the problems of the deafened individual. Other sensory systems also play significant roles in space discrimination. Visual and tactual discrimina- tions of space can be made because postural adjust- ment and orientation of the body are maintained. The kinesthetic and vestibular systems are essential for the maintenance of posture and orientation. Moreover, visual and tactual perceptions of distance and position of objects are in part at least learned; tactual and particularly kinesthetic cues are critical in this learning. The remaining sensory systems, taste and olfaction, do not, as far as we know, con- tribute significantly, to space discrimination and will not be considered further in the present discussion. At the periphery, visual and somesthetic space are represented as two-dimensional maps. Objects at a distance project upon the retina in an organized spatial pattern. Objects broughl into contact with the body produce spatial patterns of excitation in the skin receptoi - Evidence from anatomic, il and electrophysiological studies shows clearly that for all species of animals that have been studied, there is topographic projec- tion of the retina upon higher centers up to and in- cluding the cerebral cortex (jj, 33, 34, 42, 138, 160, 1711, 202) Likewise, for the somesthetic system, different regions oi the body are topographically projected upon the cerebral cortex and this organiza- tion is maintained in pathways and centers inter- vening between skin and cortex (7,4, 55, 139, 149, 226). In the auditory end organ, outer space is nol represented l>\ a spatial map. As we have seen (p. 147)8), it is frequency of the sound stimulus that is represented spatially along the basilar membrane. Sounds at different angles from the axis through the two ears differ in their time of arrival, in their rela- tive intensity and in their complexity at the end organs. Psychophysical studies (126) have shown that for tones of low frequency (1000 cps and below), time differences (time of arrival or phase) at the two ears provide the principal cues for space localization; intensity differences are more important for tones of high frequency (above approximately 4000 cps). There is a range of frequencies between 1000 and 4000 l^(')2 HANDBOOK OF PHYSIOLOGY ^ NEUROPHYSK It OGY III i ps for which localization is less accurate than for either lower or higher frequencies. For complex sounds, time, intensity and quality differences are all probably utilized. How time and intensity differences at the periphery are coded in neural centers is not as yet completely clear. A number of investigators (hiring the past 30 years have suggested that these differences become place differences in the central nervous system, i.e. that auditory space is in some manner topographically represented in the brain (18, 19, 25, 102, 1681. There is no experimental evidence to support this view, but it has not been carefullv explored. In a scries of experiments Rosen- zweig and his co-workers (182-186) have recorded electrical responses at the cochlea, cortex, and inferior colliculus of the cat when activity in the auditory system is set off by clicks which are varied in intensity and in time of arrival at the two ears. They have found that the pattern of response recorded by a gross electrode at the cortex or the inferior colliculus varies systematically as a function of the interval separating two clicks successively presented, one to each of the ears. The cortical response also varies when the two clicks arc presented simultaneously but one of less intensity than the other. Because of methodological difficulties encountered in systematically exploring different regions of the body and in restricting stimulation to the appropriate receptors, there have been few attempts to discover the organization of the pathways from kinesthetic receptors to cortex. The electrophysiological experi- ments of Mountcastle et l capacitv for pattern discrimination. \ m. ill remnant oi auditor) cortex permitted re- let 11" oi the pattei n disi rimination habit I hat the loss after total ablation was .1 loss ol capacit) to rei ognize the patterning of the physical stimulus was mown bv control tests which revealed thai the same animals which could nol do pattern discrimination could discriminate changes in frequency. Goldberg el al. (76) have also reported that ability to discrimi- nate temporal patterns of tones is affected in the cat after bilateral ablation of insular-temporal cortex ventral to the region defined as auditory cortex by evoked potential mapping. Considered only casually, the striking parallel be- tween the visual and auditor) systems with respect to cortical function in pattern discrimination may not at first be apparent. In the case of the v isual system, we speak of a loss of capacity to discriminate spatial patterns; for the auditory system, a loss of capacity to discriminate temporal patterns. But, when we look more closely at the order of events in the central nervous system, it becomes clear that the spatial patterns at the retina and the temporal patterns at the basilar membrane become, in the peripheral nerve and in higher centers, patterns of nerve impulses which differ both in space and in time. It is probably impossible to stimulate the retina in such a fashion that a pattern of nerve impulses is set off having only space, or place, differences. As Lashley has pointed out (120), even with tachisto- scopic presentation of light stimulus, the aftereffects will likely produce a train of nerve impulses. Further- more, if the activity in the first or second order neurons had only spatial patterning, it must inevitably have temporal patterning as well bv the time it arrives at higher centers because of such factors as differences in conduction rale and dillerences in number of synapses crossed. Similarly, temporal patterns of tones presented to the cochlea elicit not onlv a temporal sequence of events in the peripheral nerve but, at least when more than one tonal frequenc) is used, there will be spatial dillerences as well. Capacitv to discriminate tactual patterns has also been measured before and after cortical ablations. The experiments have been less adequate than those for vision and hearinu, both from the standpoint of stimulus control and evaluation 11! the locus and extent of the expei inientallv placed lesions. Neverthe- less, the results tend to support the evidence from visual and auditor) experiments that the cortical projection aie.i pl.iv .1 critical role in discriminations of spatial and temporal patterns. Rats and eats show deficits in ability to discriminate differences in rough- ness a I let lesions ol somatic areas I and II I J ;j, 2 Monkeys are less able to discriminate forms such as pyramid versus cone by tactual cues (|-j In man, likewise, discriminations <>l temporal and spatial patterns are severel) affected bv lesions of the parietal cortex I |l>, (JO, <>t' SENSORY DISCRIMINATION 1465 Relation of the Primary Sensory Systems to Other Systems In considering the neural mechanisms of sensory discrimination, our first emphasis has been on the structure and function of the main afferent systems leading from the peripheral sense organs to higher centers of the brain. Without knowledge of the sensory input, we cannot begin to talk about the neurophysiol- ogy of sensory discrimination. On the other hand, even with more exact and detailed knowledge of the main afferent systems than we now possess, we would still be far from an understanding of sensory discrimi- nation. We must also consider the control of sensory input, the effects of activity of nonsensory systems on this input, and the manner in which the sensory input leads to motor response. The living organism at any given moment i^ subjected to a variety of external stimuli which may act upon sense organs and arouse activity in sensory pathways. For the higher animals, behavior does not appear to be controlled simply by the total pattern of stimuli acting upon the sense organs. Careful observation of behavior leads to the inference that there is some control exercised over the sensory systems so that a systematic, not random, switching occurs, thus allowing the input of one sensor) channel to guide the organism's responses at one point in time and another sensory channel to take over when appropriate. Or, perhaps, two or more channels maj be open simultaneously or in such rapid alternation that the organism's actions are seemingly guided by .1 combined input. Psychologically, we talk about this control of sensory input under the heading of atten- tion. Introspective psychologists found it necessary to add clearness or attensity as an attribute of sensation along with such other attributes as intensity, qualiiv, extensity and duration. Physiologists, psychologists and neurologists studying brain function cither in the experimental laboratory or in the clinic have also felt impelled, or at least have found it convenient, to use the concept of attention in describing the discrimina- tory behavior of both normal subjects and of subjects in which damage has been done to brain centers. (Attention is considered in Chapter LXIY by Lindslcy in this Handbook.) Any investigator who has tested the sensory dis- criminations of animals in a multiple choice situation is familiar with a pattern of behavior which often occurs when the discrimination is made very difficult. The experimental animal which has been making appropriate responses in less difficult discriminations suddenly appears to ignore the stimulus cues and adopts a position habit such as always selecting the stimulus on the right. This behavior can be described most readily by saying that the animal acts as if it does not 'attend' to the stimulus cues which the experi- menter is presenting. This and other kinds of behavior which can, likewise, be conveniently described as deficits in attention are seen in animals with brain lesions in test situations in which the normal animal makes appropriate responses. It is an old and well-substantiated finding that animals with frontal lobe ablation are deficient in ability to perform on a delayed response test (97, 98). (See Chapter LIY by Pribram in this Handbook.) Evidence from studies in which improvement in performance has been shown to occur if animals are tested in the dark or arc given mild sedation suggest that the animal with frontal lobe lesions is more distractablc; he is less able to maintain attention to the relevant cues in the test situation (136, 212). Cats with lesions of the auditory cortex show a loss in ability to perform well in a test which requires localizing a sound in space and approaching the source of the sound in order to obtain a food reward (157). It has been suggested (hit the poor perform- ance of these cats as compared with that of normal animals ma) be due in part to an inability to 'main- 1.1111 attention' to the sound cues. Not only in the reports of animal experiments but also in the clinical literature one finds man) studies in which loss in some aspect of the ability to attend to sensory cues is described in patients who have suffered brain damage. Only a lew examples will be cited here. In their studies of the somesthetic sensibilities of patients with parietal lobe damage, Head & Holmes (90) describe a number of phenomena which they attribute to deficits in ability to attend to tactual stimulation. When a series of von Frey hairs is used to stimulate the skin in .1 region of the body affected by damage to the parietal lobe of the opposite side, the patient's responses are often quite erratic. He may respond frequently to stimuli as weak as those per- ceived on the normal side of the body but still fail to respond consistently to much stronger stimuli. Head offered the very interesting suggestion that such phenomena may be explained as defects of local attention. He thought of local attention, attention as related to a given sensory modality, as being a function of the cortical projection area of that modality. He recognized that there might also be a more general faculty of attention which would be altered by a variety of conditions affecting the brain. I (l.l. II.WDHoeiK ()K F'HYSIOLOGY NEUROPHYSIOLOGY III In liis summary of the literature on the function of the parietal lobes, Critchley has cited numerous examples from studies of his own and of other investi- gators (46). Phenomena observed during 'double stimulation,' usually of bilaterally symmetrical parts of the body, bring out in a clear fashion disturbances in sensory discrimination that can be most easily described as losses of capacity to attend. In some pa- tients with a unilateral parietal lobe lesion such that the projection area of one hand is presumably dam- aged or destroyed, tests are made in which the affected and unaffected hands are stimulated simul- taneously. The patient when blindfolded may report feeling an object only in the normal hand if two like objects are simultaneously placed in both hands. If the object is removed from the normal hand, the pa- tient may then report the presence of an obiect in the affected hand. These and similar phenomena suggest that the deficit may be one in control of attention. While such observations as those cited in the above examples were reported frequently, researchers in the field of sensory discrimination have been reluctant to deal directly with the problem of attention. This re- luctance is understandable. It is due to a lack of knowl- edge of any neural system which might interact with the main sensor) systems in such manner that ap- propriate control of input channels might occur. The discovery and clarification of the functions of the as- cending reticular system and of the diffuse thalamo- cortical projection systems have not only dissipated the scientists' reluctance to face the problems of attention but has broughl these problems to the fore. It is somewhat premature as yel to propose a neural theory of attention, but is is possible to cite a few experiments which bear upon the nature of the inter- action between the specific and unspecific afferent sy stems. It has been show n that in the brain stein the ascend- ing reticular system (as discussed in Chapter I, II of this Handbook) receives collaterals I mm the specific afferent pathways of the visual, auditory, somesthetic vestibular systems (199). When the peripheral sense organs are stimulated, impulses are condui b d therefore nol only to the primary projection areas of the cortex via the specific afferent pathways but, after .1 longer latent ) due to delay in passage through many synapses in the reticular formation, to wide-spread .us ol the cortex 1 198) .is well .is to the projection II the specific afferent pathways are 11 .msec ted, sen I organ stimulation fails to produce the ivpic.il short latency evoked responses in ihe primary projei Hon areas; activation of the BEG does occur in unanesthetized preparations (60, 62, 129, 130). In the anesthetized animal, the unspecific route via the reticular system appears to be blocked while the specific afferent pathways remain open; at the cortex, good evoked responses can be recorded from the primary projection areas during sense organ stimula- tion, activation of the EEG does not occur (61 ). These and other results suggest the hypothesis that interac- tion of the activity set off in the specific and unspecific systems must occur during sensory stimulation if the sensory stimulus is to be attended to and perceived clearly. That the cortex plays an important role in this interaction might be inferred from some of the evidence cited above. Lindsley, for example, has sug- gested that during sensory stimulation the activity arriving at the cortex via the unspecific system "sets the stage for the spreading and elaboration of the effects which reach the primary receiving areas. Its relative nonspecificity would favor this. If it acted bv resetting excitability cycles of many neuronal aggre- gates so that a greater statistical availability of cellular units or patterns of units could at once be made possi- ble, it should facilitate the chances of the primary message being received and transmitted in secondary or association areas and eventually leading to a re- sponse through the motor system" (127, p. 329 A number of investigators (35, 99, 114, 127, 128, 147) have outlined in broad terms a theory which attempts to account, in ncurophvsiologic.il terms, for a general arousal mechanism and a specific attention mechanism. The results of studies of experimental abla- tions in animals and of cortical damage in man sup- port this notion of two systems, one having to do with maintenance of a general state of alertness, the other with control of direction ol attention. In relation to the general alertness system, note should be taken of the proposal of Granit as to one possible significance ol the spontaneous activity which has been recorded in peripheral sensory nerves, lie suggests that the sense organs may serve as 'energizers'; the spontaneous activity in peripheral nerves feeds into the ascending reticular system via the collaterals from the- main afferent pathways and helps to maintain a general st. He of wakefulness or alertness (82 ). This, .is ( h.init points out, is only one possible role of the spontaneous .ictivitv. As we have noted above (p. 1454) inhibition .is well .is increase ol spontaneous .utiviiv m.iv occur during sense-Organ stimulation. This kind of mecha- nism has greater information-handling capacitv than one which is silent except during stimulation. Another important recent development in neuro- physiology which must be considered in relation to SENSORY DISCRIMINATION [467 both the specific and unspecific afferent systems is the discovery of evidence as to the functioning of cortico- fugal connections from the primary projection areas to subcortical centers of the specific afferent systems (31, 101, 143, 158, 175, 176), to the reticular forma- tion (31, 59, 92, 101), and from both the reticular system and subcortical centers of the specific afferent systems to peripheral end organs (67, 79, 81, 83, 93, 1 10, 173, 174). For many years there has been some anatomical evidence for centrifugal fibers in the primary sensory pathways but until recently the evi- dence as to structural connections was rather sketchy and there was little or no functional evidence to indi- cate the possible physiological role of these recurrent tracts. (This field is the subject of Chapter XXXI by Livingston in this Handbook.) Electrophysiological experiments have provided direct evidence that the input of the sensory svstcms may be controlled at a number of levels, including the peripheral end organ itself, by feedback through the centrifugal pathways. For example, control ol muscle spindle discharge has been demonstrated by Granit & Kaada (83); of retinal discharge bv Granil (81) and of cochlear discharge by Galambos (68). It appears that the centrifugal activity may affect sen- sory input at the end organ in several ways: a) b) influencing the level of spontaneous discharge, b) by facilitating the peripheral excitatory process; or c) by inhibiting the peripheral excitatory process. Similar control may be exercised by higher centers upon those lower in the main afferent systems; by specific afferent centers on the reticular system; and by the reticular system on centers of the specific afferent systems. The total picture of interacting specific and un- specific afferent systems and of centrifugal control of pathways within each system and between the two is one of great complexity but one which must be under- stood for a neurological theory of sensory discrimina- tion. But this is still only a part of the total problem. By sensory discrimination we mean a discriminatory response to sensory stimulation. In the present chap- ter we have confined our attention to the input or sen- sory side of the picture. To account for motor re- sponse, at least two other subsystems within the central nervous system have to be considered: a) the centers and pathways which are involved in motivation of behavior and b) the motor system. Furthermore, the manner in which a connection is established between the sensory and motor systems is the most puzzling and perhaps the most important problem of all. From one viewpoint it can be regarded .is the central theme of the subsequent chapters. REFERENCES 1. Ades, H. W. J. Neuropath. & Exper. Neurol. 5: 60, mi1' 2. Ades, H. W. and D. H. Raab. ./ Neurophysiol 1 -■ mi, ■949- 3. Adrian, E. D. J. Physiol. 61 : 49, 1926 4. Adrian, E. D. Acta physiol. aandinav. 29: 5, 1953. 5. Adrian, E. D. and Y. Zotterman. J. Physiol, in 151, 1926. 6. Adrian, E. D. and Y. Zotterman. J. Physiol. In 1,65, 1926. 7. Allen, W. F. Am. ./. Physiol. 128: 754, 1940. 8. Allen, W. F. Am. ./. Physiol. 132: 81, 1941. 9. Allen, W. F. Am. J. Physiol. 147: 454, 1946. 10. Allen, VV. F. Am. J. Physiol. 151 : 325, 1947. 11. Amassian, V. E. Electroencephalog. & Clin. Neurophysiol. j: 415. 1953- 12. Andersson, S. and B. E. Gernandt Acta ntn-laryng. Suppl. 116: 10, 1954. 13. Bagshaw, M. H. and K. H. Pribram. J. Neurophysiol. 16: 499. !953- 14. Bailey, P. and G. von Bonin. The Isocortex of Man. Urbana: Univ. Illinois Press, 1951. 15. Bard, P. Harvey Lectures 33: 143, 1938. 16. Bay, E. Brain 76: 515, 1953. 17. Be are, J. I. Greek Theories of Elementary Cognition. Oxford: Clarendon Press, 1906. 18. Bekesy, G. Physik. Ztschr. 31 : 824, 1930. ig. Bekesy, G. Physik. Ztschr. 31 : 857, 1930. 20. Bekesy, G. J. I Sim Am 27 830, 1955. Benjamin, K M \m> ( !. Pfaffman. ./. Neurophysiol. 18: 56, 1955- _•-'. 8ii.su .Mi. A ExperimenUel-anaiomisch anderzoek omlrent de corticofugale oplischt verkindigon by nap en konijn. Amsterdam P. H. Vermuelcn, 1929. 23. Bishop, G. II Physiol. Re, 26 77 1946 24. Boring, E. G. Quart. J. Exper. Physiol. 10: 1, 1916. 25. Boring, E. G. Am. J. Psychol. 37: 157, 1926. 26. Boring, E. G. A History of Experimental Psychology. New York : Century, 1929. 27. Boring, E. G. The Physical Dimensions of Consciousness. New York : Century, 1933. 28. Borinc, E. G. Sensation and Perception in the History of Experi- mental Psychology New York : Century, 1042. 29. Bovvden, R. I. M Medical Research Council Special Report No. 282. London. 1954, p 298. 30. Brazier, M. A. B. J. Neurophysiol. 20 2 i 2. 1957. 31. Bremer, F. and C. Terzuolo. ./. physiol., Paris 45: 56, '953- 32. Bromiley, R. B. J. Comp. sir Physiol. Psychol. 41 : 102, 1948. 33. Brouwer, B. Deutsche Ztschr. Nervenh. 89: 9, 1926. 34. Brouwer, B. and W. P. C. Zeeman. J. Neurol. Psychopath. 6: 1, 1925. 35. Buser, P. and A. Roger. IV Congr. Internal. Electro- encephalog. el Neurophysiol., Rapp., Discus, et Document. Brussels: Acta Med. Belg., 1957, p. 417. i4<)B IIWUHOOK OF I'HYSIOLOGY mclroi'Hysiology m 36. Butler, R. A., I. T. Diamond and W. I) Neff. ./. Neuro- 76. physiol. 20: 108, 1957. 37. Cham;. II 1. .1. Res. Nen S Mini. Dis., Proc. 30: 430, 77. '95-*- 78- 38. Chow, K L. and P.J. Hi it. lirmn 76:625, 1953. 79. 39. Clark, W I. I.i duos. Suture, London 146:558, 1940. (X) (1 \kK. \\ I. Le Gros. Docum. ophth. 3: 57, 1949. 80. 41. Clark, W E. Le Gros and I. Chacko. Suture, London 81. 160: 1 23, [947. 82. 1- Cl ark. \V. E. Le Gros and G. G. Penman. Proc. Roy. Soc, London, ser. It 114: 291, 1934. 83. I 1 1 obb, S. AM A. Arch. Neurol ^J Psychiat. 2: 505, 1919. II Tuns. R.J. Siumfiliwinl. 19:317, 1956. 84. 45. Cole, J. and P. Glees. J. Neurophysiol. 17: 1, 1954. 46. Critchley, M I In Parietal Lobes. London: Arnold, 1953. 85. 17 Davies, P. \\\, A L. Berman and V. B. Mountcastle. Am. J Physiol 126 25P, 1954. 86. 48. Davis, II F.lectioencephalog. & Clin. Neurophysiol. 8: 185, 87. 1956. 88. |'l l)\VIS. II, S K Sll VERMAN AND D. R. McAuLIFFE. J. 89. 1 oust S01 I"' 23: 40, 1951. 90. 50. DENNIS, W. Readings in the History of Psychology. New York: 91. Appleton, 1948. 92. 51. De Voss, J. C. AND R. Garson. J. Animal Behav. 5: 115, '9 •5- 93- 52. Diamond, I. T. ami \\ I) Nefp. ./. Neurophysiol. 20: 300, '957- 94 53. Donner, K. O. Acta physiol. scandinav. 21, Suppl. 72: I, 95. 1950. g6. 54. Di ssi k 111 Barenne, J. G. Quart. ./. Exper. Physiol. 9: 355, 97. 1916. 55. Dusser de Barenne, J. G. Proc. Roy. Soc. 963: 272, 1924 g8. -,ti I.rulkar, S. D., J. E. Rose and P. W. Davies. Bull. Johns Hopkins Hosp. gg:55, igs6. 99. 57. Foerster, O. and O. Gagel. Ztschr. ges. Neurol. Psychiat. ■ 1, 193a. too. 58. Franz, S. I. J. Comp. Neurol. 19: 107, 1909. 59. French, J. I)., R. Hernandez-Peon and R. B. Livino- ioi. ston. J. Neurophysiol. 18 74, 1955. 60. French, J. D. and II. W. Magoun. A.M. A. Arch. Neurol. 102. Psychiat. 68: 591, 1952. 103. 61 French, J I), M. Verzeano and II. W. Maooun. I \l A \rch Veurol Q Psychiat. 6g: 505, 1953. 104. 62. French, J I), F. K von Amerongen and II. \V. Ma- 105. .. II Davis / \ .n.phwinl. 7 2H7, 1044. 71. Gai imbos, R I I Rose, l< B. Bromilei smi J l< 112. Mi imiis / Veurophysiol 15 J59, 195a 113. Gerebtzopf, M. A. Anh. internal, physiol. 50: 59, 1940 114. 73. Gernandt, B, Acta physiol. nandinav. 21: 73, 1950. 115. 74. Girden, I. Im ./ Psychol 53 (71, ig4o. 116, ben 1 1 \ \li 1 11 1 k. < . Finch «jd E. Culler. J. 117. p Psychol. 21 367, ig |6 Goldberc, J. M., I. T. Diamond and W. D. Neff. Fed. Proc. 16:47, '957- Granit, R Sature, London 151: II, 1943. GRANIT, R. ./. Seurophysiol. 8: 195, 1945. Granit, R. Sensory Mechanisms uf the Retina. London: I Klord, 1947. Granit, R. ./. Neurophysiol. 1 1 : 239, 1948. Granit, R. ./. Neurophysiol. 18: 388, 1955. Granit, R. Receptors and Sensory Perception. New Haven: \ ale I ni\ . Press, 1955. Granit, R. and B. R. Kaada. Acta physiol. scandina: 27 '3°. '952- Gross, N. B. and \V. R. ThuRLOW. ./. Seurophysiol. 14: 4°9. '95'- Hagi-.n, I... II Knoche, D. C. Sinclair and G. Weddell. I'nic. Roy. Soc, London, ser. B 141 : 279, 1953. Hamby, \V. B. J. Seurosurg. I I : 378, 1954. Hartline, II. K. J. Cell. & Comp. Physiol. 5: 229, 1934. Hartline, H. K. J. Opt. Soc. Am. 30: 239, 1940. Head, H. Brain 28: 99, 1905. Head, H. and G. Holmes. Brain 34: 102, 1911-12. Hensel, H. and Y. Zotterman. J. Physiol. 115: 16, 1951. Hernandez-Peon, R. and K.-E. Hagbartii. J. Seuro- physiol. 18: 44, 1955. Hernandez-Peon, R. and G. Schf.rrer. Fed. Proc. 14: "9, '955- Hind, J. E. ./. Neurophysiol. 16:475, '953- Hines, M. A. Res. Sen. ss Meni. Dis., Proc. 13: 26, 1934. Holmes, G. Brain 50: 413, 1927. Jacobsen, C. F. A.M. A. Arch. Neurol. *s Psychiat. 33: 558, 1935- Jacobsen, C. F. Comp. Psychol. Monogr. 13 Serial No. 63, '93°- Jasper, II . II In: Brain Mechanisms and Consciousness, edited by J. F. Delafresnaye. Springfield: Thomas, 1954. Jasper, 11. II. and G. Ajmone-Marsan. .1 Res. Nerv. & Menl. Da., Proc. 30: 493, 1952. Jasper, II II . C. \|mone-Mars.w and J. Stoll. A.M. A Arch. Neurol. & Psychiat. 67: 155, 1952. I I 1 press, L a. ./. Comp. & Physiol. Psychol. 41 : 35, 1948. Jenkins, \V. L. In: Handbook oj Experimental Psychology, edited by S. S. Stevens. New York: Wiley, 1951. Jkrison, H.J. vndW. D. Neff Fed. Proc. 13:237, '953- Kmin. I. A \mi R. W. Rand. ./. Neurosurg. g: 61 1. Katsi'ki, \'. and II. Uchiyama Pun Jap. Acad. 32:67, '956- (Catsuki, Y., II. L'ciiiyama, 1 Waianaih \m> N. M \ki \ \m\ Proc. Jap. Acad. 32: 1, 1956. Kjbmpinski \\ II / Seurophysiol. 14:203, 1951. Ki \nedy, T., N. Kiang and S. Sutton. Am. J . Physiol. 183: 634. 1955. Kerr, I). I. B, and K..-I. Hagbarth J Neurophysiol. 18: 362, [955. Kjmura, K \mi I . M Bi idi 1 r. Am. J. Physiol. 187: 610, tg ,6 Ki 1 \ 1 r, 1 1. Psychol. Bull. 24 si''. 11127. Ri 1 vi r, II. J. Psychol. 2 49, 1936. Ki 1 \ 1 h. 1 1. ./. Psychol. 4: 383, 1937. kl 1 VI R, I I. ./- l'*\ill(il. I 1 23, I i|| I KlOver, II Symp. Soc. F\/'cr. Biol. 7: 253, mi- Lantj r, I. M I Ra Vtn t Went Du . Pro* 15: 437, 1934 SENSORY DISCRIMINATION 1469 118. Lashley, K. S. J. genet. Psychol. 37: 461, 1930. iig. Lashley, K. S. Symp. Soc. Exper. Biol. 4: 454, 1950. 120. Lashley, K. S. A. Res. Nerv. & Merit. Dis., Proc. 30: 529, '952- 121. Lele, P. P., G. Weddell and C. M. Williams. J. Physiol. 126: 206, 1954. 122. Lennox, M. A. J. Neurophysiol. 19: 271, 1956. 123. Lennox, M. A. and A. Madsen. J. Neurophysiol. 18: 412, '955- 124. Licklider, J. C. R. Doctoral Dissertation. Rochester: Univ. of Rochester, 1942. 125. Licklider, J. C. R. Experientia 7: 128, 1951. 126. Licklider, J. C. R. In: Handbook of Experimental Psychol- ogy, edited by S. S. Stevens. New York: Wiley. 1951. 127. Lindsley, D. B. Ann. Rev. Psychol. 7: 323, 1956. 128. Lindsley, D. B. In: Electrical Stimulation of the Brain, edited by D. E. Sheer. Austin: Univ. Tcx.is Press, 1958. 129. Lindsley, D. B, J. Bowden and II. W. Magoun. Electro- encephalog. & C.lm. Neurophysiol. 1:475, '949- 130. Lindsley, D. B., L H. Schreiner, W. Knovvles and II. VV. Magoun. Electroencephalog. Ci Clm. Neurophysiol. 2:493. i95°- 131. Lipman, E. A. Am. ./. Psychol. 62: 215, 1949. 132. LONCO, L. P. S., I". M. FORSTER AND T. L. L.AUTH. Neurology 7: 655, 1957. 133. Lowenstein, O. and A. Sand. ./ Physiol. 99: 89, 1940. 134. Lowenstein, (). and A. Sand. Proc. Roy. Sue, London. Set. B 124: 256, 1940. 135. Madsen, A. and M. A. I.i s\c.\. ./. Neurophysiol. 18: 574, '955- 136. Mai.mo, R. B. ./. Neurophysiol. 5: 295, 1942. 137. Marquis, D. G. A. /. Nerv. Or Merit. I>i> , Proc. 13: 5,58, 1934. 138. Marshai.i , \\ '., S. A. T\i urn and 1 1 W. Adi s. J. A physiol. 6: 1 , 1 943. 139. Marshall, W., ( :. \ Woolsey and P Bard ./. Neuro- physiol. 4: 1, 1 941. 140. Matthews, B. II. ( !. ./. Physiol. 71 : 64, [931 141. Matthews, B. II. ( :. ./. Physiol. 7a 1 53, K131. 142. Matthews, B. II. ( '.. ./. Physiol. 78: 1, 1933. 143. Meyer, I). B., R. C. Mills and P. Ratoosh. ./. Neuro- physiol. 17: 289, 1 954. 144. Meyer, D. R. and C. N. Woolsey. J Neurophysiol. 15: '49. '952- 145. Mickle, W. A. and II. W. Ades. Am. ./. Physiol. 176: 243. '954- 146. Miller, G. A. and W. G. Taylor. ./ Acoust Soc. Am. 20: 171, 1948. 147. Morrell, F. IV Congr. Internal. Electro-encephalog. ei Neurophysiol. Rapp., Discus, ft Document. Brussels: Acta Med. Belg., 1957, p- 377- 148. MOUNTCASTLE, V. B., M. R. CoVIAN AND C. R HaRRI- son. A. Res. Nerv. Or Mint Dis., Proc. 30: 339, 1952. 149. MOUNTCASTLE, V. B. AND E. Henneman. ./. Neurophysiol. 12: 85, 1949. 150. Myklebust, H. R. Perkins School /or the Blind Publication. No. 19, 1956. 151. Nafe, J. P. Am. J. Psychol. 39: 367, 1921. 152. Nafe, J. P. ./. Gen. Psychol. 2: 199, 1929. 153. Nafe, J. P. In: A Handbook of General Experimental Psy- chology, edited by C. Murchison. Worcester: Clark Univ. Press, 1934. 154. Napier, J. R. Brain 75: 147, 1952. 155. Neff, W. D., G. P. Arnott, and J. D. Fisher. Am. J. Physiol. 163: 738, 1950. 156. Neff, W. D. and I. T. Diamond. In: Biological and Bio- chemical Bases of Behavior, edited by H. F. Harlow and C. N. Woolsey. Madison: Univ. Wisconsin Press, 1958. 157. Neff, W. D., J. F. Fisher, I. T. Diamond and M. Yela. J. Neurophysiol. 19: 500, 1956. 158. Niemer, W. T. and J. Jimenez-Castellanos. J. Comp. Neurol. 93: 101, 1950. 159. Patton, H. D. and V. E. Amassian. J Neurophysiol. 15: 245. 1952 160. Peele, T I.. ./. Neurophysiol. 7: 269, 1944. 161. Penfield, W In Buun Mechanisms and Consciousness, edited by J. F. Delafresnaye. Springfield: Thomas, 1954 i»>-' I'l sni 1 d. U and H. Jasper. Epilepsy and the Emotional Anatomy of the Human Brain. Boston : Little, 1 954. 163. Penfield, W. and T. Rasmussen. The Cerebral Cortex of Man. New York Macmillan, 1950. 164. Pennington, I.. A. J. Comp. Psychol. 25: 195, 1938. 165. Pennington, L. A. J. Comp. Neurol. 66: 415, 1957. 166. Pfaffman, C. J. Cell. & Comp. Physiol. 17 243, io|i 167. Pfaffman, I 1 Neurophysiol. 18:424, 1955. 168. Pieron, H. Annie Psychol. 23: [86, 1922. 169. Polyak, s .1 /,', 1 \.,, .. Ment. Dis., Proc. 13: 535, [934 170. Poi.yak, S v\n R Hayashi. Brain 59: 51, 1936. 171. Raab, P. II and II \\ Adi s .1". ./ Psychol. 59: 59, 1946. 172. Ramsuiii. D. A. In: Hearing and Deafness, edited bj II Davis New York: Murray I fill. 1947. 173. Rasmussen, G. L. / Comp. A urot '■'■ \ 141, 1946. 174. Rasmussen, G. L. J. < c i< 1 in. 1953. :;-, Rasmussen, G. I- !»«' Rec. n-, )6i, 1953. 176. Rasmussen, <;. I.. Am. ./. Physiol. 183:653, 1955. 177. Rivers, W. II. k. um II Hi \d Brail 178. Rose, J. E. and V. B. Mountcastle. Bull. Johns Hopkins II ■ M, 2 |S 1954. 179. Rom, J. I \\d (: N Woolsey. In: Biological and Bio- chemical Bases W Behavior, edited by II. F. 1 Iaiiow and C. N. Woolsej Madison Uni> Wisconsin Press, 1958. 180. Rosenblith, V\ \ Progress Report II PNM-6), Psycho- \. (mi-ik Laboratory, Harvard University, 1947. 181. Rosenzweic, M R. Am. I Psychol 59: 127, 1946 i8j. Rosenzweig. M. R. Am. J. Phyi 6 17, 1951. 183. Rosenzweig, M. R. ./. Comp. ^ Ph) ot I' 269, '954- 184. Rosenzweig, M R \m> \ II Amon Experientia 11:498, '955- 185. Rosenzweig, M. R. and W A Rosi.NHinH. I' Monogr. 67: No. 13 I Whole No. 363), 1953. 186. Rosenzweig, M R. and E. ) Wyers. ./. Comp. & Physiol. Psychol. 48: 4.1 1955 187. Ruch, T. C. In: Textbook of Physiology, edited by J. F. Fulton. Philadelphia: Saunders, 1955. 188. Ruch, T. C. and J. F. Fulton. A. Res. Nerv. & Ment. Dis., Proc, 15: 289, 1935. 189. Scholl, D. A. The Organization of th, Cerebral Cortex. London: Mcthuen, 1956. igo. Sharpey-Schafer, E. Brain 50: 538, 1927. 191. Sinclair, D. C. Brain 78: 584, 1955. 192. Sinclair, D. C, G. Weddell and E. Zander. J. Anat. 86: 402, 1952. '47" iiwiiiiiioK "i i'in Mui nii'i ■ ^ NEi'Rorin-sim.i «;v m '93- SjOQVlST, <) ash I.. A WeiNSTEIN. ./. Neurophysiol. 5:69, 94 104. Smith, K U. J '. %enel. Psychol. 51: 329, 1937. ig5. Spiegel, E. a i Nen S Ment. Dis. 75:504, 1939. 196. Spiii.i 1. E A. .1 M .1 Arch. Neurol. & Psychial. 31: 469, ■934- 197. Spiller, W. G. and E.J. Mariin J.A.M.A. 58: 1489, [ 9 1 a 198. StaRZL, T. E. and II W. M agoun. J. Neurophysiol. 14: 133. «95>- 199. Star/i, I E , C. W. Taylor and 11 W. Magoun. ./. Neurophysiol. 14: 479, 1951. 200. Svvann, H. G. J. Com/, Neurol. 59: 175, 1934. 201. Swann, II. G. Am. ./. Physiol, in: 257, 1935. Talbot, S. A. and \V. H. Marshall. Am. J. Ophth. 24: 1255, 194'- ...; I isaki, I. J. Neurophysiol. 17:97, 1954- 204. Tasaki, I. Ann. Rev. Physiol. 19: 417, 1957. 205. Tasaki, I. and H. Davis. J. Neurophysiol. 18: 151, 1955. 206. Tasaki, I., E. H. Polley and F. Orrego. J. Neurophysiol. 17:454. '954- 207. Teuber, II. L. and M. B. Bender. Fed. Proc. 7: 123, 1948. 208. Thurlow, \V. R., N. B. Gross, E. H. Kemp and K. Lowy. ./. Neurophysiol. 14: 289, 1951. 209. Trotter, W. and H. M. Davies. J. Psychol, u. Neurol. 20: 2, 1913. 210. Tunturi, A. R Am. J. Physiol. 141 : 397, 1944. 211. Tunturi, A. R. Am. J. Physiol. 160: 395, 1950. 212. Wade, M. J. Neurophysiol. 10: 57, 1947. 213. Walker, A. E. The Primale Thalamus. Chicago: L'niv Chicago Press, 1938. 2 1 4. \\ mjcer, A E. A.M. A. Arch. Neurol. & Psychial. 43: 284, 1940. 215. Walls, G. L. Univ. Calif. (Berkeley) Publ. Physiol. 9: 1, 1953. 216. \\ ulzl, E. M. and V. B. Mountcastle. Am. J. Physiol. ■59:595. '949- 217. Weddell, (-'•. Ann. Reo. Psychol. 6: 119, 1955. 218. Weddell, (•., W. Paiiie and E. Palmer. Quarl. J. Microsc. Sc. 95: 483, 1954. 219. Weddell, G., D. A. Taylor and C. M. Williams. J. Anal. 89: 317, 1955. 2211 Wi ver, E. G. Theory of Hearing. New York: Wiley, 1949. 321. White, J. C. A.M. A. Arch. Neurol. & Psychial. 71 : 1, 1954. 222. White, J. C. and W. II. Sweet. A.M. A. Arch. Neurol. & Psychial. 67: 315, 1952. 223. Wing, K. G. Am. J. Ptychol. 59: 583, 1946. 224. Wing, K G. Am. J. Psychol. 60: 30, 1947. 225. Wing, K. G. and K. U. Smith. J. Fxper. Psychol. 31 : 478, '942- 226. Woolsey, C. N., W. H. Marshall and P. Bard. Bull. Johns Hopkins Hosp. 70: 399, 1942. 227. Woolsey, C. N. and E. M. Walzl. Bull. Johns Hopkins Hosp. 71: 315, 1942. 228. Woolsey, C. N. and E. M. Walzl. Am. J. M. Sc. 207: 686, 1944. 229. Zander, E. and G. Weddell. J. Anal. 85: 68, 1 951 . 230. Zotterman, Y. J. Physiol. 95: 1, 1939. 231. Zotterman, Y. Ann. Rev. Physiol. 15: 357, 1951. 232. Zubek, J. P. J. Comp. & Physiol. Psychol. 44: 339, 1951 . 233. Zubek, J. P. J. Neurophysiol. 15:401, 1952. CHAPTER L X I The neural basis of learning ROBERT GAL AM BOS CLIFFORD T. MORGAN Department of Neurophysiology, Waller Reed Army Institute of Research, Walter Reed Army Medical Center, Washington, D. C. Department of Psychology, The John Hopkins University, Baltimore, Maryland CHAPTER CONTENTS Introduction Psychological Methods and Terms Conditioning Trial-and-Error Learning Perceptual Learning Phenomena Related to Learning Reinforcement and extinction Discrimination Other methods Summary Ablation Studies Spinal Conditioning Subcortical Factors in Learning Conditioning in Decorticate Animals Maze Learning Problem-Box Learning Discriminative Learning Visual discrimination Interocular transfer Auditory discrimination Somesthetic discrimination Delayed reaction Conditioned inhibition Anticipation and perseveration Summary EEC Correlates The Alpha Block CR Cortical Evoked Potentials New Electrical Waves Subcortical Structures Summary Brain Stimulation Brain Shocks "Produce" Behavior Brain Shocks as CS and US Self-Stimulation Brain Shocks Influence Learned Behavior Electroconvulsive Seizures Summary Psychopharmacology Summary Ncurophysiological Theories Change in Central Synapses Anatomical theories Biochemical theories Glia cell possibilities Rearranged Neural Circuits Russian ideas Reverberating chains Theories from 1.1. (J Studies Other Neural Possibilities Mathematical Models Summary I liscussion and Summary One Neural Correlate? Structures involved Complexity Phylogenetic evidence 'Emotional' learning Ease of learning Maturation and Learning "Connections' The specific change in learning Motivation and attention Mechanisms in motivation Mechanisms in attention Limbic-midbrain circuit Summary Concluding Remarks INTRODUCTION we are concerned in this chapter with what takes place in the nervous system during learning, and our goal is a precise description of its mechanisms. It will however become apparent as we proceed that while 1471 '47-' IIWDHiiiiK OF PHYSIOLOGY " NEUROPHYSIOLOGY III relevant data abound in profusion, these often do not fit together in an orderly pattern, and they sometimes raise more questions than they answer. Besides this, the available methods for studying learning are frequently inadequate, and the ones best suited to the task have not been in use long enough to fulfill their promise. We will therefore be able to provide only a partial description of the mechanisms, this being derived from a statement of the problems as we see them, a summary of the data obtained by various methods and a discussion of some of the possible ways in which the nervous svstem functions in learning. The study of the neural basis of learning is inter- disciplinary; it draws on the facts and methods of a number of different fields. Experimental psychology, the discipline most directly charged with precisely defining the behavior called learning and con- ditioning, has made a substantial contribution, and familiarity with its techniques and terminology is required for any serious student of learning. For this reason, we devote the next section to 'Psychological Methods and Terms.' The several biological sciences, and particularly neuroanatomy, neurophysiology and biochemistry, arc also heavily represented in the work on learning. What they offer in the way of methods and facts at the present time appears in the sections 'Ablation Studies,' 'EEG Correlates,' 'Brain Stimu- lation' and 'Psychopharmacology.' From a study of this material, the reader will lie prepared, we trust, lor the theoretical explanations for learning that are treated in the final sections. But first .1 word about the references. Material published alter June i<)")7 has not been cited with .1 lew notable exceptions. The selected list here w.is chosen furthermore so that the reader could, by con- sulting their bibliographies, uncover practically every- thing known about the problem. We feel fairly certain of this with respect to the western literature despite the absence ol .1 single modern German citation. I he Russian . 1 1 11 1 eastern European literature is another m. iin 1 I he language diffi< ulty here has been formi- dable, and we have worked either from brief summa- ries of obviously large research programs under way for many years, or from the translation <>l as few as One paper in .1 much larger series bv the same author, Besides this, out Eastern colleagues in mam instances approai h learning problems in an interestingly diffei enl way 1 our own, and we are not entirely sure that, as reporters, out rendering of their facts and elusions his always been wholly accurate. Certainly our report is not complete, much as we would have liked it to be.1 PSYCHOLOGICAL METHODS AM) TERMS Although philosophers and scientists have long speculated about the nature of learning and have observed it in pets, zoo animals, children and fellow men, the first systematic observations of learning and memory were made by Ebbinghaus in i88~, (551. With nonsense syllables, which he invented, he carefully measured the course of memorizing and forgetting in human subjects. Then in 1906 Pavlov described (in English) his now famous discovery of the conditioned reflex, describing various factors governing its acqui- sition and extinction (180). About the same time, the maze was first employed for the measurement of learning in animals (221); and the problem box, in which an animal must discover through trial and error the solution to a problem, was used for a similar purpose (235). Thus were launched the two general kinds of methods, the conditioning method and the trial-and-error method, which have since provided the great bulk of information about learning and the various factors that affect it. It is necessary to have a clear understanding of these psychological techniques and the terminology that goes with them before turning to the neural correlates of learning and con- ditioning. Conditioning As Pavlov used the term, 'conditioning' applied to a situation in which two kinds of stimuli are presented to an animal. One stimulus, called the 'unconditioned stimulus3 (US), is any stimulus that evokes some definite response, called the 'unconditioned response' 1 I R), without any prior learning. In Pavlov's case, the L'S was food and the UR was salivation. Another stimulus, the 'conditioning stimulus' (or after Learning has occurred, the 'conditioned stimulus' 01 CS 1, is one that prior to learning evokes no significant response. In many of Pavlov's experiments, this was a bell. By pairing the CS and the US, the one presented just before the other, and doing this for a number ol trials, 1 iin concern on this point has been considembh allayed li\ the publication In Rusinov & Rabinovich (211) of an admirable brief summary "i much material relevant to this ■ hapter. THE NEURAL BASIS OF LEARNING '473 Pavlov observed that the CS (the bell) came to evoke the UR (salivation) which he now called the 'con- ditioned response' (CR). This Pavlovian method of studying learning has in recent years been called 'classical conditioning' or conditioning of the 'first type.' We shall refer to it as Type I conditioning. Type I conditioning procedures were subsequently modified by several investigators so as to employ a noxious stimulus, usually a shock applied somewhere to the skin of an animal, as the US. In this case an animal is presented with a CS (light or bell) followed by a shock (US), but only if the animal fails to make some specified response like lifting its leg (CR). In the United States, this conditioning procedure is usually called 'avoidance conditioning,' while others refer to it as defensive conditioning or conditioning of the 'second type.' Along with many others we shall call it Type II conditioning. Since what the subject does in the Type II procedure is instrumental in avoiding shock, it falls into the more general category of 'instrumental conditioning' — a term suggested by Hilgard & Marquis (98). There is evidence, however, that avoidance conditioning also involves the classical conditioning of autonomic reactions and consequently is not purely instrumental. Skinner's 'operant con- ditioning," described in the paragraph below, is a better example of instrumental conditioning. Pavlov's classical conditioning is not instrumental because the animal's response (salivation! has nothing to do with whether or not the unconditioned stimulus (food) is presented. Many years later, Konorski (121, p. 418) and Skinner (220) applied the term conditioning to a still different learning procedure. They put hungry rats in a box and delivered a pellet to them each time they pushed a lever. In this case, food was the US and pushing the lever became the CR. This situation differs, however, from both Pavlovian and avoidance conditioning in that: a) it contains no specifiable CS and h) the CR (pushing the lever) is not originally a UR to the food (US). For this reason, Skinner called it 'operant conditioning' in contrast to the first two kinds of conditioning, which he called 'respondent conditioning.' However, since the animal's response is instrumental in obtaining the food (US), this learning procedure must also be classified as a form of instrumental or Type II conditioning. Trwl-and-Error Learning Skinner's Type II procedure, however, does not differ in principle from other learning methods that have been called 'trial and error.' Thorndike's cat in a puzzle box (235) indeed must do practically the same thing as Skinner's rat, namely push a latch to get out of a box to reach food. Similarly, animals required to run through a maze to receive a food reward must learn to make a number of turns correctly to reach the goal of food. Although the animal must make a series of responses, rather than just one, these are still instrumental responses which, like the rat in Skinner\s box, are not themselves UR's and are not evoked by any identifiable CS's. Hence, it is only a matter of convention, not one of fundamental differ- ence, to call maze learning and learning in puzzle boxes trial-and-error learning while calling Skinner- box learning conditioning. Perceptual Learning In addition to the learning studied in Type I and Type II procedures, organisms also can learn some- thing about the relation of stimuli and of objects in their environment without necessarily making any overt responses. The general name for such learning is 'perceptual learning' because it invokes a change in the perception of the environment. For our purposes two varieties of ^uch learning may be distinguished, although there are others that need not be mentioned. One is 'sensory-sensory learning.' In this case, one stimulus becomes associated with another by being regularly paired with it. An auditory stimulus, for example, might be presented immediately preceding a visual stimulus. The learning that occurs in such a procedure is not immediately available for scientific study because it does not involve a response 1 [owever, l>v conditioning some response to one of the paired stimuli and then testing with the other stimulus, one can establish that sensory-sensory learning actually takes place because the tesl stimulus evokes a response with which it has not previously been paired (23). By use of electrical methods of recording brain changes during learning;, which will be described later, it is possible to measure changes in sensory-sensory learning directlv. Another kind of perceptual learning is that which is commonly called "insight.' Its central feature is that the organism learns suddenly and without trial-and- error responses, as emphasized by Kohler (119). By- attending to a problem and perceiving the relation- ships in it, the organism gains insight into its solution. Since we do not yet have any appreciable information '474 ll\M)H(«)K OF I'llYSIOLOGY NEUROPHYSIOLOGY III about the neural correlates of such learning, insight will not be discussed further in this paper. Psychologists have sometimes tried to include perceptual learning with Type I or Type II learning. Main oilier schemes for classifying learning have been proposed, and considerable research and much controversy have revolved around them. There is no reason in uo into these matters here, but the interesied reader can delve into them b\ referring to excellent m sources (45, 96, 97, 224, 236, 237). For our purposes, it is well to keep in mind the distinction between Type I and Type II learning procedures since the neural correlates of the learning phenomena ,i"i iciated with them are si imetimes different Phenomena Related to Learning In addition to the three general classes of learning, (here are three important phenomena closely related to learning thai nevertheless differ in certain respects from what we ordinarily consider to be learning. One is 'habituation' (236, p. 3881, the gradual waning of a response in the presence of continued stimulation ( aol 10 be confused with 'extinction* which will be discussed below). lire a gun and it at first startles or excites the organism. Upon the repetition of such stimulation, however, the organism makes less and less response to it. Similarly, a dog placed in a harness is al first excitable and restless, but in time it settles down and bei omes habituated to the situation. Another related phenomenon is "sensitization' (98, p. 41 I. This is an increase in responsiveness to a stimulus because of prior excitement by another stimulus. Shock an animal a few times and it is much more likely to respond excitedly to a bell or light than it would normally. Still a third phenomenon, recently studied in some detail b\ ethologists, is 'imprinting' (200, 236-238). I his is a relatively rapid 'learning' that takes place optimally al a certain time in the growing organism. \ duckling exposed at a particular time after hatching to the call of the mother duck follows her thereafter, since imprinting develops relative to her. One can, however, experimentally evoke imprinting l>\ the diK kling relative to any other soun e oi mo\ ing sound. I lii s 1- .1 permanent change in the behavior ol the inismand thus qualifies as learning, yet ii involves so far as we know neithei an unconditioned stimulus inn any iiisliunienl.il response and lienee is nol quite like any ol the three basic classes of learning. REINFORCEMENT AND EXTINCTION. Both Type I and Type II learning require a 'reinforcement.' In classical (Type I) learning, the reinforcement is the US: in instrumental (Type II) learning, it is the reward or punishment that follows some specified response. The process of providing the reinforcement, it should be noted, is also called 'reinforcement.' Alter a subject has learned a habit, the experimenter can institute an 'extinction' procedure. This is the omission of the reinforcement while maintaining all other aspects of the training routine. In Type I conditioning, extinction consists in presenting the ( S without the US: in Type II learning, it consists of presenting the CS (in avoidance conditioning) or of permitting the animal to make its learned responses (in a Skinner box, puzzle box or maze) without providing any reward for such responses. The result of an extinction procedure is, of course, an increasing tendency for the organism not to give the CR until eventually the organism is not responding at all. This process, like the procedure, is also called 'extinction.' Extinction is to be distinguished from 'forgetting' and its converse "retention." Forgetting is a diminution in the tendency to respond in the absence of further training or of experimental extinction. Teach an animal some particular response, rest him for a few days (or even hours) and he is likely to forget in some degree what has been learned. A lew more reinforce- ments will be required before the animal again responds as consistently as it did at the end of the learning trials. The usual incisure of such forgetting, however, is not the amount forgotten, but rather what is retained. If, for example, an animal was responding correctly 100 per cem of the time at the end of a series of learning trials, and then responds only 70 per cent of the time when tested after 2 weeks of rest, we say that the retention was 70 per cent. In neurophysi- ological studies of learning, we are interested not only in the neural basis of acquisition of learned responses but also in the brain events explaining retention of things previously learned. It sometimes happens, as we shall see, that acquisition and retention are separately affected by alterations in neural structure .Hid functioning. discrimination By .1 simple modification in learning procedure, cue 1 an studv the acquisition and retention ol .1 'stimulus discrimination.' This is done by pre- senting two (or inorei stimuli and reinforcing one but not the other \n animal will, for example, acquire a Conditioned discrimination between the sounds of a bell and a metronome if the bell only is reinforced THE NEURAL BASIS OF LEARNING 1475 with food (in salivary conditioning) or with shock (in avoidance conditioning). Responses eventually appear of course, to the bell but not to the metronome. It is possible to set up a still more complex discrimination between two stimuli. One might shock the left leg, for example, if a dog does not lift it upon hearing the bell, but shock the right leg if he does not lift that one upon hearing the metronome. In this case, a con- ditioned differential response is learned. In differential conditioning procedures, the re- inforced stimulus is known as the 'positive stimulus' and the one not reinforced is the 'negative stimulus.' On the other hand, the response the subject is trained not to make is regarded as being inhibited, and when the subject learns not to respond to the negative stimulus, the subject is said to have developed a 'conditioned' or 'internal inhibition.' Furthermore the response made to a positive stimulus may be called an 'excitatory reflex' while the response that is not made to a negative stimulus may be termed an 'inhibitory reflex.' These procedures and terms just described are most in vogue among the Russian and Pavlovian workers. American psychologists, on the other hand, in general prefer to develop discriminations in animals by using 'avoidance conditioning' and 'discrimination learning1 methods. In using 'avoidance conditioning1 to set up a discrimination, some stimulus, sa) .1 tone of 1000 cycles, is presented continuously .mcl the animal does not respond to it. When the tone is changed, however, say to 1020 cycles, the animal must learn to lift its paw (or run from one end of a box to another) to avoid shock. It is the change in the signal to which the animal is conditioned to respond. In the type II situations, discrimination learning is actually another type of trial-and-error learning. The animal is presented with two stimuli and required to choose between them. Animals can be trained to make such choices by rewarding correct responses with food and punishing incorrect ones with shock. Many such techniques for discriminative learning, however, involve only reward, not punishment. In the study of learning, it is often desirable to have problems of such difficulty that they tap the 'higher mental processes' of the animal — imaginal and ide- ational processes of the kind human beings use in language and thinking. For this purpose, the discrimi- nation technique just described can be altered so that the animal must make a 'delayed response." In this, the animal is presented with two stimuli, shown which one is correct and then is forced to wait for some interval, usually 10 sec. or more, before being permitted to make a choice. The position of the correct stimulus must of course be varied from trial to trial or the problem is no more than one of making a simple discrimination. Primates solve this kind of problem much more readily than do subprimate animals. The ability also depends, as we shall see, on the functioning of certain neural structures. other methods. One of the earliest devices to be employed in learning studies was the maze. Its chief advantage is that it can be varied in difficulty from a simple T, in which the animal makes one single choice, to extremely complicated patterns entailing 20 or 30 choices in a series. Mazes can also be planned in such a way as to constitute very ditlicult problems that can be solved only by using 'higher mental processes.' Some of the more recent techniques for studying behavior are the so-called operant (Skinnerian) methods. These are Type II procedures that permit extremel) delicate objective measures of the learned responses, the central correlates of which are our chief concern; lor this reason we ma\ expect to see them used increasingly - As an example of the niceties of the technique, consider the following (2171. Rats can be taughl i" |nr^s .1 lever to turn off a shock that would otherwise be delivered to the feet. If a clock in the circuit determines that the shocks will be delivered every 20 see. unless the lever is pressed (in which case the clock is reset to zero), the rat soon learns to space its responses reasonably accurately at nearly 20-sec. Intervals. This learned 'timing behavior" iv stable over man) hours and from day to day. Automatic re- cording of the responses frees the experimenter from many unrewarding features of experimentation while at the same time providing a clear objective account of the ongoing behavior. The reader interested in this technique will wish to consult Ferster & Skinner (60). Summary This account of methods and terms used in the study of learning is necessarily quite brief. Learning has been studied in literally hundreds of other ways. We have selected only those terms, methods and con- cepts that provide the necessary background for the data to be presented on neural correlates of learning. For a more detailed discussion of methods, the reader should consult Hilgard's chapter in the Handbook of Experimental Psychology (96) and other useful sources of information (98, 170). For a treatment of the phe- nomena and concepts of learning, see Deese (46), '476 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY II! Hilgard (97), Hilgard & Marquis (98), McGeoch & Irion (153), Pavlov (181-183) and Konorski (120, 1 2 1 I . ABLATION STUDIES The literature contains so many studies of the effects of damage to the brain upon learning and retention that any selection such as this one will provide limited coverage at best. This section attempts to summarize general trends by citing what are believed to be the most significant articles. More extensive treatments of the subject may be found in Morgan & Stellar (163), Morgan (161 ), Milner (157), Chow & Hutt (41), and in the chapters on physi- ological psychology appearing regularly in Annua/ Review of Physiology and Annual Review 0] Psychology. It is also considered in the chapters in this Handbook dealing with various regions of the brain, especially Chapters LIV and LXVII. Sjunal Conditioning It still is an open question whether the basic phenomena of conditioning and learning can be demonstrated in the spinal segments of the mam- malian nervOUS system. Attempts by several dilfcrcnl investigators (25, 26, 47, 115, 198) to obtain con- ditioning in the spinal cord have mel with failure. On the other hand, Shurrager and his colleagues have published a scries of experiments (213-216) which the) interpret in favor of spinal conditioning. These experiments, however, arc open to criticism on at least two counts: a) whal seems io be the CR may in fact be a L'R (47); and b) what appears to be con- ditioning may be sensitization (114, ii-,)- The question of spinal conditioning is therefore one that remains to be setded by future research (25, 26). Subcortical Factor i in Learning Relatively little has been done with lesions below the level of the cerebral cortex. Ghiselli & Brown (27, 77) made stereotaxic lesions in the thalami ol .1 large number ol rats, testing them thereafter cm .1 i\ ol discriminative and maze learning tasks. Almost an) thalamic injur) outside the anterior nuclei significantl) affected rate of learning, but on the whole the effects were not related to an) sel oi nuclei. Allen I ). j), working with the don, found no change in .1 conditioned diileienii.il response to olfactory stimuli after ablations in the hippocampus, amygdaloid nuclei, pyriform areas, or in combinations of these areas. Recent advances in our knowledge of the limbic system and its importance in emotion have turned attention to the role of this system in emotionally motivated learning (see Chapter LXIII by Brady in this volume). Septal lesions do not affect the learning of a classical CR to shock but do impair the retention of such a response learned before operation (21). However, learning of a conditioned avoidance re- sponse is more rapid in rats with septal lesions than in normal rats, and slower in rats with amygdaloid than in normal animals (116). On the other hand, retention of such a response is not impaired by amygdaloid ablation in cats (22). Much research of this sort is currently being reported, but it is not possible as yet to state accurately what the results imply about the role of the limbic system in conditioning and learning. Conditioning in Decorticate Animals Pavlov's theory (181) originally explained con- ditioning in terms of cortical processes. However, many experiments in recent years have demonstrated that learning can take place in decorticate animals. If the learning task is one of simple classical con- ditioning, such animals show little or no impairment (44, 78, 189). However, if the task is one of instru- mental learning, e.g. a conditioned avoidance response, animals have difficult)' with it (44). Such responses can nevertheless be acquired, if the experimenter is careful to see that the animal does not become too 'emotional' (187), but the process is .1 slow one and under some circumstances will never be completed. Apparently then, the cortex is not necessary for classic. il conditioning (Type I), but it is important for avoidance conditioning (Type 11) and more complex trial-and-error learning. .l/). The discus- sion of this problem, therefore, belongs under the heading of visual capacities rather than learning. The possible role of the so-called association areas of the cerebral cortex in the learning and retention of visual form discriminations has been investigated. The prestriate areas lying adjacent to the striate cortex have often been regarded as the 'visual associa- tion areas.' Animal experiments, however, cast serious doubt on this supposition. Although results are not entirely consistent (1,2 ), ablation experiments indicate little participation of these areas on the learning or retention of form (130, 203) or color discrimination (58) — at least when only these areas are removed. On the other hand, when relatively large lesions are made in the 'poster'or association areas,' involving some combination of the prestriate, parietal and 1478 II WDHtlOK OF l'HYSIl II i ICY Nl I Rl il'IIYSKJl.OCY III temporal sectors, there is usually some deficit in an animal's ability to discriminate forms (2, 86, 203, 205, 242, 243). In recent years, there has been considerable interest in the effect of ablations of the temporal (39, 40, 158-160, 188, 195, 204) and frontal (86, 167, 205, 24J, 243) lobes on the learning and retention of form discriminations. Both lesions can produce impairments with, on the whole, injuries to the temporal pole being the more serious. On the other hand, depending on the type of discrimination required of an animal and on the size and locus of the lesion, there may In- little or no impairment. Despite the great volume of research of this type, the results do not yet form a clear picture. [NTEROCULAR transfer. In contrast, a definitive and interesting result is provided by some recent work on the interocular transfer of visual discrimination learning. Interocular transfer refers to the ability to recognize with one eye what has been seen and learned with the other. In this work, done with cats (171, 172, 227), the animals were first subjected to operations in which both the optic chiasma and the corpus callosum were sectioned in the midsagittal plane. This operation has the result of dissociating each eye from its opposite visual cortex and the two sides of the visual cortex from each other. After operation, one eye of the animal was blindfolded and training was carried out in the discrimination of uch visual forms as a cross vs. a circle, horizontal stripes vs. vertical stripes, etc. After the animal had mastered the problem, the blindfold was transferred to the other e\e. The result in ever) case was no retention; the animal showed no memory at all of what it had learned using the other eye. It was in fact possible to train the animal to learn conflicting habits with the two eyes with, s,i\ , .1 circle to the li-ti eye positive and to the right eye negative. Normally , of < ourse, the mammalian retina projects 1., I mill v isual cortices. These experiments demonsti it that, when this overlapping connection is eliminated, ( alios. il association of the two cortices is essential to remembering 'with one eve what has Keen learned with the other.' This interesting result makes it clear that the tnemorj ol .1 visual discrimination habit is, undei these circumstances, confined to one side ol the cert bral coi tex. \UDITOR> DISCRIMINATION. Studies ol .iiuhiorv dis- ( rimination have mosl often employed a conditioned avoidance technique (156, 186, igg . although there are some cases in which an animal has been required to give a conditioned differential response to stimuli (6, 239) or to locate the source of a sound in space (185, 206). The last named technique yields measures of auditory localization, whereas the former methods are better suited to intensity and frequency discrimina- tions. No matter what technique is employed, if an animal is trained in a discrimination and then is subjected to lesions of the auditory cortical areas, it usually exhibits some loss of memory for the auditory habit (59, 185, 186, 199, 206, 232). The amount of loss varies with the size and placement of the lesion and with the difficulty of the task. Unless the lesions are too large, however, the animal usually can be retrained to the preoperative level of performance (156). An animal with a large ablation of auditory and associated cortex may, on the other hand, behave differently in two auditory tasks (66, p. 518). Thus, if two tones, A and B, are employed, cats will, after such an operation, relearn to respond appropriately to A as opposed to B (32). Relcarning of the dis- crimination between the pattern of tones ABA and BAB, however, is not possible (",1 I. This fact illustrates an idea often expressed, namely, that the cortex is required for discrimination learning when the task is 'complex' but not when it is 'simple.' soMESTHETic DISCRIMINATION. The results obtained for somesthetic discriminations .ire generally com- parable to those summarized above for auditory discriminations (208, 209, J39, 251 -253). In this case, however, the relevant areas are somatic areas I and II and, more generally, the posterior parietal lobule. Sizable lesions anywhere in these areas usually cause temporary impairment of habits acquired prior to operation, but this typically can be effaced by some refraining. It, however, the discrimination is made relatively difficult, il the lesions Include .1 large por- tion of both prim.irv and associative areas, or if both of these conditions are present, the impairment 111. iv be severe and m.iv not he ell, iced In ,mv amount of training. One of the relatively lew studies on the effects ol cortical injur) in man on learning has been done with somesthetic discrimination (76). Individuals suf- fering unilateral penetrating injur) of the cerebral hemispheres win 11. mud in making .1 tactual dis- crimination ol different forms Normal individuals using either hand and individuals with brain injur) using the li. mil on the same side .is the lesion were THE NEURAL BASIS OF LEARNING '479 able to make progress in learning the discriminations. Individuals using the hand on the side opposite to their injuries were not, however, able to learn the discrimination. This failure to learn was not related to any sensory defect. Hence, we may conclude that at least some kinds of learning depend on the hemi- sphere receiving the main sensory projections of the somesthetic system. This finding is related to a recent study of the function of the corpus callosum in the contralateral transfer of a somesthetic discrimination in the cat (228). Cats were trained to use one paw to push the correct one of two levers on the basis of tactual form, softness or roughness (three different habits). After the discrimination had been thoroughly learned, the cats were required to make the same discrimination with the forepaw not used in the original training. Normal animals did this with relatively little additional train- ing, but cats in which the corpus callosum had been sectioned prior to the experiment took as many trials to learn it as they had in learning the original discrimination. Hence, the corpus callosum seems to be essential to the transfer of somesthetic habits from one side to the other just as it is in the experiments cited above on intcrocular transfer. delayed reaction. As described earlier onl\ ;i slight change in procedure is necessary to convert the conventional discrimination problem into a test of delayed reaction (or response). In this test, the animal is shown the correct stimulus, then required to wait for a few seconds or minutes before being allowed to make its discrimination between two stimuli. As first employed by Jacobsen (106), it was regarded as a test of immediate memory. The factor, however, that often determines success or failure in the test is whether or not the animal pays attention to the correct stimulus when it is first shown. Jacobsen first demonstrated that monkeys with frontal lesions have difficulty in the delayed reaction test (106). This fact has been confirmed many times in subsequent studies (16, 65, 86, 155), but it is not a universal finding for such monkeys sometimes can succeed in the test (33, 167). Moreover, by making such minor changes in the procedure that the animal's chances of attending to the presentation of the correct stimulus are increased, substantially all these animals can succeed in passing the test (63, 151, 223). This fact indicates that the impairment is more one of attention than of immediate memory (85, 87). There is some evidence that an area lying in the dorsolateral frontal region anterior to the precentral motor cortex is especially important in delayed response per- formance (194, 197). There is also evidence that areas outside the frontal lobe are not important (I07> !54)» although a deficit in delayed reaction is reported occasionally in animals having posterior lesions (130). Closely related in principle to the delayed reaction are tests of double alternation or delayed alternation (144, 164). In double alternation tests, the animals must respond in a pattern of RRLLRRLL. In delayed alternation, it must respond RLRL after a delay between each response. In all such tests, the important element is that the animal must remember what it has previously done or experienced in order to know what to do next. In all such tests, too, frontal abla- tions are usually followed l>\ substantial impairment (1 14, 144, 164, 196). conditioned inhibition. Konorski has recently re- ported (28, 122) a series of experiments on the role of the frontal lobes in conditioned inhibition. Dogs and cats were taught a conditioned discrimination b\ setting up an excitatory reflex to one stimulus and an inhibitory reflex to another. Following this training, ablations were made in frontal areas and in parietal areas, usually in different animals but sometimes in the same animal. The general result of frontal ablation was to impair inhibitory reflexes with- out affecting excitatory reflexes. That is to say.frontal animals made correct responses to positive stimuli but also made these responses, which they had been trained not to make, to negative stimuli. Parietal lesions, however, had no significant effect on the retention of conditioned inhibitory reflexes. These conclusions apply about equally well to salivary responses (classical or Type 1 conditioning) and to motor foreleg flexion (Type II conditioning). anticipation and perseveration. Learning tasks can be so designed that they involve a series of responses leading up to a goal. A maze constitutes such a task, but so also do some problem boxes in which the animal is required to do two or three things in a particular order. If the animal is unable to perform the task correctly, two kinds of mistakes are fairly common: one is to anticipate a correct response, making it too soon in the series of the response; the other is to perseverate, making the same response two or three times in succession. Disorders of serial learning appear with almost any large lesion of the cerebral cortex, but they are especially prominent after frontal ablations. Thus I 4811 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III rats (34, 57, 150, 231) and monkeys (106) with such lesions tend to make more perseverative errors and anticipatory errors than normal animals or animals with more posterior lesions. This impairment is probably closely related to the deficit seen in the delaved reaction. Summm v Since psychologists have not seen fit to agree on any single set of procedures to be employed in studies of conditioning and learning, many different learning tasks more than iliere is room to describe here — have been devised for the study of the effects of ablations. Furthermore, different experimenters often do not make the same lesions even when they intend to do so. As a result, ablations may produce impair- ment on some tasks and not on others, and experi- ments do not always agree with each other because of minor differences in procedure. Despite these and related problems, certain facts seem clear. When the habit requires a sensory capacity (e.g. form vision ) abolished by the ablation, no amount of retraining ever effaces the deficit produced, as might be cxpciied. On the whole, however, cortical ablations have little or no effect on classical CR's; animals learn these just as easily after injury as be- fore and retain them after lesions are made. But it differential CR's arc called for (or if the learning task involves instrumental responses), retention is often impaired, thus, typically, removal of the pri- mary sensor) area relevant to the habit causes a par- tial or complete memory loss, but the animal re- learns in a reasonable number of trials. II a number ol sensorv modalities are concerned, as in maze learning, then one sees .1 mass-action effeel a cor- relation between amount of deficit and size of the cortical lesion. And, finally, lesions that include all primary and association areas oi one modality may cause considerable, lasting impairment. The posterior assoi iation areas ly ing in the parietal- ipital-temporal sector are proving to have con- siderable importance, particularly in discrimination learning, Certain parts Ol the temporal pole seem 1 pecially implicated. The frontal areas are clearly involved in learning where the ordering ol responses in time is .1 critical feature. When learning involves the cortex oi one side only, the corpus callosum plainly participates in its transfer n> the opposite side, al least foi some h.il.ils. EEC CORRELATES Promptly after Berber's rediscovery of the brain waves, serious attempts were made to put these electrical events to work in uncovering the neural events of learning and conditioning. In the following description of these efforts it must be taken for granted that the reader is familiar with certain general propositions about brain waves, neurophysi- ology and neuroanatomy. The reader may wish to consult Chapter XI by Walter of this Handbook in which autogenous brain waves are discussed. The search for electrical brain events reliably related to learning began with the discovery that exposing the eyes to light leads to the disappearance of alpha waves. It was soon found (54, 143) that simply pairing a sound (CS) with the light (US) led rather promptly to disappearance of the alpha waxes in response to the sound alone. Thus there started some 20 years ago a new and potentiallv fruitful era of research that is currently in full swing. The Alpha Block CR The cortex of most animals generates a more or less continuous series of waves in the region of 5 to 20 per sec. The exact frequency range varies with the species studied, and for man the figure 9 to 11 is ordinarily given as normal. For convenience these all can be called alpha waves. Experiments in which their disappearance is conditioned to a stimulus typically proceed as follows. The scalp EEG is recorded from the subject at rest and in the dark. From time to time a light is turned on and off; the alpha waves disappear as long as illumination continues. From time to time also a sound is turned on; at first this stimulus, like the visual stimulus, temporarily blocks alpha activity, but eventually it does not do so, an example ol habituation (70, p. i<>; 165). Finally the stimuli are paired, with sound (CS) preceding light (US) In .1 brief interval. Shortly there. liter the alpha rhythm disappears as soon as the sound conies on and the alpha block Ck has been established. Where no motor response is employed, the term sensory-sensory conditioning is applied to these EEG studies. The blocked ('desynchronized,' 'arrested,' llallened'l EEC J response is, in addition, a common event in I v pe I and fype II conditioning procedures. Morrell & Jasper ( 165) contribute .1 recent example of the so-called sensory-sensory experiment. Tones paired with light produced conditioned alpha block THE NEURAL BASIS OF LEARNING [481 in an average of 11.1 trials in 8 monkeys with im- planted cortical electrodes. When instead of sound a second visual stimulus or mild shocks to the skin were used as CS, averages of 13.2 and 9.5 trials, respectively, were required. A remarkable increase in trials is required when epileptogenic lesions (pro- duced by alumina cream) exist in the cortical projec- tion area specific for the CS or in the amygdala (166). In studies on man, Motokawa (168) paired sound (CS) with light (US), recording, in addition to the multiple scalp EEG, the galvanic skin response (GSR). Both the alpha block and the GSR became conditioned to the sound alone. This finding, besides portraying the EEG details of a typical 'sensory-sensory' proce- dure, shows that the alpha block CR may be highly correlated with an autonomic CR. These observations have been confirmed (103), and similar results were found in a Type I situation where salivation instead of GSR was conditioned to a sound (102). The alpha block CR may be preceded by a period of enhanced alpha amplitude if the interval between application of CS and US is sufficiently prolonged (101). On the other hand, Motokawa & Iluzimori ( 1 69) have shown that more than one alpha block CR may occur during the conditioning procedure. They used a bell (CS), and a shock to the fool (US) in man, recording both the GSR and EEG. If the bell was sounded for some 10 sec. before application of the US, alpha block occurred both at the onset of the CS and again just prior to its termination. A GSR response coincided in time with each alpha block when conditioning was well established. Both periods of alpha block seem to meet the criteria for CR's (102, p. 352). Kogan (118), finally, has used microelectrodes implanted in cats to study the conical e\ cuts occurring during the conditioning of a motor response to an acoustic stimulus. Alpha block seems to be .1 constant feature of the process, occurring first in the auditory area and spreading later to the motor region. In addition, beta waxes and 'special forms of electrical activity' (not further defined in the brief English report seen) appear. In the early stages of condition- ins;, changes are prominent in cortical layers III and IV, but later a shift to 'the deeper layers' is noted. According to this report, very large oscilla- tions of potential appear in the auditory cortex when a differential CR is set up with auditory stimuli. The use of microelectrodes is just beginning in non- Russian hands, but it has already revealed some valuable information (29, 202). The alpha-block CR obeys, more or less, the general rules established for conventional CR's. It appears only after a number of pairings of CS and US, and some stimuli are more effective than others (165). Furthermore, when only one of two different sounds is reinforced, the reinforced sound alone evokes the CR, an example of differential conditioning (165, 207). Finally, if the CS is not reinforced, the CR disappears; this is the phenomenon of extinction. Some attention has been given to the cortical regions blocked in the response. Prior to its habituation, the CS preferentially blocks the parietal region (165) or the brain in general (70), but after establishment of the CR it specifically blocks the occipital region (165) or the cortical region to which the sensory system of the US projects (70, p. 16). If the CS is applied to one side of the body, the conditioned EEG response may even be limited to the contralateral cortex (70). The reliability with which alpha block appears in conditioning experiments leaves much to be desired. Certainly it docs not Invariably occur and in one length) and important study, for example, 60 per cent success is reported in a large group of human subjects (70). Consequentlv the alpha block CR may be of limited usefulness in understanding the basic neural events in conditioning. The phenomenon is related to action of the reticular formation discussed in this Handbook in Chapter LI I by French and elsewhere 1 ;, Ik), j .18 250 1 ; it appears when the animal is merely alerted as well as during the learning process. Nevertheless, .is we have seen, the alpha block CR satisfies many of the criteria for a cortical event that accompanies learning. Cortical Evoked Potentials In addition to its regular rhythmical activity, the cortex generates brief electrical waves corresponding to the arrival within it of impulses from the various sense organs These so-called evoked responses follow- sound, light, touch and similar stimuli. Their latency, magnitude and duration have been examined during the conditioning process. In an important variation of the sensory-sensory conditioning procedure, for example, a flashing light is substituted for the continuous one. Not only does the regular cortical rhythm disappear, but evoked cortical waves harmonically related to the frequency of the flashing light are recorded ('photic driving'). If a sound (CS) is now made to precede the flashing light (US), the sound alone eventually produces an EEG response remarkably similar to the characteristic 1482 IIVNDBOOK OF PHYSIOLOGY M.I KOI'IIYSIOI.OGY III 'photic driving1 pattern. Analogous results arc re- ported in Type I learning as well as in sensory- sensory learning (70, 210, 248). On<- ol the earliest studies of cortical evoked po- tentials as influenced by learning is the work of Livanov & Poliakov (140) on the rabbit. Shocks to the leg (US), paired with light flashes, led in time to leu flexion (CR) upon presentation of the light alone it Si. Both light and shock occurred at 3 per sec. During the learning process a 3 per sec. rhythm appeared first in 'certain' cortical regions, later in the whole of it; when the CR was fully developed 3 per sec rhythms of large amplitude tended to be present during the CS and to be absent at other times. I [owever, every motor response, whether spontaneous or provoked by painful or other stimuli, was now accompanied by the ; per see. rhythm. The authors suggest from this that "the process of formation of the conditioned reflex defense reaction in response to the rhythmic stimulus is developed on the basis of .1 system of rhythms reflecting periodical changes in the excitability of the motor centers of the cortex" (see also 80, H'",, 190). Tin- brain wave CR pattern evoked by the CS approximates the frequency of the photic stimulus IS beiter and better as conditioning proceeds, and the CR appeals progressively earlier in time (165). It is limited to the occipital leads, is unstable as compared to the alpha block CR, and it quickly extinguishes (165). This CR appears not only at the cortical level but also in thalamic and mesencephalic his where it is reported to lie large and stable lip evoked response to sound also undergoes in- teresting changes. lis magnitude has been shown to \.n\ somewhat from one stimulus to the next in the untrained animal, but a Type 1 conditioning proo dure Stabilizes such responses at large amplitude in the auditor) region (l>7, llo). looked responses also appeal in ical areas not previously involved and the Complexity of all of them increases, with those features that follow the click by 30 to 100 msec. econdary' responses) being particularly influenced. Artemyev & Bezladnova (10) undertook in cats "to follow the dynamic changes in the nervous processes ol die cerebral hemispheres arising oul oi the development ol conditioned rellex links." A I ke- tone, 60 db above die human threshold and lasting ei constituted the CS, with shocks to the hind leg as die l s llo EEG onsel response to die tone was recorded on .1 cathode raj oscillograph. Electro- myograms oi die leg muscle showed thai to of 12 cats developed the CR ("defensive conditioned re- flexes'). In all animals a positive correlation is re- ported between the CR and the occurrence of EEG change in the auditory area as follows. "Where the percentage of occurrence of the electrical reaction in the cerebral cortex is low, the conditioned reflex is absent," and "in the course of forming the conditioned reflex in the auditory projection area with sound stimuli, the primary electrical reaction occurs in a markedly greater number of cases than before and at the beginning of the combinations." Fluctuation in magnitude of the EEG response was seen during the conditioning process and during extinction. The re- sponse disappeared in sleep and with simultaneous "strong stimulation of other analyzer systems." The findings are discussed as follows: "if the animal is passive to the stimulus the level of excitability of the neurons of the cortex is lowered and the primary electrical reaction is decreased in amplitude and disappears. When the ineffective stimulus is linked with another which is biologically important to the animals, the excitability of the nervous elements increases, of which we can judge by the more frequent occurrence of the primary electrical reaction." Confirmation of these observations has in the main been reported by others (e.g. 67, 1 ml. Since, however, similar results have also been obtained at the level of the cochlear nucleus (by, 90, 91, 1 10, 142), the changes seen at the auditory cortex may merely reflect events that occur at subcortical levels. New /'.'/<< in, ,1/ II fives Among the least understood of die central correlates of die CR are die "new potentials1 thai have been reported to be associated with the conditioning process. Many published records show high- or low- frequency events not characteristic of the normal record; these either appear during the alpha block or, when alpha is present, constitute distortions of the base line These new events bear no relation to the stimulus frequency as is true of the evoked potentials jus) discussed. In man they may be beta (20 to 30 per set I, kappa llo per see I (()), or dicta (5 per seel waves. Motokawa & llu/iinori (ili<|i, Motokawa (if>8) and lw.1111.1 (loll consider the high amplitude waves at 3 to 5 per sec. ('excitation potentials'] to indicate increased activity of die area in which they appear, this presumably being a necessary precursor of the cortical events in conditioning. Popov ( 190), however, has shown the occurrence of such waves in the parietal THE NEURAL BASIS OF LEARNING 1483 cortex of rabbit to be the major electrical event paralleling the development of Type I conditioning to a sound CS followed by shocks (US) to the paw. If the shock to the paw is very weak, on the other hand, the sound CS produces alpha block that un- masks bursts of high frequency waves, but no 3 to 5 per sec. waves whatever appear. This report raises the important but unsolved question of what sys- tematic EEG differences, if any, are associated with variation in strength of both CS and US. Popov & Popov have further been concerned with long-lasting cyclical alterations of alpha amplitude that accompany light flash CS; the original papers should be consulted for the relation of these to visual afterimages and regarding their conditioning by sounds (191-193). Siihi ortical Structures Much recent experimentation has dealt with elec- trical activity in subcortical structures during Type I learning. Behind this work lie the three new concepts about the nervous system thai render obsolete so much of the neurophysiology of 5 or 10 years ago. These are a) the demonstration in unanesthetized animals of the widespread influence of the reticular formation upon the organizing and integrating func- tions of the brain, b) the possibility that the descending sensory pathways constitute "feed-back' loops for control of afferent input to the brain, and c) the concept that limbic system structures arc intimately concerned with emotional behavior and thus with any learned activity that has an emotional component. The reader will find these ideas discussed ai length elsewhere in this volume. Our interest is in those experiments which have been done to relate them specifically to the learning situation. Il is clear that a CS (e.g. auditory clicks) not only evokes the expected responses throughout the ap- propriate sensor) system but in addition activates parts of the brain to which anatomical projections from thai modality have not been described or are poorly understood (e.g. limbic system, reticular formation). At a given recording site, the responses vary wide!) in amplitude, latency and duration, but these features have not yet been properly analyzed. Thus far, the only reasonably consistent finding is that some elec- trical response to the CS tends to be large and stable in the conditioned animal, while being small, labile and recordable at fewer brain locations in the un- conditioned or extinguished animal. As in the case of cortical electrical correlates, however, agreement has not been reached as to which subcortical events are invariably related to learning. Important information on events in the lower nuclei of classical afferent pathways comes mainly from the auditory system of cats. Hernandez-Peon and associates (90) studied the response of the cochlear nucleus in Type I learning to tonal stimulation (CS) paired with shock (US) to the hind leg. The cochlear nucleus response, having diminished in size during habituation of the animal to the apparatus, attained large size and increased duration when the conditioned leg withdrawal was fully developed. Similar results have been reported in cats that received shocks irregularly when click stimuli were being presented (67); correlated with the behavioral response de- veloped to clicks, the evoked response in the cochlear nucleus and medial sjeniculate (and auditory cortex as well 1 bee. inie large and regular in comparison with responses in the habituated and extinguished state (see also 30, 3 1 , 9 1 , no) A small amount of information is available about the activities of the limbic swem in conditioning. Waves of ;; 10 5 per sec. from a presumed caudal hippocampal location have been reported to fill die CS-US interval in a Type I situation; but such waves were absenl from septal and other hippocampal locations (149). Evoked responses appear and become stable in the hippocampus and amygdala .is well as in the caudate nucleus when a s d I S is employed in Type I conditioning (67). The amygdala of the cat seems unique among limbic structures in that 40 to 45 per see. waves of high amplitude appear in it during conditioning of a shock to sound (135); these diminish or disappear with extinction and can be re-established by repetition of the conditioning procedure. As for the reticular formation, some remarkable electrical events within it appear to be related to the learning process 1250). Cats were prepared with electrodes in the cortex, reticular formation, and ventral anterior and center median nuclei of the thalamus. A tone (CS) was presented for 5 to 10 sec, followed shortly after its ousel by a brief train of light flashes at 5 per sec. (US). Alter many pairings, the tone alone produced both the alpha block CR and new waves in the cortical leads. In addition, a 5 per sec. rhythm appeared in the subcortical loca- tions which, at the reticular formation leads in particular, developed promptly upon the tone-onset and persisted long after the tone was turned off. After very many pairings, the 5 per sec. rhythm might last for more than a minute when the animal ■4»4 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III was located in the experimental room, but they failed to appear if the recording was done elsewhere in the laboratory. The authors were impressed by the durability and persistence of the subcortical rhythms as contrasted to the transitory characteristics of the cortical ones. Summary Certain generalizations about learning derived from the EEG responses can be made with the reservation that they do not always agree with all the data. At the cortex the CS comes to block alpha waves in a new way that parallels the development of the CR in time. Besides this, the CS can arouse new cortical waves whose existence seems to depend upon the learning process. And finally, the cortical responses evoked by the CS can be larger, more complex and more stable during acquisition and retention than they arc during habituation and extinction. But entirely similar electrical changes are also taking place in many other brain locations. Before conditioning, what is to become the CS causes a variable evoked response in the afferent sensory tracts and in limbic Structures; these, like their cortical counterparts, become larger, more complex and more stable in learning. The new cortical waves arc matched bv novel events in limbic and reticular structures that are exactly similar in principle and perhaps even in kind, such waves in the reticular substance may actually be far less fickle and evanescent than the cortical ones. And, finally, the cortical alpha block probably reflects mainly the action of the reticular lui maiion. Thus limbic, reticular and cortical struc- tures arc all involved in learning, if the EEG results an- to be believed. 1 low <>ne can ol ijeciiv elv determine from this evidence alone that one part of the brain is more involved in learning than another is difficult to see. Nor is it possible at the present time to do more than guess at how the three major systems cortex, reticular formation and limbic system in- teract with one another. There seems to be little question lli.tt lliev do so, ,m\ the CS are the weaker of the two excitations. For this reason they flow toward, or are drawn toward, the center of stronger excitation, i.e. the place where excitations from the L'S are generated. Then, as a consequence of the repeated presentation of CS and US, a path is worn, so to speak, from the CS center to the US center and the CS comes thereby to bring about the same neural effects as the US. A modern Russian theor) cm be illustrated by the ideas of Beritofl 11 ;. 1 jo, p. j(>). This will be re- counted here in detail because it is not likely to be readily available elsewhere t<> the reader, although Konorski (120, p. 56) has discussed it. Like Pavlov's notion, its central concept is that excitation irradi- ates in all directions from the cortical areas excited by L'S and CS and preferentially along the shortest line between them. However, according to Beritoff, "two-way" connections come to be formed, specifi- cally between "star cells' of the two cortical areas and l>\ way of subcortical white matter through the pyramidal association neurons. Ultimately the motor pyramidal cells are influenced to produce move- ments. "All the star and other neurons with short axons and also all the small (internuncial) and medium sized (association) pyramidal neurons with descending and ascending axons form closed chains of neurons both vertically and horizontally. . . . Neu- rons forming the pyramidal and extrapyramidal tracts . . . are joined together by internuncial neurons . !<,,, HANDBOOK OF l'l 1YS101.0GY \I IROPIIYSIOLOGY III [and] take part in the formation of neuron chains joining the cortex and the underlying parts of the brain. . . . The entire action of the cortex during both unconditioned and conditioned reflexes is caused by the excitation of certain neuron chains and the more or less considerable inhibition of all the rest MTerent impulses (thalamic or association) activate the star and other neurons with short axons, on the one hand, and the internuncial and associa- tion pyramidal neurons, on the other. ... In the formation of the neuron chains of the cortex the dendrites of the pyramidal neurons do not take part because they do not conduct excitation to the cells . . . [but] more or less considerable and prolonged potentials may arise [in them]. "By combining: two stimulations acting on one and the same or on different analyzers ■two-was' tem- porary connections an- set up . . . among the star neurons . . . by means of the internuncial and asso- ciation pyramidal neurons. . . . The formation of temporal-) connections supposes both functional and morphological changes in the cellular elements and in the synaptic apparatus. ... A cell of one neuron and the synaptic knobs of another neuron lie closer to one another . . . their excitability is noticeably in- creased and the processes of excitations proceeding in them are intensified. These morphological and physiological changes take place as a result of the interaction of the excitation in all parts of the nervous system but in the cerebral cortex they arise more quickly and last for a longer time, sometimes for life, while in other parts of the brain they are always transilorv and pass ,iw.i\ soon after cessation of the interaction." Cognizance does not seem to have been taken in BeritofPs theory oi the experiments by Sperry and collaborators showing that knife cuts and the im- plantation ol win- .mil sheets oi mica -devices thai destroy mii.icortu.il connections and distort what- ever el© 1 1 H .il fields i ortii il neurons generate inter- fere little "i not ,u all with performance of even complicated CR's (226). Konorski (120) has still other objections to it. As aln-.nK mentioned, the idea thai stellate ('star1 cells in the cortex .in- ol especial important e for < IR's is supported by s.ukiso\ (212, p tig) who holds that theii structural peculiarities "show their specific role in tin- cortical processes and prim. 11 lis 111 the interconnections between the cellular elements of the- cortex " He holds, further, "thai the cells of the hinder parts of the central nervous s\siem and pri- marily ol the cerebral cortex are characterized by considerable lability of form, changing under the in- fluence of external and internal stimulations" (212, p. 120). A report of intracellular changes, twisting of apical dendrites and coarsening ol libers in cortical layers I and II following electrical stimulation is actually available (38). The collection of experimental evidence related to both the anatomical and the neuron chain theories of learning marks much of the current Russian work. REVERBERATING CHAINS. Another concept of how nerve circuits are rearranged in learning has been called the reverberating chain theory (74, 98, 120). Proceeding from the anatomical fact that neurons are at least potentially connected to other neurons in a reciprocal manner, it supposes that acquisition of a OR consists of setting up a closed circuit of neuronal activity in which neuron A fires neuron B which in turn fires A again. Such chains can include very many neurons. Retention of the CR (or memory) is explained by perseveration of activity in the chain. Hebb (881 supposes that short-term retention may invoke reverberating chains only but, if reverbera- tion persists sufficiently in such a chain, the anatomi- cal synaptic changes required to explain long-term memories occur. Many other variants of these popu- lar schemes have been advanced but space will not permit their discussion (lit, 131, 132 Theories from EEG Studies The most recent theories come from the EEG studies which, as we have seen, arc beginning to re- veal important new facts. We will consider only one of these, the theory of Gastaut t 1 al. (69, 70 1. It re- quires rearrangements in at least six separate cir- cuits to explain habituation and conditioning. Both the CS and the L'S are thought to activate the reticu- lar formation as well as their specific cortical areas, an inference iustilicd by recent ncurophysiological findings. Each stimulus, furthermore, activates both a mesencephalic and a thalamic locus in the reticular formation. Habituation consists in the inhibition of both these loci, with c c insequenl disappearance of the alerting response in the scalp EEG. Pairing of a (habituated 1 aiidilnrv CS with ( unhabiluatcd I somesthetic I S leads to formation of .1 temporary link in the thalamic reticular formation thus "pet mining thalamic collaterals borrowed by the sound signal 10 act on the neurons previously only acti- vated by the collaterals ol the somesthetic signal" (70, p. 31). The mechanism by which CS input THE NEURAL BASIS OF LEARNING '491 'borrows' the thalamocortical circuit of the US in the thalamus is not elucidated. The theory, despite its emphasis on thalamic changes in learning, is said by its author not to minimize the part played by the cortex. We may consider it to be a somewhat elabo- rate early attempt to account for the rather meager information that the EEG studies have thus far pro- vided. Other Neural Possibilities There is a group of theories that explain learning in terms of brain events for which the all-or-none law of nerve action is not particularly relevant. For example, it was once proposed that an increase in the "conductance' of nerve fibers, that is the amount of message each fiber transmitted, might account for learning. This notion was dismissed when nerves were shown to convey all-or-none signals. Recent neurophysiological research, however, establishes the importance of graded events where neurons arc in synaptic contact with each other, and so the notion may again deserve further consideration. In this connection, Bishop (14) points out that the all-or-none activities in the brain serve merely to con- vey messages from one location to another; they are set up by graded events in the first place and in turn they produce their actions through graded-response processes at the synapses upon which the) converge. Grundfest (84) also emphasizes tlii^ idea. Graded processes can be maintained at steadv siaics for long times as compared to nerve impulses which vary abruptly from all to nothing. What is needed to ex- plain learning;, or at least retention, is long-continued neural events. Since the graded responses in dendrites have appropriate characteristics, it would therefore be in, on or around them that one might look with particular care for neural events peculiar to learning. Herrick (92, 93) sees in the neuropile an answer to wrhere the integrative processes of learning take place. This neuropile is "a fabric of relatively unspecialized nerve cells and very thin fibers'" in which, as an anatomist sees it, the complex events of learning might well occur. Unfortunately he cannot specify what these might be, and so this idea, like so many others, cannot be subjected to experimental test. By contrast, the cortical 'electrical field' theories that have been advanced from time to time to ex- plain retention in learning (as well as perception in general) can be tested. Electrical conductors have been implanted in and upon the cortex of some ani- mals while insulators have been implanted in others; such devices must have distorted or destroyed very effectively any existing cortical electrical field, yet the complex learned behavior suffered minimally, if at all, as a consequence (226). Mathematical Models There is one final class of theories to be considered. Certain physicists and mathematicians oxer the vears have been challenged by the complexities of the learning process to develop explanatory formulations for it. Besjinnin" perhaps with Rashevskv in 1938 (201), new contributions at a rate of at least one every year have been made in this area. Recently, with the advent of digital and analogue computers and theory, the rate has been stepped up with the idea, perhaps, that our advancing knowledge of complex switching circuits in machines may have application to the brain. The reader interested in these models of the brain will want to consult the available contributions (11, 56, 97, 100, 240, 24] I. Summary A large number of speculations have been ad- vanced to explain the neural correlates lor learning. Some of these are based upon a certain amount of objective data about the brain. The most popular schemes incorporate hypothetical changes at svn- apses with hypothetical reverberating activity in neuron chains. In the large collection of speculations on record, the one (or ours! that will finally har- monize with the facts may well be present but, if so, there is no compelling experimental evidence for it (or them) at this time. DISCUSSION AND SUMMARY This is the appropriate place in our exposition for the authors to propose a comprehensive theory which, without violating an) of the data, will explain what happens in the brain during the process of learning. In our opinion, however, this cannot be clone at the present time. New^ information is currently being de- veloped rapidly and, as this happens, the large gaps in knowledge that still exist stand out more and more clearly. Until some of these gaps are filled only the most general of formulations seem warranted. In this discussion, therefore, we shall merely point first to some of the basic questions implied by the material in the preceding sections and then consider one of 1492 HANDBOOK OF PHYs JOY NEUROPHYSIOLOGY III the ideas thai is presently under active experimental study. One Neural Correlate? In the literature repeated references arc found to the 'neural event' that accounts for learning, as if a single one was envisaged. Yet many theories, as we have seen, postulate two or more to occur (e.g. a synaptic change and a new neural circuit). There seems to lie no way to settle the point of single vs. multiple possibilities except by further experiments that will define where in the brain, and when in time, the essential changes occur. Such explanations must harmonize a number of disparate facts about CR's which we will enumerate here. STRUCTl RES INVOLVED. Some Type I habits can be abolished by decortication and then be relearned. The pre- and postoperative ( 'R's seem to be the same and, if we make the plausible assumption that the cortex participated in the original habit, it is evident that the second one is mediated by extracortical structures. The cortex, however, is clearly the site of the durable change in certain kinds of Type II habits. I his conclusion emerges, for example, from the ex- periments in which the corpus callosum was sec- tioned: in that situation learning is localized to the cerebral hemisphere, and to one side only, for habits involving both tactile and visual discriminations. Consideration of the time course of acquisition raises another set of problems. A monkey learning to avoid shocks bv pressing a lever at a signal appears in pass through a series of behavioral stages in the process. At firsl the signal arouses much 'emotional' activity, such ,1' piloerection and vocalization. Later, when learning has progressed to 50 per cent correct responses, ibis 'emotional' behavior can be partly replaced bv an 'alert or attentive' attitude. The fully trained animal, in final complete command of the situation, seems undisturbed by the signal and often del. iv- making the comet response until the \ei\ las) moment. rhese successive behavioral stages presumably re- flect -i progressive reorganization of brain structures or processes during acquisition. II this is true, brain events measured al a particular place in the earl) - ol learning might be absent there later, while .11 another brain locus, characteristic brain events might appear Old) when learning has become com- plete. COMPLEXITY. Different amounts of brain seem to be required according to the degree of complexity of Un- learning problem. If a two-tone pitch discrimination and a three-tone pattern discrimination are taught to a cat, removal of its auditory cortex abolishes both and only the "simple" pitch discrimination can be re- learned. Some neural events responsible for the 'com- plex' tone-pattern CR have been eliminated Im- partial decortication but the regions necessary for a 'simple' CR remain. phylogenetic evidence. Learning is common to the octopus and the cat despite the large differences in their neural apparatus (17). In particular the mam- malian cerebral cortex is obviously not needed for learning per se. 'emotional' learning. In the case of the rat trained both 11) to press a lever to get a drop of water and 6) to expect' a shock at the termination of a signal, the animal loses only the second of these habits after experiencing a number of convulsive seizures. This experiment defines a clear operational difference between CR's in the intact animal, and it suggests that a corresponding difference exists in the neural b.isi- of "emotional' as opposed to other CR's. ease of learning. Ordinarily many or verv many combinations of CS and US are required to establish a CR, but in the case of imprinting, a single exposure to CS alone produces lifetime retention. The neural events in this exceptional instance, where for a few hours the brain is 'primed' to make a particular set of functional connections might, if understood, also serve for the general case. Perhaps instinctual be- havior, like that of newly hatched birds which scatter for cover at the first presentation of specific sounds or moving shapes (238), represents simply the ultimate with respect to such neural processes, namely build- ing them into the organism at the outset so that they need not be formed bv le. lining at all. Maturation and Learning This brings us to tin- question of whether or not the neural changes taking place in the normal em- bryological and postnatal growth of an organism dif- fer from those taking place in learning. Is it possible ih.it 1 1 1<\ .in- basically the same and require only slightlv different environmental conditions to bring I liem about? Innate behavior is distinguished from learned be- THE NEURAL BASIS OF LEARNING ■493 havior on the basis that the first of these emerges in the normal development of the organism while the second appears only through appropriate learning experiences. We will express this dichotomy in terms of maturation vs. learning since many items of un- learned behavior appear after birth and, strictly speaking, are not "innate' even though they are surely not learned. The important point is that basic patterns of behavior are laid down in the nervous system in the normal development of the system while others are acquired only through learning. In general, the behavior appearing early in the life of the organ- ism is the result of maturation and what emerges later is the result of learning, but the two stages overlap and interact so that learning occurs in some sectors before maturation is completed in others. 'connections'. In both maturation and learning, some change in the nervous system is necessary for 'new' patterns of behavior to occur. It is reasonable to suppose that this change consists of new 'connec- tions' formed between different organs, centers and neurons in the system. The new connection or link, whatever its exact nature, makes possible the How of messages between points that were formerly 'uncon- nected.' In maturation certain parts of the nervous system, as well as the muscles, undoubtedly exert "attractive influences' on the growth of neural fibers. Somehow or other, motoneurons are guided to their respective muscles and other neurons connect up appropriately with the motoneurons. In addition, sensor) neurons will connect up in exactly the 'correct' constellations with nuclei in the central nervous system to perform their functions. And we know that certain portions of the nervous system, for example the medulla, have a controlling and directing influence on the growth of pathways formed in nearby structures. Some pathways laid down in maturation, when once established, are so rigidly fixed that they can- not be altered. The numerous studies of Stone and of Sperry (225) clearly demonstrate this point. If, for example, a motor nerve of the left leg of the rat is surgically crossed to the right leg, and vice versa, a noxious stimulus to the left leg reflexly evokes lifting of the right leg and no amount of training of the animal corrects this maladaptive response. The animal continues to lift the 'wrong' leg. Similarly, in amphibia, rotating the eyes or transplanting them to the opposite sockets, thus twisting or reversing the animal's visual field, causes the animal to respond to visual objects in a direction opposed to the normal one. Prolonged learning experiences with such visual fields, however, do nothing to alter these inappro- priate responses. We are led therefore to conclude that many connections or pathways, once established through maturation, are not easily changed or altered. Those established through learning, in contrast to this, lend themselves easily to formation in the first place and are more or less impermanent. Although the precise conditions for the formation of the links are different in maturation and learning, there is so far no reason to believe that the processes are fundamentally different. That is to say, it is probably our best assumption at present that the neural changes taking place in maturation and learn- ing are essentially the same and that only the condi- tions or immediate causes are different. We might say that maturation rigidly fixes some routes so that unlearned reflexes and responses can be altered little or not at all, and it also prepares other routes m> that they can become fixed through the processes of learning. But there is no reason to believe that the central processes that fix behavior in maturation differ from those that fix them in learning. the specific change in ii \K\iNo. What, then, is the specific event thai fixes the connections in matura- tion and learning? So far as maturation goes a num- ber of such factors have been considered, among these are a) mechanical, b) electrical and c) metabolic or biochemical influences upon the direction and State of growth (225). As for the changes in learning, factors of the same general sorl have also been postu- lated, but none of the evidence is critical as we have already seen. We musl conclude, therefore, that both the precise mechanism of guidance and control of neurons in maturation, and the exact nature of the specific event in learning are --till a mystery. MOTTA VTION WD ATTENTION. Psychologists and ph\ si- ologists have, from time to time, been led by their observations to infer the existence of many events going on within the mammal during learning. We have considered one of these in detail, namely the idea that some durable change— at synapses or else- where— results from pairing CS with US to produce the CR. There are two others, which we will call here 'motivation' and "attention,' winch also appear repeatedlv in formal treatments of the problem. An animal "motivated' to learn (e.g. by reward for suc- cess or by punishment lor failure) does better than one not so motivated. Similarly, an animal (or a '494 HANDBOOK OF I'lIYSIOl-dGY NEUROPHYSIOLOGY III student) that pays 'attention' to a problem is more likeK to solve it than one whose attention wanders. What is the role of these poorly defined and somewhat elusive concepts in the learning process? mim:ii wisms in motivation We know thai a hungry rat readily learns to run a maze for food, and that a well-fed one learns the maze imperfectly or not at all. It is generally accepted that both of these observa- tions are related to the fact that, so far as food and water at least are concerned, the animal possesses a built-in mechanism that detects what substance is required for continued well-being and institutes motor activity appropriate for restoring an equilibrium with respect to it. So far as learning goes, we are concerned not with the detection aspect of this process but rather with the way in which such innate mechanisms promote motor activity, for in doing so, they also somehow produce conditions that are favorable for learning to occur. The hypothalamus is known to contain 'centers' for hunger and thirst. These are defined by the fact that certain hypothalamic lesions produce animals that are continuously ravenous while other lesions produce animals that starve to death in the presence of abundant food (162, 230, 233). For thirst, on the other hand, stimulation of the hypothalamus elec- trically (83) or with salt solutions (7) elicits dramati- cally increased drinking behavior. Such observations make it clear that hypothalamic structures connect with the motor apparatus and, under certain condi- tions, organize and control behavior. Animals are motivated to learn, of course, other- wise than by hunger or thirst. Electric shocks, for instance, offer powerful motivation and such shocks act, if we can believe our introspections, through pain and the emotional response to pain that they produce. Learning is a No commonly accomplished when a pleasant emotional experience is associated with the termination of the training procedure. Thus the emotional repertoire of the animal seems to be implicated in learning in a manner similar to that of "simple' factors such .is hunger and thirst. The ques- tion for which we would like an answer is how .ill these motivating factors hunger, thirst, emotional mechanisms and the like prepare the brain for the specific 1 hanges of learning. mechanisms in \iiinhmn. Before proceeding to fur- ther consideration ol ibis, however, le( us examine the concept ol 'attention.' Ii is a common observa- tion in human behavioi that we can deliberately pa) attention to one stimulus rather than another in a constellation of stimuli. The related ability to empha- size relevant stimuli and exclude irrelevant ones is also clearly an important item in some human learn- ing at least. Similar attentive functions appear to operate in animals as well, and selection of the rele- vant stimuli and objects in the learning situation may be the first thing an animal does in learning a maze, a discrimination or even a simple classical CR. In recent times this capacity to attend has been brought into the experimental realm by the studies on the reticular system which are described in detail in Chapter LI I by French in this Handbook (see also 139, 141). Excitation of the ascending reticular acti- vating substance exerts a general alerting or activat- ing effect on many parts of the brain and particularly on the cerebral cortex. Depression by sleep or drugs leads, among other things, to the dropping out of learned reactions, although unconditioned reactions may still remain (113). Many other examples could be given to show its apparent direct involvement in attentive processes. The reticular formation is not the only part of the brain involved in the reactions we call attention, nor is there any reason to believe that 'causing attention' is its sole function. There is not much question, however, that it plays a major role in the process limbic-midbrain CIRCUIT. How then do the motivat- ing and attentive functions prepare the brain for the specific changes in learning? Recent experiments on the limbic system provide what may be a key to the answer. It is becoming increasingh clear that emo- tional mechanisms are largely the consequences of activities in the limbic system and hypothalamus (e.g. 146, 1471. According to anatomical data re- cently analyzed by Nauta (173, 174), these structures receive .1 substantial afferent supplv from the nuclei of Gudden and Bechterew in the midbrain reticular formation; the limbic structures project, in turn, back to these same midbrain regions. Thus a major input to, and outflow from, the limbic system involves the midbrain reticular substance which is somehow con- 1 nned with attention and alertness, and which also transmits, and modifies the transmission of, impulses passing to and from the cortex. II, .is serins reason- able, the limbic and hypothalamic structures are concerned with innate- mechanisms in behavior — emotional and Otherwise then the midbrain reticu- lar substance into which they discharge is potentially the place where such neural aclivilv is brought into contact with the neural consequences of current THE NEURAL BASIS OF LEARNING [495 environmental events. How activity of the neurons of the limbic midbrain circuit would select only some afferent impulses for transmission, which is the proc- ess required to explain 'attention,' is by no means clear (but see io8, 142). Neither, unfortunately, can any critical suggestion be advanced at the present time as to how an influence upon, say, the alerting function of the reticular formation would create con- ditions in the cortex favorable for the specific changes of learning which is our central problem. The simple scheme outlined here is, however, sup- ported by a sufficient amount of anatomical, physio- logical and behavioral evidence to make it worth con- sideration. It does not, of course, account for all the known facts about learning gleaned from any of the experimental sciences. Its main virtue, if it has any at all, is that many experimenters are concentrating their attacks upon it at the present time (see 108, 142, 145, 174). Figure 1 summarizes the anatomical plan that forms the basis of the ideas under discus- sion here. Summary At least two classes of events appear to provide the neural basis for learning. One of these includes tin- durable neural change that constitutes the new link between previously unconnected parts of the brain. Whether this durable change is to be explained l>\ .1 synaptic change or a new neural circuit, or in some other terms, is still a matter for speculation. Its locus, too, is unsettled. Certainly the cerebral cortex is not exclusively the place where such changes occur. The second class of neural events in learning con- sists of those that prime or prepare the brain for the durable change it will undergo. Among these arc the so-called 'motivational' and 'attentive' states that commonly precede and accompany the learning process and without which learning is unlikely or im- possible. Study of the motivational and attentive mechanisms can be expected to supply at least some of the answers to our questions, for the brain changes they produce underlie the brain change we wish to understand. Learning is consequently best conceived not as a particular event in a particular place but rather as a sequence of events that involves various organ sys- tems of the brain in a certain order. The end result, to be sure, is the production of a more or less per- manent change somewhere, but antecedent events determine where, and even whether, it will occur. fig. 1. Simplified anatomical plan of the neural connections involved in learning. The classical afferent systems disseminate information about current environmental events to the cortex and to the reticular formation. Efferent sensory tracts originating from these latter structures terminate in the afferent nuclei and even in the sense organs themselves. The limbic-midbrain hi, ml, which starts in, and distributes to, the reticular sub- stance brings the phylogenetically oldest parts of the cortical mantle, as well as the hypothalamus, into functional contact with the rest of the brain structures at a highly strategic point. 1 Ixperimental information on the vertebrate, while fragmentary as we have seen, is consistent with the idea that each neural structure and circuit outlined here plays its special part as the durable brain change in learning is produced. [U. S. Army photograph. CONCLUDING REMARKS People in the past have ruefully commented upon the primitive state of our knowledge about the neural basis of learning and, as this review makes it clear, such comments are fully justified. Our ignorance, however, is not wholly due to lack of industry, as this summary also clearly demonstrates. What, then, have been the main obstacles to progress and what is being done to overcome them? The major obstacle has undoubtedly been the inability to break through the barrier of the cal- varium in order to expose the brain of normal tin- anesthetized animals to direct experimental investiga- 1496 HANDBOOK OF PHYSIOLOGY M I Rl il'lIYSIOLOGY III lion. Within the last few years this barrier has been effectively overcome. Wires can now be placed in practically any desired brain structure, and conse- quently the study of electrical responses of the brain is no longer limited to the pale and distorted picture provided by scalp electrodes. Similarly with other measurement devices; so long as they are small in size, they can be accurately placed within the neural tissue where the supposed changes occur. This obvi- ously represents a great step forward. On the behavioral side the development of new techniques of measurement (e.g. the operant methods ol Skinner) and the attention being paid to types of learned behavior not previously examined (e.g. im- printing) have appreciably enlarged both the preci- sion and scope of the analysis. Coupled with these factors are the findings from basic neurophysiology and anatomy that expand the possibilities for investigation and provide a solid base for new experimentally testable learning hypotheses. Put another way, there has been a shift of emphasis in thinking about the neural basis for learning. It is no longer fashionable to conceive of the 'temporary con- nections' as occurring exclusively in the cortex. At least the reticular substance and the limbic system can in addition be presumed to play important roles, and thus two entirely new parameters are provided along whicli experimental attacks can be made. The present era in neurophysiology is therefore unlike any other with respect to the neural basis of learning. As recently as 5 years ago an observer not ordinarily given to pessimism might well have con- cluded from the available evidence that our search for the answer was permanently doomed. But today, when obviously relevant measurements of many types are being made in the brain substance of un- anesthetized animals, an avalanche of entirely new information is about to descend upon us and such a conclusion is hardly justified. R E F E R E NCES 1 1. 1 3 13' '4 1 ')■ 16 '7- 18 '9- :n -• I Am s, II. W. ./. Neuropath. & Exper. Neurol. 5: 60, 1946. \m , II \\ \nd I). H. Raab. J. Neurophysiol. 12: 101, '949- Adrian, E. D., F. Bremer and II H. Jasper (editors). Brain Mechanisms and Consciousness. Springfield: Thomas, ■954- Vixen, W. I'. Am. ./. Physiol. 128:754, 1940. \i 1 1 s, U I Am I. Physiol. 132: 81, 1 94 1 . \i 1 1 -. U I Am. J. Physiol. 144:415, 1945. Anderson, B. Acta physiol tcandinav. 28: 188, 1953. Anichkov, s V, Fi&ol. Zhur. S. S. S. R. 38: 395, 1952. Akmincion. J (.:. and L. L. Mitnick. Incidence and properties of ill. kappa EEC rhythm. Report to the Eastern Psycho- logical Association, New York, April 1957. Vrtemyev V V. and N. I. Bezladnova. Trudy Inst. l-,-l,>l I (All/.. I I 228, III",-! \ mi-, \\ K Design foi u Brain. New York Wiley. 195-'. Beck I. C uidK W. Doty. J. Comp.& Physiol. Psychol. 0 Beritofp, I Trans. Inst. Physiol {Tbilisi) 10: 66, 1956. Bishop, c;. II I'hwml 1: ■>■ -~h, iq-,t>. Blouoh 1 1 S Inn Weu York \cad. v i>l>: 733, 1957. Bi.i-m, J. S., k. I. Cilow \m> R. A. Hi im ./. Xruro- physiol. 14 107, 1951 Boyi oi 1 B B and J. /.. Youno. Proc. Roy. Soc., London. ' B 143 449, 1955. Brady, J V. Ann. New York Acad S 66 719, 0157. Brady, .1 V In: Brain Mechanisms and Drug Iction, edited by V\ S Fields Springfield rhomas, 1957, p. m. Brady, J. V. \\n II. F. Hunt, I Psychol |0 ; ; 1955. Br m.-. I V vmi W J II Nauta I Comp S /" . Psychol |i. Brady, | \ L Schreiner, 1 Geller sm> \ Kn\r. p. & Physiol Psycho \ 179 195 1 23. Brogden, W. J. J. Exper. Psychol, 25: 323. 1939. 24. Brogden, W. J. and \V. II. (i\sii I M A. Arch. Neurol. & Psychial. 48: 437, 1942. 25. Bkomh i.v, R. B. ./. Comp. I) W. Neff. ./. Neurophysiol. 20: 108, 1957. Hi. Campbell, R.J. \m> II. I Harlow / Exper. Psychol. r, 1 10, 1945. 14 Carpenter, J. A. J. Comp. & Physiol. Psychol. 45: 413, '95-=- 35. Cere, J. A. and I. S Oris. Fed. Proc. 16: 20, 1957. |6 Cherkes, V. A. / 1 •: ■ Dejai Pavlova 5 415, ■955- 37. Cuius. W. I). ./. Comp. & Physiol. Psychol. 17 412, 1954 38. CHOLOKASHVILI, I. S \mi A. I Riiiikvk Finns. Inst. Physiol. I TbiliHi in -<< ; mV> ig < Ihi iw, K I I Comp & Physiol Psycho: 1 , 1 40. Chow, K I. J Comp. & Physiol. Psychol. 47: 194. [954 41, ( !how , K I \m. I' I Ihii Brain 76: 625 953 !_• Conrad, D. G. \m> I V. Brady. Am 1 12: !'•; "r,7 1 1 1 iiimih, (;. Am. .1. Physiol. 94: 553, ig |0 THE NEURAL BASIS OF LEARNING '497 44. Culler, E. and F. A. Mettler. J. Comp. Psychol. 18: 78. 291. '934- 45. Dashiell, J. F. Psychol. Bull. 32: 261, 1935. 79. 46. Deese, J. The Psychology of Learning (revised ed.). New 80. York: McGraw-Hill, 1958. 47. Deese, J. E. and VV. N. Kellocg. J. Comp. & Physiol. 81. Psychol. 42: 157, 1949. 48. Delgado, J. M. R., W. W. Roberts and N. E. Miller. 82. Am. J. Physiol. 179: 587, 1954. 49. Desmedt, J. E. and G. La Grutta. J. Physiol. 136: 20, 83. 1957- 84. 50. Dews, P. B. and B. F. Skinner (conference co-chairmen i. 85. Ann. New York Acad. Sc. 65: Art. 4, 1956. 51. Diamond, I. T. and D. W. Nefp. J. Ncurophysiol. 20: 86. 3°°. "957- 52. Doty, R. VV., L. T. Rutledge, Jr., and R. M. Larsen. 87. J. Ncurophysiol. 19:401, 1956. 53. Duncan, C. P. J. Comp. & Physiol. Psychol. 42: 32, 1949. 88. 54. Durup, G. and A. Fessard. Annec Psychol. 36: 1, 1935. 55. Ebbinghaus, H. Vber das Gedachtnis: Untersuchungen zur 89. experimentelen Psychologic. Leipzig: Dunckcr und llumblot, 90. 1885. 56. Eccles, J. C. The Neurophysiologies Basis of Mind. Oxford: 91. Clarendon Press, 1953. 57. Epstein, M. A. and C. T. Morcan. J. Exptr. Psychol. 92. 32:453. '943- 58. Evarts, E. V. J. Ncurophysiol. 15: 191, 1952. 59. Evarts, E. V. J. Ncurophysiol. 15: 435, 1952. 93. 60. Ferster, C. B. and B. F. Skinner. Schedules of Reinforce- ment. New York: Appleton, 1957. 94. 61. Fessard, A. E. In: Brain Mechanisms and Consciousness, edited by E. D. Adrian, I- Bremer and H. II. Jasper, 95. Springfield: Thomas, 1954, p. 200. 62. Fields, W. S. (editor). Brain Mechanisms <»«/ Drug Action. 96. Springfield : Thomas, 1 957. 63. Finan, J. L. Am. J. Psychol. 55: 202, 1942. 97. 64. Finley, C. B. J. Comp. Neurol. 74: 203, 1941 65. Fulton, J. F. Frontal Lohotomy and Affective Behaviour. 98. New York: Norton, 1951. 66. Galambos, R. Physiol. Rev. 34: 497, 1954. 99. 67. Galambos, R., G. Sheatz and V. G. Vernier. Science 123: 376. '956- '°o. 68. Gantt, VV. H. .1. .1/. .1 Arch. Neurol. & Psychiat. 37: 848, 1937. 69. Gastaut, H. Electroencephalog. & Clin. Ncurophysiol. Suppl. 101. 6: '33. '957- ">2- 70. Gastaut, H., A. Jus, C. Jus, F. Morrill, \V. Storm van Leeuwen, S. Dongier, R. Naquet, H. Regis, 103. A. Roger, D. Bekkerinc, A. Kamp and J. VVerre. Electroencephalog. & Clin. Ncurophysiol. 9: 1, 1957. 104. 71. Gellhorn, E. Physiological Foundations of Neurology and 105. Psychiatry. St. Paul: Univ. Minnesota Press, 1 953. 72. Gengerelli, J. A. and J. VV. Cullen. J. Comp. & Physiol. 106. Psychol. 47: 204, 1954. 107. 73. Gengerelli, J. A. and J. VV. Cullen. J. Comp. & Physiol. Psychol. 48 : 3 1 1 , 1 955. 1 08. 74. Gerard, R. VV. Am. J. Psychiat. 106: 161, 1949. 75. Gerard, R. VV. Science 122: 225, 1955. 109. 76. Ghent, L ., S. Weinstein, J. Semmes and H.-L. Teuber. J. Comp. & Physiol. Psychol. 48: 478, 1955. 77. Ghiselli, E. E. and C. VV. Brown. J. Comp. Psychol. 110. 26: 301, 1938. Girden, E., F. A. Mettler, G. Finch and E. Culler. J. Comp. Psychol. 21: 367, 1936. Giurgea, C. Stud. Cercet. Fiziol. Neurol. 4: 41, 1953. Giurgea, C. and J. Drogan. Stud. Cercet. Fiziol. Neurol. 5:367, 1954- Giurgea, C. and N. Raiciulescu. IV Congr. Internal. Electroencephalog. ei Ncurophysiol., Abstr. 98, 1957. Grastyan, E., K. Lissak, F. Kekesi, J. Szabo and I. Vereby. Kiserletes Onostudomany 9: 88, 1957. Greer, M. A. Proc. Soc. Exper. Biol. & Med. 89: 59, 1955. Grundfest, H. J. Ncurophysiol. 20: 316, 1957. Harlow, H. F. and J. Dagnon. J. Exper. Psychol. 32: 35'. '943- Harlow, H. F., R. T. Davis, P. H. Settlace and D. R. Meyer. J. Comp. & Physiol. Psychol. 451 419, 1952. Harlow, H. F. and T. Johnson. J. Exper. Psychol. 32: 495- '943 Hebb, D. O. The Organization of Behavior. New York: Wiley, 1949. 1 1 mm ad, G. J. Comp. & Physiol. Psychol. 48: 482, 1955. Hernandez-Peon, R., H. Sciierrer and M. Jouvet. Science 123: 331, 1956 I ll.R\AM)l./-Pl.<)N, R, II SdlERRER AND M. VELASCO. Acta neurol. latmoam. 1: 256, 1955. IIikkick, C. J. In: Publication* m Honor of J. lieritoff. I iilis I'ulil I lonsr. Acad, hi Sciences, Georgian SSR, i95<>. P 335 IIikkick. < | Evolution of Human Nature. Austin: Univ. I exas Press, 1956. Hess, R., Jr., VV. P. Koella and K. Akert. Electroen- cephalog. & Clin. Ncurophysiol. 5: 75, 1953. Hess, W. R. Vegetative Functionen und Zwischenhern. Basil S, liu.ihe, 1948. Hilgard, E. R. In: Handbook of Experimental Psychology, edited by S. S Stevens New York Wiley, 1951. HlLOARD, E. R. Theories of Learning (2nd ed.). New York: Appleton, 1956. Hiloard, E. R. and D. G. Marquis. Conditioning and Learning. New York: Appleton, 1940. HOLT, 1'.. B. Animal Drive and the I. cat rung Process. New York: 1 lolt, 19 31 , p. 26. Householder, A. S. and H. D. Landahi Mathematical Biophysics of the Central Nervous System. Bloomington: Principia Press, 1 945. Iu wis, K. Tohoku ./. Exper. Med. 52: 53, 1950. [WAMA, K, AND M Am / ohol.u ./. Exper. Med. 56: 345, ■952- I\\ \\i\, K. and M. Abe. Tohoku J. Exper. hied. 57: 327, ■953- Jacobsen, C. F. ./. Comp. Neurol. -,-• 271, 1931. Jacobsen, C. F. A. Res. Neri & Merit. Dis., Proc. 13: 225, 1932. Jacobsen, C. F. Physud. Rev. 19: 303, 1939. Jacobsen, C. F. and J. H. Elder. Comp. Psychol. Monogr. 13: 61, 1936. Jasper, H., P. Gloor and B. Milner. Ann. Rev. Physiol. '8: 359. '956- Jasper, H. H., L. D. Proctor, R. S. Knighton, VV. C. Nosh ay and R. T. Costello (editors). Reticular Forma- tion of the Brain. Boston: Little, 1958. Jouvet, M. and R. Hernandez-Peon. Electroencephalog. & Clin. Ncurophysiol. Suppl. 6: 39, 1957. ,498 IIWDIKKiK OF PHYSIOLOGY NEURi)l'H\M()Ioi;V III ««3- i'4 lit.. 117. 118. "9- 123 124. 125. 126. 127. 128. 129. 130. '31- 1 i- "33- "34- '35- 136. '37' .38. '39- 1 p. 141. 142 1 1! [Cappers, C I A Brain 44 125, 1921. Km/. I J AND W. C. Halstead. Comp. Psychol. Monogr. 20: 1, 1950. Kellogg, W. N. ./. Exper. Psychol. 29: 334, 1941. Kellogg, U V J. Exper. Psychol. 37: 263, 1947. Kellogg, \\ N., I Debse, N. H. Pronko and M. Feinberg. J. T.xfiet. Psychol. 37: 99, 1947. Kin(;. F. A I Verv. S \4eni Dis. 126:57, '958- Kluver, II. J. Psychol. 2: 49, 1936. Kogan, A. B. XX Internal. Physiol. Congr., Communic. 253. '956- KOMI ER, W. The Mentality of Apes. New York: Ilarcourt, ■925- KONORSKI, J. Conditioned Reflexes and Xcuron Organization. London: Cambridge, 1948. Komirski, J. In: Physiological Mechanisms in Animal Behaoiota (Soc. Exper. Biol. Symp. No. 4). New York: Acad. Press, 1950, p. 409. Konorski, .(., 1. Stepien, S. Brutkowski, \\ . Lawicka and I. Stepien. Hull. Soc. Sc. Lettr. Lodz in: 1, 1952. Krech, I)., M. R. Rosenzweio and E. L. Bennett. J. Cell. & Comp Physiol. 49: 261, 1956. Kupalov, P. S. XX Internal. Physiol, Congr., Communic. 34. 1956- Lashley, K. S. Brain Mechanisms and Intelligence. Chicago: Univ. Chicago Press, iQ2g. Lashley, K. S. J. Comp Neurol. 53: 419, 1931. Lashley, K S. Comp. Psychol. Monogr. 11: 5, 1935. Lashley, K. S. Comp. Psychol. Monogr. 11: 43, 1935. Lashley, K. S. ./. Comp. Neurol. ;q: 431, 1943. Lashley, K S Genet Psychol. Monogr. 37: 107, 1948. Lashley, K S. In Physiological Mechanisms in Animal Behaviour (Soc. Exper. Biol. Symp. No. 41. New York Acad. Press, 1950, p. 454. Lashley, K S. In: Patterns of Organization in the Central Mervous System, edited by P. Bard. Baltimore: Williams & Wilkins, [952, p. 529. Lashley, K. S. and L. E. Wiley. ./. Comp. Neurol. 57: 3. '933 I.i u Ii In Nebraska Symposium on Motivation, edited by M K Jones. Lincoln: Univ, Nebraska Press, K|-,7, vol. v. Ijvamiv, M N \m> K I. Poliakov Hull. Acad. Sc. USSR ) 286, 1945 Ltvinoston, K B In Brain Mechanisms and Drug Action, edited l>\ U S I ii Ida Springfield ["nomas, 1957. Ltvinoston, K B, In Handbook 0/ Physiology Neuro- physiology, edited in J. Field, II. W. Magoun and V 1 1 1. ill Washington, D. C Am. Physiological S01 '959 vol. 1, p. 741. Looms, A I.. I. N Harvey wot; Hobari J. Expei Psycho, Loucks, R. B. J , ; j 1 1 . 1931 . /. Neurosurg. 1 1 : 29, 1954. A.M. A. Arch. Neurol. & Psychiat. 78: 45. MacLean, P. D. 4b. MacI.I AN, P I). "3. '957- 47. MacLean, P. I). AM A. Arch. Neurol. & Psychiat. 78: 128, 1957. 48. MacLean, P. D. and J. M R. Delgado. Electroen- cephalog. & Clin. Ncurophysiol. 5: 91, 1953. 49. MacLean, P. I), S. Flanigan, J. P. Flynn, C:. Kim and J. R. Stevens. Yale .1 Hud Si Med. 6: 380, 1955. 50. M mu k. B. A. ./. Comp. 0: Physiol. Psychol. 48: 102, 1955. 51. Malmo, R. B. ./. Ncurophysiol. 5: 2q->, 10,42. 52. Marquis, D. G. and E. R. Hilgard. ./. Comp. Psychol. 22: '57. '936- 53. McGeoch, J. A. \m> A. L. Irion. Psychology of Human Learning (revised ed.). New York: Longmans, [952 -,4. Meyer, I). R.. II 1 Harlow and II. W. Ades. Am. ./. Psychol. 64: 391, 1 95 1. 55. Meyer, D. R., H. F. Harlow \nd P. II Settlage. J. Comp. Ci Physiol. Psychol. 44: 17, 1951. 56 Meyer, D. R. and C. N. Woolsey. J. Ncurophysiol. 15: '49. >952- 57. Milner, B. Psychol. Bull. 51 : 42, 1954. 58. Mishkin, M. ./. Comp. Ci Physiol. Psychol. 47: 187, 1954. 59. Mishkin, M. and M. Hall. ./. Cornp. II II [asper. I encephalog. l DS, J. J Comp. & Physiol Psychol. 49: 281, 77 < >i 1 >-,, J In: Reticulai Formation oj the Brain, edited by II II Jasper, 1. I) Proctor, R S. Knighton, U C Noshay and R. T. Costello Boston Little, 1958, p. 671. 78 1 h os, |. K 1 K111 \m \nd P Bach-y-Rjta. Science 12) 2l.-„ IC|V 79. Olds, J. and P. Milner. .1. Comp. ss Physiol. Psychol. 47: 419, 1954 80. Pavlov, II' Si -■ 1 61 ;, 1906 81. Pavlov, 1. P. Conditioned Reflexes. New York Oxford, "127 182. Pavlov, IP/ I onditioned Reflexes, translated THE NEURAL BASIS OF LEARNING '499 by W. H. Gantt. New York: Internat. Univ. Press, 1928. 183. Pavlov, I. P. Experimental Psychology. New York: Philos. L'br., 1957. 184. Penfield, W. In: Brain Mechanisms and Consciousness, edited by E. D. Adrian, F. Bremer and H. H. Jasper. Springfield : Thomas, 1 954. 185. Pennington, L. A. J. Comp. Neurol. 66: 415, 1937. 186. Pennington, L. A. J. Comp. Neurol. 74, 169, 1941. 187. Pinto, T. and R. B. Bromilev. J. Exper. Psychol. 40: 121, 1950. 188. Poirier, L. J. J. Comp. Neurol. 96: 209, 1952. 189. Poltyrev, S. S. and G. P. Zelionv. Ztschr. Biol. 90: '57. !93°- 190. Popov, C. Electroencephnlog. II \v. Ades. J. Comp. & Physiol. Psychol. 46: 145, 1953. 205. Riopelle, A. (., II. F. Harlow, P II Si 1 n age and I I VV. Ades. J. Comp. & Physiol. Psychol. 44: 283, 1951. 206. Rosner, B. S. and VV. D. Neff. Am. J. Psychol. 4: j;-,, '949- 207. Rousinov (Rusinov), V. S. and G. D. Smirnov. Electro- encephalog. &.Clin. Neurophysiol. Suppl. 6: 9, 1957. 208. Ruch, T. C. A. Res. New. & Mint. />/.., Proc. 15: 289, ■935- 209. Ruch, T. C, J. F. Fulton and W. J. German. I U.I. Arch. Neurol. & Psychial. 39: 919, 1938. 210. Rusinov, V. S. XX Internat. Physiol. Congr., Communit. 346, 1956- 211. Rusinov, V. S. and M. Y. Rabinovigh. Electi oencephalog. & Clin. Neurophysiol. Suppl. 8, 1958. 212. Sarkisov, S. A. XX Internal. Phsyiol. Congr., Communis. 1 12, 1956. 213. Shurracer, P. S. and E. Culler. J. Exper. Psychol. 26: ■33. '94°- 214. Shurracer, P. S. and E. Culler. J. Exper. Psychol. 28: 287, 1 94 1. 215. Shurracer, P. S. and H. C. Shurracer. J. Exper. Psychol. 29: 217, 1 94 1. 216. Shurracer, P. S. and H. C. Shurrager. J. Exper. Psychol. 36: 347, 1946. 217. Sidman, M. Science 122: 925, 1955. 218. Sidman, M. Ann. New York Acad. Sc. 65: 282, 1956. 219. Sidman, M., J. V. Brady, J. Boren, D. Conrad and A. Schulman. Science 122: 830, 1955. 220. Skinner, B. F. The Behavior of Organisms. New York: Appleton, 1938. 221. Small, VV. S. Am. J. Psychol. 1 1 : 133, 1899. 222. Smith, K. U. J. genet. Psychol. 51 : 329, 1937. 22 ; Spaet, T. and H. F. Harlow. J. Exper. Psychol. 32: 4J4. '943- 224. Spence, K. W. In: Handbook of Experimental Psychology, edited by S. S. Stevens. New York: Wiley, 195 1. 225. Sperry, R. VV. In: Handbook of Experimental Psychology, edited by S. S. Stevens. New York: Wiley, 1951. 226. Sperry, R. W. and N Miner. J. Comp. <£? Physiol. Psychol. 48: 463, 1955. 227 Sperry, R. VV., J. S. Stamm and N Miner. J. Comp. & Physiol. Psychol. 49:529, 1956. 228 Stamm, J. S. and R. W. Sperry. J. Comp. & Physiol. Psychol. 50: 138, 1957. 229. Stein, L. Science 124: 1082, 1956. 230 STELl sr, E. Psychol. Rev. 61 : 5, 1954. Stellar, E., C. T. Morgan ami M Yarosm. J. Comp. Psychol. 34: 107, 194. :jj Stewart, ). W. wo H. W. Ades. J. Comp. © Physiol. Psychol 11 479, 1951. 233. 1 1 iiei.baum, P. J. Comp. ss Physiol. Psychol. 48: 156, 1955. 234. Thompson, R and VV. Dean. J torn/' ^- Physiol Ps) 48:488, 1955. 235. Thorndike, E, I- Psychol. Rev. Monogr. 4, Suppl. II: i, 1898. j jti Thorpe, \\ II. In: Physiological Mechanisms in Animal Behaviour iScx-. Exper. Biol. Symp. No. 4). New York: Acad. Press, 1950, p. 387. 237. Thorpe, W. II Learning and Instinct in Animals. Cambridge: 1 1. >i \ ard, 1956. 238. TlNBERGEN, N. The Study of Instinct. London : Oxford, 1 95 1 . 239. Tunturi, A. R. Am. J. Physiol. 181 225, 1955. 240. vus \iim\\n. J. In: Cerebral Mechanisms m Behavior, edited by I. A. Jeffress. New York Wiley, 1951, p. 1. 241. Walter, W. <; Tht Living Brain New Yoik Norton, '953- 242 Warren, J. M. and H. F. Harlow. ./. genet. Psychol. 81: 45. '952- ^43. Warren, J. M. and 11. F. Harlow. J. Comp. £ / Psychol. 45: 119, 1952. 244, Wiki.er, A. Thi Relation of Psychiatry to Pharmacology. Baltimore: Williams & Wilkins, 1 957. 245. Winc, K. G. Am. J. Psychol. 59: 583, 1946. .'4t). Wing, K. G. Am. J. Psychol. 60: 30, 1947. 247. Wing, K. G. and K. U. Smith. J. Exper. Psychol. 31 : 478, '94-*- 248. Yoshii, N. Elei tiocn, ephalog. & Clm. Neurophysiol. Suppl. 6: 75- >957- 249. Yoshii, N., J. Matsumoto and Y. Hori. IV Congr. Internal. Electro-encephalog. et Neurophysiol., Abstr. 79: 1957. 250. Yoshii, N., P. Pruvot and H. Gastaut. Electroencephalog. & Clm. Neurophysiol. 9: 595, 1957. 251. Zubek, J. P. J. Neurophysiol. 15: 401, 1952. 252. Zubek, J. P. Canad. J. Psychol. 6: 183, 1952. 253. Zubek, J. P. J. Comp. & Physiol. Psychol. 45: 438, 1952. CHAPTER LXII Drive and motivation ELIOT STELLAR1 Institute of Neurological Sciences, University of Pennsylvania Medical School, Philadelphia, Pennsylvania CHAPTER CONTENTS Classical Instinct Doctrine Cannon's Local Theories Homeostasis and Self-Regulatory Behavior Multifactor Central Neural Theory Behavioral Definition of Motivation Need Drive Goal Goal-Directed Behavior Satiation Motivated Behavior Behavioral Measures of Motivation General Activity Consummatory Behavior Obstruction Method Choice, Preference and Competition Among Drives Learning and Learned Performance Summary The Neurophysiology of Motivation Diencephalic Mechanisms Other Central Mechanisms Sensory Factors Internal Environment Factors Interaction of Factors The Role of Learning Conclusions concerned themselves with the basic question in motivation by trying to explain the arousal and selective direction of behavior; but their use of theo- logical, teleological and vitalistic conceptions of the 'forces' operative in behavior threw the whole question of motivation into scientific disrepute. Such was the Zeitgeist 30 years ago, for example, that Boring was able to write .1 History oj Experimental Psychology (25) in i()2(j without mention of motivation. It is now possible, however, to look back over history and trace the lines of development of modern approaches to motivation through the contributions of the stu- dents of instinct, the experimental physiologists and the physiological psychologists. What emerges is an historical process showing a) a gradual replacement of imaginary, explanatory 'forces' by objective, operational definitions of motivated behavior, and b) a shift of physiological emphasis from peripheral sensors and hormonal mechanisms to the central neurophysiologic.il mechanisms underlying motiva- tion. CLASSICAL INSTINCT DOCTRINE the notion that behavior is motivated and that the scientific study of motivation might be a profit- able approach to the understanding of behavior arose only recently in the history of experimental psy- chology. From the earliest times, the philosophers 1 This chapter was written while the author held National Science Foundation Grant G-1372 for the study of phys- iological mechanisms of motivated behavior. Thanks are due to Dr. J. M. Spraguc and Dr. Alan Epstein for their helpful comments on the manuscript. The earliest scientific thinking about motivation developed with the instinct doctrine. [Beach (19) may be consulted for a recent critical history of instinct.] Unfortunately, much of the idea of motiva- tion was lost in the philosophical effort to maintain the view of man as a unique and free creature of reason and, by sharp distinction, to relegate animal to the control of nature's predetermined instinctive- forces. Not until Darwin's emphasis of the role of adaptive behavior in evolutionary survival did the instinct doctrine begin to receive full scientific atten- 1501 I i02 II VNDIlooK OF PHYSIOLOGY NEUROPHYSIOLOGY III tion and did ii become possible to think of human motivation in terms of the simple paradigm provided by the growing study of instinct in animals. Interest- ingly enough, it was through the efforts of students of personality and social processes, in the work of Freud (58) and McDougall (9(1), that the concept of motivation re-entered psychology. Freud's concep- tions of the 'id,' 'libido,' 'pleasure principle,' 'anxiety1 and McDougall's 'propensities' and 'hormic forces' all focused attention on the arousal and direction of behavior, and rooted the notions of instinct and motivation deeply in modern psychological thought. But even when laid in the foundations of biology by Darwin and pushed into psychology by Freud and McDougall, the concept of instinct still brought objections, lor it was all too often used as an explana- tory force, and once behavior was labeled instinctive, little was done to investigate it. It was sufficient to s.iv that animals made adaptive responses like build- ing nests and mating because of nest-building in- stincts and mating instincts, and that man fought and banded together in societies because of instincts of aggressiveness and gregariousness. In their revolt against the mentalism of the earlier philosophers, the first behaviorists, under Watson (5 I, ''^ ;. 165), rejected the instinct doctrine and with it a large part of the concept of motivation. First, they objected to the use of instinct as an imaginary explanatory force. Second, they rejected instinct because the doctrine implied physiological processes inside the organism, while the behaviorists were trying to account lor behavior solely in terms of environmental stimuli and responses and considered instinctive acts merer) complex chains of reflexes Third, the instind doctrine definitely assumed, con- trary to behavioristic theory, that some behavior was not derived from experience but rather stemmed from the organisms' inherited biological charac- teristics. I In- contribution of the behaviorists to psychology is undeniable, fur they firmlv established objective experimental methods in the stud) of behavior, But the) proceeded in the tradition of their mentalistic predeo sors 10 build a psychology withoul motiva- tion, without hcredit) and with nothing more than liji service to the physiological basis ol behavior. As I till "l" put il, the behaviorists 'threw the babv cjiii with the bath' when the) quite rightly rejected instil i in m force and quite w rongl) ignored the biological foundations of important kinds of tivated behavior, simplv because they had on< >■ been called instinctive. I wo important contributions emerge from this controversy over instincts, a) The behaviorists devel- oped an objective operational analysis of motivated behavior without resort to imaginary explanatory forces. /)) The students of instinct called attention to the fact that motivated behavior is more than a complex response to external stimulation by empha- sizing the role of internal physiological states which could determine whether or not external stimuli would be effective. ( VNXON S LOCAL THEORIES The first really scientific stride in our understanding of the physiology of motivation came with the efforts ol ( .iiiiKin 1 |n 1 .mil his co-workers in the investigation of hunger and thirst. Theirs was an approach from the strict experimental and analytic view point with the definite aim of elucidating physiological mecha- nisms. On the behavioral side, Cannon was concerned with only two facets of motivated behavior. One was the initiation of the behavior, and the other was the accompanying sensation that could be reported In- human beings. Since he found a correlation between gastric contractions and reports of hunger in his stomach-balloon experiments on man, he coin hided that eating was aroused by local gastric contractions which, moreover, provided the physiological basis for the sensation. Similarly, he concluded that local dryness of the throat and mouth was at the basis of thirst sensation. Neurophysiologically, of course, these local theories must mean that receptors are stimu- lated in the stomach and throat, providing a basis lor .liferent impulses from these peripheral structures. Other workers, following Cannon's lead, proposed local theories for other kinds of motivation, for example local irritation or pressure arising from the genitals in sexual behavior and local changes in taste-receptor sensitivity in salt hunger. These local theories were attractive because they seemed to lii human personal experience and because the) were so simple in pulling their main emphasis on peripheral structures which could be dealt with experimentally. Furthermore, they provided a verv e.isilv understandable model for the behaviorists who were Irving lo understand all behavior in simple stimulus-response terms. When motivation began to enter behavioristic thinking, the theories went some- thing like this: llie hungT) animal works lor food or learns on the basis ol food reward because lood DRIVE AND MOTIVATION l5°3 reduces or removes the peripheral local stimulation. While the behaviorists decried any reference to subjective experience, it seems unlikely that this theory was hurt by the fact that it implied the removal of 'unpleasant' peripheral stimulation with consequent 'relief or 'pleasure.' Despite the general acceptance of Cannon's local theories and their extension and oversimplification by others, in time many cogent arguments against this viewpoint developed. In the first place, human cases were found where sensations of hunger and thirst existed apart from local stimulation. Some people never seemed to experience gastric contrac- tions, yet they ate; people with congenital absence of salivary glands distinguished between their chroni- cally dry mouths and thirst and were able to drink appropriately to their water deficits (150). Further- more, it was shown that denervation or surgical excision of the stomach did not destroy hunger in man (74, 163) or the regulation of food intake in animals (159) or the animal's ability to work and learn for food rewards (11, 1 08). Finally, as further investigations were made into the physiological basis of hunger, thirst, sexual be- havior, etc., it became apparent that local stimula- tion could only be one of several factors contributing to these kinds of motivated behavior. By no means can gastric contractions or local throat dryness, etc. be considered essential in the arousal, maintenance and satiation of such motivated behavior. On the other hand, we have no evidence at present enabling us to deny that local factors make some contribution or even to argue against the possibility that they might provide a most important basis of the sensa- tions accompanying motivation; or that an extremely important part of the motivation of animals working and learning for rewards might not be the reduction or removal of peripheral stimulation. These are matters to be assessed experimentally, and we will discuss some of the relevant studies later. HOMEOSTASIS AND SELF-REGULATORY BEHAVIOR The second important physiological advance in the study of motivated behavior was made in the bold conceptualizations and extensive investigations of Richter (129, 130). A behaviorist with an organ- ismic point of view, he saw that motivated behavior could be of adaptive value in the survival of the organism because of its essential contribution to the maintenance of the internal environment. Starting with the conceptualizations of Claude Bernard and with Cannon's homeostasis, Richter conceived of motivated behavior as self-regulatory behavior in the sense that it may correct deviations of the internal environment in cooperation with the more automatic physiological mechanisms. For example, the warm- blooded animal can regulate its temperature by dietary selection, nest building or simply moving from a hot to a cooler environment or vice versa, as well as by shivering, piloerection, panting, sweating, and vasomotor and metabolic changes. Quite clearly such behavioral responses as these have all the characteristics of motivated behavior, and their investigation has given valuable insight into the physi- ology of motivation. In his extensive investigations, Richter was able to show that the organism actually is sensitive to main of its own physiological needs and will develop motivated behavior appropriate to the correction of those needs and the maintenance of the internal environment. For example, he showed that rats were able to select their own diets, cafeteria-style, from a complete array of dietary components (129). Further- more, shifts in amounts of different substances ingested followed in accordance with prior dietary restrictions, changes in environmental temperature, endocrine gland extirpations, pregnancy, etc. For example, the parathyroidectomized rat ingests abnormally large amounts of calcium and abnormally small amounts of phosphorus, quite in keeping with its physiological mills (127). Similarly, the adrenalectomized rat keeps up its sodium level by ingesting excessive amounts of sodium chloride solutions (126), and so on through main strikini; examples. As a behaviorist, Richter saw no advantage in invoking instinct to account for this remarkable behavior. Instead, he sought an explanation in terms dI some local expression of the state of physiological need, after the fashion of Cannon's theories, in chansjes in peripher.il receptor mechanisms. Thus, he attributed the strong salt hunger of the adrenalecto- mized rat to an increase in sensitivity of the salt receptors in the mouth, for the adrenalectomized rat shows a preference for concentrations of sodium chloride solutions so weak that the normal rat fails to select them over water (10, 128). The fact that, following taste-nerve section, adrenalectomized rats failed to select salt and died (129) was taken to support the role of the peripheral receptor change in this motivation, but the delect may have been in the capacity to detect salt rather than in the motivation. As a matter of fact, electrical recording from the '5"4 HANDBOOK OF PHYSIOl nl.Y NEUROPHYSIOLOGY III chorda tympani (116) and a study of conditioned responses to salt (41) both show that the normal and adrcnalcctomizcd rat have the same very low sensory thresholds for salt in solution. Therefore, the greater salt ingestion of the adrenalectomized rat, even at threshold concentrations but without a change in sensory threshold, suggests that the physiological need must reflect itself elsewhere in the nervous system than in the local, peripheral sensory mechanisms of taste. Regardless of the mechanism of its influence on the nervous system, the role of the internal environ- ment in motivated behavior is obviously an important one. Vet a number of questions should be asked. a) Are there some physiological needs and deficits that do not lead to adaptive motivated behavior? b) Are there changes in the internal environment which are not needs or deficits but which can lead to motivated behavior? c) Can there be motivated behavior in the absence of internal environment influences? The answer to all three questions is 'yes.' For example, all rats do not grow well in cafcteria- type feeding experiments where they select their own diets (117), and thus far, no specific hungers have been shown for vitamins A and D (67, 168). Actually of course, there is no logical reason why organisms should have specific hungers for every identifiable nutrient in order to survive. In regard to the second and third questions, it is quite evident that motivation ran occur without specific deficit or need. There is no interna] deficit in sexual and maternal behavior, and the survival of the individual is not at stake, although there are important changes in the internal environ- ment associated with these motivations. Finally, in Cases such as the motivation which an animal shows for a nonnutritive substance like saccharin (20) or the motivation to avoid pain, or to manipulate objects or explore .1 new environment (65), there are no asSO- c iati d 1 hanges in the internal environment known to be important in the arousal and maintenance of the motivation. Thus, not all motivated behavior is self-regulators behavior, nor is it always of particular adaptive sig- nificance in the survival of the individual. The in- iriii.il environment, important .is it is in main re- markable cases of sell-regulation, is onl\ one of the factors contributing to the control ol motivated be- havior. When it does operate, the pi i\ siological tion becomes: how do changes in the internal environment influence the nervous s\siem and, there- Ion', behavior? Although local peripheral mechanisms may be the critii al targets in certain cases, as Richter and Cannon suggest, it is clear we must look elsewhere in the nervous system for the major effects, presumably in some central neural mechanism. MIT IT FACTOR CENTRAL NEURAL THEORY It was Lashley (86) who first put the problem of motivation on a modern neurophysiological basis in his classic paper, 'The experimental analysis of in- stinctive behavior.' Unlike Cannon and Richter, Lashley gave no special emphasis to local sensory factors in motivation and did not concern himself directly with homeostatic mechanisms, regulatory behavior or needs. Rather he approached the problem of motivation as a student of the central nervous sys- tem with the multifactor theory that motivation was the outcome of the joint contribution of many sensory and humoral influences to some central neural mecha- nism. Although he offered no direct suggestion as to the locus and nature of the central neural mechanism, he made a thoroughgoing analysis of motivated be- havior, departing radically from the simple stimulus- response theories of the behaviorists; thereby, he described what the major properties of this central neural mechanism might be. In his behavioral analysis, Lashley made three im- portant points, a) Instincts and motivated behavior are not simply complex chains of reflexes and arc not represented by stereotyped acts. I'he detailed re- sponses involved in mating, nesting, retrieving, etc. vary from individual to individual and occurrence to occurrence. One cannot, therefore, specify a particular motor sequence thai characterizes the behavior, for the same result inav be achieved bv different be- havioral means on different occasions. In Similarly, motivated behavior is not dependent upon anv single stimulus, confined to a particular receptor locus. Usually, a number of stimuli are elle, live in a par- ticular sensory Or perceptual pattern across several modalities. While a single stimulus might be sufficient to arouse motivated behavior, the adequacy and the intensity of the response are determined bv the com- pleteness of a complex pattern of stimulation the ani- mal receives. , i Whether stimuli are effective and how effective they are may depend upon sensitization of the organism to particular stimuli bv changes in its internal environment. For example, the chronically castrated male rat will not be aroused by the usually effective pattern presented bv the female in heat until injected with SfX hormones. 1 hat I .ash lev w as on t lie right track is shown bv the DRIVE AND MOTIVATION :5°5 subsequent development and extension of his theories by Morgan (107), Beach (13) and Stellar (151), and by the parallel theories developed by the European ethologists, Tinbergen (158) and Lorenz. In his ex- tensive work on sexual motivation, Beach was able to bring together much experimental evidence in support of Lashley's thesis and did much to make his analysis of sexual behavior a model for the under- standing of the physiology of motivation. He was able to show, for example, that sexual motivation actually was under multifactor control, presumably through the joint effects of a number of variables on a central excitatory mechanism which he felt had many of the properties of Sherrington's central excitatory state. Beach (14, 16, 17) drew the following conclusions from his extensive research and surveys of the litera- ture, a) No one sensory avenue is indispensable for the arousal of sexual behavior, for any two sensory s\ 5- lems can be interrupted by peripheral nerve or tract section (auditory, visual or olfactory) or by partial denervation (of genital areas or face and mouth 1 without destroying sexual motivation in the naive rat. Rather, it appeared that it is a nonspecific mini- mum of sensory input that is important in determining sexual arousal, just as it seems to be in the activation of locomotion (62). h) The neocortex plays a role, but no one part of it is critical in sexual arousal of the male rat, for example, since experiments show that, regardless of locus, the larger the cortical lesion, the greater the deficit in arousal, c) Sex hormones also add their effects, for without them, sexual motivation may be greatly reduced or absent, yet it can be re- stored by hormonal injection, d) Learning also makes an important contribution to the arousal of sexual behavior, for previously ineffective stimuli may, through experience, facilitate the arousal of sexual motivation. In sensory deprivation experiments, for example, it may be necessary to interfere with three sensory systems peripherally in the experienced male rat before motivation is eliminated, compared to two in the naive animal. Furthermore, there is evidence that in the male primate, some sexual experience may be essential for the appearance of adult sexual motiva- tion, e) That these various factors interact in a com- mon subcortical neural mechanism is suggested by tin- fact that sexual behavior, lost as a result of neocortical lesions, may be restored by hormone injections. (See p. 1520 on interaction of factors.) Particularly important in this analysis of sexual be- havior are the changes which take place in the con- trol of sexual motivation in phylogeny. Comparing animals from rat to man, there are a decreasing de- pendence upon hormones, and an increasing depend- ence upon sensory factors, learning and the neocortex. Sex differences are also important and instructive. The female, for example, is much more dependent upon hormones through the phylogenetic series Beach compared than the male. The male, on the other hand, is much more influenced by changes in sensory stimuli, cortical lesions and learning than is the fe- male. A somewhat independent development of these views of Lashley and Beach is seen in the contribu- tion of the European ethologists, Tinbergen and Lorenz (cf. 89, 158), in their study of instinctive behavior. This theory of instincts perhaps suffers because it is not based on modern neurophysiological principles but rather is cast in hydrodynamic terminology. Thus, Tinbergen speaks of neural mechanisms controlling instinctive acts which build up a "reservoir' of 'action specific energy' until released l>\ some appropriate Stimulation. No direct effort at localization or experi- mental manipulation of the neural mechanism is as yet apparent in this work. But this criticism is of only minor concern at the moment, for it would be quite possible to recast Tinbergen's terminology and make a direct experimental approach to the neurophysio- logical problem. Like Beach and Lashley, the ethologists propose that changes in the internal environment, such as those produced by deprivation or an increase in sex hormones, contribute, along with sensory influences, to the arousal ol .1 central neural mechanism. In some instances, internal changes may be intense enough to yield instinctive patterns, in the absense of sensory stimuli, in which case the ethologists speak of 'vacuum reactions.' But usually, sensor) stimuli play two essen- tial roles: a) they contribute to the level of excitation in the central neural mechanisms, and b) they 'trigger' the response by releasing the excitatory mechanisms from the control of a postulated inhibitory mecha- nism; in this latter case, the stimuli are called 'sign stimuli' or 'releasing stimuli." On the basis of their analysis of complex instinctive acts into behavioral hierarchies, the ethologists fur- thermore assume that there is a hierarchy of neural mechanisms, each built up by internal and sensory influences and each selectively released from inhibi- tion by appropriate stimuli. The release of each neural mechanism not only results in a behavioral expression but also 'primes' the next lower neural mechanism in the hierarchy by contributing to its excitation along with sensory and humoral factors. Thus in the re- productive behavior of fish, for example, the sequence i v>»> IIWDBOOK OF PHYSIOLOGY XF.t KOPHYSIOI.OGY III CHEMICAL 8 PHYSICAL HORMONES 9L000 TEMP OSMOTIC PRESS OHUGS REGULATION OF INTERNAL BALANCE FEEDBACK FROM 1SUMMATORY BEHAVIOR fig. I. Schematic diagram of the physiological factors con- tributing to the control of motivated behavior. Description in text. [From Stellar (151).] is started b) hormonal triggering of the highest neural mechanism which yields migrator) behavior. Then, as new stimuli in the environment are encountered, excitation in a succession of lower neural mechanisms is built up and released by sign stimuli specific to selection of territory, nest site and nesting materials, fighting in territorial defense against intruders, mating and care of the young. While the neurophysiologicaJ propositions of the ethologists are quite speculative, their behavioral analyses have been excellent and will provide a basis for direct physiological investigations "1 the mecha- nisms underlying motivation. Their work makes possible the extension of Beach's phylogenetic com- parisons to infrainammalian species and to kinds of motivation other than sexual. Furthermore, they oiler rich insights into the organization of motivated behavior because of their insistence upon relatively completi descriptions of patterns of motivated be- havior as observed in naturalistic settings, in contrast to the American psychologists' relatively artificial laboratory testing oi isolated segments of behavior. One thing should lie apparent now about all of thes< 1 entral neural theories; the) are based larger) on inferences from behavior, for the) fail to take into 11 1 1 1 1 direct studies ol the central nervous system which might provide experimental evidence relevant to the locus and properties of the postulated central neural mechanism. Fortunately, there is now a large and growing body of experimental data on the role of certain central neural structures in the arousal and integration of motivated behavior, and it is possible to use this information in a specific and physiologically concrete extension of the theories of Lashley and Beach. This was done originally in i<)r>4 by the present author (151) in an effort to arrive at a unified multi- factor theory of motivation, general enough to apply to many different kinds of motivation, across many different species of animals. A brief summary of this theoretical view will be given here in order to provide an up-to-date physiological framework for the re- view, later in this chapter, of the experimental evi- dence we now have available on the physiological basis of motivation. A schematic diagram of the physiological mecha- nism believed to underlie motivation is shown in figure 1. Look first at the diencephalic mechanism in the middle of the diagram. As far as we can tell from present experimental evidence, the major focus of the neural system or the integrating mechanism respon- sible for the arousal, execution and satiation of mo- tivated behavior lies in the diencephalon, probabl) the hypothalamus. Since ablation and stimulation of restricted foci in this region of the brain result in either increases or decreases in motivated behavior, it ap- pears, furthermore, that it may contain two kinds of functional areas which can be described operationally as excitatory and inhibitory mechanisms. The basic assumption here is that the arousal ol motivated be- havior is determined directly by the output of the ex- citatory mechanism, and the satiation of motivated behavior b) the output of the inhibitor) mechanism. Thus, there is believed to lie .1 reciprocal mechanism which provides a basis for refined and graded control ol motivated behavior. Whether the inhibitory mecha- nism acts onl) on the excilalorv one as suggested in some experimental work, and shown in the diagram, or whether the two mechanisms exert their effects on a common mechanism for the execution of motivated behavior is still an open question. In any case, starting with this dual diencephalic mechanism, the question then becomes: what controls its .letivitv and therefore the arousal, execution and satiation of motivated behavior? As I.ashlev and Beach suggest, the behavioral evidence implies that three classes of factors ate operative. "' Sensory in- fluences, operating through afferent pathways, physio- logicall) defined as specific and nonspecific, ma\ DRIVE AND MOTIVATION r507 arouse these diencephalic mechanisms directly or in- directly. On the one hand, as Beach suggests, these sensory influences are additive in their effects, so that it is the sum total of sensory input to these dien- cephalic mechanisms that determines the amount of arousal or satiation. On the other hand, sensory in- fluences must contribute highly specific information, for motivated behavior may be highly discriminative and selective. Furthermore, as Beach points out, sensory influences may be classed as learned or un- learned, for previously ineffective stimuli can come to arouse or satiate as a result of past experience, b) Chemical and physical properties of the internal en- vironment, operating through the circulators system and through the cerebrospinal fluid system, pre- sumably can contribute directly to the activity of these excitatory and inhibitory mechanisms and, therefore, to the arousal and satiation of motivated behavior. Again, while these humoral influences may have general arousing and depressing effects, there is evidence to suggest that their effects may be highly specific to particular types of motivated behavior, implying the possible existence of highlv selective central 'receptors,' sensitive to changes in the internal environment, c) Finally, central neural influences, arising elsewhere in the nervous system, particularly the neocortex and the rhinenccphalon, may contrib- ute excitatory or inhibitory effects to the control of the diencephalic mechanisms and, through them, play a role in the serial organization and patterning of mo- tivated behavior, as well as its arousal and satiation. While the factors operating in the control of motiva- tion can be listed simply, their mode of action is undoubtedly complex. In the first place, all of these factors probably interact in their influences, being interdependent and perhaps equipotential in the arousal of the basic diencephalic mechanisms. Thus, their combined influences might be thought of as addi- tive such that an increase in one influence would com- pensate for a decrease in another. For example, how effective a given sensors stimulation will be in arousing motivation will depend upon the concom- itant influence of the internal environment or upon the nature of previous stimulation whether learned or unlearned. In the second place, it is possible that each of these factors could have two kinds of influence on the dien- cephalic mechanisms; thev could either activate or depress either excitatory or inhibitory diencephalic structures. This results in an enormous increase in the possibilities for refined control and for complexity of mechanism. A sensors stimulation, a neocortical or rhinencephalic influence, a hormone, or a drug, for example, could decrease motivated behasior by activating the inhibitory mechanism or depressing the excitatory mechanism. A third complexity in this mechanism derives from the fact that the execution of the motivated behavior itself can provide important sensory changes and changes in the internal environment (see the lower part of fig. i ). Where materials are ingested by the organism as in hunger and thirst, there is a continuous change in the internal environment due to absorption once the behasior starts, and throughout, there is obviously a new source of stimulation of the ali- mentary tract upon ingestion. In addition, even svhere there is no ingestion, as in pain asoidance or in nest building, there is feedback of response-produced stim- ulation occasioned l>s the execution of the behavior as well as changes in stimulation resulting from changes in the ens ironment produced by the behavior. As you ssill see later, there is much experimental csidence for many parts of this conceptualization, but there are still many gaps in our knowledge and it is not clear that all the details of this general mecha- nism appls tn .ill cases of motivated behasior. For ex- ample, we cannot be certain esen at present that the major focus of the neural system controlling motiva- tion actually is in the hypothalamus (69). Esen in cases where the hypothalamus is implicated our in- formation is still incomplete. There seems to be no humoral factor in the motivation to avoid noxious stimuli; no inhibitory mechanism in the diencephalon has set been discovered for thirst, although one has been shown for hunger; very little is known about cortical mechanisms in hunger, although a great deal has been worked out relative to sexual behasior, maternal behasior and emotions. Furthermore, there is a real question of boss- much of this theoretical mechanism mas be operatise in learned motivation, and in certain complex instances of human social and personal motives. As Beach has shown in the case of sexual motivation, it is clear that the rclatise contribu- tions of the factors controlling motivated behasior change in phylogeny. Obviously, the specific physio- logical mechanism for each kind of motisation, and for each species, svill be unique in some way; but the point here is that all motisation should ha\-e the same general, multifactor physiological mechanisms at its basis. Before going into the experimental evidence sum- marizing our knowledge of the physiology of motisa- tion, it will be helpful to digress in order to specifs more precisely sshat sve mean by motisation in be- 1508 II WllllllIlK OK PHYSIOLOGY NEUROPHYSIOLOGY III havioral terms and how we may measure motivation experimentally. !'.!■ HAVIORAL DEFINITION OF MOTIVATION In order to approach this difficult task of analysis and definition of motivated behavior, it will he help- ful to start with two simple examples of motivation. This first is an instance cited by Lashley (Hfi) of the motivation to ingest Hydra seen in the round worm, Microstoma. This worm uses stinging cells to capture prey and in its defense, but it must acquire these cells from Hydra. When Microstoma has lost its stinging cells, it ingests Hxdra voraciously. As stinging cells are accumulated, they are evenly distributed around the surface of the body, and triggering mechanisms are grown out to them. A point is reached, however, when Microstoma has a full complement of stinging cells and it no longer ingests Hydra, even in the com- plete absence of all food sources. Illustrated here are a number of basic concepts of motivation. The first is the concept of 'physiological need,' represented in this case by a lack of stinging cells. Second is 'drive1 shown by the increased activity of Microstoma in the face of its deficit. Third, there is 'goal' and 'goal-directed activity' represented, re- spectively, by Hydra and by the specific selective ac- tivitv M/tttntoma shows in the approach to and inges- tion of Hydra. Fourth, is "satiation' shown by the failure of Hydra to elicit further specific goal-directed activity, and the reduction of general activity in Microstoma, or its quiescence, once it has corrected its defii it of stinging cells. I he same concepts derive from a second example, a case of s.ih hunger in a three-year-old child unable to retain -..h because adrenal tumors caused an in- sufficiency of adrenal cortical hormones (169). Here the physiological need is the salt deficit; this child was restless and agitated, especially during feeding, in- dicating an undirected drive. The goal of salt became apparent in the child's marked preference for bacon and sod. i crackers from which he licked the salt. I he goal-directed behavior became eve ire obvious when the child accidentally discovered the salt shaker ind ate sail by the teaspoonful. He then learned to ask for salt in his prelanguage years by si reaming and nting to the cupboard where -ah was kept; it is perhaps significant, moreover, thai his firs! word was 'salt.' hollow iny bonis of salt ingestion, the child showed satiation; he was no longer interested in the -.ill shaker, he no longer preferred salt) foods, and hi- general appetite improved. Of course, as salt was lost from the body, the need returned and the whole cycle repeated itself. In this way the boy managed to keep himself alive until he was brought to a hospital for observation and unfortunately was placed on a normal salt diet which kept him alive for only 7 days. These two examples are particularly interesting because they show how the same kind of behavioral analysis applies over a wide range of the phvlogenetic scale, up to and including man. Furthermore, in con- trast to the first, the second example illustrates quite clearly the role of learning in the organization of motivated behavior, particularly in the selection of goals and in the use of specific responses, among them language, as instrumentalities in the development of goal-directed behavior. Not all instances of motivated behavior fit this pattern completely nor is it a simple matter to go from this behavioral analysis to experimental investigation of underlying physiological mechanisms. But it is possible to make some headway in this direction and, at the same time, to evaluate these behavioral con- cepts in the light of physiological data and theories. Need The concept of physiological need was a valuable one in the past, for it gave the study of motivation biological roots. Classically, it has been specified as some physiological deficit or imbalance in the in- ternal environment that, in the extreme at least, would endanger the life of the organism 03°)- Unfortu- nately, however, the term has been badly abused and overworked. First of all, need has lost some of its bio- logical significance for it has been inferred by some workers in almost every kind of motivation, without any physiological reference (111). Moreover, as we have already pointed out, not all physiological needs lead to motivation and not all motivation derives from needs or deficits. From a physiological point of view, furthermore, specification of an internal en- vironment factor in the control of motivation is far more inclusive than a concept of need. And finally, as you will see in a moment, the concept of drive can easily be substituted for need at the behavioral level. Thus, il appears that the historical concept of need is superfluous, and perhaps misleading, in the analysis of motivation and should be dropped from use. Drivt Drive is a purely behavioral concept 1l1.1t refers to the inteiisiiv of motivated behavior, regardless of how DRIVE AND MOTIVATION '5°9 it is measured, whether it is the amount of general activity an animal shows, the amount of ingestion, the amount of work done to overcome a barrier to a goal, or the frequency, speed or magnitude of a response instrumental in reaching or avoiding a goal. Although drive has been used in the past to refer to animistic forces which propel or energize the motivated animal, its value in modern usage is that it can be specified quantitatively and operationally by a variety of ex- perimental measures which we shall discuss below. It should also be mentioned at this point that drives may be learned in the sense that previously neutral stimuli may elicit an increase in any of the measures of drive after appropriate experience. Thus, a rat will learn to work hard to escape a weak visual or auditory stimulus that has previously been followed by electric shock (97). Or a chimpanzee will work for a poker chip that can later be used to obtain food (170). From a physiological point of view, drive, whatever its origin, must be equivalent to the degree of activation or arousal of the excitatory neural mechanism operating in motivation. While there is some evidence supporting this assumption, much experimental work is needed before the mechanism of drive can be speci- fied in any detail. Goal The goal of motivated behavior is not always easy to specify. In a simple case, it may be some specific ob- ject which the organism approaches or avoids selec- tively and acts upon with a particular pattern of be- havior, like a mate or a noxious stimulus. In some cases, however, there may be no specific identifiable object as in sleep, a burst of running activity or ex- ploration; and even where there is a goal object, part of the goal, at least, may be simply the stimulation arising from the execution of a pattern of response. Experimentally, a goal is identified when, following attainment of the goal object or execution of the goal response, there is a reduction of drive and eventually satiation. In addition, where specific goal objects are involved, the goal is also indicated by the organism's selective orientation to it, cither approach or avoid- ance. We will discuss both satiation and goal-directed behavior in a moment. At this point, it is worth noting that once a goal is reached or an animal has had pre- vious experience with it, it may function to increase drive, depending upon the nature and pattern of stim- ulation the goal provides. For example, a 10 per cent glucose solution will elicit more consummatory be- havior and more work, even when ingestion is rela- tively negligible, than a 5 per cent glucose solution (64, 95). Behaviorally, then, we may speak of the 'incentive-value' of the goal; physiologically, this is presumably correlated with its relative contribution to the arousal of the excitatory neural mechanism in- volved in motivation. A suitable goal may be used to promote learning or maintain performance in a motivated animal, as in the case of the animal learning to run a maze to reach food. In this instance, we refer to the goal as a 're- ward.' Rewards and 'punishment' in the case of noxious stimuli are also called positive and negative 'reinforcements' to signify their effects in strengthen- ing learning and performance when appropriately ap- plied in an experiment. Furthermore, as you might expect, goals may be learned, as in the case where the poker chip becomes (he goal of the chimpanzee's efforts. Then we speak of acquired goals' and second- ary rewards and secondary reinforcement in contra- distinction to primary rewards and primary rein- forccmenl which require little or no learning to be effective. We know relatively little about the physiol- ogy ol learning but, in eases like these, presumably the central neural mechanism in motivation plays an important role. Later in this chapter, we shall dis- cuss briefly some direct experimental approaches to the physiology of reinforcement. Goal-Directed Be ha mot This is the specific pattern of behavior, associated with .1 particular kind of motivation and elicited by specific environmental stimuli, from an objective point of view, this is the appetitive behavior of the students of instinct; the appetite of the physiologist and psychologist; the pleasure seeking and pain avoidance of the philosopher and the psychologist, the preference, choice, the approach, avoidance, the selective behavior of the experimental psychologist. Orientation to a goal involves two major aspects. One is the selective perception of the goal. In the motivated state, the organism is particularly sensitive to certain facets of the environment, as in the response of the maternal organism to its young or the male rat to the female in heat. The other is that, in the moti- \ .1 ted state, certain response patterns are facilitated, in the extreme case of very strong motivation, to the point where the goal-directed response may be emitted 'spontaneously' or at least upon minimal stimulation, the so-called "vacuum-reactions' of Tinbergen and Lorenz. 1510 II Willi!" >K (IF PHYSIOLOGY XF.l'ROPHYSIOLOGY III As we have already pointed out, the same coal m.i\ be reached by different behavioral means at different times. This variability in goal-directed be- havior is ureatlv increased l>y learning new instru- mentalities to attain goals and in learning new goals. Thus, a rat may learn to escape or avoid electric shock by depressing a lever, turning a wheel, standing on its hind legs, leaping over a hurdle, running through a white rather than a black door (146). Or it may be taught one arbitrary response to obtain food and another to obtain water (88), showing not only specific instrumentalities appropriate to each motivation but also a precise discrimination of its own motivational states. In addition to such selective learning of new instrumentalities, the animal, of course, may also learn new goals, thereby expressing its motivation in new ways, not involving consumma- tory behavior or other natural patterns of expression. Satiation Satiation or satiety refers to the reduction ol moti- vated behavior following the achievement of a goal. Subjectively, this is relief from pain, satisfaction or pleasure; or it may be loss of interest or indifference. Operationally, it is the reduction of the organism's drive, its general activity and restlessness, and its specific goal-directed behavior. Since an animal is typically under the influence of many motivations at once, reduction in the expression ol one motivation ma) be due, in part, to the interference resulting from the expression of other motivations. The thirsty ani- mal, for example, drinks steadily until it is presumabl) tun ed to rest in response to fatigue; after a brief rest, il returns to drink some more, and then rests an, tin. \ the tendency to drink reduces even more, the ani- mal ma\ turn to food and eat instead of drink. Then it may drink In iefl\ again, then groom itself. Eventu- ally, ii ma\ sleep, presumabl) a response to still another motivation (152). So satiation must be thought of not only in term- ol direct reduction of drive and goal-oriented response, but also in terms of ill- 1 ompetition oi motives. Physiologically, we are no) sure whal satiation is. \ ugge ted earlier, it can be thought of most simpl) .1- .in mi re.ise in the activit) oi inhibitory mechanisms in the diencephalon which lead direcd) or indirecd) to a redui nun in the activity of the central excitatory 11 hi nanism 01 1 1 ounteraction <>l iis effects. Presum- ably, correction of .1 deficit in the internal environ- ment ma) lead in atiation; some evidence suggests that menu the execution ol a pattern ol motivated behavior feeds back .stimulation which contributes to satiation, fatigue and sensory adaptation may also enter as factors, and learning may possibly enter. The problem of what factors contribute to satiation is something on which we have some evidence, but it is still a major experimental question and we have yet to work out the details of the ncurophysiological mechanism involved. Motivated Behavior In terms of this analysis, then, motivated behavior includes three major behavioral processes: the arousal of drive, goal-directed activity and satiation. Physio- logically, these are the problems of: a) activation of an excitatory mechanism by internal states, and learned and unlearned sensory influences; b) the basis of the change in sensitivity and reactivity of the organism to patterns of sensory stimuli, and the mechanisms responsible for the organization and facilitation of the patterns of response involved in the specific execution of the motivated behavior; and c) the activation of an inhibitory mechanism by internal states, and learned and unlearned sensory influences. Behaviorally, the experimental study of motivation requires: a) the measurement of drive or the intensity of motivated behavior; 6) the analysis and measure- ment of goals, their effectiveness and their modifi- ability through learning, and the measurement of selection, choice or preference in the execution of specific goal-oriented behavior and the modification of these specific responses by learning; and < I the specification of the conditions under which satiation will occur and the measurement of its magnitude, especially taking cognizance of the competition among motives. Some of these problems will be taken up in the next section, then we will <_;o on to address the physiological problems. HI II \V1( >R.M Ml \sl RI-.S ( il M< il IV \ I K1X As you can see from the Complexity of main of the concepts arising in the anal) sis of motivated behavior, it is extremel) important to anchor them experi- mentally. b\ use of operational definitions. Io a large degree, this is the problem of the experimental meas- urement of behavior Therefore, at this point, it will be helpful to review the methods that have been used in specif) motivated behavior and to measure its various |,n etS DRIVE AND MOTIVATION General Activity The most general measure of motivation is to re- cord gross bodily activity. This measure is simply a quantitative record of the intensity of behavior and as such serves primarily as an index of drive. While many factors influence bodily activity, and not every change in activity is a change in drive, it is nonethe- less possible under appropriate experimental condi- tions to show that changes in activity do reflect changes in drive. Two somewhat different kinds of general activity have been recorded. The one most often used is gross locomotor activity. This is the measure obtained when a rat runs in a rotating drum or activity wheel (no), or when an animal is tethered to a counting device (84). The second measure is of restless activity, recorded in a tambour-mounted (125) or spring- suspended cage (73), the tilting cage (26), or the cage divided by a photoelectric beam (142), where small as well as large movements may activate the recording system. In addition, there are several other kinds of activitv which can be recorded and which have been seen particularly following certain brain lesions: stereotyped pacing back and forth (133), forced circling (78), forced following of visual stimuli (77, 144) and obstinate progression (7). A number of studies show a systematic relationship between changes in drive and changes in activity. In the female rat, for example, activity increases during estrus, when estrogen levels and sexual re- ceptivity are high, and decreases during diestrus ( 162). Where a rat is deprived of food or water or both, or of certain vitamins, activity increases steadily as dep- rivation proceeds, reaching a peak and then falling off as the animal becomes physically impaired (23, 161). This peak may be as much as five times basal activity and is reached alter 5 days of food or water deprivation, 2 days of both food and water depriva- tion, and 10 days of vitamin deprivation. That the factor of physical impairment may be specific to the locomotor activity measure is suggested by the finding that gonadectomy reduces running activitv by go per cent (72), while it results in only a 10 per cent decline in restless activity (73). In addition to physical impairment, there are other variables, not necessarily related to drive, which also influence activitv. In a number of different experi- ments, it has been found that: a) activitv is related to daily light-dark cycles, the rat, for example, dis- playing most of its activitv at night (37); b) activitv is high in low environmental temperatures and low in high temperatures (38, 39); and c) it is possible to breed selectively for activity (134). Accordingly, while the association between activitv and drive is an empirical fact, successful use of gross bodily activity as an index of drive depends upon care- ful control of conditions of measurement. This is particularly true because learning and other sources of individual differences make the activity measures very variable. For example, it may take some animals from 10 to 25 days of constant opportunity to run in wheels before they run at all, and it may be over a month before animals kept on a restricted feeding schedule reach a stable base line of daily running (124). Furthermore, these individual differences may be quite large in absolute magnitude, some rats running over 20,000 revolutions a day (over 10 miles) while others may never exceed 200 revolutions a day. Despite all these shortcomings, we may conclude that gross bodily activitv is a function of the amount of drive and that, under appropriately controlled con- ditions, increases in drive arc reflected in increases in activitv up to the point where the animal is physically impaired. Consummaiory Behavior The most commonly used measure of motivation in physiological studies is the goal-directed consum- mately response, idealized in the examples of lood ingestion and water intake. In these two instances, the experimental procedure is to measure the amount consumed in some standard test situation where the animal is, ideally, but not necessarily actually moti- vated in just one way (e.g. deprived of water) and is limited in its choice of responses to one goal (watei In addition, the animal is usually given the oppor- tunity to habituate to the test situation cmotionallv and to adapt to the physical and temporal circum- stances of testing. Under such conditions, it has been shown that the amount of water ingested is a function of the amount of deprivation. Rats, for example, will drink more and more water following longer and longer deprivations up to the 8th day of deprivation when they will succumb (152). No in-between point of physical impairment where water intake falls off alter reaching a peak has been reported as in the case of other measures of motivation, like running activity or performance in the obstruction box (see below I. Detailed analysis of the course of water intake gives further insight into thirst motivation and shows the complexity of even simple consummatory response 1 5 1 -■ II \M)Hi 11 iK (ll I'HYSKH (l(, Y NF.l'ROI'HYSIOl.OGY III JS JO 2S 20 15 10 — 5 - RAT N 3 1 It J 8 HOURS / r J tu WATER INT 48 HOURS j 24 HOURS f— / 6 HOURS f Bi TIME IN Ml NUTES J 1 20 40 60 80 100 120 in. 2. Continuous records of the drinking behavior of one rat after different amounts of water deprivation, showing how this type of consummately behavior tends to be an all-or- nothing response with alternations of constant rates of drinking and resting. From Stellar & Hill I 152).] measures ( 1 -,_> ). Rats lap water .11 I he a ins! a nt rate of 6 to 7 laps per sec, netting 0.004 to 0.005 cc Per 'aP- The onl) thing that changes during drinking, and as 1 function of deprivation, is the duration of bursts of lapping and the length of pauses. At the outset of drinking, lapping is steady lor .1 maximum of about 8 min. , then it is interrupted by a brief pause ( fig. 2 1. As lime '_M"v 1 in, the liursis cil steady drinking decrease in length and llie pauses increase. Observation shows thai the pauses are due to the interference of other motivational tendencies with drinking: fatigue, grooming, exploration, sleep. The weaker the thirst, the mure readilv these other tendencies interlere and, therefore, the sooner and more frequent the pauses and the "i eatei their length. The magnitude of the consummatory response is determined l>\ the nature of the goal as well as the amount oi deprivation. This is illustrated most clearl) in the case of specihe hunger tin sodium chloride. Here, the 1 lunt ol salt solution ingested .^ well as iln absolute amount of sodium chloride is a function ul tin 1 on< entration of the solution (see fig, ) 1 as well .is the animal's salt deficit (55). The concept of consummatory behavior changes considerably where the motivated behavior does not involve ingestion. In sexual and maternal behavior, for example, the consummatory response is often specified as an "innate reflex pattern,' such as the execution of the copulatory response or nest building, retrieving and care of the young, for these are bio- logically useful goal-directed patterns that presumably lead to drive reduction or satiation. In the sexual be- havior of the male rat, the consummatory response is measured in terms of the latency and frequency of mounting, pelvic thrusts, intromissions and ejacula- tions. The strength of these responses is determined by the concentration of sex hormones, the adequacy of the stimulus animal and previous experience (14, 16). With the female rat, the acceptance of the male, the degree of completeness of the pattern of lordosis, rump elevation, tail deflection and ear wiggling, as well as the latency and frequency of these responses to the male serve to indicate the strength of consum- mation (14, 16). In this case, hormonal concentra- tion is critical. In maternal behavior, the frequency and persistence of retrieving, the promptness in moving a nest from a blast of cold or hot air, the com- pleteness of nest building, etc., all are measures of tin- strength of drive or consummation. The strength of these responses is determined primarily by the stim- lus provided by the young and, to some extent per- haps, by hormones (18, 87). Specification of the consummatory response be- comes still different in cases like the avoidance of noxious stimuli, sleep, exploration, fighting, etc w here a specific response pattern is often more difficult to de- scribe. In fact, there is some question .is lo whether or not it is possible to use the term consummatory be- havior meaningfully where there is no ingestion. The argument in favor of extending the concept is essen- tially that Consummatory responses are the 'natural,' adaptive, goal-directed responses of the animal, not Critically dependent upon learning, although modifiable by experience. However, as Miller (99) has pointed mil, oilier measures of motivation do not always give the same results .is the consilium. not v incisures and, therefore, we must be careful not to base (00 111. m v of our conclusions about the phv siolog) of motivation on consilium. Hon incisures alone. Quite clearly, the different incisures get at different facets "f motivation and probably will reveal differ- ent aspects of the underlying physiology. Ibis does not detract, however, from the great value of the consilium. ituiv behavior incisure DRIVE AND MOTIVATION [5'3 Obstruction Method Because even the simplest expression of motivation involves work, fatigue and other competing motives, it is possible to look at the consummatory response and other measures of motivated behavior as the strength of the tendency to overcome some inter- ference in reaching a goal. This principle is put di- rectly to work in the obstruction method. Here the motivated animal is required to cross some barrier to reach its goal. In the case of the rat, the barrier might be a physical obstruction, such as a tunnel filled with sand, a doorway blocked with many thicknesses of paper or an electrified grid which the animal must traverse. In the first two cases, the animal must work to reach its goal; in the last, it must take punishment. In all of these methods, the frequency with which the barrier is crossed in unit time, the speed of response and the size of the barrier necessary to inhibit goal- directed behavior completely are all measures of the strength of drive. The obstruction box with an electrified grid was developed by the Columbia workers under Warden (164) and used as a standardized test for (he measure- ment of many kinds of motivation in the rat. The pro- cedure was to use rats 185 days old, presumably sati- ated in all drives except the one being tested. First, the animals were familiarized with the goal object and a standard electric shock; then they were placed in the starting chamber, and the number of ap- proaches, contacts and crossings of the grid were re- corded in a 20-min. period. After each crossing, the animal was allowed brief exposure to the goal object and was returned to the starting chamber before any significant amount of consummation. In different experiments, the goal objects were food, water, a receptive female, a male, another rat of the same sex, a litter of infant rats, a new area to be explored or, as a control, an empty, familiar goal box. Thus, it was possible to measure the relative strength of different drives against the standard of the electrified grid as a function of such things as amount of deprivation, de- gree of deficit or hormonal change in the internal en- vironment, and the type of goal object. A summary of the major results obtained with this method is shown in figure 3. In terms of the number of crossings of the grid in a 20-min. period, the rank order of drives from the strongest to the weakest is: maternal, thirst, hunger, sex and exploratory. This ordering may only be taken as suggestive, however, 1 2 3 4 5 6 7 8 "28 DAYS OF DEPRIVATION fig. 3. A summary of obstruction-box data from the rat showing the mean number of crossings of an electrified grid in a 20-min. period to reach different goal-objects. The abscissa shows the periods of deprivation of food and water for both sexes and of sexual experience for male rats. Points at the extreme right show the mem number of crossings made bj estrous and diestrous female rats and by multiparous females as well as by rats of both sexes when the t;oal object was a new compartment that could be explored. [From Warden ul>|> since there was considerable variability in the scores of individual animals, and there is some question as id whether .ill the conditions of lestiiiL; were truly com- parable. In addition to this rank ordering, several other findings are worth) of note. In sexual behavior of the emale rat, there is a clear association between the frequency of grid crossing and the estrous cycle, the maximum occurring during estrus and the minimum in diestrus, just as in the case of the activity measure. In contrast, the male shows a constant daily rate of crossing, even after varying amounts of sexual dep- rivation from 12 hr. to 28 days. In the cases of hungei and thirst, on the other hand, as deprivation proceeds, there is a surprisingly early point when the amount of drive decreases. After 1 to 2 days of thirst and 2 to 4 days of hunger, grid crossing decreases in frequency. Perhaps this decrease actually represents a true meas- ure of one aspect of motivation, different from that in the activity and ingestion measures. While we have no experimental evidence illuminating these differ- ences, the obstruction method, nevertheless, is a valuable standard test for the measurement of motiva- tion. i -, i 4 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III , Prejerenu and Competition Among Drives As pointed out earlier in connection with thirst, much of the strength of motivation is determined by the competition anions; different drives and the choice among alternative goals. As animals become partly satiated in one motivation, other motivations become prepotent and contribute, by competitive interference, to the decrement of the measured drive. These rela- tionships are seen most clearly in situations where two drives arc pitted against each other. The obstruction box, of course, is a commonly used method for the stuck of competition between hunger, thirst, etc., and pain avoidance. Similarly, experiments have been done where animals are shocked upon eating in order to determine the amount of shock necessary to in- hibit eating (98). Or, in other cases, food or water have been adulterated by quinine or some other sub- stance with aversive properties to determine the interference necessary to reduce ingestion or abolish it completely (100). A few attempts have been made to develop tests to compare the strength of different drives by the direct choice method. In one test, the rat is given the choice between going in one direction to a food goal and another direction to a sexual part- ner (1 mi. In another test, it is determined how much shock lias to I"' delivered to a rat to make it leave a grid and enter cold water, or how cold water has to be to make the rat enter the electrified grid (1 to). Several different kinds of methods have been used to study specific hungers and food preferences. 1 he i l.issical method developed by Richtcr (129) allows the animal to drink from two bottles, one containing water and the other some test solution. The rats have access to the bottles for 24 hr. with food ad libitum, and then, to evaluate position habits and solution preference, the hollies are switched from one side to thr other. Thus, an average of 2 days of ingestion gives the preference for the substance tested. For ex- ample, the normal rat may take roughly 15 cc from each bottle if both contain water, but it will take 80 to 100 cc from one containing 10 per cent glucose and practically nothing from the one containing watei Preference may also be measured by the single- stimulus method (jo, 166) in which .1 different solu- tion is given for .1 i -hr. period on each day of testing after 1 5 hr. of water deprivation. Food is not avail- able during the test but is given later in the day in .1 j in maintenance period along with plain water. This method has the disadvantage ol introducing .1 constant factor of thirst in the study of specific hun- but it has two great .idv antages : a) it makes con- tinuous records at the time-course of ingestion quite feasible, and b) it makes it possible to test the effects of short-term physiological variables on preference, like the effects of drugs, intubation, etc. In the single- stimulus method, the measure of preference is simply the relative amount ingested, as opposed to the two- bottle method where choice as well as ingestion make up the preference. Choice is even more important in the two-bottle test where two test substances are offered simultane- ously (154). In this test, animals may drink almost exclusively solutions which they can ingest in only small amounts (30 per cent glucose) in preference to solutions they can ingest in large amounts (10 per cent glucose). Thus, preference indicated by choice and by ingestion do not always agree. The differences among the four methods may be illustrated by the rat's preference for sugar. In Rich- ter's two-bottle method, 10 per cent glucose is the most preferred, for that is ingested in the largest quantity over water (131). In the one-bottle test, where thirst is a factor, 5 per cent glucose is the most preferred (95). Where two sugars are presented simultaneously, 30 per cent glucose is taken in prefer- ence to all other solutions (154). Finally, in Young's situation, where ingestion is not a factor at all, the peak preference is for the most hypertonic solutions (173). Obviously, in assessing food preferences it is necessary to make the distinction between how much an animal can ingest and what it chooses. Learning and Learned Performance While it might be expected that the rate .11 which an animal learns a new response should be .1 function of its motivation, it actually turns out that rate of learning is a poor measure of motivation. The rate at which an animal reaches the asymptote of learning is roughly the same over .1 wide range of strengths of drive and over a wide range of incentives (171 ). The performance of .1 learned response, however, is greatly affected by motivation in that the latency of the new response, the number of errors an animal makes, etc., at the asymptote of learning are higher with weak motivation and poor incentives than with Strong motivation and highly effective incentives. Thus, it has been shown in a number of studies that performance is affected I mt rale of learning unchanged with variations in degree of hunger (49), the kind of reward (54), the amount of reward (174) and the number of items of reward with amount held con- stant (45) DRIVE AND MOTIVATION l5l The value of the learning technique in the studv of motivation, then, lies primarily in the fact that it of- fers a large variety of ways of measuring performance as a function of motivation where investigations may be made of such variables as the amount of work re- quired, the amount of ingestion or consummation, the nature of the punishment or reward used as in- centive or reinforcement, and the strength of drive varied by deprivation, hormonal changes, etc. Thus, an animal learns to press a lever to get food and its rate of pressing or performance, once it has learned, is a function of its hunger, what kind of reinforcement or reward it gets from pressing, and the amount of work required for each reinforcement (143). For example, the marmoset will perform at a high rate of responding when 10 per cent glucose is given as a reinforcement for every lever-pressing and will hardly work at all for a 40 per cent reward. On the other hand, when 80 lever-pressings are required for each reinforcement, the animal will not work for 10 per cent glucose but will give the highest rate of responding for 40 per cent (unpublished observa- tions). Similar measures of motivation may be ob- tained from the latency, strength and frequency oi response in such learned performances as running down a runway to reach a goal box containing an incentive, lifting the lid of a food cup, jumping from a platform through a window to a goal platform, turning a wheel to get food or getting off a grid that is electrified or can be electrified, etc. Summai v It is possible, then, to define many of the concepts in motivation operationally and to measure drive, goal-directed behavior and satiation quantitatively. Gross bodily activity yields a measure of undirected drive. Directed drive is measured by strength of con- summatory behavior, preference or choice and learned performance (including the obstruction method) as shown by the latency, frequency, speed and magni- tude of response. Where goal-directed behavior is involved, it is possible, furthermore, to assess the in- centive value of goals, their role in orienting and directing the execution of motivated behavior and their function as rewards and punishments. In all of these cases, satiation may be measured as drive re- duction and the competitive interference of other motivation with the expression of the motive being measured. Drive can be produced by deprivation, deficit, hormonal changes and certain other fluctuations in the internal environment as well as by goal stimuli functioning as incentives. The greater the drive, the greater the activity, the greater the consummatory response, the greater the energy expended and the work done, the greater the barrier that will be over- come to reach a goal, the greater the preference, and the greater the level of learning reached. Increasing drive beyond a certain point, however, may result in a decrement in motivated behavior, in some cases, presumably as a result of physical impairment of the animal. But whether the decrement shows up or not, and at what intensity of drive, will depend upon the measure of motivation used. THE NEUROPHYSIOLOGY OF MOTIVATION In reviewing the literature on the physiological mechanisms underlying motivated behavior, we shall use as a framework the general multifactor mechanism schematically described earlier in this chapter. While this approach loses some of the ad- vantage of discussing each kind of motivation sepa- rately, it allows us in focus directly upon the physio- logical problem, to make useful comparisons among the various kinds of motivated behavior and to point up sharply the gaps in our knowledge of physiological mechanisms. [For a separate discussion of the major kinds of biological motivation sec Morgan & Stellar (109).] To be sure, we shall eventually learn that each kind of motivation lias its own, unique physiological mechanism, but at this stage of knowledge, the simi- larities brought out by a unified approach are more important than the differences, for what we can learn about one kind of motivation can be most helpful in our understanding of other kinds. Diencephalic Mechanisms The evidence to date strongly points to a major focus of the physiological mechanism in control of motivated behavior in the dicncephalon, primarily the hypothalamus. (This is considered in Chapter XXXYII of this work by Ingram.) Physiological studies of this region of the brain implicate it as an important integrating mechanism in the control of autonomic and somatic adjustments of the kind that are important in motivation. For example, Hess (70; cf. 61), in his extensive stimulations of the diencepha- lon of waking cats with chronically implanted elec- trodes, was able to produce: changes in arterial pressure, respiration and pupil size; salivation, vomit- i5i6 IIWDBOOK OF l'IIVSIiilii[;V NEUROPHYSIOLOGY III ing, micturition and a normal pattern of defecation, including crouching and covering; a normal sequence of responses leading to and including sleep; rage and defensive reactions, flight and changes in feeding. While some of these responses may he produced upon stimulation of other parts of the central nervous system, the importance of the dicncephalon is striking. While Hess has emphasized the view that func- tional areas of the dicncephalon overlap and thus are poorly localized, the investigations of others offer evidence for much more discrete localization of function. Two general conclusions seem clear from these studies, a) There are discrete regions within the hypothalamus where marked changes can be experi- mentally produced in different types of motivated behavior, b) Some of these regions are excitatory in that they yield increases in motivated behavior upon stimulation and decreases in motivated behavior upon ablation; others, by the same criteria, are in- hibitory. Both excitatory and inhibitory regions are found in studies of hunger and sleep. In hunger, bilateral lesions in the vicinity of the ventromedial nucleus of the hypothalamus produce marked hyperphagia, a doubling or a tripling of food intake (30); similar lesions more lateral in the lateral hypothalamus re- sult in starvation (2, 157). Confirmation of the con- clusion that the medial region is inhibitory and the lateral excitatory is brought by the fact that stimula- tion of the medial region through chronically im- planted electrodes in the rat markedly depresses food intake whereas lateral stimulation elevates it sig- nificantly (85, 145). (See Brobeck's presentation of this matter in Chapter XIA'Il of this work.) In the case of sleep, discrete bilateral lesions in the pos- lerior hypothalamus in the region of the mammillary bodies cause somnolence (112, 121) while anterior hypothalamic lesions in the preoptic region yield persistent wakefulness (112). Thus, there is a pos- terior excitatory mechanism and an anterior in- hibitory mechanism for wakefulness. Stimulation studies so far have revealed successful induction of sleep only when electrodes are located in the massa intermedia of the thalamus (70), suggesting an addi- tional inhibitory area or another route for reaching the inhibitor) mechanism physiologic-lily.- (This sub- • It is interesting to note, however, tli.it stimulation of the medial reticulai formation in tli<- lower medulla of the 1 .ii will elicit the postural adjustments <>l sleep without any sign 1li.1i i is ,K tually asleep I 1 171 Thus il sccrns possible to eparati thi responsi mechanism ft the mechanism me- diating the arousal and satiation of motivated behavioi in ex- pei imental prot edures of this 1 n 1 ject is also discussed by Lindslcy in Chapter LXIY of this Handhook.) Several additional points of interest have come from further studies of these excitatory and inhibitory mechanisms, a) In both sleep and hunger, the at- tempt has been made to ablate the excitatory and inhibitory regions in one preparation (2, 112); the result in each case was that the symptoms were the same as those produced by destruction of the excita- tory mechanism alone, namely somnolence and starvation. These findings have suggested the possi- bility that the main influence of the inhibitory area is upon the excitatory mechanism, but they do not rule out the possibility that both the excitatory and in- hibitory mechanisms function through a common structure below the dicncephalon. h) The hyper- phagia produced by ventromedial hypothalamic lesions is permanent whereas there is recovery from the starvation produced by lateral lesions if the ani- mals are maintained for several weeks by forced feeding (157). In the case of somnolence following posterior hypothalamic lesions, there is also recovery, with some capacity to maintain wakefulness return- ing gradually over a period of several weeks (121). These findings suggest additional mechanisms for maintaining wakefulness and feeding behavior. So far only an excitatory mechanism has been re- vealed in the study of thirst. Stimulating the dicn- cephalon of the goat electrically or with minute quantities of hypertonic saline directly injected through an implanted pipette will promptly produce great drinking in water-satiated goats (3, 4, 5). The critical area is dorsal to the infundibulum between the fornix and the mammillothalamic tract, just lateral to the paraventricular nucleus, midway between the dorsal and ventral hypothalamus. Similar results have been reported in two investigations using the 1. u Hi 3, in-,). In addition, ablation of the same area in the dog produced adipsia from which there was recovery after 14 da\ s (hi. In the case of sexual behavior, bilateral lesions in the ventral portion of the anterior hypothalamus be- tween the optic chiasm and the st.dk of the pituitary will abolish sexual behavior in the female guinea pig, even following the injection of gonadal hormones , 52). Similar data have been reported on the male guinea pig (34) and on rats ol both sexes (42). Furthermore, upon stimulation of the lateral pre- optic area in male and female rats l>\ injection of minute amounts nl appropriate sex hormones through .111 implanted pipette, long-lasting sexu.il beh.i\ioi w.is elicited (56). Thus, there is evidence for an ex- citatory mechanism fot sexual behavior in the hypo- DRIVE AND MOTIVATION '01 / thalamus. (See also Sawyer's account of this topic in Chapter XLIX of this Handbook.) The role of the hypothalamus has been less exten- sively investigated in other kinds of motivation, but the findings available are quite suggestive. Increased emotionality in cats has been demonstrated following lesions in the vicinity of the ventromedial nucleus of the hypothalamus (167), suggesting an inhibitory mechanism. Posterior hypothalamic lesions result in reduced emotionality and placidity (121) and stimu- lation of these posterior regions produces many of the signs of rage (70, 123), suggesting an excitatory mechanism. Strong and persistent maternal behavior has been elicited in both male and female rats by injection of male and female sex hormones respec- tively into the medial preoptic area (56). Reductions of gross bodily activity have been reported following lesions near the ventromedial hypothalamus of the rat, even without concomitant hyperphagia (71 ). Lacking in all of these experiments, unfortunately, is any precise anatomical work specifying precisely what structures must be involved to give these various effects on motivated behavior. It is not clear in either the ablation or the stimulation studies whether ii is cell bodies or fiber tracts or both thai must be de- stroyed or stimulated. The most we can say is that the lesion or the tip of the stimulating electrode or pipette must be in a particular region of the brain to yield significant effects on the particular kind of motivated behavior that is measured. A further criticism of many studies is that only one kind of motivated behavior is measured upon stimu- lation or ablation of a particular locus. The physio- logical studies by Hess show a great deal of overlap- ping and intermingling of points as far as the kinds of effects he found upon stimulation are concerned. It has also been found that lesions in the vicinity of the ventromedial hypothalamus, designed to produce hyperphagia, will also produce hypoactivity in some rats and emotionality in others. But examination for these additional effects has only been casual in most studies, and we know nothing about whether such lesions would also produce noteworthy changes in sexual behavior, pain avoidance, etc. A number of studies give some idea of the multiplicity of functions that can be revealed from experimentation with one locus. For example, lateral hypothalamic lesions eliminate hunger and thirst simultaneously and often produce transient somnolence (157). In his experi- ments, Fisher (56) found that injection of sex hor- mones into the anterior hypothalamus could elicit maternal and sexual behavior simultaneously, and in addition could sometimes produce changes in respira- tion, exploratory behavior, digging, leaping, etc. Thus, only the crudest questions about localization of function within the diencephalon can be answered on the basis of the present data on motivated be- havior. At present, it appears that there are different foci in the diencephalon which can be manipulated experimentally by stimulation and ablation to produce marked changes in different kinds of motivated be- havior. Some of these foci can be characterized as excitatory and others as inhibitory. But much more has to be done experimentally, on the anatomical side and on the behavioral side, before we can con- clude much beyond these two points. Other Central Mechanisms Investigation of other central neural structures out- side the diencephalon has revealed much about their role in motivation. Again, it is possible to characterize the function of many of these regions of the brain as excitatory or inhibitory in terms of the effects of lesions and stimulations on the arousal and satiation of motivated behavior. Perhaps the most thoroughly studied kind of motivation is emotional behavior, although the picture we have at present is not en- tirely clear. (Its present si. mis [s considered in Chap- ter I. XIII by Brady in (his work.) Decortication, as Bard showed, leads to "sham rage,1 suggesting an inhibitory role of the cortex in emotion (8); while Bard & Mounteastle (()) found that ablation of parts of the rhineneephalon, particularly the amyg- dala and transitional cortex of the mid-line, produced great increase in the ferocity of cits. Removing only the neocortex, on the other hand, resulted in extremely placid cats. In other studies on monkeys and cats, rhinencephalic ablations, including parts of the tip of the temporal lobe and the hippocampus in some cases, led to placidity rather than ferocity (81, 135). So far this disagreement is unresolved, but it is clear that certain parts of the cortex may exert an inhibitory effect and others an excitatory effect on emotional behavior. Contributions from other parts of the brain to emotional behavior have been revealed in a number of other studies, a) An inhibitorv role is suggested for the septal area where lesions produce a transient, but greatly exaggerated, emotional response to tactile stimuli in rats (27). b) There is evidence for an excita- tory contribution from the brain-stem reticular system and mid-line thalamic nuclei (76, 91, 92). c) There is an inhibitory contribution from the anterior nuclear complex of the thalamus, ablation of which causes [8 II VXDBOOK OF l-MYMol I k.\ SI I ROI'IIVSIOLOGY III i .ii> id solicit petting and respond to it with heightened 'pleasurable1 reactions (137). d) There have been re- ports of reduced emotionality following cingulectomy 1 ii(|, ijoi and the frontal lobes have been impli- cated, although their contribution is not entirely clear (59). The situation is somewhat clearer in the case of sexual behavior. Here it has been found that neo- cortical lesions lead to a reduction in sexual motiva- tion, the larger the lesion in the male rat the greater the effect without regard to locus (12); females are less affected than males by cortical ablations; and the higher the animal on the mammalian scale, the greater the effect (15, 17). In addition to this excita- tory role of the neocortex, there is evidence for an inhibitor) role of the amygdala (135, 136) and over- lying pyriform cortex (43 1, in male cats at least. After some delay following ablation of these regions, there may be exaggerated sexual responses in male cats and monkeys characterized by strong drive and poor discrimination of sexual objects. (See also Chapter XI.IX on reproductive behavior and Chap- ters LVI, LVII and LVII1 on the limbic system in this Handbook. 1 In the case of sleep, it has been found that decorti- cation in dogs is followed by an inability to postpone sleep and maintain wakefulness for long periods of time with the result that such dogs sleep and wake in short cycles over the 24-hr. period (80). Similar excitatory contributions are revealed in human cases with restricted cortical and thalamic lesions (46-48). Most striking of all, however, is the excita- tory contribution from the brain-stem reticular forma- tion (91, 92). Lesions here are followed by marked somnolence, and central stimulation leads to prompt arousal from sleep. Little is known about the role of central neural structures outside the diencephalon in hunger and thirst, although overeating has sometimes been re- ported as a result of frontal lobe damage (132, 13 ;) In maternal behavior and in food-hoarding, it has been reported thai destruction of mid-line cortex in the ral leads to a reduction in motivation .is well as a disintegration of the organization of these patterns ol behavior (148, 149). Finally, some suggestion about an excitatory role oi the frontal cortex in the motivation to avoid pain (nmes from the psycho mrgii al studies dune for the relief ol intrai table pain In these cases, the patients reporl postopera- tively that while thev still feel the pain, it no longer bothers them .1- b Because <>l the lack "i good anal ical data rele- vant to these problems, it is impossible to say defi- nitelv by what route these structures exert their ex- citatory and inhibitory effects on motivated behavior. I he most common hypothesis is that thev work through the integrating mechanisms of the hvpo- thalamus, for many pathways to this structure have been described. Thus it has been suggested that the contribution of the rhinencephalic cortex to emotional behavior is mediated by the amygdala which in turn functions through the ventromedial nuclei of the hypothalamus (9). While such hypotheses are reason- able in the light of a concept of hierarchical organiza- tion of the central nervous system and are part of the theoretical viewpoint followed here, it will remain for future experimental work to supply the detailed evidence needed for their support. For example, although we may conclude clearly that a part of the cortex has an excitatory or an inhibitory effect on motivated behavior, we do not know whether its physiological role is excitatory or inhibitory, for under the present viewpoint, it may exert its influence on either an excitatory or inhibitory diencephalic mechanism, or have a reciprocal influence on both. Or by alternative anatomical and physiological routes, it may bypass these mechanisms. Sensory Factors From both anatomical and physiological evidence, it is quite clear that the hypothalamus is under ex- tensive sensory control, receiving afferents directly from all the modalities over the specific pathways and, through collaterals, over the nonspecific path- ways ol the reticular system (60, 90, [51). Further- more, on the basis of extensive behavioral evidence, a prhnarv role must be given to sensorv factors in the control of motivated behavior. It has already been pointed out that mammalian sexual motivation mav be destroyed bv peripheral surgical reduction or elimination of two or more sensorv avenues (16). In this case, the sensorv contribution to arousal is non- specific since any one sensorv path can be eliminated without significant effect. In his analvsis ol' sleep, Kleitman (79J draws the similar conclusion that it is the sum total of afferent input that controls wakeful- ness rather than anv specific sensorv svstem, and this notion is inherent in Magoun's conception (9a) of the role of the brain-stem reticular formation. Direct support lor these v iew s coines from Bremer's prepara- tions in which waking EEG patterns survive brain sei linn until there is a sufficient reduction in alter- ent input I 28, -' of the restora- tion, bv sex hormone injection, of sexual motivation losi through cortical lesions is one good example. Another is the reporl ol Brooks (36) that neither de- cortication nor olfactory bulb ablation in die male rabbit eliminates sexual behavior, but a combination of the two does. A third is (he report of Schreiner & Kiing (136) that castration destroys the hypersexu- ality induced in male Cats bv amv gdalectomv . Perhaps one of the better siudies showing the joint contribution of two or more factors to the control of motivated behavior is the work of Teitelbaum (156) who investigated the changes in the sensory control ol eating produced by ventromedial hypothalamic lesions, lie found that hvperphagic rats reject pow- dered laboratory food, adulterated with nonnutritive DRIVE AND MOTIVATION I 52 I 35 20 ?5 i0 15 LOG MOLAR CONCENTRATION NoCL . A 1 1 / \GLUCOSE /• \ FRUCTOSE ^ */ y\ i\ R O- HOH INTAKE \\ -'0 ■ 0 i5 .10 .15 LOG PERCENT CONCENTRATION OF SUGAR IN TEST SOLUTION I I II I III 1 1 I II MM 1 1 I llllll I 0 0001 J 1 I I I Mil I I II I llll I 1 1 MOLAR CONCENTRATION OF SACCHARINE WATER TESTS LOG MOLAR CONCENTRATION OF ALCOHOL fig. 4. Preference-aversion curves lor different foods offered in solution, showing fluid intake as a function of the concentration ■ >! the solution. The sodium chloride, the glucose and fructose, and the saccharin functions are based upon 1 hr. of drinking in the single stimulus method, the alcohol function is based on 24 hr. drinking of alcohol ami watei in the two-bottle method. [The sodium chloride data are from Stellar el al. (153); the sugar data, from MeCleary (95) ; the saccharin and alcohol data, from Stellar, unpublished observations cellulose, at much lower percentages of adulteration than normal rats. This result suggested the importance of changes in the operated animal's reactivity to the sensory stimuli provided by the diet, possibly even in the original overeating of regular laboratory food. This hypothesis was, in part, confirmed by the finding that hyperphagic rats ingest more of the powdered diet when 50 per cent glucose is mixed with it, and less when a low concentration of quinine is added or when the food previously given in the form of pellets is presented in powdered form — neither of which changes influence the intake of normal rats. Thus, whether a lesion of the ventromedial hypothalamus leads to overeating or undereating, relative to normal, is a function of the nature of the sensors stimuli pro- \ ided by the diet. Two additional results are of note here, a) Rejec- tion of cellulose-adulterated food was also observed in rats with similar lesions that did not become hyper- phagic, suggesting the possibility that the refusal to eat certain foods might be dependent upon structures outside of the inhibitory feeding mechanism in the ventromedial hypothalamus, b) Hyperphagic rats that are not allowed to become obese do not respond '522 IIANDHiiiiK HI- l'lIYSlnl.lKiY NEUROPHYSIOLOGY III abnormally to the addition of quinine and glucose to the diet or to the change in texture. But they arc sensitive to cellulose adulteration. It is possible, then, that obesity makes some indirect contribution to the 'finickiness' Teitelbaum reports, although this finding may he partly the result of the "heedless' eating the nonobese hyperphagics do. In similar studies, rats with lateral hypothalamic lesions accept no food postoperatively. But as recovery occurs, mi a regime of forced feeding, it comes about in a very specific way. Alter a week or two of complete starvation, the operated rats will accept chocolate and evaporated milk, but they cannot be aroused to drink water or eat powdered food, glucose, meat, etc. Within a few days to a few weeks later, they will accept water and, last of all, they will eat powdered food. While there are undoubtedly many chemical and nutritive differences among these substances in addi- tion to sensory ones, the suggestion is strong that these rats are changed in their reactions to the stimuli pro- vided by the diet following operation and throughout the course of recovery. The Role nf Learning Our discussion of the physiological mechanisms of motivation would not be complete without considera- tion of the role of learning. Unfortunately, we are hampered in our thinking by our ignorance of the physiology of learning, but there are experiments on motivation, involving learning, which are particu- larly instructive at the behavioral level and give us some valuable ideas about the general nature of the underlying mechanisms. Three different kinds of ex- periments are of interest here: a) studies of the influ- ence of learning in the modification of biologic, illy adaptive motivated behavior, /» ) comparisons of the effects of a number of physiological variables on mo- tivation in which learned performance .is opposed to unlearned eonsummalnrv response is used as the measure of motivation; and r) direct physiological studies <>l reward and punishment, and the reinforce- ment of learning and learned performance. Illustrative of the first kind of stud) is the postop- erative ingestion of sodium chloride bv the adrenal- ectomized rat. Epstein & Stellar (-,-,) have shown thai the rat requires no special postoperative experi- ence to increase ils sodium chloride intake following adrenalectomy. < >n the other hand, Harriman (66) has demonstrated thai rats given experience with glucose and sodium chloride solutions before opera- tion, unlike naive rats, will prefer glucose to sodium chloride postoperatively and will die because they fail to ingest sufficient amounts of salt. In a similar type of experiment. Young (172) has shown that protein- deficient rats will prefer sucrose to protein in a test situation where they had been in the habit of selecting sucrose before any deprivation; if placed in a new test situation, however, where they have no previous habits, they show an immediate preference for protein; if tested in the two situations on alternate trials, they alternate between sugar and protein preference. Similar interfering and facilitating effects of previous habits have been shown in many other studies (67, 138, 139), and they hint at how complicated the mechanism of motivation can become when the op- portunity for learning is introduced. This problem has been solved in most of the experiments we have discussed so far by keeping the possibilities of learning at a minimum by using as a measure of motivation a simple, uncomplicated consummatory response, natural to the animal. But as you will see in a mo- ment, this may be a misleading practice, and our con- clusions about the physiology of motivation should properly be based on measures of learned performance as well as consummatory measures. A number of experiments have been done compar- ing results with the measure of learned performance and the consummatory response. In a striking study of rats with lesions of the ventromedial hypothalamus, Miller et al. (102) were able to show that while hyper- phagic rats ate more than normals in a free-feeding situation, they showed much less hunger motivation than normals when required to perform learned re- sponses, to work or to overcome the taste of quinine to get food. In the face of these contradictory findings, the authors suggest that ventromedial lesions mav not release hunger motivation for easier arousal but simplv interfere with the mechanism for stopping eating. However, it is well to point out that the hyperphagics used in this study were obese and, judging from Tcitelbaum's findings cited above, the results reported here mav be as much a matter of obesity as they are of hy pothalamic lesions. In a series of different experiments bv Miller and his students (22, 82, 101, 103, 104), much better agreement was found between the consuiumatorv measure and the learned-response measure. In these studies, iliev used rats with gasiric fistulas so that they could compare the effects of saline, milk or water introduced directly into the stomach or taken bv mouth. Using both hungrv and thirsty rats, they found that taking fluid bv mouth had greater satiating ellcels whether they used as .1 measure of motivation DRIVE AND MOTIVATION <523 later consummation or performance of a learned re- sponse involving very little consummation. They were also able to show that injection of fluid directly into the stomach could be used as the reward to pro- mote the learning of a simple maze, but that allowing animals to drink milk or water by mouth was an even more effective reward. In another study, modeled after Andersson's work on the goat, this same laboratory was able to confirm the finding that intraventricular injections of minute quantities of hypertonic solutions increased the thirst of cats, and to demonstrate further that injections of water and hypotonic solutions led to decreased thirst. This result turned up with the use of a consummatory response as a measure and also with the use of a learned response in which the animal worked for rela- tively few opportunities to drink (ioi). In similar studies with food, there has been some question as to whether stimulating the medial hypothalamus elec- trically elicits hunger motivation or merely evokes re- flexes of seizing, chewing and swallowing, for stimu- lated animals sometimes seize and chew nonedible objects. That there is an actual increase in motivation upon stimulation of the ventromedial hypothalamus, however, is confirmed by the fact that following stimulation these animals will press a bar many times to get an occasional tiny bit of food, even though they are otherwise satiated (ioi). Obviously, agreement among the various measures of motivation in determining the effects of a physio- logical variable strengthens the conclusions we can draw about the mechanism of motivation. Disagree- ment may mean one of two things, however: that there is interference with something specific to the behavior invoked in the affected measure, not neces- sarily of a motivational nature, or that there are dif- ferent facets of motivation measured by the different tests and subserved by somewhat different physio- logical mechanisms. Clearly, it will be profitable to extend the measures of motivation used in physiologi- cal studies to include techniques that are not solely dependent upon consummation. The third and most striking kind of experiment in- volving learning is the use of electrical stimulation of the brain to serve the function of a reinforcement, much like a food reward for a hungry animal, or much in the way that escape or avoidance of an elec- tric shock reinforces a new response. The first study along these lines was by Delgado et al. (50). By im- planting electrodes in the vicinity of the medial lemniscus and posteroventral nucleus of the thalamus, these workers were able to elicit clear negative moti- vation in cats of the sort observed upon administering electric shocks to the feet. They reported four major results, a) Cats learned to rotate a wheel to turn off this central stimulation, b) They learned to respond to an auditory signal by turning the wheel in order to avoid the stimulation that had alwavs followed the signal previously, c) They learned to escape immedi- ately when placed in a compartment in which they had received central stimulation. d) Central stimula- tion administered at the time of feeding inhibited eating for long periods of time, despite strong hunger. Thus by stimulation in the vicinity of the pain path- ways centrally, these workers were able to elicit strong negative motivation, sufficient to reinforce a variety of kinds of learning. On the positive reinforcement side, Olds & Milner (1 14) have reported that rats will work in order to be Stimulated at a number of points within the brain. If an electrode is chronically implanted in the septal area or the mamillothalamic tract and activated every time after the rat presses a lever, the animal will change its rate of pressing the lever from several times an hour to as high as almost 750 times an hour. Self- stimulation of the cingulate gives similar but less marked results, while corpus callosum, caudate and hippocampal stimulation yielded no effect. Stimula- tion of the tegmentum was equivocal, and medial lemniscus and medial geniculate stimulation, if an) thing, produced avoidance of the lever. follow in» this sainc procedure of sell-stimulation reinforcement with rats and cats, Sidman et al. (141) used the technique of administering stimulations only after certain lever pressings. When the schedule of stimulations called for activation of the implanted electrode at irregular intervals of time, averaging id sec, low rates of responding were obtained. When the schedule called for seven pressings for each stimu- lation, the rate of responding was very high. These results were typical of those ol it. lined with such varia- ble-interval and fixed-ratio schedules of reinforcement with food or water as the rewards. Thus, it appears that under appropriate experimental conditions, stimulation of the brain can be rewarding and can reinforce the performance of a learned response in much the same manner as food and water do for the hungry or thirsty animal. In this work, Sidman et al. found stimulation of the septal area most effective for the rat and the caudate for the cat. In a later study on the rat, Olds (113) concluded that the stimulation of the amygdaloid complex and the anterior hypothalamus were as effec- tive as stimulation of the septal area in yielding high 'VI il Wlilli K iK i il I'll1! S .\ NEUROPHYSIOLOGY III rates of responding, while moderate rates were pro- duced upon stimulation of the cingulate, hippocam- pus, posterior hypothalamus and anterior thalamus. The distinction between effective and ineffective points, and rewarding and punishing [joints, however, may not be simple. Miller and his colleagues (mi | have reported that a single point may he positively reinforcing in the sense that an animal will work to have it stimulated, but it may also be negatively re- inforcing, for the animal will escape from it if it has the opportunity. From a purely behavioral point of view, it looks as though direct stimulation of the brain should have the same physiological effect as the various rewarding consummatory responses that eventually lead to satia- tion. This notion is supported by the fact that so many of the effective points for self-stimu'ation reinforce- ment arc in areas of the brain that are known to play .in important role in motivation. However, we lack the specific experimental information as yet to tell whether the reinforcing effects are tantamount to transient drive reductions or satiations of specific motivations through appropriate activation of their specific mechanisms; or whether these reinforcing effects are not specific to any of the mechanisms of motivation, but rather evidence for a general mechanism for 'pleasure' or, put more operationally, reinforcement. Preliminary reports on human patients with elec- trodes implanted in the septal area indicate "feelings of comfort' upon stimulation (68). But it is too early lo speculate beyond these general possibilities, for at present there are too few facts at hand. CllNCI I'SIONS A number of conclusions can be derived from this historical and experimental analysis of the physio- logical basis of motivated behavior. It is possible to siud\ motivated behavior with ob- jei tive methods and to make an operational analysis of the factors important in i t^ control without resort in ieleuliMjie.il and vitalistic concepts. This analysis indicates that various kinds ol moti- vated behavior, like hunger, specific hungers, thirst, se\, maternal behavior, emotion, sleep, etc., are under the s.ime general kind of multifactor control, receiving influences from the various sensor) avenues, the in- lern.il environment and the central nervous system. An often neglected, but nevertheless important factoi is learning, for previously neutral stimuli can, through experience, come to contribute i<> the arousal and satiation of motivated behavior, and various new instrumentalities and goals may be learned in the execution of motivated behavior. The relative contribution of the various factors con- trolling motivation and presumably the underlying physiological mechanisms change in the course of phylogeny. Judging mainly from the example of sexual behavior, there is with ascending phylogenctic posi- tion an increasing dependence upon sensory factors, learning and the cerebral cortex, and a decreasing dependence upon the internal environment. The behavioral analysis of motivated behavior divides it into three main aspects: drive, the intensitv of the arousal and maintenance of motivation; goal- directed behavior, the enhanced perception of selected stimuli in the environment and the selective execution of a pattern of behavior in respect to them; and satia- tion, the reduction of drive once the goal is sufficiently attained. Experimental methods are available to measure drive and satiation quantitatively and to specify, in some cases, the choice and selection involved in goal-directed behavior. The possibilities for the meas- urement of motivation are greatly increased by the learning of new drives, new goals and new instrumen- talities for attaining goals. There are, however, two important limitations in the present methodologies: a) the various measures of motivation are not always in good agreement, and b) there is too much depend- ence at present upon simple measures of consumma- tory response. Physiologically, drive appears to be a function of the activitv ol a general excitatory mechanism having ils major central control in the hypothalamus, and satiation is a function of a similar inhibitory mecha- nism, for increases and decreases in many kinds of motivated behavior can be produced by ablation and stimulation of restricted hypothalamic foci. There are two serious limitations in the conclusions we can draw about localization, however, a) We do not know, with im\ anatomical precision, the Structures in the hypo- thalamus which must be involved to produce these changes in motivation, we do not even know whether lesions and stimulations are effective because they affect nuclei or liber tracts or both. /)) We do not know whether a particular effective locus subserves onlv one kind of motivation or several kinds, although there is much evidence suggesting overlap of function within the hypothalamus. Central neur.il structures outside ol the hvpothala- uuis also contribute cm italoiv and inhibitory inllu- ences to the control of motivation as main ablation DRIVE AND MOTIVATION 'O'D and stimulation studies show. The current hypothesis is that neocortical and rhinencephalic structures exert their influence directly through the hypothalamic integrating mechanism; but we have very little direct evidence on the anatomical and physiological path- ways involved. Analysis of the sensory contribution to the arousal and satiation of motivation shows that it is typically a multisensory matter in which the influence from the various sensory pathways is additive. As yet, however, there is only a small amount of information showing direct sensory contribution to the hypothalamic mechanisms involved in motivation. The important role of the internal environment in motivated behavior is thought to be mediated largely through the hypothalamus. Particularly relevant are the arousal of motivation by direct injection of sub- stances into the hypothalamus and the failure of sys- temically injected hormones to elicit motivation fol- lowing hypothalamic lesions. The complexity of the neurophysiologies mecha- nisms described here is greatly increased by two facts. a) These various influences on the hypothalamic mechanisms interact in the arousal and satiation of motivated behavior in the sense that one kind of influence may add its effects to another, b) The possibilities for the arousal, expression and satiation of motivation are greatly expanded by learning, undoubtedly altering the underlying physiological mechanism in some way as yet unknown. Electrical stimulation of the brain can serve to reinforce learning and learned performance in much the same way that food reinforces the hungry animal. It is too early to speculate about the nature of the mechanism involved here, but it is perhaps significant that many of the reinforcing points within the brain are known to be important parts of the physiological mechanism involved in motivated behavior. Little has been said and little is known about the mechanism for the execution of motivated behavior. The hypothalamus has been thought of as a major integrating mechanism in the expression of motiva- tion, but not much is known about the spinal and brain-stem reflex mechanisms invoked. Some good evidence points to the role of the cortex in the spatial and temporal organization of motivated behavior (151 ), but space has not permitted discussion of this very important question. A second neglected problem Ins to do with per- ceptual changes in motivated behavior. We have dis- cussed only the relatively nonspecific role of afferent systems in the arousal of motivation, but the moti- vated organism is highly specific and selective in its perception of the environment, as many ethological and psychological studies show. Unfortunately, we have very little physiological evidence relevant to this problem. \\ e have made onl) a start on the elucidation of the physiological mechanisms underlying motivated be- havior, but a basic core of knowledge and a general orientation have been established. REFERENCES 1. Adolph, E. F. Am. J. Physiol. 151: 110, 1947. 2. Anand, B. K. and J. R. Brobeck. Yale J. Biol.& Med. 24: i*3. '951- 3. Andersson, B. Acta physio!, tcandinav. 28: 188, 1953. 4. Andersson, B. and S. M. McCann. Acta physiol. scandinav. 35: '9'. '955- 5. Andersson, B. and S. M. McCann. Acta physiol. scandinav. 33:333. '955- 6. Andersson, B. and S. M. McCann. Acta physiol. scandinav. 35:312, 1956. 7. Bailey, P. and E. W. Davis. Proc. Soc. Exper. Biol ^ \l 51: 307, 1942. 8. Bard, P. A. Res. Ner< < \A~cnt. His., Proc. 19: 190, 1939. 9. Bard, P. and V. B. Mountcastle. A. Res \,i. ^ Went Dis., Proc. 27: 362, 1947. 10. Bare, J. K. J. Comp.& Physiol. Psychol. 42: 24 2, 1949. 1 1. Bash, K. W. J. Comp. Psychol. 28: 109, 1939. 12. Beach, F. A. J. Comp. Psychol. 29: 193, 1940. 13. Beach, F. A. Psychosom. Med. 4: 173, 1942. 14. Beach, F. A. Psychol. Bull. 39: 200, 1942. 15. Beach, F. A. Psychosom. Med. 6: 40, 1944. 16. Beach, F. A. Physiol. Rev. 27: 240, 1947. 17. Beai ii. I A Psychol. Rev. 54: 297, 1947. 18. Beach, F. A. In: Handbook oj Experimental Psychology, ed- ited by S. S. Stevens. New York: Wiley, 195 1 , p. 410. 19. Beach, F. A. Psychol. Rei: 62 : 401, 1955. 20. Beebe-Center, J. G., P. Black, A. C Hoffman \\i> M Wviu . ./ ( 'omp. & Physiol. Psychol. 41 : 239, 1948. si. Bellows, R T. Am. .1. Physiol. 125:87, 1939. 22. Berkun, M. M., M. L. Kjessen vmi N I Miller. J. Comp. & Physiol. Psychol. 45: 550, 1952. 23. Bloomfield, A. and M. L. Tainter. J. Lab. <) /' ■. / A 6a 3 13. 1955. 70. Mess, W, R. Dienccphalon: Autonomic and Extrapyramidal /•'unctions. New York: Grune, 11134. 71. Hetiierington, A. \V. and S. \V. Ranson. Am. ./. Physiol. 136:609, 1942. 72. HoSKINS, R G. Am. J. Physiol. 72:324, 1925. 73. Hunt, J. McV. and II SchloSBERO. J. Comp. Psychol. 28: 23. >939- 74. Ingelfinger, F.J. New England J. Med. 231 321, 1944. 75. Janowitz, II D. and M. I. Grossman. Am. ./. Physiol. ■59: '43. >949- 76. Jasper. H. II In Brain Mechanisms and Consciousness, edited by J. F. Delafresnaye. Springfield: Thomas, 1954, P- 374- 77. KENNARD, M. A. A. M. A. Arch. Seurol. & Psyehiat. 48: 227, 1942. 78. K.ENNARD, M A \nd I.. EcTORS. J. Neurophysiol. I: 45, 1938. 79. Kleitman, N. Sleep and Wakefulness. Chicago: L'ni\ Chicago Press, 1939. 80. Kleitman, N. and N. Camille. Am. J. Physiol. 100: 474, '932. 81. Kluver, H. and F". C. Bucy. A. M A. Arch. Neurol. & Psyehiat. 42:979, 1939. 82. Kohn, M. J. Comp. & Physiol. Psychol. 44: 412, 1951. 83. Kuo, Z. Y. Psychol. Per. 31 : 427, 1924. 84. Langworthy, O. R. and C. P. Richter. Am. J. Physiol. 126: 158, 1939. 85. Larsson, S. Acta physiol. scandinav. 32, Suppl. 115:8, 1954. 86. Lashli y, K. S. Psyi hoi. Pec. 45: 445, 1938. 87. LeBlond, C. P. ./. genet. Psychol. 57: 327, 1940. 88. Leeper, R. W. ./. s;cnet. Psychol. 46: 3, 1935. 89. Lehrman, D. S. Quart. Pec. Biol. 28: 337, 1953. 90. Lindsi i.y, D. B. In: Handbook of Experimental Psychology, edited by S. S. Stevens. New York: Wiley, 1951, p. 491. 91. Lindsley, D. B., J. Bowden and H. W. Macoun. Electro- encephalog. & Clin. Neurophysiol. 1 : 475, 1949. (>2. Magoun, H. VV. In: Brain Mechanisms and Consciousness, edited by J. F. Delafresnaye. Springfield: Thomas, 1954, p. 1. 93. Mayer, J. Bull. New England Med. Centei 14: 43, 1952. 94 \1 vyer.J. Ann. New York Acad. Sc. 63: 15, 1955. 95. Mi Cm \rs, K. A. ./. Comp. & Physiol. Psychol. 46: 411, '953- 96. McDougali., W. An Introduction In Social Psychology. London: Methuen, 19118. 97. Miller, N. E. In: Handbook of Experimental Psychology, edited by S S Stevens. New York: Wiley, 1951, |>. 435. 98. Miller, N. E. ./. Personality 20: 82, 1951. iim M11 1 1 k, N. I. Ann. New York Acad. Sc. i>j 141, 1955. 100 M N I Inn \ ... York Acad. Sc. 65: 318, 1956. 101 Miller, N I Sri 126 1271, 1957. 102. Miller, N E., C. J Bailed vnd J. A I Stevi nson. S ience 112: 256, 1950. 103 Miller, N. E. snd M. I.. Kessen J. Comp *s Physiol. Psychol. 45: J55, 1952 104. Miller, N. I.. R. I Sampuner vnd P Woodrow. ./. Comp.& Physiol. Psychol 50 1, it,;. 105. MONTEMURRO, D. G. AND | \ I STEVENSON. Canad. J. Bio, hen PA) 15 ;i, 1957. 106. Montgomery, A V vnd J II Holmes. Am. .1 . Physiol. 182: 227. 1955. I.. Morgan, <■'.. I. Physiological Psychology. New York Mi ( rraw-1 [ill, 194 i. |> 294. DRIVE AND MOTIVATION '527 108. Morgan, C. T. and J. D. Morgan. J. genet. Psychol. 57: ■53. >94°- iog. Morgan, C. T. and E. Stellar. Physiological Psychology. New York: McGraw-Hill, 1950, chap. 16, 17, 18, 19, 20. 1 10. Munn, N. L. Handbook of Psychological Research on the Rat. Boston: Houghton, 1950, p. 52, 85. III. Murray, H. A. Explorations in Personality. New York : Oxford, 1938. 1 12. Nauta, W. J. H. J. Neurophyswl. g: 285, 1946. 113. Olds, J. ./. Comp. & Physiol. Psychol. 49: 281, 1956. 1 [4. Olds, J. and P. Milnf.r. J. Comp. & Physiol. Psychol. 47: 4 '9. '954- 115. Pfaffmann, C. J. Neurophysiol. 18: 429, 1955. 1 16. Pfaffmann, C. and J. K. Bare. J. Comp. & Physiol. Psychol. 43: 32°. >95°- 117. Pilgrim, F. J. and R. A. Patton. J. Comp. & Physiol. Psychol. 40: 343, 1947. 1 18. Porter, R. W., E. B. Cavanaugh, B. V. Critchlow and C. H. Sawyer. Am. J. Physiol. 189: 145, 1957. iig. Pribram, K. H. and J. F. Fulton. Brain yy: 34, 1954. 1 20. Pribram, K. H. and L. Weiskrantz. J. Comp. & Physiol. Psychol. 50: 74, 1957. 131. Ranson, S. W. .1. M. .1 Arch Neurol. & Psychiat. 41: 1, ■939- 122. Ranson, S. VV. A. Res. Nerv. & Ment. Dis., Proc. 20: 343, 1940. 123 Ranson, S. W., H. Rabat and H \V. Magoun. A. M .1 Arch. Neurol. & Psychiat. 33 : 467, 1 935. 124. Reid, L. S. and F. \V. Finger. J. Comp. & Physiol. Psychol. 48: no, 1955. 125. Richter, C. P. Quart. Rev. Biol. 2: 307, 1927. 126. Richter, C. P. Am. J. Physiol. 115: 155, 1936. 127. Richter, C. P. Am. J. M. Sc. 198: 9, 1939. 128. Richter, C. P. Endocrinology 24: 367, 1939. 129. Richter, C. P. Harvey Lectures 38: 63, 1942-43. 130. Richter, C. P. J. Comp. & Physiol. Psychol. 40: 129, 1947. 131. Richter, C. P. and K. H. Campbeli J. Nutrition 20: 31, 1940. 132. Richter, C. P. and CD. II \uki s. ./. Neurol. & Psychiat. -'" -'3'. '939- 133. Ruch, T. C. and H. A. Shenkin. ./. Neurophysiol. 6: 349, '943- 134. Rundquist, E. A. J. Comp. Psychol. 16: 415, 1933. 135. Schreiner, L. and A. Kling. J. Neurophysiol. 16: 643, '953- [36. Schreiner, L. and A. Kling. .1. M. .1. Arch Neurol. & Psychiat. 72: 180, 1954. 137. Schreiner, L., D. McK. Rioch, C. Pechtel and J. H. Masserman. J. Neurophysiol. 16: 234, 1953. 138. Scott, E. M. Tr. Am. A. Cereal Chemists 6: 126, 1948. 139. Scott, E. M. and E. L. Verney. ./. Nutrition 34: 471, 1947. 140. Share, I., E. Martyniuk and M. I. Grossman. Am. J. Physiol. 169: 22g, ig52. 141. Sidman, M., J. V. Brady, J. J. Boren and D. G. Conrad. Science 122: 830, 1955. 142. Siegel, P. S. J. Psychol. 21 : 227, 1946. 143. Skinner, B. F. The Behavior of Organisms. New York: Appleton, ig38, p. 341. 144. Smith, K. U. Psychol. Bull. 39: 493, 1942. 145. Smith, O. A. Anat. Rev. 124: 363, 1956. 146. Solomon, R. L. and E. S. Brush. In: Sebraska Symposium on Motivation, edited by M. R. Jones. Lincoln: Univ. Nebraska Press, 1956, p. 212. 147. Sprague, J. M. and VV. W. Chambers. Am. J. Physiol. 176:52, 1954. 148. Stamm, J. S. J. Comp. & Physiol. Psychol. 47: 21, 1954. 149. Si \mm. J. S. J. Comp. & Physiol. Psychol. 48: 347, 1955. 150. Stegcerda, F. R. Am .1 Physiol. 132: 517, 1941. 151. Stellar, E. Psychol. Rev. 61:5, 1954. 152. Stellar, 1.. and J. H. Hill. J. Comp. & Physiol. Psychol. 45: 96, 1952. 153. Stellar, E., R. Hyman and S. Samet. J. Comp. & Physiol. Psychol. 47 2211, m",) 154. Stellar, E. and R. A McCi f.ary. Am. Psychologist 7: 256, 155. Stowe, F. R. and A I Miller Experientia 13: 114, 1057. 156. Teitelbaum, P. J. Comp.& Physiol. Psychol. 48: 156, 1955. 157. Teitelbaum, P um E. Stellar. Science 120:894, 1954. 158. TlNBEROEN, N I h, Study of Instinct. Oxford: Clarendon Press, I 95 1. 159 rsANG, V. C. J. Comp. Psychol. 26: 1, 1938. 160. Verney, E. B. Proc. Roy. Soc, Lone/. Ser. B [35: 24, 1947. 161. H'ald, G. and B.Jackson. Proc. Nat. .lor/. Sc., Washington 10 255. '944- 162. Wano, G. H. Comp. Psychol. Monogr. 2: No. 6, 1923. 163. Wangensteen, O. H \mi A J. Carlson. Proc. Soc. Exper. Biol & Med. 28: 545, 1931. 164. Warden, C. J. Animal Motivation Studies- The Albino Rat. New York: Columbia, 1931. 165. Watson, J. B. Behaviorism. New York Norton, 1925. 166. WEINER, I II. and E. Stellar. J. Comp. & Ph) Psychol. 44: 394, 1951. 167. Wheatley. M I) A. hi. A. Arch. Neurol. & Psychiat. 52: 'Ml 168. \\ 11 in R, ( :. E. J. Comp Psychol. 24 547, 1937. 169. Wilkins, L. and C. P. Richter. ./. A. M. A. 114: 866, 1940. 170. Wolfe, J. B. Comp. Psychol. Monogr. 12: No. 60, 1936. 171. Woodworth, R. S. and H. Schlosberg. Experimental Psychology. New York Holt, 1 9",4, p. 664 172. Young, P. T. J Comp. Psychol. 37: 327, 1 044 173. Young, P. T. Comp. Psychol. Monogr. 19: No. 5, I94g. 174. Zeaman, I). ./. Exper. Psychol. 39: 466, 1949. CHAPTER LXIII Emotional behavior JOSEPH V. BRADY Walter Reed Army Institute of Research, Washington, D.C., and I 'niversity of Maryland, College Park, Maryland CHAPTER CONTENTS Some Psychological Considerations Some Historical and Methodological Considerations Neurophysiological Developments and Brain-Behavior Relationships Midbrain Reticular Inlluences Diencephalic Participation The Limbic System Neocortical Function Some Recent Developments the subject matter which provides the title for the present chapter has long occupied the attention of biological scientists concerned with the general problems of behavior and the related events of an organism's physiology. Our understanding of the relationships between organismic-environmental inter- actions referred to as 'emotional' and the 'structure- function' properties of the nervous system, in particu- lar, has evolved slowly and somewhat haltingly, however, amidst a host of psychological and phys- iological complexities. Introspective emphasis upon the phenomenological or "feeling' aspect of the 'emo- tion' problem has occupied a prominent place in the development of many highly speculative theories. Such basic scientific descriptions of this subject matter as we do have at present and are likely to attain in the future, however, would seem to depend upon the experimental analysis of expressive or behavioral phenomena, objectively and operationally defined. Similarly, although the measurement of more peripheral bodily changes related to such affective processes has been extensively described, the direct analysis of neurological mechanisms associated with emotional behavior has only recently begun to take firm anatomical and physiological form. In approaching the general topic of emotional behavior within the framework of this Handbook, emphasis will most appropriately focus upon related neurophysiological events Some brief consideration will first be given to the psychological or behavioral aspects of the problem, however, as they relate specifically to the defining properties of the or- ganism's interaction with the environment and the participating events of the nervous system. Against this background, we can then proceed to consider some of the methodological approaches which have characterized the half century or more of biological research in this problem and the contributions made toward the elucidation of central organization in emotional behavior. With this perspective, more re- cent developments in the anatomical, physiological and behavioral analysis of affective processes can be considered with a view to elaborating observed rela- tionships and assessing the present status and future outlook for this problem. Certainly, we shall not hesitate to pause along this charted course to spend some time with an interesting finding or a new development, but the general direction and scope of our coverage will adhere reasonably well to this broad outline. SOME PSYCHOLOGICAL CONSIDERATIONS In probably no other domain of psychological science has so little empirical data provided the oc- casion for so much theoretical speculation as in the general area of the 'emotions.' For the most part, the i5*9 •53° HANDBOOK OF I'llVsli i[ ( lf;Y NEIROI'IIYSIOLOGY III voluminous literature on this topic reveals a phenom- enological emphasis upon the 'affective' or 'feeling' aspects of the problem, and the wide range of diges- tive, respiratory, secretory and cardiovascular changes presumably related to emotional experience. Cer- tainly, the classical James-Lange formulation (ill, 201, 234) and even Cannon's 'neural organization' theory of emotion I by, 72) can be seen to share this experiential emphasis. And indeed, the ever-popular Freudian view of emotional processes as "mental states' or 'psychic phenomena' (128) continues to pervade even the most sophisticated treatments of the topic. The history of psychological speculation, however, has not been devoid of attempts to deal with t he emotions bchav iorisucally, if, at times, somewhat introspectively. The roots of such ap- proaches arc to be found even as far back as Darwin's early attention to the facial musculature in his con- sideration of the evolutionary aspects of emotional expression (93) and Wundt's emphasis upon the emotions as 'conscious contents' (402). Furthermore, Watson's classical treatment of the emotions as con- ditioned phenomena emphasized interactions be- tween the organism and its environment as the focus for his behavioristic views (383, 384). More recently, this descriptive behav ioristic tradition has found ex- pression in Skinner's analysis of emotion in terms of the probability or predisposition to change of a more or less broad range of behavioral response patterns (347), and in the related view of Keller & Schoenfeld that emotional behavior represents widespread changes in "reflex strength' as a function of specific environmental contingencies (210). Many other more or less extreme theoretical views have charac- terized psychological speculation in this area, and the serious student of emotional behavior will not want to overlook these efforts (3, 20, 26, 43, 51, 106- 109, 1 ;<,, [64, (65, 172174, 196, 205, 235, 237, 245, 248, 265, 273, 274, 276, 277, 280, 281, 292, 207, ;o (24, 325, 129, 152, 354, 368, 403). Experimentally, attempts to define emotional be- havior for laboratory investigative purposes have u 11. i!h fo< used both upon antecedent stimulus events which appear to produce or provide the occasion for a given response pattern, and upon (he charac- teristics of the response pattern per se. Typically, for example, both conditioned and unconditioned n,n lions associated with aversive stimuli have been re- gard 1 •emotional,' and a broad range ol inlet nal (e.g, epinephrine administration 1 and external (e.g. electric shock) environmental changes have become traditional!) identified as the antecedents which define such behavioral events (51, 232). More fre- quently, however, the properties of a given response pattern per se appear to serve as the basis for classify- ing behavior as "emotional." Certain characteristics of an organism's muscular activity (e.g. vocalization, trembling) have been conventionally identified with affective phenomena, and extensive physiological processes (e.g. autonomic changes) have received wide acceptance as 'indicators' of 'emotional' par- ticipation in a behavioral situation. Despite extensive research effort in this direction (70, no), however, it has not been possible to distinguish reliably be- tween emotional activities either on the basis of specific antecedent stimulus events or observable muscular and autonomic response patterns. Even the identification of emotional behavior in general by these criteria has presented problems, since such stimulus and response events are frequently observed to occur under circumstances not conventionally asso- ciated with emotion, such as temperature changes or heavy exercise. Faced with such difficulties, many psychological studies have focused upon the consequences of emo- tional situations for a broad spectrum of behavioral processes in an attempt to define emotion. Com- monly, the disruptive or suppressing effects of emo- tional disturbance upon ongoing activity have received both clinical and experimental emphasis (6, 43, 114, lib, 129, 206. 236, 240, 243, 265, 273, 330, 403). The defining properties of emotional be- havior segments, however, may as often involve an increased frequency or probability of adaptive response patterns, particularly when the contingencies of the situation require avoidance of aversive stimuli (281, 343, 352) or emergence of appetitive conse- quences (46, 1 1 -„ 1 7I1, 378). In the rather obvious absence ol an) completely satisfactor) theoretical or experimental formulation of emotional behavior, the task of defining and de- limiting the psychological subjeel matter for such .1 neurophysiological survey can be seen to present many difficulties. Conventional criteria lor identify- ing emotional activities appear far from adequate, and the classification of emotions or differentiation of subtle 'strength' phenomena (from the 'milder effects' (o (he "violent emotions') continues tu elude definitive analvsis. Certainly, (lie choice of material for (he present treatment will, of necessity, appeal arbitrary in man) instances, and (here will be legiti- mate questions concerning the appropriateness of much that has been included as well as much that has been omitted Since our present level of psy- EMOTIONAL BEHAVIOR [53i chological sophistication would hardly seem to justify a more restrictive or provincial approach the dictates of conventional or traditional usage will determine, in large part, the character of our be- havioral coverage. There can be little doubt that the results of future research efforts are likely to re- quire extensive modification and reorientation of past and present thinking in this area. But the some- what descriptive behavioristic framework within which much of the laboratory work on emotional behavior has proceeded should permit some realistic assessment of the past history, present status and future perspectives of a most complex psychophys- iological problem. SOME HISTORICAL AND METHODOLOGICAL CONSIDERATIONS For the most part, the measurement and analysis of bodily changes associated with emotional be- havior have focused upon peripheral response mechanisms related principally to integration by the autonomic nervous system, the cerebrospinal system and the endocrine system. Dunbar's recent revision of her volume on Emotions and Bodilx Changes (110) provides an exhaustive 4,717-itcm classified bib- liography in this area. Several oilier authors have also described techniques for recording such periph- eral changes and presented detailed experimental analyses of the characteristic physiological processes associated with more or less broadly defined af- fective phenomena (5, 70, 89, 95, 97, 146, 191, 228- 230, 245, 248, 262, 266, 324, 392, 403, 404). Typi- cally, studies in this general area have emphasized the relationship of the emotions to such peripheral phenomena as the electrical response of the skin (88, '^3, 358, 390, 391), arterial pressure and blood volume (88, 239, 317, 337), electrocardiogram and heart rate (2g8, 317, 389, 396), respiration (66, 123) skin temperature (285, 286), pupillary changes (21, 253), salivary secretion (399, 400), pilomotor effects (250), dermographia (392), skin sweating (344, 393), changes in the chemical composition of blood, saliva and urine (104, 105, 140, 146, 149, 188, 261, 305, 399), gastrointestinal activity (27, 70, 261, 263, 264, 367), metabolic rate (248), muscle tension (65, 84, 94, 318), tremor (24, 254), and even eye blink and eye movements (249). In addition, of course, au- tonomic, endocrine and neurohumoral relation- ships have been extensively and somewhat more directly analyzed in the quest for a better under- standing of emotional processes (69, 71-76, 146-148, 353). Certainly a comprehensive view of the affective process must not fail to take account of the intimate relationship between such peripheral physiological response expressions of emotion and the more central neural participants which provide the focus for the present neurophysiological analvsis. Methodologically, clinical and experimental ob- servations related to central neural organization in emotional behavior have emerged from four major sources: a) laboratory ablation studies involving central nervous system structures, h) direct electrical or chemical stimulation of central nervous system structures, c) electrical recording from central nervous system tissue and d) clinical, including neurosurgical, observations. For many centuries, of course, more or less informal accounts of clinical changes in emo- tional behavior associated with pathological in- volvement of central nervous system function (e.u. epilepsy i have emerged from both medical and literary sources. It was not until the latter half of the nineteenth century, however, that the clinical acumen of Hughlings Jackson and the consequent dedication of men like Ferrier and Sherrington began to provide the Inundations for systematic analysis of neurological mechanisms in emotional behavior and for subsequent emphasis upon de- velopment of the more experimental approaches which will receive primary emphasis in our present treatment. Since that time, a somewhat voluminous clinical literature on the psychological consequences of brain damage and related neurological disorders (83, 96, 126, 131, 168, 214, jib, 255, 278, 362) may be seen to have contributed both directly and in- directly to our understanding of central nervous system participation in emotional behavior. Neces- sarily, coverage of this clinical material will be highly selective within the relatively limited experi- mental scope of the present treatment, although related sections of this volume on autonomic activi- ties and transactional mechanisms should provide a somewhat broader factual analysis of data pertinent to this general source. Historically, ablation techniques can be seen to have provided the earliest laboratory approaches to the experimental analysis of neural mechanisms in emotional behavior. In the course of their more broadly conceived neurophysiological inquiries into cortical function, for example, Brown & Schafer (67), reporting in the Philosophical Transactions as early as 1888, described 'emotional' changes in rhesus monkeys following temporal lobe lesions involving relative!) ■532 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III selective subcortical structures. The early observa- tions of Goltz (156) toward the close of the last century on 'emotional' responses to mere handling in the decerebrate dog can be seen to provide the experimental beginnings for a laboratory analysis of neurological mechanisms in affective expression. With the turn of the century, ablation studies pri- marily concerned with only indirectly related neuro- phvsiological problems continued to (Joint up the ill-defined role of central neural processes in the organization of emotional behavior. In 1904, an investigation by Woodworth & Sherrington (401) of the spinal pathways related to pain revealed what iIh \ described .is pseudoaffective' behavioral changes in the decerebrate cat. Within the first two decades of the century, the further observations of Dusser de Barennc (112) following acute decortication in the cat had begun to locus more directly upon emotional behavior changes associated with experimental manipulation of the nervous system. Starting in the early 1920's, a rapid succession of experimental observations by Bazett & Penfield (19), Rotlmiann (323) and the now classic investigations of the decorticate preparation's 'sham rage' response by Cannon & Britton (73) further elaborated the specific character of central nervous system involve- ment in emotional expression. These early studies can be seen to have set the stage effectively for a host of experimental and theoretical efforts in this general direction which were to follow Over the next three decades or more. L'ntil well into the [930's, however, ablation techniques combined with gross observation of expressive phenomena in the experimental animal provided virtually the only laboratory methods available to the investigator of central participation in emotional behavior. Indeed, Berger's pioneering work on the electrical activity ol the nervous system (22) and even thai of Hess on direct electrical stimulation methods (179) may be seen to have their origins at .1 somewhat earlier dale, but the direct application of these important methodological developments tO the analysis of central organization in emotional behavior was, understandably, to follow onl) a dela) ol several years The striking changes in electrical activit) ol the brain, which were observed even by Berger to accompany 'attention' or 'anticipator) responsive- ness' 10 sensor) stimulation, earl) suggested the role electrical recording methods were to play in the elaboration oi neural events associated with af- fective States II was not until the later work ol Lindsle) (247), Darrow (90) and others (160, iH<>, 189, 364, 380, 397), however, that the more direct application of these methods to the problem of emotion was to make its firmest contribution. The studies of emotional behavior changes following direct stimulation of selective brain structures by Ranson & Magoun (315), Masserman (271, 273) and others (45, 99, 100, ib8, 3381, which were to follow Hess' fruitful lead (180, 182), did not appear in definitive form until after the late iujo's and the early [940's. Clearly, however, the 'modern era' in laboratory brain-behavior research related to the problem of emotions and the central nervous system can be seen to date from the now classical presentation before the 1937 meetings of the American Physio- logical Society by Kliiver & Bucy (217) reporting dramatic emotional behavioral alterations produced bv rather extensive temporal neocortical and palco- cortical lesions in the rhesus monkey. A few months later, Papez's speculative paper on "A Proposed Mechanism of Emotion" (3021 appeared with its emphasis upon primarily paleocortical, juxtallo- cortical and related subcortical structures; and in the following two decades, numerous anatomical, neuro- phv siological and behavioral studies have testified to the truly remarkable perspicacity of these early efforts. This enduring interest in the neural substrata of emotional behavior has had the not inconsiderable advantage of important technical and investigative advances in neuroanatomy (150, 295, 296, 369) and neurophysiology (98, 138, 199, 259, 291 ) over the past decade or more, not to mention the main more recent developments in behavioral control tech- niques (121, 122, 346). Such combined methodologi- cal skills have been profitably applied to the experi- mental analysis of affective processes 1 1 -, j .;, .)-,, |h, 51, 53, -,4, 170, 274, 342). Indeed, some reflection of the degree and direction of this progress in (his in- terdisciplinary approach is to be found in the recent appearance of several excellent comprehensive re- views directly related to this subject (1 ;_>, 216, 255, 311). NEUROPHYSIOLOGIOAl DEVELOPMENTS v\n BRAIN-BEHAVIOR REt VTIONSHIPS Midbrain Reticular Influences Probabl) the lir-t systematic experimental efforts to suggest differential participation of nervous system components in the elaboration of emotional behavior EMOTIONAL BEHAVIOR '533 can be safely attributed to Goltz (156) and to Wood- worth & Sherrington (401) in their analysis of the 'pseudoaffective' behavioral reactions which are observed to follow transection of the brain stem at the intercollicular level in laboratory carnivores. Nociceptive stimulation of the skin in such prepara- tions can be seen to elicit mimetic expressions of apparent 'anger' and 'rage' in the form of growling, barking, opening of the mouth, retraction of lips and tongue, snapping of jaws, snarling, lowering of the head as if to attack, and increase in arterial pressure. All such responses are extremely brief and do not outlast stimulation, however, and the normal emotional repertoire of such preparations appears severely circumscribed, especially with respect to the more positive types of affective expression usually present in unoperated members of these species (ap- parent 'joy' or 'satisfaction,' and sexual behavior). Indeed, many later observations have confirmed the general appearance of a poorly organized 'rage' response of but brief duration in the decerebrate preparation, along with a number of associated sympathetic reactions, although in not all cases does there seem to be a complete absence of pleasurable reaction (19, 208, 323, 328). The main point of in- terest to be derived from these observations, how- ever, is that at least some primitive 'pseudoaffective' behavioral expressions of emotion can be readily elicited at this midbrain level even in the absence of all other forebrain structures. Section of the brain stem below this level, however, has been reported by Bard (11) to abolish such behavior and strongly suggests involvement of the midbrain reticular forma- tion in the mediation of at least these rather funda- mental aspects of emotional expression. Certainly, a host of important subsequent experimental observa- tions (16, 96, 127, 211, 251) have continued to draw attention to reticular influences in many basic features of the emotional behavior pattern and in the maintenance of 'aroused affective states.' Linds- ley (248) has recently proposed an 'activation theory of emotion' which assigns critical executive functions to the reticular formation of the brain stem. Diencephalic Partit ipation The more classic theoretical conceptions relating emotional behavior and the nervous system can be seen to have developed around experimental empha- sis upon cortical-diencephalic interrelationships. The early experiments of Dusscr de Barenne (112) and Cannon & Britton (73), using decorticate prepara- tions and the now famous 'sham-rage' phenomena characteristic of such animals, did indeed provide an early focus for this continuing emphasis upon fore- brain mechanisms in emotional expression. By comparision with the more drastic Sherringtonian decerebrates, the decorticate preparations can be observed to respond even more readily and in a somewhat more intense, better organized fashion to ordinary handling and care, although the behavior remains poorly directed and short-lived. As a matter of fact, subsequent experimental analysis by Bard and Rioch (10-13, 17, 319) clearly demonstrated that such sham-rage behavior (described in some detail by these authors as involving lowering of the head and body in crouch, raising the back, drawing back the ears, loud angr\ growling, hissing, biting, striking with claws unsheathed, erection of hair, pupillodilata- tion, retraction of nictitating membrane, and widen- Lng of palpebral fissures) could still be elicited after removal of all cerebral tissue rostral, dorsal and lateral to the hypothalamus (including virtually all of the paleocortex, juxtallocortex and major portions of the related subcortical structures). This vigorously patterned but poorly directed activity involving both somatic and visceral components failed to de- velop, however, following truncation of the brain stem it ,m\ level below the caudal hypothalamus, a consideration suggesting both the central executive and facilitatory functions of these diencephalic regions in the elaboration of emotional behavior, as well as the possible inhibitory role of more rostral neocortical and paleocortical forebrain structures. It is interesting to note in this connection that even a somewhat broader range of affective expression, including 'fear' and 'sexual excitement,' in addition to the sham-rage response, was also described by Bard and Rioch as being elicitable from such prep- arations with extensive forebrain damage. More re- tent observations by Bromiley (61), in the course of discrimination learning experiments with the de- corticate dog, seem to indicate that the capacity for affective expression in such preparations can be maintained over relatively long periods of time even in the absence of extensive forebrain influences. Barking, growling, snarling and snapping as charac- teristic components of the sham-rage response to mere cage manipulation or gentle handling persisted over an extended period of almost 3 years in the Bromiley decorticate preparation. Clearly this im- portant combination of experimental inquiries points up the striking facilitatory role exerted by the addi- tion of limited (although obviously critical) hv- li wiiiii ii ik ill- pir. sii i| i >i;v NEIROPHYSIOLOCIY III pothalamic influences to the more primitive reticular ai n\ ating mechanisms. The continuing emphasis upon hypothalamic mechanisms in the development of neurophysiological approaches to the experimental analysis of emotional behavior has shown no diminution over the several decades that have intervened since these early ex- plorations. Subsequent efforts have provided con- vincing evidence of the central but somewhat compli- cated role played by hypothalamic portions of the forebrain in the elaboration of affective phenomena. The monumental studies of Hess and his collaborators (152, 155, 1 81—183) have demonstrated the broad range of emotional response patterns which can be selectively elicited from discretely localized electrical stimulation of carefully mapped diencephalic regions. The behavior changes observed following such direct stimulation of the hypothalamus have been com- pared with alterations seen in emotional responses of normal animals conventionally associated with 'fear," 'anger' and 'pleasure,' as well as with such phenomena as 'exploratory tendencies, feeding tendencies, cleaning tendencies and continuous rest- lessness.' In general, the results of these studies have indicated that more anterior and lateral portions of the diencephalon, including the basal septal nuclei, the preoptic area, the lateral hypothalamus and part ol the basal medial thalamus, may be associated with hostile and aggressive 'rage' responses, or 'affective defense reactions' as Hess prefers to call them. As stimulation is carried more posteriorly, changes in oral behavior, increased restlessness and escape responses appear, although no clear-cul topographical arrangement of hypothalamic nuclei with specific functional significance for affective processes has \<-i been discerned. Indeed, Ranson ;i|, ;i-,i, Ingram (197, '98), Masserman (271- l) and others (1H7, 207) have also demonstrated the emotionally exciting effects ('fear' and 'rage1 responses with multiple sympathetic manifestations) of direct hypothalamic stimulation in cats and monkeys, and al least some partial confirmation of hypothalamic involvement in emotional activities has been obtained b) White (395) with electrical stimulation methods in conscious human patients under local anesthesia. In addition, Grinker (159) also recorded selective electrical activity from deep- lying hypothalamic electrodes in man in response to 'emotional probing.' And, of course, the recent emergence of experimental emphasis upon results 1. Li. imed with intracranial self-stimulation tech- niques, following the interesting demonstration by Olds & Milner | [Ol 1 of the rewarding effects asso- ciated with direct electrical stimulation of selective forebrain structures, has suggested the intimate par- ticipation of hypothalamic influences, anions; others, in the presumably affective components of this phenomenon in a wide variety of species (46, 299, 300). That the hypothalamic role in the mediation of affective behavior is not to be regarded as unitary or uncomplicated, however, has been even more clearly emphasized by the results of numerous ablation studies involving relatively discrete de- struction of selected diencephalic nuclei and the analysis of behavioral changes. Bilateral lesions in the caudal part of the hypothalamus of cats and dorsolateral to the mammillary bodies in the monkey were early reported by Ingram et al. (198) and by Ranson (314) to produce complete loss of emotional responsiveness (masklike faces, stolidity), and some- times somnolence and sleep. Similar results were also obtained by Masserman (270), although reports of his own experiments in this area stress metabolic and homeostatic changes in accounting for the appar- ently transient emotional effects of such lesions. In contrast, Kessler (213) and Wheatley (394) have shown that destruction of the more medial aspects of this diencephalic region can result in dramatic rage reactions. In the more extensive study by Wheatley, relatively small lesions in the ventro- medial hypothalamic nuclei were observed to pro- duce "extremely, chronically and incurably savage' behavior in cats. Interestingly, however, the rage reactions in these animals (which appeared 'not unmixed with tear') were not blind or senseless, but well-directed and coordinated, complete with all the normal autonomic phenomena and well- calculated defensive and offensive activities. Nor were these ventromedial rage reactions altered appreciably b) superimposition of partial or total frontal lobectomy, removal of the temporal neo- cortex, or destruction of the mamrnillothalamic tracts, tin- fornices at the septal level, the dorsomedial thalamic nuclei or the mammillary bodies (197). Differential changes in cerebral cortie.il potential patterns have also been recorded with implanted electrodes from such ventromedial hypothalamic preparations 1 1 99)- I'r.icticallv all the behavioral alterations observed in follow ablation and stimulation hi the hypothala- mus in experimental animals have also been reported in man alter trauma, operative manipulation, tumor, vascular lesions and infections of the hypothalamus. EMOTIONAL BEHAVIOR !535 Although precise anatomical localization of specific diencephalic regions tends to be far from satisfactory under such conditions, various manifestations of affective changes including "terror,' "rage," "anxiety" and even some of the more "pleasant moods' ('witty,' 'jocular,' 'obscene') have been reported following hypothalamic involvement in the human (1,80, 82, 85, 87, 124, 134, 385, 395). And indeed, both experimental and clinical observations over the past three decades have made it abundantly clear that many other important biological motivations intimately related to emotional expression, including hunger, thirst, sleep, sex and activity, bear a critical dependence upon the functional integrity of rela- tively specific hypothalamic components (2, 14, 58, 62-64, 85, 134, 184, 294, 314, 316, 375, 385). This diencephalic emphasis has, as a matter of fact, found most recent expression in Stellar's (360) presentation of a 'physiological theory of motivated behavior' which places a heavy explanatory burden upon hypothalamic excitation in accounting for a wide range of motivational-emotional behavior patterns. The weight of available evidence, then, would certainly seem to indicate that at least some primi- tively organized, relatively undifferentiated patterns of emotional behavior may be elaborated within limited reticular and hypothalamic levels of neural organization. The emergence of homeostatic and adaptive autonomic functions, as well as important somatomotor activities basic to such affective proc- esses, would seem to depend critically upon the unique and direct integration of such brain-stem components with peripheral effector mechanisms. But the functional limitations of such gross react ton patterns contrast sharply with the more delicately balanced and restrained discriminative emotional behavior of which the normal organism is seen to be capable. Quite obviously, important influences from more advanced forebrain levels of integration con- tribute significantly to the elaboration and refine- ment of complexly organized and finely differentiated emotional response repertoires. Indeed, the early writings of Head (167) and the subsequent theo- retical formulations of Cannon (69, 72) suggested an important role for the more rostral thalamic nuclei in the elaboration of these affective processes, and several clinical and experimental inquiries over the past two decades have clearly justified this specula- tive focus. Spiegel and his collaborators (356, 357), for ex- ample, have reported changes in emotional behavior in both experimental animals and human patients following various thalamic lesions, involving prin- cipally the dorsomedial nuclei. Such ablations appear to reduce "anxiety,' 'tension,' 'agitation' and 'aggressive or assaultive behavior' in psychiatric patients, and at least a transitory reduction in emo- tional reactivity was presumably observed in simi- larly operated animals. More extended observations, however, by Schreiner et al. (335) on such animal preparations (cats) with lesions rather carefully restricted to the thalamic dorsomedial nuclei, demon- strate emotional changes in the direction of 'increased irritability and rage,' even though there has been some confirmation of an 'amelioration of neurotic patterns' in some of the same animals (275, 303). Of course, many of these same experimental studies have implicated more extensive regions of the thala- mic, including the anterior and intralaminar nuclei (275, 335)1 and even some of the more posterior nuclear groups 1 156) in the elaboration of emotional change, although reports of negative findings have likewise made important contributions (79). Del- gado (99), however, lias recently reported elicitation of 'conditioned anxiety, defensive and offensive movements, vocalizations, and autonomic mani- festations' in both cats and monkeys electrically stimulated in the posteroventral nucleus of the thalamus. Olds (300) also finds at least some mildly rewarding effects .1- .1 result of intracranial self- stimulation in the rat from such diverse thalamic placements as the habenula, and the lateral, ventral and ventromedial nuclei. The anterior nuclei, however, appear to have invited at least some special attention in the ex- perimental quest lor thalamic participants in emo- tional behavior. For the most part, lesions in this region of the thalamus in the cat are reported to produce marked reductions in emotional respon- siveness (8, 275, 335), while the effects of direct electrical stimulation in the anterior thalamus ap- pear to be at least 'alerting' in the cat (8) and highly rewarding in the rat (300). The range of these emotional changes following thalamic involve- ment suggests the possibility of a limited modulation of affective procc^^ .it this diencephalic level, even though the intimate relationship of these tha- lamic nuclei with more advanced paleocortical, juxtallocortical and neocortical systems must pro- vide for the more refined and integrative behavioral expression. Certainly, both the structural and functional interaction between the mediodorsal thalamus and the frontal neocortex on the one hand, >536 HANDBOOK OF l'HYSlol.OGY NEUROPHYSIOLOGY III .mil between the anterior thalamus and the cingu- late gyrus on the other, would seem to suggest the framework within which this forebrain integrative process in emotional behavior may be best understood. Ii is through the latter of these two systems that the anterior nucleus of the thalamus can interact with the 'limbic system' circuits receiving so much contem- porary emphasis in the analysis of emotional be- havior (45, [32, 133, 155, 216, 255-260, 311). Th 1 .1 nihil S) item The early reference by Broca (59) to "le strand lobe limbique' as a common denominator in all mammalian brains, and Papez' (302) subsequent theoretical speculations on a possible anatomical ■mechanism of emotion,' long ago suggested the potential mediating role of the "limbic system' in affective processes. In general terms, Papez' pro- posal focused upon the more medial aspects of the cerebral hemispheres and emphasized the transmit- tal of the 'central emotive process of cortical origin built up in the hippocampal formation' (hippocam- pus, hippocampal gyrus, dentate gyrus and amyg- dala) via the fornix to the mamillary bodies. Lifer- ents from this hypothalamic center were then pre- sumed to course both downward to the brain stem and lower effector mechanisms, and upward through the mammillothalamic tract to the anterior thalamic nuclei, and onward to the cingulate gyrus, Papez' candidate for the 'cortical receptive and association area' lor affective behavior. It was Papez' view "thai the hypothalamus, the anterior thalamic nuclei, the gyrus cinguli, the hippocampus, and their connections constitute ,1 harmonious mecha- nism which may elaborate the functions of central emotion, as well as participate in emotional ex- pression." Certainly, Papez' delimitation of these Structures as bearing an important relationship to emotional behavior, and his concomitant prediction 11I symptomatic changes associated with involve- ment ol this 'anatomic circuit," can in the light of subsequent clinical ,\ih\ experimental developments he seen 111 represent a considerable tour de force. The morphological and functional characteristics ui this 'limbic system' have been more precisely defined and elaborated over the two decades since this original proposal, and several systematic attempts have been made in order the obviously complex inter -relationships between these stun inn and othei nervous system components rding to both anatomicophysiological and behavioral principles 14-,, 155, 161, 175, 215, 216, 255, 256, 258, 259, 299, 311). Despite these recent efforts, however, there is no general agreement as to the definition of morpho- logical formations to be subsumed under the several presumably synonymous terms ('rhinencephalon,' visceral brain,' "paleocortex," etc.) used to refer to these groups or systems of functionally related fore- brain structures associated with emotional behavior. Figure 1 illustrates diagrammatically some of the more prominent anatomical interrelationships which characterize the medial aspects of the hemisphere (45, 295, 3201 and which provide at least some basis for considering the functional properties of these "limbic system' structures within three general groups or classes as follows. First, the paleocortical or allocortical portions of the system can be distinguished as those surface structures which meet the criteria for 'cortex' sug- gested by Rose & Woolsey (320) (a composition of at least three layers with the superficial layer con- stituting a fiber layer I and which also have clear phylogenetic primacy. These structures include the hippocampus (Amnion's horn and the dentate gyrus), the pyriform lobe (prepyriform cortex, pcriamygda- loid cortex and entorhinal area) and the olfactory bulb and tubercle. Secondly, the juxtallocortical portions of the svs- tem define that group of cortical regions which are intermediate in position between the phylogene- tically old paleocortex and the phylogeneticall) young neocortex, most of which have demonstrable anatomical connections with paleocortical structures. Such juxtallocortical regions include the cingulate gyrus or 'limbic cortex' (bo, 349), the presubiculum and the 'frontotemporal' cortex, as recently defined by Pribram & Kxuger (311). Finally, a third group of subcortical structures (not meeting the criteria for 'cortex') which have been show rr to lie iiitiin.itelv related both anatomi- cally and functional!) to the paleocortex and jux- tallocortex must be considered as pan of the 'limbic system,' most broadl) defined. These would include the amygdaloid complex, the septal region (septal nuclei and the nucleus of the diagonal band), cer- tain thalamic and hypothalamic nuclei, and possibly even the caudate nucleus .urd midbrain reticular formation, Unfortunately, most of these structures were liiimeilv believed lo be involved irr mechanisms of olfaction and are even presentlv referred to irr rrrost standard didaclic sources as the 'olfactory lirain' EMOTIONAL BEHAVIOR •537 BOIf fic. i. Semidiagrammatic representation "I the principal anatomical relationships between the 'paleocortex,' the 'juxtallocortex' and the several 'subcortical structures' considered in recent treatments of the 'limbic system.' The brain-stem portions of the system have been schematically displaced from the hilus of the hemisphere and represented in the Iowa halt of the figure in order to facilitate visualization of the numerous anatomical interconnections involving these structures. .1, anterior nucleus of the thalamus, Am, amygdaloid complex, Ar, arcuate nucleus, B. 01/., ol- factory bulb; C.I, .ulterior commissure, Ch, optic chiasm, Corp. Call., corpus callosum; DM, dorso- medial nucleus of the thalamus, En, entorhinal area; Fx, fornix, H, habenular complex; HP, habenulointerpeduncular tract; IL, intralaminar thalamic nuclei, IP, interpeduncular nucleus; L, lateral thalamic nucleus, MB, mammillary bodies, MT, mammillothalamic tract; Periam, periamyRdaloid cortex, Pit, pituitary; Prepxr, prepyriform cortex, Preuib, presubiculum; S, septal region; Teg, midbrain tegmentum, TO, olfactory tubercle; I*, ventral nucleus of the thalamus. [From Brady (45).] (311). This view, which tends to be perpetuated by the broad usage of the term rhinencephalon (216), would presently appear to lie untenable in the light of recent studies (81, 125) which have clearly demon- strated 1 more restricted distribution of primary olfactory afferents (i.e. of fibers originating from (he olfactory bulb). On the basis of this evidence, the term rhinencephalon should not be used synony- ■538 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III mousl) with limbic system in referring to forebrain participants in affective processes lout should be re- Stricted to those structures subserving; olfactory func- tions, including the olfactory bulb and tubercle, prcpyriform and periamygdaloid cortex, some of the nuclei of the amygdaloid complex, and the bed nucleus of the stria tcrminalis. Almost simultaneously with the 1937 publication of Papez' theoretical effort, Kliiver & Bucy reported their most striking demonstration of the important participant role which rather extensive "limbic sys- tem' components could be expected to play in the balancing, integration and elaboration of critically profound motivational-emotional behavior patterns. These now classic experiments (217-219), defining dramatic behavioral changes in monkeys following temporal neocortical and paleocortical lesions in- volving the frontotemporal cortex, pyriform lobe, amygdaloid complex, presubiculum and hippocam- pus, are now too well known to require detailed review. The broad range of behavioral alterations observed in these preparations, however, can be seen to bear directly upon the central problem of nervous system participation in affective processes. The formerly wild and intractable rhesus macaques used in these experiments became tame and docile, showing signs of neither fear nor anger, following this rather extensive involvement of the limbic sys- tem. They would not light or retaliate when abused by other monkeys and also displayed what the authors refer to as 'psychic blindness,' oral tendencies and hypermetamorphosis, a kind of compulsive behavior. They behaved as if they could no longer discrimi- nate between objects that were either potentially dangerous or useful to them Such an animal would, as il by compulsion, smell and mouth everything (dirt, feres, nails, food) thai captured its attention. Unless the objeel wen- edible, it was immediately dropped. If presented with a nail 100 times in suc- cession, the animal would smell and mouth it each time as though he had not examined ii before. Finally, these animals showed striking changes in sexual behavior; thev appeared hypcrsexed, mastur- bated excessively, sought partnership with male or female indiscriminately, and manifested bizarre oral sexual behavior. ( )l particular interest, too, was the fa< 1 that when unilateral excision of only one lobe was accomplished, or when bilateral lesions were restricted to the temporal neocortex and spared the limbic system structures, the animals failed to show ;ni\ of these dramatic changes in behavior. Vgainsl the background oi these earl) experimental and theoretical efforts, a host of important subse- quent neurophysiological and neuropsychological developments have continued to broaden the base for a more thorough understanding of extensive limbic system components involved in the mediation of emotional behavior. Within a few short months of these first reports, Spiegel and his co-workers (355) clearly demonstrated the dramatic participation of the more rostral portions of the limbic system (in- cluding the olfactory tubercle and septal region) in rather complicated motivational-emotional behavior patterns. 'Sham-rage' reactions were observed by these authors following bilateral lesions confined to the olfactory tubercle and septal region in both cats and dogs, while similar effects resulted from involvement of the anterior amygdaloid nuclei, parts of the hippocampus, and the fornix. Indeed, a somewhat earlier report by Fulton & Ingram (135) had described similar "rage" reactions in cats follow- ing bilateral prechiasmal lesions at the base of the brain, and more recent reports by Brady & Xaut.i (53, 54) confirm these findings in the rat. In addition, a recent publication by Heath et al. (168) further reflects the participant role of these more rostral limbic system components in the elaboration of emotional behavior, and the most dramatic demon- strations of rewarding effects consequent upon direct electrical stimulation of these same anteriorly placed limbic structures (including the orbital surface of the frontal lobes, the head of the caudate nucleus and the anterior hypothalamus I (4b, 300, 301, 342 1 would seem to establish firmly their involvement in affective processes. The most extensive and systematic research pro- gram initiated by Bard & Mountcastle (15, 16) within the decade following the report by Kliiver & Bucy represents an important landmark in the further experimental analysis of limbic system rela- tionships and emotional behavior. Concerned pri- marily with the role of forebrain mechanisms in the expression of "rage' and 'angry behavior,' their initial experiments with cats showed that removal of all neocortex, while sparing the paleocortical, juxtallocortical and related subcortical compo- nents of the limbic- svstcm, produced a markedly placid and emotionally unresponsive animal. In the authors' view, these results indicated (hat por- tions of the limbic system, either singly or in con- cert, could exert a restraining influence upon lower brain mechanisms ol demonstrated prepotence in the mediation of gross affective expression. More- over, subsequent experiments in this same scries EMOTIONAL BEHAVIOR J539 strongly suggested that the amygdaloid complex or cingulate gyrus, or both, might be specifically involved in the mediation of this restraining influence in the absence of the neocortex since rather striking in- creases in emotional reactivity followed removal of these subcortical and juxtallocortical structures in the previously neodecorticate preparation. More recently, Rothfield & Harman (322) have confirmed the placidity and emotional unresponsiveness result- ing from neocortical ablation sparing the limbic sys- tem in cats, and have further demonstrated that the fornix (distributing important hippocampal fibers to the septal region, diencephalon and rostral mid- brain) may figure prominently in the mediation of such restraining influences. Interruption of the fornix in their neodecorticate preparations resulted in a significant lowering of the rage threshold. Bard & Mountcastle (15) further analyzed the effects of relatively discrete paleocortical, juxtallo- cortical and related subcortical lesions in otherwise intact cats in an attempt to delimit more precisely the character of limbic system participation in such behavioral phenomena. Although bilateral removal of the hippocampus and presubiculum produced little demonstrable alteration in affective expression (with the possible exception of slight increases in pleasureable reactions), bilateral removal of the pyriform lobe and amygdaloid complex (sparing other limbic structures) resulted in dramatic, if somewhat delayed, behavioral changes. In this par- ticular case, the cats were observed to develop a markedly lowered rage threshold within (> to 8 weeks following surgery, although subsequent reports by Gastaut and his collaborators (141, 288) and by Schreiner & KJing (332, 334) describe changes more in the direction of the Kluver-Bucy effect (218, 219) consequent upon similar (but obviously not pre- cisely the same) destruction of the pyriform-amyg- daloid complex in the cat. Certainly, these rather gross differences in observed emotional changes (pre- sumably related to what appear to be only minor variations in the anatomical substrate involved in these somewhat conflicting; studies) may serve to point up even more sharply the delicately balanced relationships which characterize limbic system par- ticipation in affective processes At least one additional contribution of the Bard and Mountcastle research program, however, was to call attention to the somewhat complicated role of the cingulate gyrus in the elaboration of such ob- viously complex motivational-emotional behavior patterns. Although the removal of this juxtallocor- tical structure in neodecorticate preparations was observed to lower the rage threshold in their earlier experiments, these investigators found that cingulate ablation in otherwise intact cats tended to raise this threshold and produce emotionally less responsive animals (15). Indeed, several authors (151, 350, 381) have referred to similar consequences of cingu- lectomy ("loss of fear,' 'social indifference') in an attempt to define the role of this juxtallocortical portion of the limbic system in emotional behavior, and although Pribram & Fulton (310) have empha- sized the rather limited extent and duration of such changes, their observations generally confirm the character and direction of these effects. Undoubtedly, however, what has now come to be known as the 'Kluver-Bucy syndrome' (363) appears to have suggested the most stimulating lead for a host of subsequent research efforts to unravel the somewhat complicated role of limbic system com- ponents in the elaboration of emotion.il behavior. In a paper presented 10 the American Neurological Association in 1949, for example, Fulton and his colleagues (1371 reported that bilateral ablation of the frontotemporal portion of the juxtallocortex in monkeys can produce alterations in emotional be- havior similar to, but apparently nut ,is extensive .is, those found bv kltivcr and Bucy following temporal neocortical and paleocortical lesions Compulsive oral behavior and apparent lack of emotional re- sponsiveness to aversive stimuli were observed in these Irontotemporal preparations, but no alterations in sexual behavior seem to have appeared. As a matter of fact, the affective character of such changes .ire ol .111 even more limited scope when the lesions are restricted to the lateral surface of the temporal lobe and preoccipital cortex, although Blum el al. (29) have observed following such lesions in monkeys some deficits in complex visual tasks and learned discriminative performance which they believe to be at least somewhat related to motivational-emotional effects. When, however, experimental lesions are carefullv restricted to the pyriform lobe, amygdaloid complex and hippocampus (sparing neocortical regions for the most parti in the monkey, Smith (351 ) has con- firmed the appearance of very striking portions of the Kluver-Bucy syndrome (loss of 'fear' and 'anger' responses, docility, compulsive oral behavior) with- out gross motor or sensory deficits. In addition, this same report would seem to indicate that the docility and loss of fear can be produced in such animals without the compulsive oral behavior bv selectively 1540 HANDBOOK 1)1 PHYSIOLOGY NEUROPHYSIOLOGY HI ablating specific components of this 'pyriform- amygdaloid-hippocampaJ complex," although the precise delimitation of these particular structures has not as yet been satisfactorily accomplished. Indeed, a number of subsequent studies with mon- keys have succeeded in demonstrating varying por- tions of the Kliiver-Bucy complex as a function of equally varied placements of limbic system lesions. Thompson & Walker (365, 377), for example, have confirmed the 'taming' effects of bilateral lesions of the medial surface of the temporal lobe apparently restricted to the amygdaloid complex and hippocampus, although they have empha- sized the temporary character of such changes (4 to 5 month duration) and affirm the fact that lesions in other parts oi the inferior temporal cortex do not produce these effects. Apart from this increased docility and reduction in 'fear' responsiveness (all these animals could still express 'anger' and 'rage' to appropriate stimuli I, as well as a somewhat sur- prising decrease in sexual activity, however, none of the other Kliiver-Bucy symptoms could be demon- strated, even though histological analysis of their data seemed to support the implication that the amygdala is primarily involved in the changes result- ing from such lesions. Poirer (506) has also reported finding fragments of the Kliiver-Buc) syndrome in monkeys with somewhat more restricted lesions of the temporal pole, although these observations have been limited largely to the apparent 'apathy' and 'drowsi- ness' of the operated animals without mention ol other behavioral changes. Significantly, continuing attempts to analyze the specific relationships in this temporal lobe-amygdala- hippocampus syndrome have led investigators to a wide variety of different animal species in their quest for some basic understanding of this rather complex motivational-emotional phenomenon. Pribram, Mish kin and their collaborators (130, 283, 284, 309, 312), foi example, have reported upon the effects ol lesions involving the frontotemporal cortex, temporal pole .mil amygdaloid complex in baboons and iIoljs, the results confirming, for the most part, the 'taming' and oral effects which follow involvement of these structures. And Schreiner & Kling (332, 334) have di cribed similar behavioral changes associated with the amygdaloid complex and pyriform lobe in c mis .mil K n\. In the cat preparations (which were most extensively studied 1. refractoriness to rage 01 anger-producing stimuli, exaggerated 01. il and vocal behavior, and marked hypersexuality were observed, a iplex of changes similarly re- ported by Gastaut and his co-workers (141, ->88) following amygdala lesions in the cat. The most striking observation in the notably wild and intrac- table agoutis and lynx studied by Schreiner & Kling, however, was the dramatic conversion to virtually complete (if somewhat temporary in the case of the lynx, at least) docility following bilateral lesions in the amygdala and pyriform lobe. It may be of some interest to note that at least the state of hypersexuality produced by amygdalectomy in Schreiner & Kling's cats can be abolished hv castration and restored with substitution therapy (333), suggesting the role of important neuroendo- crine relationships involving the limbic system in the elaboration of such motivational-emotional be- havior patterns. Furthermore, these same authors (332) have also reported that the characteristic. tllv placid amygdalectomized cat can be readily con- verted into a 'vicious,' 'rageful' animal by additional superimposed lesions in the ventromedial nucleus of the hypothalamus. And, in fact, one of the first efforts to analyze quantitatively the behavioral effects ol such limbic system lesions upon the acquisition and retention of a conditioned avoidance response by Brack et al. (55) involved many of these same amygdalectomized cats. The avoidance technique used in these experiments consisted of the animal's passage through an open doorway separating the two compartments of a conventional •double-grill' box in response to the presentation of a 30-sec. conditioned clicker stimulus, thus terminating the clicker and avoiding the shocks which followed failure to respond within the 30-scc. stimulus interval. Acquisition of such a conditioned avoid. nice response was found in lie significandy impaired in bilaterally amygdalec- tomized cats, although virtually identical lesions appeared to have little effect upon the retention of precisely the same avoidance behavior in cats con- ditioned prior to operation. Weiskrantz & Pribram (386 388) have also undertaken similar efforts to approach the "temporal lobe problem' in this some- what quantitative fashion, and correspondingly significant decrements were observed in conditioned avoidance behavior following bilateral lesions in- volving not only the amygdaloid complex in the monkey, but other portions of the limbic system as well (frontotemporal cortex, cingulate gyrus). \ considerably broader conception of limbic svs- tem participation in emotional behavior would seem to be reflected, however, m the host of clinical and experimental observations which continue to emerge from more recent psychological and physiological EMOTIONAL BEHAVIOR '54' analyses of affective processes. Not only has the conditioned behavior of the individual animal with limbic system involvement begun to come under careful scrutiny in specifically controlled testing situations, but even the integration of social be- havior and its dependence upon these neural systems has now been explored. Rosvold el al. (321) observed that the effects of amydalectomy in eight male Rhesus monkeys generally changed their hierarchical position in a group-cage situation from dominant to submissive, even though they appeared somewhat more 'aggressive' when in individual cages. Cer- tainly, the results of these experiments would seem to suggest that the postsurgical social environment and the length of time preoperative relationships have existed can be as important a consideration as differences in the extent and location of the lesion in evaluating the consequences of limbic system ablations for emotional behavior. And indeed, clinical reports of observations following temporal lobe and amyg- daloid lesions in man by Terzian & Ore (363) and Sawa et al. (326) have emphasized this same diminu- tion of 'social aggressiveness.' Gastaut and colleague (144) have also pointed out that discharging lesions in these limbic system structures, as seen in psycho- motor epilepsy, apparently produce a lowered 'rage' threshold since these patients frequently show violent temper outbursts in social situations. Ii ma) also be significant that Gastaut & Collomb (142) have observed a decrease in sexual behavior in these patients with irritative lesions of the temporal lobe- amygdala region, while Gastaut & Mileto (143) have further elaborated upon the disturbances in sexual behavior which follow involvement of the hippocampus in both human and animal cases of rabies. Many physiological studies (4, <), (>8, 77, 78, 132, 141, 202-204, 221-226, 257, 260, 307, 311. 336> 380 in both animals and man, using chemical and elec- trical stimulation as well as electrical recording methods, have also demonstrated limbic system in- volvement in a wide variety of somatic and auto- nomic phenomena closely related to the broad range of behavioral activities conventionally associated with emotional expression. Significantly, it has been difficult to discern any clear-cut topographical or- ganization for specific behavioral components, even though the observations of Kaada et al. (203) would seem to suggest that such delineation may be possible. For the most part, however, the striking features of such correlative data would seem to be the extensive overlap of all sorts of behavioral re- sponses in their representation at this limbic system level (203, 221-226), and the remarkably broad spectrum of psychological activities in which these structures can be presumed to participate (132, 141, 202, 257, 311). Heath and his collaborators (169, 238) have even recently proposed extensive involve- ment of these specific neural systems in the elabora- tion of 'thought' and 'psychological awareness.' The intimate relationship of these limbic structures (particularly the amygdala) to the mechanisms of neuroendocrine integration has been convincinglv demonstrated by both stimulation and ablation studies (166, 190, 223, 225, 267, 308, 327, 331-333). Finally, electrophysiological methods have con- tinued to define the characteristic functional inter- relationships within the limbic svsiein and subcor- tical regions basically involved in the elaboration of emotional behavior (68, Il8, 119, [53— 155, 157, 241, 242). Of particular importance in this respect would seem to be the extensive studies of Macl.ean and his collaborators (2 , 260), and of Gloor (153- [55), carefully delineating the limbic system role in affective processes. Imi with this host of clinical and experimental observations, and the rapidly accumulating body of anatomical, physiological and psychological in- formation, however, no completely satisfactory in- tegration of the limbic system with the neeessarilv broad range of central neural participants in emo- tional behavior has .is vei emerged. There never has been ,mv shortage of speculative efforts assigning specific functional roles to the various components of this anatomical complex, and a significant thread of similarity is indeed discernible among the many neurological hypotheses which have characterized the multidisciplinary theorizing in this area. Almost 30 years ago, for example, Herrick ( 1 7 ~, ) , on a com- parative an.itomie.il basis, suggested that the limbic system mav serve as .1 nonspecific activator for all cortical activities, influencing "the internal appara- tus of general bodily attitude, disposition, and affec- tive tone." Even Kleist's (215) speculations of the same era about the 'inner brain,' as he referred to the more medial aspects of the hemisphere, can be seen to emphasize the fact that these limbic struc- tures were not only basic for 'emotional behavior,' 'attitudes' and 'drives,' but were also instrumental in correlating 'visceral receptions' from the oral, anal and genital regions, as well as the intestines, thus subserving functions related to the search for food and sexual objects. And clinical observations of human patients with limbic system involvement led >542 HANDBOOK OF PHYSIOLOGY NEl'ROPHYSIOLOGY III Grunthal (161) to propose that the hippocampus, as the virtual 'hub' of the limbic system, may repre- sent a 'catalytic activator' which, although not necessarily participating in specialized functions itself, is nevertheless basic for the proper functioning of affective and neocortical activity. More recently, MacLean (255) has reviewed and elaborated Papcz' (302) earlier theoretical views on emotional behavior and limbic system mechanisms (or "visceral brain,' as is MacLean's reference), suggesting the basic importance of these forebrain structures not only for affective processes, but also for correlating 'oral and visceral sensations' as well as 'impressions from the sex organs, body wall, eye and ear.' Even Pribram & Kruger (311) in their com- prehensive review of the 'olfactory brain,' have speculatively assigned 'olfactory-gustatory,' 'meta- bolic' and "socioemotional" functions to the various 'limbic' components comprising their three 'systems.' And Gloor's (155) recent analysis of telencephalic influences upon the hypothalamus has assigned to the limbic system the role of "modulator of func- tional patterns integrated at the level of the hypo- thalamus and the brain stem tegmentum," even though in his view, "the limbic system docs not fundamentally integrate the functions it is capable of influencing by its activity." Indeed, the weight of available anatomical, physiological and psycho- logical evidence would certainly seem to support at least some generally similar concept of the 'inter- mediary' role of the limbic system in the integration of brain-stem and neocortical participation in emotional behavior. Neot hi tii n/ Fum tion Despite this icon! experimental and theoretical emphasis upon limbic system relationships, however, the long-enduring quest for 'localized functions' at the level of the neocortex continues to exert im- portant influences upon both clinical and laboratory contributions to the neurophysiological analysis oi emotional I >» - 1 1. 1 v i < »i F01 the mosl part, attention has traditionally focused upon the frontal lubes with specific reference to affective processes (83, 117, 126, 131, 132, 171, 200, 214, 231, 23 |, 244, -'78, 279, [66, 376, 382), although some additional con- (11 n with the participant role ol more exten- sive neocortical regions (7, 86, 158, 212, 282, 293, 104, 313) has recently been in evidence. 1 ndei tandably, clinical observations can be seen in have contributed the lion's share to the available literature in this area, although the laboratory analy- sis of ablation consequences and selective changes in electrically recorded potentials from the neocortex has more recently suggested important neural- behavioral relationships. Long before systematic treatment of such problems was fashionable ob- servations on the behavioral consequences of the 'sacred disease' — epilepsy — included both literary and professional descriptions of affective changes presumably related to neocortical involvement. And as early as 1875, David Ferrier (120) provided a provocative description of behavioral changes closely related to emotional phenomena in monkeys following experimental frontal ablations. Somewhat later, Bianchi (25) made similar observations, and the story of the classic report by Fulton & Jacobsen (136) before the London meetings of the Second International Neurological Congress in 1935 and the subsequent adoption of frontal ablations as a therapeutic procedure by Moniz & Lima (2871 is now too well known to require detailed repetition. Significantly, the presumed therapeutic emotional changes observed to follow such prefrontal lesions have frequently been rationalized in terms of the intimate anatomical and functional relationship of these more or less specific portions of the cerebral mantle with the affective integrative mechanisms of the diencephalon (via principally the dorsomedial thalamus). It has now become abundantly clear, however, that extensive limbic system influences doubtless exert important mediating effects on such diencephaliconeocortical interactions, and thai the assessment of emotional changes consequent upon neocortical involvement must be considered within this integrative relational framework. Indeed, the wide variety of behavioral changes which have been observed to follow such frontal neocorlie.il ablations would appear comprehensible nnlv within the broad framework of such an integrative analysis. For the most part, the consequences of frontal lobe lesions appear to involve changes in the direction of diminished 'emotional responsiveness.' Both clinical and laboratory reports, however, have also con- firmed the frequent appearance of increased 'emo- tional lability' in man and animal following at least some therapeutic and investigative efforts to alter affective behavior patterns with frontal neocortical ablation. Clearl) we are a considerable distance from a satisfactory understanding of the participant role of such specific neoeortie.il regions in the elab- oration oi emotional processes, although the evi- dence for such involvement seems unequivocal. EMOTIONAL BEHAVIOR '543 An important recent emphasis upon electroen- cephalographic studies in relation to "affective states' can also be seen to hold considerable promise for a more thorough understanding of neocortical functions in emotional behavior. Only within the past decade have the first systematic treatments of this neurophysiological approach by Lindsley (247) and Darrow (90) begun to appear although, as we have already seen, even earlier explorations of EEG phenomena had suggested their relationship to the 'emotions' (22, 160, 189, 364, 397). Characteris- tically, changes in the EEG pattern "under condi- tions involving some degree of emotional arousal, as in apprehension, unexpected sensory stimulation, and anxiety states," as summarized by Lindsley (247) in his 1948 review, can be reflected in "a reduction or suppression of alpha rhythm and an increase in the amount of beta-like fast activity."' Although the observations which provided the basis for these general conclusions did not focus upon any selective neocortical regions in particular, they can be seen to form at least part of the foundation for Lindsley's 'activation theory' of emotion (248I with its emphasis upon neocortical arousal in affective processes. Subsequent reports by Ulett el a/. (370— 372) and others (18, 23, 30, 91, 92, 113, 220), how- ever, have suggested involvement of more specific cortical areas, and Walter (379) has even recently reported that emotional disturbances arising during flicker stimulation experiments can be associated with rather selective EEG changes in the temporal neocortex. Certainly, the close anatomical and func- tional association of the temporal lobes with limbic system structures intimately involved with the elab- oration of emotional behavior would seem to fit well with such a suggested delineation of neocortical participation in affective processes SOME RECENT DEVELOPMENTS Probably the most striking feature of this long- enduring neurophysiological interest in the problem of emotion has been the slow pace at which ex- perimental analysis of the critical behavioral phe- nomena has proceeded. Needless to say, this failure of psychological science to keep abreast of anatomical and physiological developments is clearly reflected in the obviously primitive, phenomenological and conspicuously prescientific descriptions and defini- tions of emotional behavior which can be seen to characterize most of the research in this area. And indeed, one cannot help but wonder about the econ- omy and parsimony of elaborate speculative efforts to develop a comprehensive neurophysiological theory of emotion in the absence of a sound behavioristic account of those presumably affective interaction processes between organism and environment. The analysis of such functional relationships at this de- scriptive behavioristic level must provide the founda- tion for any adequate treatment of neurophysio- logical participation in "emotional' events. Within recent years, however, the emerging out- lines of an objective psychological science have begun to provide precise and reliable techniques for con- trolling the behavior of the individual subject as a basis for interdisciplinary neurophysiological analy- sis. For the most part, the dedication of B. F. Skinner, his co-workers and others to the experimental analysis of behavior in its own right has been responsible for the development of these methods in a variety of applications and for their recent extension to the investigation of the problem which provides the subject matter for this chapter (6, 28, 31-44, 47-54, 36, 57, 101-103, 114-116, 121, 122, 145, 162, 177, 178, 185, 192-195, 209, 240, 252, 268, 269, 289, 290, 339-342. 345-348, 359, 373, 374, 388, 398). The application of these so-called 'operant con- ditioning' techniques to the experimental analysis of both behavioral and neurophysiological problems can be seen to rest upon a simple principle, namely that the characteristics of an organism's behavior are, to a considerable extent at least, determined by what the environmental consequences of that behavior have been in the past. Thus the term 'oper- ant behavior' has been used to refer to behavior which operates upon the environment in this fashion, and the process of manipulating such behavior as a func- tion of its environmental consequences has been termed 'operant conditioning' (345). The systematic analysis of orderly relations among behavior seg- ments within this framework has been accomplished, first, by selecting for measurement and manipulation a response having a topography congenial to the or- ganism, and one that the organism can perform and immediately be in a position to repeat. Secondlv, this kind of analysis has been enhanced by selecting an environmental consequence, or 'reinforcement,' that is appropriate to the particular individual and by utilizing motivational levels that are strong enough to minimize the effects of many experi- mentally irrelevant variables. Finally, an additional aspect of this approach to behavior science is the systematic limitation of the experimental environ- '544 II VNDBlKlK (il- I'HYMul ' K.X M I KOPHYSIOI.OGY III a typical output duping /5 minute period B- FIPST CONDITIONING T0IAL C CONDITIONED EMOTIONAL RESPONSE 1 EABLY STAGES 2.FULLY ESTABLISHED LE GEND CLICKER INTQODUCEO AT C, TEQMINATED BY SHOCK AT S AFTER 5 MINUTES. TIME in. 2. The conditioned emotional response as it appeals typically in the cumulative response curve. Abscissae, time; ordinate*, cumulative number of responses. [From Hunt & Brady 1193).] men! to permit at least sonic reasonable degree of stimulus control and specification. Typically, the subject in such a study may be a hungry pigeon pecking at a lighted spot on the wall for a small amount ol grain as reinforcement; or .1 thirsty rat ma\ press .1 lever to obtain a small drop of water, ipi .1 monkey ma) push a panel in order to postpone .1 painful electric shock; or even Homo sapiens may pull a plunger to acquire a candy or cigarette reward. The point, ol course, is not the investigation of eating, drinking or smoking behavior per se, or even pil pain. Most importandy, the object of these tech- niques is to place some arbitrary sample ol behavior mill. 1 experimental control so that behaviorial processes may be investigated as .1 function of a wide variety of operations, including neurophysiological manipulations and even 'emotional' disturbance Estes & Skinner (116) first suggested the basis of an approach to at least one aspect of the 'emotion' problem within the framework of this developing behavioi science, and several recent extensions ol these methods to both the psychological and physio- logical analysis of affective processes ( ;_>, 36-41, 43-47. 51. 53. 54. '93. '94. -">-'. 268, 269, 290, 3:59- 34'. 34**) justify their enthusiastic accep- tance by a broad interdisciplinary audience. Much of this work has had its origin in the rather common clinical and experimental observation that 'emotional' disturbance can, as one of several possible effects, disrupt or interfere with an or- ganism's ongoing behavior. Experimentally, the fundamental empirical fact which has provided the cornerstone for such an approach to the problem ul emotion is the suppressing effect of anticipated pain upon an animal s ongoing lexer pressing be- havior. This conditioned suppression phenomenon is readily produced by pairing some previously neutral stimulus with pain shock. The typical consequences of such a procedure in the rat (involving repeated presentations of a clicking noise for 5 min. followed by pain shock to the feet during a lexer pressing session for water reward) are shown in figure 2 (193). The clicking noise is introduced at point C on the cumulative lever pressing curves, continues for 5 min. and is terminated continguously with shock at point .S'. Within a few trials, the anticipatory 'emo- tional' response to the clicker begins to appear as a perturbation in the lexer pressing curve, accom- panied by crouching, immobility and usually defe- cation. Xow this emphasis upon a rather primitive and possibly somewhat molecular psychological phe- nomenon as the starting point for an experimental analysis of emotional behavior has several clear-cut advantages from the standpoint of a neurophysio- logical analysis. First, focus directly upon this con- ditioned suppression response per so eliminates a major source of error attributable to variables that affect the instrumental behavior from which 'emotional' effects are usually inferred in the more conventional observational or even 'escape-avoid- ance' learning approach to this problem. Secondly, this simple relatively uncomplicated response is clickable under .1 wide range Ol conditions and ap- pears in quite consistent form or topography in all animals. Thirdly, this response is remarkably stable oxer time, surviving without apparent diminution in the absence ol exercise or further reinforcement virtually throughout the entire life span of the or- ganism, Finally, and probably most importantly in. in the standpoint of a relational neurophysio- logical analvsis, the technique of superimposing the eiiHiiiiin.il response upon a well-established stable levei-pressing li.il.it makes it possible to approximate EMOTIONAL BEHAVIOR '545 an objectively quantifiable definition of the be- havior in terms of changes in output during various segments of the lever pressing curve (195). In many respects, this rather restricted behavior sample appears to be prototypical of the most primitive aspects of an organism's emotional repertoire, and the investigation of its vicissitudes — the conditions which determine its increases and decreases in strength, etc. — has already contributed significantly to the differential experimental analysis of both the physiological and psychological variables upon which the organization of emotional behavior de- pends. The potential implications for a neurophysio- logical analysis of emotional behavior within this rather broad operant conditioning framework first became apparent as a consequence of a series of studies on the effects of electroconvulsive shock (.33. 34. 49- 50, 52> 54. 5». i4.> '93. '95> The results of these experiments showed that ii was pos- sible to separate and measure selectively the effects of such physiological manipulations upon the '<■ - tional' components of a behavior segment inde- pendently of any gross effects upon the simple repetitive lever-pressing habil which provided the control base line. Subsequent applications of this approach to the analysis ol more direct neurophysio- logical participation in alfcclive processes have made it possible to show, for example, thai the elab- oration of even such basic aspects of emotional be- havior depends heavily upon the integrit) of quite specific portions of the forebrain and brain stem, notably the limbic system. Although large neocor- tical lesions were found to produce little or no effect upon the acquisition, retention or extinction of the conditioned suppression pattern, lesions in the septal region and hippocampus produced significant decre- ments in the maintenance of such behavior and most dramatic changes in gross affective expression (36, 53). In addition, lesions of the habenular com- plex of the thalamus appear to reduce resistance to extinction of the conditioned suppression response, although cingulate lesions have no apparent effect on such behavior (40, 43, 54). More recently, this operant conditioning approach to the neurophysiological analysis of emotional behavior has found most dramatic application in the exploration of reinforcement or "reward' effects produced by intracranial electrical self-stimulation of the nervous system. Olds & Milner (301) first reported that rats, electrically stimulating themselves in various portions of the limbic system by pressing a WATER REWARD STIMULATION REWARD {24 HOURS DEPRIVATION ) ( SEPTAL ELECTRODE Ld PLACEMENT ) £ 400- S - £ 200" cr UJ > 01 iV ^ jy X 1st . 3rd . 8th . Is' J 8th 16th "ANXIETY" CONDITIONING TRIALS n_._ ^200 ' no. 3. Sample cumulative response curves showing acquisi- tion trials for the conditioned 'anxiety' response superimposed upon lever pressing for water and intracranial electrical stimu- lation reward. The oblique solid arrows indicate the onset of the conditioned auditory stimulus, and the oblique broken arrows indicate the termination of the conditioned stimulus contigu- ously with the brief unconditioned grid-shock stimulus to the feet during each trial. [From Brady | ", bar, would maintain high lever-pressing rates over long periods of time without an) other reward. This same phenomenon has been systematically reproduced and analyzed in the cat and monke) bj Brack 144, (6) and others (246, 142), and investiga- tive efforts have begun to delimit some of the critical variables in this area. I In- s| critic anatomical locus of the Stimulating electrodes, schedules of intra- cranial electrical stimulation reinforcement, food and water deprivation, stimulus intensity, temporal factors, and (he like have all been shown to constitute critical determinants of this effect (44, 4-,, 46, 48, 299, |00, 142). Of particular interest from the standpoint of a primary concern with neurophysiological relation- ships .md emotional behavior, however, would seem to be the- recent demonstration of interaction effects between the conditioned suppression response illus- trated in figure j and this intracranial -elf-stimula- tion phenomenon (40, 44, 4",, 4(1). The consequences of repeated pairings of the clicker and shock accord- ing to this conditioning procedure have been con- sistently reported to include suppression of the lever- pressing rate, crouchinij, immobility and usually defecation in response to presentation of the audi- tory stimulus. Although this relatively stable condi- tioned 'fear' pattern was reported by Brady (40, 43-46) to be readily elicited by presentation of the clicker when both rats and monkeys were pressing a lever for food or water reward, substitution of brain stimulation through chronically implanted limbic system electrodes (septal region, medial forebrain bundle) for the food or water on the same reinforce- ment schedule resulted in failure of the auditorv 1546 HWDBOOK OF PHYSIOLOGY M I R( il'IT, SII '1 M,.\ III stimulus to elicit the emotional suppression of lever- pressing behavior. Figure 3 shows the development of the conditioned "fear' response in a rat with repeated pairings of clicker and shock superimposed upon the water-reinforced lever-pressing curve, and illustrates the striking failure of the suppression behavior to appear in the same animal with the same clicker and shock when lever pressing is re- warded with brain stimulation in the septal region rather than water. With the monkey, this same phenomenon has been demonstrated with the self- stimulation electrodes in the anterior forebrain portions of the limbic system (medial forebrain bundle), although an extensive analysis of rewarding electrode placements which do not show this inter- action effect with the conditioned emotional re- sponse has not yet been accomplished. It is clear that even this developing refinement in the experimental analysis of neuropsychological relationships has not as yet provided more than the most preliminary framework within which a satis- factory formulation of emotional behavior is to be sought. An almost infinite complexity remains to be unraveled at the level of organismic-cnvironmental interactions, and an adequate neurophysiological analysis of affective processes would seem to bear a critical dependence upon the systematization of such behavioral relationships per se. Indeed, many attempts have been made to order these behavioral diversities to single broad principles within the con- text of contemporary psychological emphasis upon 'acquired drives' and similar motivational con- structs. And physiological focus upon specific 'neural mechanisms' in speculative accounts of the emotion problem has at times appeared equally restrictive. The promise of a more empirical relational analvMs of independently definable psychological and physio- logical events, however, may be seen to reflect some concern that such monolithic ordering, prematurely embraced, might serve to obscure important neuro- behavioral relationships basic to an understanding of what we conventionally regard as emotion. REFERENCES 1. Alpers, B. J. A.M. A. Arch. Xeurol. 0? Psychiat. 38: 291, '937- 2. Anand, B. K \si) J- R. Brobeck. Yale J. Biol, ^s' Med ->4: y C. Murchison. Worcester: Clark Univ. Press, '934. P- 264. 13. Bard, P In: Tht Inter-Relationship of Mind and Body, edited by \ li.nn/ Baltimore: Williams & Wilkins, ig 19, p 190. 1 Baku, P In Tht Hypothalamus and Central Levels oj Auto- nomic Function, edited by J. Fulton. Baltimore. Williams & Wilkins, 1940, |) 551. Bard, I'. In: Feelings and Emotions, edited by M. L. Reymert. New York McGraw-Hill, 1950, [> 211 iti Bard, P and V. B. Mountcasiii In Tht Frontal Lobes, edited In J, Fulton. Baltimore Williams & Wilkins, mi 17. Bard, P. and I) M Rioch Bull. Johns Hopkins //»>/>. 60 73, ig 18. Barker, W\, S. Burgwin and D. J. Simons. J. Nerv. & Mint. Dis. 112: 187, 1950. 19. Bazett, H. C. and W. G. Penfield. Brain 45: 185, 1922. 20. Beebe-Center, J. G. The Psychology of Pleasantness and I'npleasantness. New York : Van Nostrand, 1932. 21. Bender, W. R. G. Psychol. Monogr. 44: 1, 1933. 22. Bercer, H. Arch. Psychiat. 87: 527, 1929. 23. Berlin, I. N. and C. L. Yeager. Am. J. Dis. Child. 81 : 664, 1 95 1. 24. Berrien, F. K. ./. Expei Psychol 24: 485, 1939. 25. Bianchi, L. The Mechanism of the Brain and the Func- tion of the Frontal Lobes, translated by J. H. Mac Donald. New York: Wood, IU22. 26. Bi.ndra, I). Can, ul. ./. Psychol. 9: 161, 1955. 27. Bindra, D. and W. R. Thompson. J. Comp. & Physiol. Psychol. 46: 43, 1953. 28. BLOUGH, D. S. Science 121: 703, 1955. 29. Blum, J. S, K. I. Chow and K. H Pribram. ./. Comp. Neurol. 93: 53, 1950. 30. BoNVAi.i.ET, M., P. Dell and G llu 11 1 Electroencephalog. & Clin. Neurophysiol. 6: 119, 1954 31. Boren, ]. J. ./. Pharmacol. & Exper. I'herap. 119: 134, 1957. BOREN, J. J. and M Sidman. (.'amid. ./. Psychol. II: 185, ■957- ; 1 Brady, J. V. ./. Comp ^- Physiol. Psychol. 44: r)07, 1951. 34. Brady, J. V. J. Comp. & Physiol. Psychol. 45: 9, 1952. 35. Brady, J. V. ./ Comp & Physiol Psychol. 46 107. 1953. 1 Brady, J. V. XIV Internal Congr. Plychol., Proc. Amster- dam: North-Holland, 1954, p- 148. 37. Brady, J. V. J. Psychol. 40: 25, 1955. 38. Brady, I V Science 123: 1033, 1956. 39 Brady, J. V. Ann. New )'»»< Acad. Sc. 64: 632, 1956. 40. Brady, J V. Tr. Neu ) ■ [cad. Sc. 18: 601, 1956. EMOTIONAL BEHAVIOR '547 41. Brady, J. V. In: Brain Mechanisms and Drug Action, edited by W. S. Fields. Springfield: Thomas, 1957, p. III. 42. Brady, J. V. Ann. New York Acad. Sc. 66: 719, 1957. 43. Brady, J. V. In: Experimental Psychopatho/ogy, edited by P. Hoch and J. Zubin. New York: Grune, 1957, p. 20. 44. Brady, J. V. In: Reticular Formation of the Brain, edited by H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay and R. T. Costello. Boston: Little, 1958, p. 689. 45. Brady, J. V. In: Biological and Biochemical Bases of Be- havior, edited by H. F. Harlow and C. N. Woolsey. Madison: Univ. Wisconsin Press, 1958, p. 193. 46. Brady, J. V. In: Electrical Stimulation of the Brain. Subcorti- cal Integrative System, edited by D. Scheer and D. C. Kroeger. Austin: Univ. Texas Press, i960. 47. Brady, J. V. Psychopharmacology : Problems in Evaluation, edited by J. O. Cole and R. W. Gerard. Nat. Acad. Sc- Nat. Res. Council Publ. 583, Washington, 1959. 48. Brady, J. V., J. Boren, D. Conrad and M. Sidman. ./. Camp. & Physiol. Psychol. 50: 134, 1957. 49. Brady, J. V. and H. F. Hunt. J. Comp. & Physiol. Psychol. 44: 204, 1951. 50. Brady, J. V. and H. F. Hunt. J. Comp. & Physiol. Psychol. 45: 180, 1952. 51. Brady, J. V. and H. F. Hunt. J. Psychol. 40: 313, 1955. 52. Brady, J. V., H. F. Hunt and I. Geller. J. Comp. & Physiol. Psychol. 47: 454, 1954. 53. Brady, J. V. and \V. J. H. Nauta. J. Comp. & Physiol. Psychol. 46: 339, 1953. 54. Brady, J. V. and W. J. H. Nauta. J. Comp. & Physiol. Psychol. 48: 412, 1955. 55. Brady, J. V., L. Schreiner, I. Geller and A. Kling. ./. Comp. & Physiol. Psychol. 49: 381, 1956. 56. Brady, J. V., W. C. Stebbins and R. Galambos. J- Comp. & Physiol. Psychol. 46: 363, 1953. 57. Brady, J. V., W. C. Stebbins, and H. F. Hunt. J. Comp. & Physiol. Psychol. 46: 368, 1953. 58. Brobeck, J. R., J. Tepperman and C. N. Long. Yale J. Biol. & Med. 15: 831, 1943. 59. Broca, P. Rev. Anthropol. 1 : 385, 1878. 60. Broca, P. Rev. Anthropol. 2: 385, 1879. 61. Bromiley, R. B. ,/. Comp. & Physiol. Psychol. 41: 102, 1948. 62. Brookhart, J. M. and F. L. Dey. Am. J. Physiol. 133: 55'. '941- 63. Brookhart, J. M., F. L. Dey and S. W. Ranson. Endocrinology 28: 561, 1941. 64. Brooks, C. M. New Zealand M. ./. 46: 243, 1947. 65. Brothers, J. D. Genet. Psychol. Monogr. 53: 75, 1956. 66. Brower, D. J. Gen. Psychol. 34: 47, 1946. 67. Brown, S. and E. A. Schafer. Phil. Trans. B 179: 303, 1888. 68. Cadilhac, J. Hippocampi el Epilepsie. Montepellier: Dehan, '955- 69. Cannon, W. B. Am. J. Psychol. 39: 106, 1927. 70. Cannon, W. B. Bodily Changes in Pain, Hunger, Fear and Rage. New York: Appleton, 1929. 71. Cannon, W B. Lancet 218: 1109, 1930. 72. Cannon, W. B. Psychol. Rev. 38: 281, 1931. 73. Cannon, W. B. and S. W. Britton. Am. J. Physiol. 72: 283, 1925. 74. Cannon, W. B., H. F. Newton, E. M. Bright, V. Menkin and R. M. Moore. Am. J. Physiol. 89: 84, 1929. 75. Cannon, W. B. and A. Rosenblueth. Autonomic Neuro- Effector Systems. New York: Macmillan, 1937. 76. Cannon, W. B. and A. Rosenblueth. The Supersensitivity of Denervated Structures: A Law of Denervation. New York: Macmillan, 1949. 77. Chapman, W. P., H. R. Schroeder, G. Geyer, M. Brazier, C. Fager, J. Poppen, H. Solomon and P. Yakovlev. Science 120: 949, 1954. 78. Chapman, W. P., H. Solomon, H. Schroeder, M. Brazier, G. Geyer, C. Fager, J. Poppen and P. Yakovlev. Ill Internal. Congr. Intern. Med., Proc. Uppsala: Almqvist, 1956, p. 66. 79. Chow, K L. A.M. A. Arch. Neurol. & Psychiat. 71: 762, '954- 80. Clark, W. E. LeGros, J. Beattie, G. Riddoch and N. M. Dott. The Hypothalamus : Morphological, Functional, Clinical and Surgical Aspects. Edinburgh: Oliver, 1938. 81. Clark, W. E. and M Meyer. Brain 70: 304, 1947. 82. Claude, H. and J. Lhermitte. Presse mid. 25:417, 1917. 83. Cobb, S. In: Personality and the Behavior Disorders, edited by J. Hunt New York : Ronald, 1944, p. 550. 84. Courts, F. A. Psychol. Bull. 39: 347, 1942. 85. Cox, E. B. Tumours of the Brain: Their Relation to Patho- logical Sleep and other Changes in the Conscious State. M. H. Australia, 1937. 86. Critchley, M The Parietal Lobes. Baltimore: Williams & Wilkins, 1953. 87. CrsiiiNG. H. Papers Relating In the Pituitary Body, Hypo- thalamus and Parasympathetic Net am System. Springfield: Thomas, ig ;-■ 88. Darrow, C. W. Psychol. Bull. 33: 73, 1936. 89. Darrow, C W, Physiol. Rev. 23: 1, 1943. 90. DARROW, G W. In: Feelings and Emotions, edited by M. L. Kivmcrt. New York: McGraw-Hill, 1950, p. 247. 91. Darrow, C. W. Comp. Psychol. Monogr. 20: 51, 1950. 92. Darrow. C W. Ann. New York Acad. Sc. 56: 289, 1953. 93. Darwin, C. The Expression of the Emotions in Man and Animals. London: John Murray, 1872. 94. Davis, R C ./. Exper. Psychol. 23: 141, 1938. 95. Davis, R. C. In: Methods of Psychology, edited by T. G. Andrews. New York: Wiley, 1948, p. 391. 96. Delafresnaye, J. F. (editor). Brain Mechanisms and Consciousness. Oxford : Blackwell, 1 954. 97. Delay, J., Ii I.aine, J. Puech and J. Clavreul. Encephale 4:289, 1953. 98. Delgado, J. M. R. Yale J. Biol. & Med. 24: 351, 1952. gg. Delgado, J. M. R. J. Neurophysiol. 18: 261, 1955. 100. Delgado, J. M. R., W. W. Roberts and N. E. Miller. Am. J. Physiol. 179: 587, 1954. 1 01. Dews, P. B. J. Pharmacol. & Exper. Therap. 113: 393, 1955. 102. Dews, P. B. J. Pharmacol. £? Exper. Therap. 115: 380, 1955. 103. Dews, P. B. Ann. New York Acad. Sc. 65: 268, 1956. 104. Diethelm, O. A.M. A. Arch. Neurol. £? Psychiat. 36: 342, >93°- 105. Diethelm, O., E. J. Doty and A. T. Milhorat. A.M. A. Arch. Neurol. & Psychiat. 54: HO, 1945. 106. Dinsmoor, J. A. Psychol. Rev. 61 : 34, 1954. 107. Dinsmoor, J. A. Psychol. Rev. 62: 96, 1955. 108. Dollard, J., L. W. Doob, N. E. Miller, O. H. Mowrer .-,.»» II WIUli II iK i 'I I'HVSIl ll ( )(A \I IROPIIYSIOI.OGY III AND R R Skars. Frustration and Aggression. New Haven: Yale Univ. Press, 1939. 109. Duffy, E. J. Gen. Psychol. 95: 283, 1 941. 110. Dunbar, H. F. Emotions and Bodily Changes (4th ed.). New York: Columbia, 1954. 111. Dunlap, K. (editor). The Emotions. Baltimore: Williams & Wilkins, 1922. 1 12. Dusser de Barenne, H. G. Arch. need. Physiol. 4:31, 1920. 113. Ei.i.incson, R. J. Psychol. Pull. 53: 1, 1956. 114. l.siis, \V. K. Psychol. Monogr. 57: No. 263, 1944. 115. Estes, W. K. J. Exper. Psychol. 38: 173, 1948. 116. Estes, \V. K. and B. F. Skinner. ./. Exper. Psychol. 29: 390, 1 94 1. 117. Kvarts, E. V. and H. W. Nissen. A.M. A. Arch. Neurol. g Psychiat. 69: 323, 1953. 118. Faure, M. J. and M. Portmann. Rev. laryng. 76: 168, ■955- 119. Feindel, W. and P. Gloor. Electroencephalog. & Clin. Ncurophysiol. 6: 389, 1954. 120. FERRIER, D. Phil. Trans. 165: 433, 1875. 191. Fkrster, C. B. Psychol. Bull. 50: 264, 1953. 122. Ferster, C. B. and B. F. Skinner. Schedules of Reinforce- ment. New York: Applcton, 1957. 123. Finesincer, J. E. Am. J. Psychiat. 100: 659, 1944. 124. Foerster, O. and O. Gagel. Ztschr . ges. Neurol. Psychiat. 149:312, 1933. 125. Fox, C. A., W. A. McK.ini.ey and H. W. Macoun. ./. Ncurophysiol. 7:1, 1 944. 126. Freeman, W. and J. \V. Watts. Psychosurgery. Spring- field : Thomas, 1942. 127. French, J. D. and H. W. Magoun. A.M. A. Arch. Neurol. & Psychiat. 68: 591, 1952. 128. Freud, S. Collected Papers. London: Hogarth Press, 1924. 129. Frick, F. C. J. Comp. & Physiol. Psychol. 46: 120, 1953. 130. Fuller, J. L., H. E. Rosvold and K. H. Pribram. ./. Comp. v I F. The Frontal Lobes. Baltimore: Williams & Wilkins, 1948. 132. 1'iT 1 '>N, I F. Frontal Lobotomy and Affective Behavior. New York: Norton, 1951. 133. Fulton, J. F. Vale J. Biol. & Med. 26: 107, 1953. 134. FULTON, J. F. and P. Bailey. J. \,i . Merit. Dis. 69: '45. '929- 135. Fit TON, J I \mi F. I). INGRAM. ./. Physiol. 67: 27, 1929. 136. Fulton, J. F. and C. F. Jacobsen. Advances Mod. Biol. / S.S.R.) 4 1:3 1 935. 137. Fulton, J. F., K. H. Pribram, J. A. Stevenson and P, 1) \\ \li li Am. Xrurol. A. 74: 175, 1949. 138. Galambos, R , G. Sheatz and V. G. Vernier. Science 576, 1956. 139. Gantt, W II Experimental Basis foi Neurotic Behai New Y01 k I loeber, 1944 140. Gaskill, II. V. Genet. Psychol. Mono:;!. 14: 177, 1933. 141. Gastaut, II J. physiol., Paris 44 431, 1952. 14.' GaSTAI i. II \mi II CoLLOMB Ann. med. -psycho!. 2: 357, i 1 Gastaut, H. and C. Mm ro Ret neurol 92: 5, 1955. II GaSTAOT H., G MuRIn \m> N. LeSEVRE. Ann med.- '»55- GELLER, 1 M SlDKAN \m> I V, Uluni J. < „mf>. fif Physiol. Psychol 48 1 jo, 1955 46. Gellhorn, E. Autonomic Regulations. New York : Inter- science, 1943. 47. Gellhorn, E. In: Anxiety, edited by P. Hoch and J. Zubin. New York: Grune, 1950, p. 205. 48 Gellhorn, E., R. Cortell and J. Feldman. Science 92: 288, 1940. 40 Gii.di \, 1. 1 '., V. I. Mmlikiusi \nd I) F. Morris. Am. ./. Psychiat. 92: 115, 1935. 50. Glees, P. ./. Seuropath. <£ Exper. Xeurol. 5 54, 1946 51. Glees, P., J. Cole, M Wiutty and H Cairns ./ Neurosurg. & Psychiat. 13: 178, 1950. 52. GLOOR, P. A.M. A. Arch. Xeurol. & Psychiat. 71 : 773, 1954. 53. Gloor, P. Electroencephalog. & Clin. Ncurophysiol. 7 293, '955- 54. Gloor, P. Electroencephalog. 938- 61. Grunthal, E. Monatsschr. Psychiat. 11. Xeurol. 113: I, 1945. 62. Guttman, N. and H. I. Kalish. ./. Exper. Psychol. 51: 79. '956- 63. Haggard, E. A. J. Exper. Psychol. 35: 46, 1945. 64. Harlow, H F. and R. Stagner. Psychol. Rev. 39: 570, '932- 65. Harlow, H. F. and R. Stagner. Psycho, Ret \><: 184, 1933- 66. Harris, G. \V. Adrenal Cortex, Transactions oj the Third Conference. New York: Macy, 1952, p. ",4 67. Head, H. Studies in Neurology. London: Ii.iwdc, Hodder & Staughton, 1920. 68. Heath, R. G. (editor). Studies in Schizophrenia. Cambridge: Harvard, 1 954. 69 Heath, R. G. Psychosom. Med. 5: 383, 1955. 70. Heath, R. G. and B. E. Leach. In: Changing Concepts in Psychoanalytic Medicine, edited In S K.ido and C Daniels New York: Grune, 19")6, p. 201 71. Hebb, D. O. A.M. A. Arch. \ Psychiat. 54: u>. 1945 72 Hum, D. O. Psychol. Rev. 53: 88. [946 Hi 1 i . D. O. Psychol. Rev. 53: 959, ml*' 74. Hebb, D. O. The Organization of Behavior. New York Wiley, 1949 7r). Herrick, C.J. Proc. Nat. Acad. .SV.. Washington ic, -. 1933. 76. Herrnstein, R I \M> W 11 Morsi Am Psychologist 1 ' 448, 1 956- 77 Herrnstein, R.J. \m> W. II Morsi S |6 956 B Herrnstein, l< I md W II. Morse. Science 195 999, ki-,7 7g Mi -s \\ R ll,i ges Physiol. 49 554, 1928. 80. Hess, \\ K /'. \lethodik det lokalisierten Reizung und Ausschaltumg subkortikalei Hirnabschnittt Leipzig ["hiemi 1932 EMOTIONAL BEHAVIOR J549 1 8 1 - Hess, W. R. Das Zwischenhirn: Syndrome, Lokalisationen, Funklionen. Basel: Schwabe, 1949. 182. Hess, YV. R. Diencephalon: Autonomic and Extrapyramidal Functions. New York : Grune, 1 954. 183. Hess, W. R. and K. Akert. A.M. A. Arch. Neurol. & Psychiat. 73: 127, 1955. 184. Hetherington, A. W. and S. W. Ranson. .4m. J. Physiol. 136:609, 1942. 185. Hill, H. E., R. E. Belleville and A. Wikler. Science 120: 153, 1954. 186. Hill, D. and G. Parr. Electroencephalography. New York: Macmillan, 1950. 187. Hinsey, J. C. In: The Hypothalamus and Central Levels of Autonomic Function, edited by J. Fulton. Baltimore: Williams & Wilkins, 1940, p. 657. 188. Hoagland, H. J. Aviation Med. 18: 450, 1947. 189. Hoagland, H., D. E. Cameron and M. A. Rubin. J. Gen. Psychol. 19: 227, 1938. 190. Hume, D. M. and G. J. Wittenstein. Proceedings of the First ACTH Conference. Philadelphia: Blakiston, 1950, P- '34- 191. Hunt, A. and G. Landis. J. Exper. Psychol. 19: 309, 1936. 192. Hunt, H. F. Ann. New Yuri Acad. Sc. 65: 258, 1956. 193. Hunt, H. F. and J. V. Brady. J. Comp. & Physiol. Psychol. 44: 88, 1951. 194. Hunt, H. F. and J. V. Brady. J. Comp. >2? Physiol. Psychol. 48: 305. '955- 195. Hunt, H. F., P. Jernberg and J. V. Brady. ./. Cum/:. & Physiol. Psychol. 45: 589, 1 952. 196. Hunt, YV. A. In: Foundations 0/ Psychology, edited by E. G. Boring, J. S. Langfeld and H. I'. Weld New York Wiley, 1948. 197. Ingram, W. R. Electroencephalog. & Clin. Neurophysiol. 4 397. '95'2- 198. Ingram, W. R, R W. Barris and S W. K wson. A. MA. Arch. Neurol. & Psychial. 35: 1 1 75, 1936 199. Ingram, \Y'. R., J. R. Knott, M. D. Wheatley and T. D. Summers. Electroencephalog. & Clin. Neurophysiol. 3: 37. '951- 200. Jacobsen, C. F. Comp. Psychol. Monogr. 13: No. 63, 1936. 201. James, W. Mind 9: 189, 1884. 202. Kaada, B. R. Acta physiol. scandinav. 183: 285, 1 95 1 . 203. Kaada, B. R., P. Andersen and J. Jansen. Neurology 4: 48. '954- 204. Kaada, B. R , J. Jansen and P. Andersen. Neurology 3 844. '953- 205. Kantor, J. R. Principles of Psychology. New York: Knopf, 1924. 206. Kaplan, M. ./ Comp. 95°- Lichtenstein, P. E. ,/. Comp. &" Physiol. Psychol. 43: 419, 1950. Liddell, H. S. Emotional Hazards in Animals and Man. Springfield: Thomas, 1956. Lilly, J. C. Science 125: 748, 1957. Lindsley, D. B. In: Feelings and Emotions, edited by M. L. Reymert. New York: McGraw-Hill, 1950, p. 238. Lindsley, D. B. In: Handbook of Experimental Psychology, edited by S. S. Stevens. New York: YY'iley, 1951, p. 473. '55° IIWDHOOK OF PHYSIOLOGY M IROPHYSIOLOGY III 249. LmDSLEY, I) B \M> W. S. HUNTER. Ptoc. Xat. Acad. Sc, 286. Washington 25: 180, 1939. 250. Lindsley, D. B. and \V. H. Sassaman. J. Sturophysiol. 287. 1 : 342, 1938. 251. Lindsley, I) B., L. H. Schreiner, \V. B. Knowles 288. and H. W. Magoun. Electroenctphalog. \, I \\ '., J \ I'.i' Mi-, wn M Sidman Endocrinology 60: 74'. '957- 3°5- 270. Masserman, J. H. A.M. A. Arch. Neurol. & Psychiat. 39: 1250, 1938. 306. 271. Masserman, J . H. Psychosom. Med. 3: 3, 1941. 307. 272. Masserman, J. H. Am. J. Psychiat. 98: 633, 1942. 308. 273. Masserman, J. H. Behavior and Xeurosis. Chicago: LTniv. 309. Chicago Press, 1943. 274. Masserman, J. H. In: Feelings and Emotions, edited by 310. M. I. Keyinert New York: McGraw-Hill, 1950, p. 40. 311. Masserman, I \m> C Pechtei Am. J. Psychiat. 112: 865, 1956. 312. j-t). McDuri; \i 1 . \V. .-1h Introduction to Social I'wchology (2nd cd.). Boston: Tiki-, m.t. J13. .177. McTebr, VV. Psychol. Rev. bo: 172, 1953. 278. Mettler, F. A Selective Partial Motion of the Frontal 314. Cortex. New York Hoeber, 1949. 279. Mettler, 1 A. (editor). Psychosurgical Problems. New 315. York : lilakiston, 195 280. Meyer, M. Psychol Rei p> 292 1953 |i6 281. Miller, N. I. In: Handbook oj Experimental Psycho, edited by S s Stevens New York: Wiley, 1 93 1 , p. 435. 317. Ml P. P Bull. 51 : 42, 1954. 183. Mimikis, M. J. ( ' Physiol. Psychol. 47: 187, 1954. 318. Mi.iimv M \m> K. II. Pribram J. Comp. & Physiol. Psychol. 17 14, 1954. 285. MlTTELUANN, B \M> II S J. KAPLAN. J. Comp. ss Physiol. Psychol. 45: 565, 1952. Pribram, K. H. and I. Wi iskrantz. J. Comp. & Physiol. Psychol. 50: 74, 1957. Ranson, S. W. A.M. A. .he//. Xeurol. & Psychiat. 41: 1, '939- Ranson, S. W. and II. W M -< ''«. Physiol. 41: 56, 1939. K ansi hum, S The Ihpothalamus and Sleep Regulation. I pps.il.i : Aluupist. 1 (i |- Reiser, M 1 '., R B Reeves \m> J Ajuhnoton. Psycho- som. Med. 17: 185, 1955. Reusch, I nn J I.. Finesinclr I U I Arch Veurol.& Psychiat. 50: 430. 1943 Rioch, I1 Mi K. Psychiatry 1 : 339. 1938. ROSE, 1 I [955- 323. Rothmann, H. Ztschr. ges. Neurol. Psychiat. 87: 247, 1923. 324. Rucknick, C. A. The Psychology of Feeling and Emotion. New York: McGraw-Hill, 1936. 325. Sartre, J. P. The Emotions. New York: Philos. Libr., 1948. 326. Sawa, M., Y. Ueki, M. Arita and T. Harada. Folia psychiat. neurol. Japonica 7 : 309, 1 954. 327. Sawyer, C. H. Am. J. Physiol. 180: 37, 1955. 328. Schaltenbrand, G. and S. Cobb. Brain 53: 449, 1930. 329. Schlosberg, H. Psychol. Rev. 61: 81, 1954. 330. Schoenfeld, W. N. In: Anxiety, edited by P. Hoch and J. Zubin. New York: Grune, 1950, p. 70. 331. Schreiner, L. and A. Kling. Medical Film, U.S. Army 8-7935. '953- 332. Schreiner, L. and A. Kling. ./ Xeurophysiol. 16: 643, '953- 333. Schreiner, L. and A. Kling. A.M. A. Arch. Seurol. & Psychiat. 72: 180, 1954. 334. Schreiner, L. and A. Kling. Am. J Physiol. 184: 486, '956- 335. Schreiner, L., D. Rioch, C. Pechtel and J. Masser- MAN. ./. Xeurophysiol. 16: 254, 1953. 336. Segundo, J. P., R. Arana, E. Migliaro, J V11 1 \k, A Guelfi AND E. Austt. J. Neurophysiol . 18: 96, 1955. 337. Shapiro, A. P. and P. W. Horn. J. New. & Ment. Dis. 122 : 222, 1955. 338. Sheer, D. and D. C. Kroeger (editors). Electrical Stim- ulation of the Brain. Subcortical Integrative Systems. Austin: Univ. Texas Press, i960. 339. Sidman, M. Science 118: 157, 1953. 340. Sidman, M. Science 122: 925, 1955. 341. Sidman, M. Ami. New York Acad. Sc. 65: 282, 1956. 342. Sidman, M., J. V. Brady, J. Boren, D. Conrad and V Schulman. Science 122: 830, 1955. 343. Sidman, M., R.J. Herrnstein and D. Conrad. ./. Comp. & Physiol. Psychol. 50: 18, 1957. 344. Silverman, J. J. and V. E. Powel. Am. J. M. Sc. 208: 297. '944- 345. Skinner, B. F. The Behavior of Organisms: An Experimental Analysis. New York: Appleton, 1938. 346. Skinner, B. F. Am. Psychologist 8: 69, 1953. 347. Skinner, B. F. Science and Human Behavior. New York: Macmillan, 1953. 348. Skinner, B. F., H. C. Solomon and O. R. Lindsley. ./ NertL vr Menl. Dis. 120: 403, 1954. 349. Smith, G. E. J. Anal. 30: 185, 1896. 350. Smith, W. K. Fed. Proc. 3: 42, 1944. 351. Smith, W. K. Fed. Proc. 9: 118, 1950. 352. Solomon, R. L. and E. S. Brush. In: Nebraska Symposium on Motivation, edited by M. R. Jones. Lincoln: Univ. Nebraska Press, 1956, p. 142. 353. Solomon, R. L. and L. C. Wynne. Am. Psychologist 5: 264, 1950. 354. Solomon, R. L. and L. C. Wynne. Psychol. Rev. 61 : 353, '954- 355. Spiegel, E. A., H. R Miller and M. J. Oppenheimer. J. Neurophysiol. 3: 538, 1940. 356. Spiegel, E. A. and H. T. Wycis. Proc. Roy. Soc. Med. Suppl. 42: 84, 1949 357. Spiegel, E. A., H. T. Wycis, H. Freed and C. Archi- nik. Am. J. Psychiat. 108: 426, 1951. 358. Staudt, V. M. and J. F. Kubis. J. Psychol. 25: 443, 1948. 359. Stein, L. Science 124: 1082, 1956. 360. Stellar, E. Psychol. Rev. 1:61, 1954. 361. Streb, J. M. and K. Smith. J. Comp. & Physiol. Psychol. 48: 126, 1955. 362. Taylor, J. (editor). Selected Writings of John Hughlings Jackson. London: Hodder & Stoughton, 1932, p. 162. 363. Terzian, H. and G. D. Ore. Neurology 5: 373, 1955. 364. Thiesen, J. W. Arch. Psychol. No. 285, 1943. 365. Thompson, A. F. and A. E. Walker. A. MA. Arch. Neurol. & Psychiat. 65: 251, 1951. 366. Tilney, F. The Master of Destiny. New York: Doubleday, '93°- 367. Todd, T. W. and M. E. Rowlands. J. Comp. Psychol. 10: 167, 1930. 368. Tolman, E. C. Psychol. Rev. 30: 217, 1923. 369. Ule, G. Fortschr. Neurol., Psychiat. 22: 510, 1954. 370. Ulett, G. A. and (. Gleser. School of Aviation Medi- cine Report. Proj. No. 21-020-0007, No. 4, 1953. 371. Ulett, G. A., G. Gleser, A Lawler \m> G Winokur. School of Aviation Medicine Report. Proj. No. 21-37-002, No. 4, 1959 ;7- Ulett, G A., G. Gleser, P. Starr, J. Haddock, L. Lincley and A Lawler. School of Aviation Medicine Report, Proj No 2 1 -0202-0007, No. 5, 1953. 373. Verhave, T. J. Pharmacol. Or Exper. Therap. 119: 164, ■957 374. Verhave, T., J I Owen and I H Slater. Progr. Newobiol., Amst. 1958. 375. Verney, E. B. Proc. Roy. Soc, London, ser. B 135: 24, 1947. 376. WADE, M ./. Comp. Neurol. 96: 179, 1952. 377. Walker, A. E., A I". Thompson and J. D. McQueen. Bull. Johns Hopkins Hasp. 93: 65, 1953. 378. Walker, K C. / Exper. Psychol. 31: 312, 1942. 379. Walter, W. G. J. Ment. Sc. 96: 1, 1950. 380. Walter, W. G. The Living Brain. New York: Norton, '953- 381. Ward, A. A., Jr. ./. Neurophysiol. 11: 13, 1948. 382. W'aterhouse, I. K. J. Comp. & Physiol. Psychol. 50: 81, '957- 383. Watson, J. B. Behavior, An Introduction to Comparative Psychology. New York: Holt, 1914 384. Watson, J. B. Psychology from the Stand/mint of a Behaviorist. Philadelphia: Lippincott, 191 9. 385. Weisenburg, T. H. Brain 33: 236, 1910. 386. Weiskrantz, L. J. Comp. . 1947- 391. Welch, L. and J. Kubis. J. Psychol. 23: 83, 1947. 392. Wenger, M. A. and M. Ellington. Psychosom. Med. 5: 241, 1943. ' V,-' I! Willie ink i iK I'HYSIUl OG\ Mil kl ll'HVSI()LOC;V III ;>i; Wengef M A \m> J (: Gilchrist. J Exper. Psychol. 38: 757. '948- 394. Win uikv, M. 1). A.M. A. Arch. Neurol. & Psychial. 52: - 944 395. Win 11, I (: In The Hypothalamus and ('.mind Levels of Autonomic Function, edited by J. Fulton. Baltimore: Wil- liams & Wilkins, 1940, p. 854. Winn horn, J. C, M. R. Kaufman and J. M. Thomas I MA Arch Neurol ^ Psychiat. 33: 712, 1935. 3(17 Williams, A. C, Jr. Arch. Psychol. No. 240, 1939. 398. Wilson, M. P. and F. S. Keller. J. Comp. & Physiol. Psychol. 46: 190, 1953. 399. WlNSOR, \ I. and B. KoRCHlN. ./. Exper. Psychol. 23: 62, 1938. 400. Winsor, A. L. and B. Korchin. J. Gen. Psychol. 22: 25, 1940. 401. WoODUORTH, R. S. AND C. S. SHERRINGTON. J. Physiol. V ■ -34. "9°4- 402. Wl-ndt, W. Outlines of Psychology (2nd ed.), translated by C. Judd. Leipzig: Englemann, 1902. 403. Young, P. T. Emotion in Man and Animal. New York: Wiley, 1943. 404. Young, P. T. In: Methods of Psychology, edited by T. G. Andrews. New York: Wiley, 1948, p. 348. CHAPTER LXIV Attention, consciousness, sleep and wakefulness DONALD B. LINDSLEY Departments of Psychology and Physiology, University of California, Los Angeles, California CHAPTER CONTENTS Neurophysiological Mechanisms Early Neurophysiological Concepts of Sleep and Wakefulness Origins of the Ascending Reticular Activating System Con- cept Hypothalamic-cortical discharge concept Origins and Characteristics of the Diffuse Thalamic Pro- jection System ARAS and DTPS Arousal and Alerting Effects Interaction of ARAS, DTPS, STPS and Neocortex Inhibition and facilitation via the reticular formation Cortical Interaction of Specific and Unspecilic Influences Microelectrodc studies 'Habituation' and Attention The Electroencephalogram in Sleep and Wakefulness Characteristics of the BEG in Wakefulness Characteristics of the EEC in Sleep The Sleep-Wakefulness Continuum EEG and Eye Movement Studies of Dreaming During Sleep Is Learning During Sleep Possible? Consciousness, Attention, Hypnosis and the EEG Consciousness Consciousness and EEG Characteristics Induced Physiological Changes Seizure Patterns with Modilication of Consciousness Temporal Course of Consciousness Attention and the EEG EEG in Hypnosis Summary the terms attention, consciousness, sleep and wake- fulness are a part of our everyday language, and we seem to understand, at least in a general way, what is meant by them. As scientific investigators, how- ever, we are often loathe to attempt a definition of these terms because there are too many unknown or variant conditions involved, or because to limit one- self to a given criterion, or even a set of criteria, may not account for all examples. We are fully aware from our own experiences, as well as those reported by others, that an attentive set or posture toward a given object in the environ- ment docs not always result in awareness and per- ception of the object. Furthermore, experiments have shown that a pre-established set or intention to respond as quickk .is possible to the onset of a specific stimulus does not lead to uniform reaction times, there may be wide variations in response lo successive stimuli. Thus, due apparently to spontaneous fluctua- tions of some process controlling attention, or due to distractions produced by competing stimuli, ideas and thoughts of past experience, one may look direct 1\ .it .111 object and not see it, or overlook the cue to respond to a given stimulus with resulting delay or complete absence of response. The following are .1 lew examples ol some of the vagaries oi attention in which the behavioral attitude orset in.i\ belie the actual attentional and perceptual attitude. The astronomer waiting lor the exact instanl of a stellar transit across his telescope may miss \\ due to a lapse of attention or daydreaming. The radar operator waiting for the unusual and unexpected 'blip' on his radar screen may not be reach' to report it when it appears among other signals, due to peripheral sensory or central ideational distractions. The sonar operator listening for the special 'ping' of the enemy submarine by sound-echo return may be- come habituated to the repeated sounds and noises and miss the critical signal. Thus the temporal course of consciousness or aware- ness may show many vicissitudes. Often these can be determined and assessed by introspection or sub- jective report after the fact, but with possible loss of information and sequence because of fallibility result- ing from attempting to 'observe' both external and internal events. If we require an immediate verbal 1553 '554 II \NDB' » IK i il PlIYMi 'I ' «.\ NEIROPHYSIOLOCY III report upon events as they happen, or a reaction to test stimuli periodically inserted while in process of attending to a series of on-going events, we may interfere with that process or destroy its efficacy. If, by preliminary instructions, we attempt to structure an attitude or set to particular events as they occur, we encourage anticipation and watchful waiting which ma\ interfere with, or seriously limit, the process. Naturalness and freedom of response are curtailed, and association between present and past experience mav become seriously restricted. For main years attention and consciousness were considered proper subjects for psychological study by the method of introspection. In the second and third decades of the present century the behavioristic move- ment in American psychology, fostered by J. B. Watson, E. B. Holt, A. P. Weiss and others, rejected the idea of subjective descriptions of experiences and abandoned the use of the concepts of attention and consciousness. The) substituted description and meas- urement of overt behavioral responses, including vocal, subvocal and other responses of semiovert nature, detectable only by special measuring or re- cording instruments. From the point of view of ob- jective measurement and the establishment of re- liable, though perhaps limited, criteria of response, this was an advance. However, it overlooked the fact that not all stimulation which is capable of ex- citation of, or influence upon, the central nervous system results in overt behavior of an immediate and measurable nature. Although behavioral criteria of sleep and wakefulness, and even of attention and consciousness, may be established, it has become apparent that one cannot depend solely upon these criteria. In fact the very essence of attention and consciousness now seems to reside in shifting proc- esses and states within the central nervous system, some of which are detectable through changes in table I . Psychological States and Their EEC, Conscious and Behavioral Correlates* Behavioral Continuum Electroencephalogram State of Awareness Behavioral Efficiency Strong, excited emotion; Desynchronized : low to mod- Restricted awareness; divided Poor: lack of control, freezing fear, range, anxiety erate amplitude; fast mixed attention; diffuse, hazy; "con- up, disorganized frequencies fusion" Alei t attentiveness Partially synchronized: mainly- Selective attention, but may Good: efficient, selective. fast low-amplitude waves vary or shift; 'concentration' quick reactions; organized anticipation, set' for serial responses Relaxed wakefulness Synchronized: optimal alpha Attention wanders — not forced; Good: routine reactions and rhythm favors free association creative thought Drowsiness Reduced alpha and occa- Borderline partial awareness; Poor : uncoordinated, spo- sional low-amplitude slow imagery and reverie; 'dream- radic, lacking sequential waves like' states timing Light sleep Spindle bursts and slow w.i\ es Markedly reduced conscious- Absent (larger); loss of alphas ness (loss of consciousness); dream state Deep sleep Large and very slow waves <)omplete loss nl awareness 1 no Absent (synchrony but on slow memory for stimulation or time bases 1 . random ir- for dreams regular pattern ( loma [soelei ii ii to irregular large (! pl.t,' loss ,,| consciousness; Absent slow waves little ni mi response to stimu- lation . amnesia Death Isoelectrii ■ adual and pei - Complete loss of awareness as Vbsent M,ii,.i,i disappeai am e "i death ensues all electrical activity 'I I.mdsles i ,<| ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS '555 electrical potentials recorded indirectly and diffusely from the brain, or directly and focally in certain regions of the brain. The electrophysiological changes in the brain, when coupled with behavioral observations and measurements and, in the case of humans, with sub- jective reports, show remarkable correspondences among these diverse criteria. The electroencephalo- gram (EEG) has proved to be one of the more re- liable and less disrupting methods of studying the transition from wakefulness to sleep, and the reverse. It reveals several distinct stages in the process of going from wakefulness to deep sleep. It permits continu- ous monitoring of sleep without disturbance to the sleeper, as may be the case with test stimuli applied to determine the threshold for behavioral or subjec- tive response. Insofar as they are applicable, the criteria of depth of sleep revealed by the latter meth- ods correspond well with those of the EEG. Further- more by combined use of EEG, behavioral and sub- jective methods, it has become possible to view the differential and gradational stages of attention and consciousness upon the same continuum with those of wakefulness and sleep (see fig. i i and table i I. The neural mechanisms which underlie these changes are gradually being revealed, and with this increased understanding have come new concepts of the functional organization of the brain. Awake or asleep, the characteristic rhythms and patterns of electrical activity in the brain appear to regulate not only the discharge patterns of individual neurons and of the aggregates of which they are a part, but of other neurons and aggregates more widely dispersed. Whether such interactions within the brain are ac- complished by direct connections, by resonance ef- fects, by Held effects or by other means yet unknown remains to be determined. Regardless of how the interactions take place, it is known that certain dif- ferential electrical patterns exist at any moment in widely dispersed regions of the brain, and that a certain amount of correlation has already been demonstrated between these electrophysiological pat- terns and the behavioral and subjective indices of attention, consciousness, wakefulness and sleep. NELIROPHYSIOLOOICAL MECHANISMS During the past 30 years electroencephalographic and neurophysiological studies have provided evi- dence that attention, consciousness and sleep depend upon a common neurophysiological mechanism. In broad functional outline this mechanism has been described as the ascending reticular activating system (ARAS), with origins in the reticular formation of the lower brain stem and with upward extensions including parts of the hypothalamus, subthalamus and thalamus. (This system is the subject of Chapter LII of this Handbook.) Integral with, or closely related to, the ARAS is the diffuse thalamocortical projection system (DTPS), with origins in the nonspecific nuclei of the thalamus (discussed in chapter LIII). These neuroanatomically and functionally related neuronal systems of the lower brain stem and diencephalon provide not only a basis for understanding sleep and wakefulness, but also make possible some meaningful correlations between neurophvsiological, behavioral and psychological events which help to define the limits of the various states or gradations of attention and consciousness. The manner in which neuro- anatomical, neurophysiological, electroencephalo- graphic, behavioral and psychological data conjoin 10 support the above statements will be presented in subsequent seetions of this chapter. It is difficult to trace the origin of the ideas which have led lo .1 new concept or theory, or to decide at what critical juncture the accumulated experimental evidence established confirmation of the theory. However, with respect to the functional concept of the ARAS and the role it plavs in sleep and wakeful- ness, as well as in clcctrocortic.il activation and be- havioral alerting, there is little doubt that Moruzzi & M.inoun (1K1), and subsequently Magoun and his collaborators (78-80, 163, 164, 172), clearly and firmly laid the groundwork lor an important new concept of brain organization in relation to behavior. Similarly, following upon the pioneering work of Morison & Dempsey (59, 60, 179, 180) in which they described the verv significant 'recruiting' and 'augmenting' responses elicited in the cortex by stimulation of certain regions ot the thalamus, Jasper and collaborators (95, 96, 126, 130-132) in a series of experiments have outlined some of the functional relationships of the diffuse thalamocortical projec- tion system (DTPS). Jasper (127, 128) has further proposed some important functions which this system, in conjunction with the ARAS, may play in relation to consciousness and attention. Tissot & Monnier (225) and Monnier el al. (176) have recently pro- vided additional information of importance to the understanding of the roles of the ARAS and DTPS. The pictures of both the ARAS and the DTPS have been painted in broad and bold strokes, but the out- lines of the figures are unmistakable. These concepts >-,v> HANDBOOK OF PHYSIOI < '< ,V NEUROPHYSIOLOGY III have Niimulated much new investigation and work goes on apace all over the world. There remains much Inn detail to fill in, and corrections and additions have already been made. There is a need to develop further extensions and modifications of these views so as to encompass more of the higher functions of which man and his near relatives in the animal king- dome are capable. Perception, memory, learning, emotion and motivation arc some of the psychological problems currently being pursued intensively by further investigation of the ARAS and the DTPS, hut especially in ancillary and related systems else- where in the brain stem, limbic svstcm and neocortex. I. ei us attempt to trace some of the earlier and alternative views of sleep and wakefulness, and par- ticularly provide some of the background which leads to the present concepts of the ARAS and DTPS. For some reason, perhaps because adult man spends only about one third of his time in sleep, wakefulness has been traditionally thought of as the natural condi- tion and sleep a deviant, but necessary, recuperative period requiring explanation. Hence the persistent search for a sleep or sleep-regulating center in the brain, a son of magic push button to turn on and off this process. Other conceptions of sleep have em- phasized .1 generalized depression or inhibition of central nervous sxstetn function. Whether a generalized or a local sleep-regulating center was envisaged, the factors responsible for depression or reduction of function in sleep have variouslv been noted as anemia, accumulation of fatigue products and toxins, periodic change in humoral and endocrine action, change in amount and rate ol cerebral circulation, development of a generalized or irradiated internal inhibition, and retraction of dendrites at synaptic junctions. These and other theories have been reviewed by Pieion (195), Ebbecke (68), von Economo (230), Gillespie (91 ), Kleitman (145), Kayscr ( 1 40I and Ploou ( I <)!>). Concise, but relatively comprehensive, overviews ..1 some of the characteristics oi sleep and theories of sleep have been presented bv Wiggers (233) and Morgan & Stellar (1781. Kleitman's (145) book, v, ;i and Wakefulness, is .1 classic in the field Early Neurophysiological Concepts "I Sleep and Wakefulness Several viewpoints have held thai there is .1 block- age "I afferent impulses at some point in the brain which prevents it from being maintained in an active and wakeful state. These have been called stimulus deficiency theories, and some of them resemble in general notion the modern neurophysiological theory based on the ARAS, but do not, of course, distinguish between the specific and unspecific sensory systems and their respective roles Following an epidemic of encephalitis lethargica, with sleep as a prominent sign, Mauthner (175) in 1890 reported clinical and neuropathological evi- dence of swelling and other lesions in the periven- tricular and periaqueductal grey matter of the midbrain. The sleep was attributed to a compression of afferent pathways, cutting off the influx of sensory impulses to the brain. Mauthner generalized upon this conclusion based on patients with encephalitis and proposed a midbrain sleep-regulating center which in some manner regulated the flow of impulses to the brain and accounted for sleep in normal per- sons, von Economo (229), after similar experiences with encephalitis patients following World War I, extended the concepts of Mauthner. Instead of a specific center, he proposed that there is an area reaching from the midbrain through the hypothala- mus to the basal ganglia which is concerned with the regulation of sleep and wakefulness. In general tin- location and structures involved correspond fairly well with the modern neuroanatomical description of the ARAS provided in the extensive review ol the reticular formation bv Rossi & Zanchetti (204). However, von Economo had rather different ideas about its function. He conceived of two centers of control, one rostrally located in the basal ganglai which was thought to be able to inhibit the activitv of the thalamus and cortex and produce disturbances in consciousness and what he called "brain sleep.' The other he located in the midbrain. It was thought to be inhibitory to vegetative and somatic centers in the posterior hypothalamus and lower brain stem, thus giving rise to 'body sleep.' Numerous facts ol normal and patholouie.il sleep lit such a concept, as do some modern concepts of disturbances of consciousness in epilepsy. However, the theory, like so many others, does not make explicit how inhibition or excitation of the centers is effected, nor in what manner these in- fluence central or perpheral functions in sleep .mil wakefulness. Many other proposals have been made concerning the so-called sleep ,md wakefulness centers, for the most part the location was in the posterior hypo- thalamus, midbrain or thalamus, and overlapped in some degree with the ARAS and DTPS sv steins de- scribed above. The details of the individual view- points have been covered in Klcitm. m's lt|V com- prehensive survev. Notable, because of one element ol similarity to modern views, namely cortical acti- vation "I the ARAS, is Kov.us' (149) conception that ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS •557 sleep is due to anabolic depression of sympathetic activity on the one hand and to reduction of psychic activation on the other. The modern neurophysio- logical conception recognizes both a peripheral sen- sory and a central influence upon the activation of the ARAS. Several other views emphasize that in the physiological accompaniments of sleep there is a predominance of parasympathetic influence, in con- trast to sympathetic dominance during wakefulness. Marinesco et al. (174) believe on the other hand that parasympathicotonia is a cause and not a result of sleep, and that endocrine secretions act as sensitizers of vegetative centers. Their view holds that sleep is a conditioned process and that the circumstances favoring sleep (reduction of stimulation) constitute a conditioning situation which gives rise to the un- conditioned humoral and vegetative responses. Tromner (227) and Spiegel (219) have advocated a sleep-regulating center in the thalamus. Spiegel proposed a primitive center of consciousness in the thalamus, and believes that inhibitory and excitatory interaction between thalamus and cortex may in- fluence consciousness as well as sleep and wakefulness. Pavlov (189, 190) has described a variety of condi- tions under which local, and subsequently generalized, inhibition of the cortex develops and leads to sleep Among these are conditions of extinction of a con- ditioned response in which the conditioned stimulus is repeatedly presented without adequate reinforce- ment, or in which delayed reinforcement i> unduly prolonged. Some of these conditions of stimulation are reminiscent of the modern concept of 'habituation1 described by Hernandez-Peon and collaborators (105, 106) but in which the explanation of the effects oi repeated stimulation seems to reside in the brain- stem reticular formation rather than the cortex as Pavlov maintained. There remains much to be learned about the conditioning phenomena described by Pavlov and the 'habituation' effects studied by Hernandez-Peon and others, particularly in terms of their neurophysiological basis. Hess (ill, 112) has described the experimental production of sleep in cats by electrical stimulation in the posterior hypothalamus and along the walls of the third ventricle. These effects were produced by d.c. stimulation which may have produced polariza- tion effects capable of blocking transmission in an area now known to lie part of the ARAS and im- portant to the maintenance of arousal and wakeful- ness. Ranson et al. (199), stimulating with faradic current in the same regions of the hypothalamus, were not able to produce sleep. On the contrary, as would now be expected, thev produced activity and excite- ment, apparently due to excitation of portions of the ARAS traversing the posterior hypothalamic region. The more recent work of W. R. Hess (113—117) and of his son, R. Hess, and collaborators (109, no) has been summarized in Brain Mechanisms and Con- sciousness (118). They now distinguish two systems represented in different areas or fields which give quite different results upon electrical stimulation. The ergotrophic or 'dynamogenic' field is located in the posterior and mesial part of the hypothalamus, and extends into the mesencephalon. Stimulation in this region, which overlaps with the ARAS, is said to produce excitement and arousal which Hess de- scribes as mobilization and preparation for defense. Others from the time of Karplus & Kreidl ( 1 39) have, of course, found this general region favorable to the production of sympathetic nervous system effects associated with arousal and mobilization (18, 138, 200). In contrast to this field of general excitability and alertness, Hess describes another which he calls the trophotropic field or 'hypnogenic' center. This region of the diencephalon lies lateral to the massa intermedia, and extends eaudally to the habenulo- peduncular tract and rostrally to the mammillotha- lamic bundle. Medially il is 1.5 to 2 mm from the mid-line. Its lateral boundaries have not been de- termined due to technical difficulties. Stimulation of (hi-, trophotropic region with slow Iv rising d.c. pulses of 1 to a v and a frequcy of 4 to 12 per sec. causes a generalized depression of activity and sleep, capable of reversal by stronger or higher frequency shocks. The electrocortical activity and behavior induced by low-frequency stimulation in this zone is said to be indistinguishable from natural sleep. As its name implies, its significance is believed to be the preserva- tion of energetic resources, and the reparation and protection of tissues from overstrain. It is thought to function as an antagonist to the arousal system. The trophotropic system of Hess overlaps with the region in which Morison & Dempsey (59, 179) and others (222, 223) have consistently produced 're- cruiting responses' in the cortex at similar frequencies of stimulation, but from which higher frequency of stimulation (50 to 300 per sec.) produces instead cortical desynchrony or activation and behavioral arousal. The fact that both the ergotropic and tropho- tropic systems overlap to some extent with the ARAS, and also the DTPS, would easilv account for the activation and arousal responses with high-frequency stimulation. The results of low-frequency stimulation in the trophotropic area are less easily accounted for. They undoubtedly produce recruiting responses simi- lar to those of Morison & Dempsey with corre- '558 HANDBOOK OF 1'IIYSIOI i >f ; Y NEUROPHYSIOLOGY III sponding synchronized cortical waves of high voltage, perhaps favoring sleep. It would also seem possible ih.it the slowly rising d.c. pulses might block the influences of the ARAS in maintaining wakefulness, by electrotonic or polarization effects. It has been argued l>\ I less that this is not the case since the same type "I stimulus has been applied in the ergotropic center with excitatory rather than depressive and sleeplike results. Nauta's (182) results on sleep in the rat in some respects agree with those of Hess and in other respects are more closely related to the work of Ramon's group. He believes there is a center for sleep in the preoptic and suprachiasmatic regions, and that there exists in the matmnillarv region a center for wake- fulness. According to Xauta the center for wakeful- ness plays a predominant role, and the sleep center of the anterior hypothalamus serves mainly a modula- tory function upon the activity of the center of wake- fulness. Origins of the Ascending Reticular Activating System Concept Our present knowledge of the ascending reticular activating swrm (ARAS) derives from the discovery by Moruzzi & Magoun (181) that stimulation of the reticular formation of the lower brain stem in the cat produces clectrocortical 'activation1 or 'desynchroni- zation' and behavioral arousal. This recognition of the ARAS as a second or unspecilk sensory system which plays a very significant role not only in the regulation of sleep and wakefulness, but as a potential integrator of other important functions mediated by the central nervous system has led to some entircK new concepts ol brain function. Such an important discovery could scarcely have been made had it not been lor certain antecedent experiments and observations by others. Among these were the pioneering experiments of Berger (19- 22) who lirsi successfully described the human elec- troencephalogram in I'l-'f). I lis general findings were confirmed by Adrian & Matthews (6), and subse- quently by a host ol Others. 'See Chapter XI of this Handbook.) His work demonstrated thai the brain has intrinsic rhythms oi its own, among them an 8 to 12 per sec alpha rhythm which ma) be blocked l>\ at- tention and sensory stimulation. He showed that waking and sleeping may !»■ distinguished in the IM, b\ their different patterns. I I\ Ranson, Hess and others for autonomic and sleep and wakefulness centers was beginning to focus attention upon new mechanisms. The insightful observations of Ranson & Magoun (200) concerning the descending and ascending influences resulting from stimulation in the brain stem and hypothalamus were also contributory to new ideas. Finally, Magoun and collaborators (165, 173, 209, 221) had studied the descending effects of stimulation and lesions in the brain-stem reticular formation upon spinal reflexes and motor behavior, and had been able to dem- onstrate a rostral excitatory area and a caudal inhibitor) region. With the importance of the reticu- lar formation demonstrated for downstream effects upon spinal reflexes and muscular contractions, there remained the exploration of possible ascending in- fluences suggested by Ranson & Magoun (200). It was precisely these that Moruzzi & Magoun were able to demonstrate. Figure 1 illustrates schematically the ascending reticular activating system (ARAS) projected upon the monkey brain. Shown here is the pathway of a somesthetic afferent, relaying in the thalamus and proceeding to its destination in the sensory cortex. This represents the specific, primary or classical sensorv system. The unspecific or secondary sensorv system is represented by the ARAS with origins in the reticular formation of the lower brain stem. This is shown by the darker arrow in the central core of the brain stem with multisynaptic relays schematized. The upward extensions of the ARAS in a diffuse .-,(»' HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III manner are illustrated by the arrows directed to all parts of the cerebral cortex. Beginning in the medulla, the reticular formation extends upward through the central core of the brain stem, through the region of the pons, midbrain, hypothalamus, subthalamic and thalamus. Some of its ascending pathways appear to reach the cortex and other forward structures via the internal capsule; others go to the thalamus, especially to the reticular, intrathalamic and other unspecific nuclei. The extra- thalamic route appears to provide the ARAS with direct access to widespread cortical areas, whereas the thalamic route brings the ARAS into relation with the DTPS and possibly with the specific relay nuclei as well. Thus the more direct pathways from the reticular formation to the cortex via the internal capsule constitute a possible mechanism subserving preliminary arousal and general alerting of the cortex to impending messages in specific sensory systems. The thalamic component of the ARAS, in conjunc- tion with the DTPS, may provide a kind of scanning and screening mechanism capable of modifying or regulating the influx of messages to the cortex via the specific thalamic relay systems. It may also aid in controlling the distribution and integration of the messages upon arrival at the cortex. As such it may constitute a specific alerting mechanism capable of sharpening and shifting the focus of attention upon a given sense modality or within a modality. Such possibilities will be further discussed in the next section which deals with the DTPS. To return to the means of excitation of the ARAS, note in figure i that the arrows branching off from tin- classical sensory pathways symbolize collaterals to the reticular formation. Evidence now indicates that even sense modalitv , apart from its specific or primary message-carrying function, also has connections with some part of the reticular formation and is capable of exciting this structure, thus gk ing rive to an unspecific or secondary sensory influence as manifested in the activity of the ARAS. Moruzzi & Magoun (iHi) demonstrated not only that electrical stimulation of the midbrain reticular formation was capable of arousing behaviorally and electrocortically a drowsy, sleeping or lightly anesthetized cat, but that natural sensory stimuli of all types would produce a similar effect. Furthermore in the waking cat the same stimuli led to cortical activation and behavioral alerting, I ile< trical recording directly from the reticu- lar formation has demonstrated that natural sensory stimuli of all types evoke potentials there. These functional demonstrations of the influence of col- laterals from sensory pathways to the reticular forma- tion are in accord with previous neuroanatomical knowledge which has recently been further elaborated by Olszewski (187 1, Brodal (40), Rossi & Brodal (203) and the Scheibels 1206). Still another mode of excitation of the reticular formation has been demonstrated which is of particu- lar importance to the topics under consideration here. Several investigators (2, 3, 38, 104, 120, 185) have shown that stimulation of various parts of the cortex gives rise to potentials in the brain-stem reticu- lar formation, but in particular the work of French et al. (77) may be referred to here. Figure 2 shows schematically both the corticifugal pathways and the collaterals of classical afferent pathways converging upon the reticular formation of the lower brain stem. These of course are by no means the mjIc afferents to this structure, for others arise in the cerebellum, basal ganglia, thalamus, hypothalamus and rhinen- cephalon. For an excellent review of the anatomy and physiology of the reticular formation and its afferent and efferent connections, an extensive article by Rossi & Zanchetti (204) should be consulted. Other valuable surveys of the reticular formation are those by Segundo (211), O'Leary & Coben (186) and French in Chapter FI1 of this Handbook. in. 2 Corticifugal pathways and collaterals "t classical ill, n in pathways convcrijini; on the reticular formation of the lower brain stem. Stimulation of widespread cortical areas gives rise to electric potentials in the reticular formation, hence functional connection l>\ assumed corticorenculai paths. Af- ferent impulses from all s us and impulses originating in the \ are capable of exciting the ARAS, which in turn main- tains the cortex and behavior in a state "l arousal and alert- ness, and perhaps selectivel) controls attention. From French tl al. I ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS .56l HYPOTHALAMIC-CORTICAL DISCHARGE CONCEPT. The point of view of Gellhorn (87) concerning the im- portance of the hypothalamus in relation to sleep and wakefulness, consciousness and attention should also be considered. He has argued on the basis of a variety of experiments that "direct and reflex ex- citation of the posterior hypothalamus is associated with a diffuse excitation of the cerebral cortex. The intensity of this hypothalamic-cortical discharge is directly related to the excitability of the posterior hypothalamus." He states further: '"The hypotha- lamic-cortical discharge is associated with the state of wakefulness. Conditions which interfere with this discharge cause somnolence and coma." Finally Gellhorn places emphasis upon the role of proprio- ceptive and nociceptive influences, hypothalamic imbalance, and corticofugal discharges in the main- tenance of hypothalamic excitability. The excitability and activation of the hypothalamic-cortical system, maintained from external and internal sources, he believes to be responsible for the maintenance ul wakefulness, consciousness and the state of awareness exhibited by perceptual discriminations. It will be recognized that Gellhorn places his main emphasis upon the hypothalamus and particularly the posterior hypothalamus for the regulation and control of states which many have attributed to the diffuse reticular system, including the ARAS and DTPS. There is no real inconsistency here, however, since the reticular substance extends into the posterior hypothalamus and the upward efferent projections from it extend to and beyond the hypothalamus (204). It seems likely that many of the results reported l>\ Gellhorn and his collaborators might well be ac- counted for in terms of the ARAS, inclusive of parts of the hypothalamus. However, the importance of many of the autonomic factors considered by Gell- horn should not be overlooked either from the point of their direct influences upon the cortex or their indirect homeostatic influences. Origins and Characteristic!! of the Diffuse Thalamic Projection System Before beginning this topic it may be well to clarify the terminology to be used. Because of the great complexity of interrelationships among diencephalic structures, and particularly those of the thalamus, it will be necessary to deal with a simplified concept of this organization. Thus, following the terminology used by Jasper (130), we will speak of specific and diffuse thalamic projection systems. The topically organized projections from classical sensory relay nuclei upon somewhat delimited cortical receiving areas will be called specific thalamic projection systems (SPTS), while those shown to have a wide- spread effect upon electrocortical activity, either in terms of activation or recruiting responses, will be referred to as diffuse thalamic projection systems (DTPS). This distinction might be made on the basis of neuroanatomical considerations as well as neuro- physiological, but with many more qualifications and much less simplicity. The dorsally placed association nuclei will be included under the diffuse system, al- though like the specific nuclei they tend to be more topically organized and delimited in their cortical influence. To use the label 'diffuse thalamic projection sys- tem' rather than 'diffuse thalamocortical projection system' takes cognizance of the fact, as pointed out by Droogleever-Fortuyn (65) and Nauta & Whitlock (183), that there is uncertainty that more than a few of the nuclei lumped under the terms diffuse or non- specific actually project directly to the neocortex. Instead, there is indication that many of them project cither to the caudate, striatum or rhinencephalon, or form connections with the dorsal association nuclei, nucleus ventralis anterior and the rostral pole of the reticular nucleus. Among the so-called diffusely projecting nuclei, the reticular complex, according to Rose (201), shows degenerative changes following neocortical ablations, but the temporal character of these is such as to suggest that they may be secondary to degeneration in the dorsal thalamic nuclei. Thus the over-all picture in the thalamus is complex neuro- anatomically, and neurophysiologically it is not as easy to categorize the effects produced in the cortex by stimulation as first supposed since there appears to lie considerable overlapping and interaction of specific and nonspecific systems. Further details of the relationships of the DTPS or 'unspecific thalamo- cortical projection system' are taken up by Jasper in Chapter LI 1 1 in this Handbook. Ramon y Cajal (198) described two-way connec- tions between thalamus and cortex, and saw in this arrangement a possible means of control by the cortex over the sensory influx via the thalamus, thus afford- ing a possible mechanism of attention. Head & Holmes (97), on the basis of clinical neuropathological data, postulated an inhibitory control of the cortex upon the thalamus which was believed to regulate attention and affect. It was some time later, however, before the method of strychnine neuronography, as developed by Dusser de Barenne & McCulloch (67) 1562 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III and used also bv others 1 185 1, was able to demonstrate functional connections between the cortex and specific and nonspecific nuclei. As early as 1933 Bartley and Bishop (16, 25) were suggesting reverberating circuits between the thalamus and cortex to account for synchronization they ob- served between alpha-like rhythms of the cortex and activity in the pathways and thalamic nuclei of the visual system. Subsequently Jarcho (124) and Chang (44) also postulated reverberating or loop circuits to account for interactions they observed between the thal- amus and cortex. Jarcho believed that the reverberat- ing circuit responsible for repetitive after-potentials was a part of the specific projection system, whereas Chang felt thai there were two independent systems, but capable of influencing one another. Chang demon- strated that the second of two click stimuli was able to abolish the recurrent after-potentials which fol- lowed the specific evoked response to the first click in the auditory cortex, and that it reset or established its own rhythmic aftereffects. Only when the interval between the two clicks was properly adjusted to the duration of the rhythmic after-potentials of the first evoked response would it give an optimal response. Thus the responses of the specific system are influenced by the recurrent rhythms of the nonspecific system, and the rhythms of the latter system can be influenced or reset by the activity in the specific projection sys- tem. This may lie an important step in the process ol attention and consciousness, and will be discussed in a later section. Another important neurophysiological clue to tin- control of attention and consciousness derives from the work of Morison & Dcmpsev (59, i;<)> who dis- covered that stimulation of the mid-line, intralaminar and dorsomedial nuclei of the diffuse thalamic pro- jection system .ii frequencies off) to 12 per see. in the cat gives rise 10 a 'recruiting response' in several widespread areas of the cortex. The first three or lour shocks to the thalamus produce a gradually increas- ing magnitude of response in the cortex, hence the term 'recruiting,' implying that more and more units or increments to the field of activity have been brought in 1. 1 synchrony. By about the fifth stimuli is the ampli- tude oi the response stabilizes. If a much higher frequency of stimulation is applied to the same elec- trodes, the recruiting response is abolished and a 1 1, million "i 'activation1 prevails, much like that from stimulation of the reticular formation of the lowei brain stem which brings the ARAS into play Ii is interesting thai stimulation of the mesen- cephalii reticulai formation with shocks ol too to ;<<<< per sec. produces clcctrocortical activation, behavioral arousal and alerting of enduring persistence, whereas similar high-frequency stimulation of the recruiting areas of the thalamus produces the same effects but without long-lasting persistence and with distinctly more limited topographic representation. This again may reflect the relatively different roles of the ARAS and DTPS in the matter of the temporal control and flexibility with respect to attention, the former per- haps determining longer-lasting general states of alertness and the latter modulating these ^i,n,s on a shorter and more variable temporal scale. Other differences in the two systems are reflected in the fact that the recruiting mechanism is not affected by barbiturate anesthesia, either in its low-frequency recruiting response or in its higher frequency activa- tion response, whereas the activation mechanism of the ARAS is seriously disrupted. Still another differ- ence is the tendency, reported by Hess (118) and Akimoto el ol. (8), for low-frequency stimulation in certain regions of the DTPS to produce sleep, whereas similar stimulation in the reticular formation of the ARAS does not. Such neurophysiological differences serve to distinguish the ARAS and DTPS as having different functional roles, although there is also reason to believe that they may work integratively and in some instances may have mutually reinforcing effects. One very important point seems to be th.u the ARAS, except when it is blocked by anesthesia, lias an activation effect which takes precedence over that of the DTPS. This would seem to have ideological Significance, from a protective viewpoint it would appear to Ik- more important for a general alerting mechanism such as the ARAS (o clear the way for any or all warning messages from a threatening en- vironment, than 10 have the central mechanism, represented bv the thai. nuns and cortex, remain pre- occupied with a specific locus ol attention and ob- livious to danger. Further distinctions between the ARAS and DTPS are seen in their reactions 10 drugs and anesthetics. Ronvallet tt al. (29) bv some ingenious experiments demonstrated that epinephrine acts principally on the upper part of the brain-stem reticular formation to produce activ alion of the ARAS, w iih aceoinpanv ing eleCtrOCOrtical activation or arousal. This etlec 1 was shown bv these authors, and by Rothballer (205), to be limited 10 the pontomesencephalic n ticular forma- tion, since lesion ol the ARAS at the junction ol the midbrain and diencephalon prevented epinephrine from having this effect. Thus the upper midbrain ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS •563 ^^v~./\/lVttA/^"AJ^v/' OA w vv^wv/n/^w^ JA^ ~My\7V\rv^ LP X<), an area less able ap- parent!) to activate the ARAN or 10 mobilize atten- tion through the DTPS, although receiving projections limn certain portions of the DTPS, results in no apparent 1 hange in behavior. h has been pointed out b) Walker (231 1 and others thai there is considerable confusion and disagreement about tin- interconnections of thalamic nuclei and particularly the pathways by which unspecific nuclei reach or influence the cortex. Some agreement ap- pears to be emerging gradually from the results of several neurophysiological studies bearing on this problem. In particular those structures in the thala- mus which, when stimulated electrically, produce recruiting responses, activation or both in widespread areas are known. Destruction of certain areas elimi- nates recruiting, natural sleep spindle bursts, and unilateral synchronization of spindle bursts in thala- mus and cortex. Some lesions appear to produce dis- tractihilitv and loss of attention. In the following studies the nucleus ventralis anterior, or at least certain anterior ventral regions, appear to be im- portanl as a possible pathwav of egress of DTPS influences. Lindsle) ti al. (163) demonstrated in acute cat preparations thai transection of the neuraxis at suc- cessivel) higher levels in the brain stem produced increasing amounts of electrocortical ^v nc 1 1 ionization, ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS AFTER PONTILE SECTION ^65 L.ANT R.ANT. L.POST R.POST AFTER BILAT HYPOTHALAMIC LESIONS I .ANT B L.POST. -^~-~>u-WV/\W\/||^j|(^^ jl^ UV^>'W'*~V^^ I R.POST. R. POST fig. 6. Cat with transection at the pontobulbar junction shows partially activated EEG (A), loss of activation and replacement with spindle bursts after caudal hypothalamic lesion v Monnier ..V Tissol (177) in the rabbit. 1 ligh-frequency stimulation ol the reti< ular formation produced the characteristic arousal and alerting sponses behaviorally and desynchronization electro- cortieallv, but with synchronization in the rhinen- cephalon which showed rhythmic 5 to 7 per sec. waves during the conical activation. Both behavioral and electrocortical arousal reactions could also be produced with low-frequency stimuli of 4 per sec. in the reticular formation, but with much longer latency, occurring toward the end or even after stimulation. This suggested to these authors a dual mesencephalic reticular system lor arousal and alerting, one of short latency and one of very long latency. Tissot & Mon- nier (225) and Monnier et al. (176) have observed a similar dichotomy from stimulation in the medial thalamus, again with short and longer latency (6 to 12 msec, and 20 to ;jb' msec). They identify the early response mechanism as ergotropic and related to the reticular system because it is a quick-acting system, with persistent response, which increases with wakefulness and is enhanced by ergotropic alerting drugs. The other slower response mechanism they identity as trophotropic in type because it decreases during arousal, is facilitated by tranquilizing drugs and is identical with the thalamic recruiting system. Mil roelectrode records from cortical neurons seem to bear out their contention that these two svsiems anchored in the medial intralaminar regions of the DTPS are antagonistic and bear a somewhat recip- rocal relation to one another. Tissot & Monnier (225) believe that this reciprocal antagonism probaMv plays an important role in the regulation of vigilance and consciousness. Interaction oj ARAS, DTPS, STPS and Neocortex Although neuroanatomical and neurophysiologic.il details are far from complete, the foregoing surve) of neurophysiological experiments bearing on the ARAS and DTPS seems to support certain generali- zations concerning sleep, wakefulness, consciousness and attention. The ARAS is ., system promoting wake- fulness lis origins are in the reticular formation of the lower brain stem, situated Strategically at the i rossroads of afferent and efferent svsiems from which it receives collaterals. Its upward extensions permit it, either directl) or indirectly, to influence the neo- cortex in a diffuse manner, and in turn to be in- fluenced by the neocortex through widespread corti- c ifugal connections. When excited through its afferent collaterals or by cortieiliig.il connections, it is capable of arousing a sleeping animal or alerting a wakeful one. In so doing il modifies the electrical aeliv itv of the cortex, shifting ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS •567 the pattern from one characteristic of sleep with large slow waves and spindle bursts to one of low- amplitude fast activity or, in the case of wakefulness, further differentiating the pattern by desynchroniza- tion and activation. If its activity is reduced by anes- thesia, or if its upward extensions are cut off from the lower brain-stem portion of the reticular formation, a state of somnolence and unconsciousness ensues, despite the fact that sensory messages may still traverse the classical pathways to the thalamus and cortex. Even under light to moderate barbiturate anesthesia such messages are ineffective so far as perceptual discriminations are concerned without the influence of the ARAS. The routes of the upward extensions of the ARAS are as yet uncertain, but it is believed that one takes an extrathalamic course, possibly by way of the sub- thalamus and internal capsule, and that another passes by way of the thalamus, presumably terminat- ing in the mid-line and intrathalamic nuclei, in the reticular nucleus or in both. Whatever its influences may be in the thalamus, upon DTPS or SIPS, its ultimate effects upon the cortex arc not to be denied. Its extrathalamic influences are believed to be rela- tively rapid ones and concerned with general arousal, its thalamic influences are more likely concerned with gradations of alerting to attention and may be related to scanning or modulating influences atlei 1 ing the STPS, or the integration of STPS information by the DTPS. The functions and relations of the DTPS are siill less clear than those of the ARAS. Some of the nuclei composing this system apparently serve intrathalamic association functions, others interact with the cortex, basal ganglia and rhinencephalic structures. Those in direct or indirect connection with the neocortex ap- pear to be capable of electrocortical activation or desynchronization, after the manner of ARAS func- tion. Another function appears to be the regulation of so-called spontaneous cortical rhythms, control of after-discharge and after-potentials following STPS action, and the regulation of temporal synchroniza- tion between thalamus and cortex in a fashion per- mitting the development of cortical recruiting waves, sleep spindles and other rhythmic phenomena of the cortex, perhaps including the alpha rhythm of the normal resting EEC As such it may control waxing and waning cycles of excitability in the cortex which could regulate rapid shifts of attention (159). Evi- dence of such excitability cycles has been put forward by Bishop (24), Bartley & Bishop (17), Chang (44-46), Lansing (153) and others. INHIBITION AND FACILITATION VTA THE RETICULAR formation. Cortical or reticular stimulation has been shown in several experiments to be capable of pro- ducing inhibition (63, 84, 89, 92, 94, 100, 141) or facilitation (36, 66, 161) in one or another of the several sensory systems. Inhibition has been demon- strated from the first synaptic relay to the final relay in the thalamus. To take one example, Hernandez- Peon et al. (107I have demonstrated that stimulation of the mesencephalic reticular formation depressed or abolished the secondary wave of the evoked poten- tial in the nucleus gracilis induced by stimulation of the dorsal column. They also demonstrated that the second component of the evoked potential in the trigeminal nucleus, elicited by stimulation of the infraorbital nerve, could be enhanced by destruc- tion of the midbrain tegmentum, thus demonstrating release of tonic inhibitor) influence from the reticular formation. Finally, and important in relation to other recent experiments dealing with the visual system, they show that photicallv evoked potentials in the optic tract, lateral geniculate bod) and visual cortex could be significantly modified by reticular stimula- tion. In both the lateral geniculate bod) and in the visual cortex the second component of the evoked complex was depressed during and following brief reticular formation stimulation. The optic tract showed both depression and potentiation, believed to result from two antagonistic effects of centrifugal dis- charges upon retinal synapses. In contrast to the above results, recent studies |6, ;;, 66, Mii I have shown facilitatory effects in the visual cortex and in the lateral geniculate bodies as a result of reticular stimulation concurrently with excitation of the visual pathways bv photic or optic nerve stimulation. Improved resolution or temporal facilitation in the visual cortex of the cat has been reported by Linclslev (161 1 following reticular stimu- lation. Figure 7 shows a single evoked potential in the visual cortex (VC) to a pair of light flashes 50 msec, apart, but alter reticular formation stimulation the same pair of flashes resulted in two evoked potentials. The dual response continued for about 10 sec. and then the response became single again. Another kind of facilitation has been described by Bremer & Stoupel (36, 37) and by Dumont & Dell (66). These investigators report increased magnitude of evoked potentials in the lateral geniculate bodies and visual cortex to optic nerve stimulation which followed stimulation of the mesencephalic reticular formation. The former also observed enhancement of cortical I568 I1WDBOOK OF PHYSIOLOGY M 1 ROI'HYSIOLOGY III AFTER IO- 12' fig. 7. Limitation of the cortex in a cat in responding to closely spaced brief light flashes, but facilitated by reticular stimulation. Two 20-fisec. Hashes of light separated by 50 msec, and presented once per sec. are responded to as one by the visual cortex 1 VC) until after a 5-sec. period of reticular forma- tion stimulation. For 10 sec. thereafter two evoked potentials appear, and then return to the original single response. This is an example of temporal facilitation. OT, optic tract, LG, lateral geniculate body; VC, visual cortex. [From Lindsley (it,,, 162). potentials to optic nerve stimulation when the centro- mcdian nucleus of the DTPS was stimulated. Thus we see that ARAS and DTPS interaction with STPS, or STPS cortical effects, is capable of facilitation, hut also in some instances inhibition. Similarly, we have noted that centrifugal discharges, due to cortical or reticular stimulation, produce in- hibition in peripheral sensor) relays and in a few instances give rise to potentiation of specific responses. What is the role of these mechanisms? Do they pro- vide a means of selective control of sensory input such as might seem to be required for restricting attention? ( ortical Interaction of Specific ami Unspeciflt Influences The role of the specific I STPS) and unspecific (ARAS and DTPS) sensorv systems is by no means clear at the present time, and yet it is certain that the) must interact if incoming sensor) messages are to be decoded and integrated with past experience in meaningful ways. One important indication of the necessit) for this is the fact that perceptual discrimina- tion will not occur in the absence of ARAS influence, when blocked b) lesion or anesthesia, or even when reduced as in natural sleep, despite the apparent de- liver) of sensor) messages over the specific system as indicated b) unimpaired cortically evoked potentials. Such interaction may be conceived in al least two general ways, one involving individual units and their local relationships, the other involving more general relationships ol ret eptive anil association /ones. With ikI to the first of these there have been hypotheses about the axosomatic terminations of STPS upon cortical units, and axodendritic terminations of the more diffusely arrayed ARAS and DTPS influences. The existence of interneurons between diffuse pro- jections and cortical units has also been proposed (197). Although far from complete, microelectrode studies of cortical unit activity is beginning to supply some information on these questions. With regard to the grosser arcal relationships, there is already some interesting evidence. Starzl & Magoun (222) and others have observed that diffuse thalamic projections, as exhibited by the cortical areas in which the most prominent recruiting responses can be elicited, tend to be limited to the associational cortex of the frontal, cingulate, orbital and suprasylvian portions of the hemisphere in the cat, but with o\ er- lapping of the motor fields. Jasper et al. (133 ) demon- strated a more labile topographic organization, and under more specific conditions observed that recruit- ing responses extended also to sensory receiving areas. Dempsey & Morison (59) originally observed that primary sensory areas, and particularly auditory and visual areas, developed poor recruiting responses. When they applied repetitive low-frequency stimula- tion to specific thalamic nuclei and adjacent regions, they observed another, but perhaps related, phe- nomenon which they called 'augmenting' responses in the cortical zones of specific projection (61, 180). Augmenting responses have a shorter latency than recruiting responses, and can coexist with spontaneous waves and spindle bursts; recruiting responses on the other hand displace spontaneous waves and spindle bursts, suggesting that they involve some of the same cortical elements. There is obviously a need to distinguish more clearly between, or to identify common features among, the following : spontaneous waves of the cortex, spindle bursts, recruiting waves, sensory after-discharge (44), secondarv responses (62, 75), and self-sustained and corticalh spreading or corticocortical responses (4, •202). Xot only must these be distinguished in terms of the elements contributing to them, but also the areas from which they arise The functions which these areas subserve must also be more clearly de- lineated The significance of these areas, their local functions and their thalamic counterparts might derive some meaning it viewed in the light ol the phv logenetic interpretation and approach suggested b) Herrick & Bishop (108) and Bishop (26). From a phylogenetic viewpoint Bishop sees the reticular System as a series of segmental, level-to-level integrating systems. The reiic ul. 11 formation is one of these, the links between ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS !569 thalamus and cortex constitute others, and within the cortex there are still others. Both specific and nonspecific systems get duplicated and reduplicated in the highest forms of evolutionary development of the brain. Thus a sense mode may be represented by three primary receiving areas and two or more asso- ciation areas. The reticular formation, the region of the centromedian nucleus and the older associational cortices are areas in which one might expect con- vergence of sensory influx with multiple sensory representation, as Albe-Fessard et al. (9, 10) have found recently. The intermixture and convergence of the older and newer specific, nonspecific and asso- ciational systems is undoubtedly responsible for some of the difficulty and confusion which exists currently in attempting to identify and separate the fields of primary and secondary evoked responses as well as other patterns of activity which exist on the cortex. According to Bishop, in primates newer association areas act as facilitators or modulators of older asso- ciation areas which he refers to as thalamocortex. It is in this thalamocortex that he believes the two sensor) systems, specific and unspecific, converge. Could it be here that a temporal coincidence and convergence makes possible perceptual identification and discrimination? If the phylogenetic concept of a 'reticular system' is one extending from cord to cor- tex, and this integrating network is connected throughout by multisynaptic junctions, but with seg- mental links with some hierarchical ordering, then throughout the neuraxis it becomes a dominating influence for afferent and efferent regulation, as fore- going sections of this chapter have indicated. But wakefulness, attention and discrimination are prob- ably some of the principal functions of major segments of 'the reticular system,' as represented in the reticu- lar formation of the lower brain stem, the unspecific nuclei of the thalamus and the associational areas of the cortex, respectively. The recent and important studies of Buser et al. (41, 42) and of Albe-Fessard et al. (.9, 10) suggest the existence of not only dual, but triple thalamocortical systems, and convergences within the thalamus and cortex of multiple sensory representations, just as Bishop's concept would seem to require. Using chlora- lose anesthesia or curarized preparations, Buser et al. (41, 42) have found what they call secondary associa- tive responses in two separate association zones in the auditory and in visual fields in response to auditory and visual stimuli. These responses are enhanced by chloralose, but are blocked by barbiturates. This would suggest that thev originate in or are a part of a reticular system. They do not depend upon their corresponding primary projection areas which can be removed or depressed by drugs without loss of response in the well-defined cortical association areas. Destruction of the reticular formation at a mesen- cephalic level did not abolish these secondary re- sponses which suggest that the mechanism is not a part of the ARAS nonspecific system, but does not rule out the DTPS. The secondary association responses are confined to the association zones and can be distinguished from primary responses by their longer latency and greater duration. With respect to alerting and wakefulness, it is especially interesting to note that the association irradiation responses are best observed not during marked activation or arousal, but during lower de- grees of vigilance, with only slight activation or alert- ness. Since these responses can be produced by stimu- lating the specific thalamic nuclei (medial and lateral geniculate bodies) for audition and vision, or from the lateral-posterior nucleai group of association nuclei, it has been proposed that thev depend upon 'collateral' elements which branch from the primary sensory pathway and enter the thalamic association nuclei. Stimulation of the primary nuclei for audition and vision gives rise to both primary and associative responses in the cortex. The appearance of these responses in a certain level of wakefulness and alert- ness implies thai they are related to the process of attention, and 1n.1v have still further significance so far .is perception is concerned. The observations of Albe-Fessard et al. (9, 10) indicate that stimulation of different limbs of the cat and monkey, irrespective of location, causes evoked potentials in the centromedian nucleus and in parietal and frontal association areas, both ipsilaterally and contralateral^. Such nonprimary responses, elicited by somatic stimulations of different origins, have not previously been described. Their latencies and dura- tions are greater than those of primary responses. However, they are individualized with silent zones between. The outstanding characteristic is the con- vergence of stimulations upon a given cortical locus, with all somatic areas having representation in this convergence response. Stimulation of the centro- median nucleus will cause responses in the superior frontal gyrus, but with latencies too long to suggest direct projections. MtCROELECTRODE studies. The response of individual cortical units provides a more detailed evaluation of specific-unspecific interaction. Jung et al. (137) have discovered several types of units in the visual cortex. Some show no reaction to light (A neurons); others '-|7" II WDIlooK OF 1'IIYSloI i ii.-i M I ROPHYSIOLOGY III show reciprocal inhibition and activation (B, C, D and 1. neurons). Of particular significance to the problem ol interaction being considered hen- is the fact that there is a convergence of specific and uii- specihe systems upon the same cortical neurons. Jung's group (7, 51, 137) has shown that the mode of discharge, and the number of light-responsive neurons is altered by stimulation of the nonspecific system. They have found aNo that mosl neurons are subject to joint influence in the form of facilitation, inhibition and occlusion. Furthermore, thalamo- reticular stimulation increases the maximum fre- quency with which visual cortical neurons can re- spond to flickering light. The flicker-fusion rate of individual cortical neurons can be raised considerably by thalamic or reticular stimulation. In contrast to these results, Jasper ( 1 29) and Li et al. (157) ha\ e not I >< en able to delect cortical unit spike discharges, ex- cept by stimulation of a specific afferent system. However, they do record an intracortical slow- nega- tive wave upon which the spike response of a single unit may be superimposed; and during repetitive stimulation of the nucleus centralis lateralis, multiple discharges of the same unit may be superimposed upon the resulting slow-wave recruiting responses. This undoubtedly means that facilitation of the unit has taken place due to stimulation of the unspecific system. Thus their results, though not strictly in ac- cord with those of Jung's group, may nevertheless indicate some of the same kinds of influences. Jung et al. (137) point to a parallel between the response of individual conical neurons to flickering light at fusion level and above, when accompanied bv stimulation of the ARAS, and the fact that human fusion level mav be elevated by increased attention and alertness. They state that: "Activation and in- creased frequency of cortical neurons in response to peripheral stimulation after excitation through the nonspecific system thus finds a parallel in subjective experience and uggests a possible neurophysiological mechanism for the regulation of attention." These workers believe thai the nonresponding A neurons constitute about hall' of the cortical neurons in ihc optic cortex. The A neurons are thought to receive thru activation not from the specific visual i' la-, nut lei, but fi assoi iation fibers or the medial thalamic nuclei. They look upon them as serving a Stabilizing background function which is capable of ind adjusting the excitation level in the cortex. This mav verve ., protective function in the of massive, vei/ute-like discharges, or il might serve .\]) attention-regulating function under lesser and more differential stimulation. Thus interaction of ARAS and DTPS with STPS at the level of the cortex might play a significant role in specific alerting and attention. What controls and regulates the specific shifts of attention still remains a problem. Is this a mass influence imposed bv ARAS or DTPS, or is there a certain degree of topographical represen- tation, if so, what particular portions of the reticular formation and what particular nuclei of the DTPS are involved? Jung and fellow workers speak of A neurons acting like inierneurons which might be both inhibitory and f'acilitatorv in their influence upon other neurons subserving STPS functions. Jasper ( 129) observes that a single cortical unit will respond at a higher frequency if preceded by a con- ditioning shock to the unspecific svstem, but reports also that from extracellular recordings he has found certain units which will manifest inhibition or facili- tation by either unspecific or specific intervention. He calls attention also to the possibility that unspecific influences mav affect the excitability of the cells possibly by changes induced upon the dendrites. This might suggest that the unspecific system (ARAS or DTPS) could have a modulating influence upon the excitability of specific cortical synapses. Again the question may be raised, which synapses and by what control? At the moment the possibilities are manifold. More and more specific mechanisms are being dem- onstrated which add flexibility to the means ol con- trol and at several different levels of sensory influx from receptors to cortex, but the manner in which the control is exerted still remains elusive. Is there suffi cienl topographic representation in the reticular for- mation or the nonspecific nuclei of the ih.il.nnus to accomplish this? <.'.n\ the degree of differentiation of individual units already demonstrated there ui, 12, 207) be modified and regulated sufficiently bv con- ditioning, learning and habituation? The next section provides some partial answers to such a question. 'Habituation' and Attention As was pointed out above in connection with the work of Jung's group, approximately 50 per rent of the cortical units studied bv microelectrodes in the visual cortex of the cal were unresponsive to light stimulation upon die retina. This group of neurons they looked upon as having a stabilizing influence upon aelivitv in the visual cortex, and perhaps plav- ing a role in visual response mainly under certain kind- ot conditions in which arous.il, alerting .m(\ attention were involved. Ii is interesting thai Erulkar ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS 1571 et al. (72) have observed a similar thing in connection with the auditory cortex of the cat where they find that about 34 per cent of the units isolated by micro- electrode methods could not be activated by sounds. Galambos and collaborators (1 19) in studying single- unit responses of the auditory cortex of unrestrained and unanesthetized cats have also encountered units which are unresponsive to sounds, unless the cat is 'paying attention' to the sounds. They have referred to these as 'attention' units since they are so obviously involved only in response to sound when the sound commands the attention of the animal. They found that about 10 per cent of the units examined were units which responded only when attention was simul- taneously manifest by the animal's behavior, such as turning of the head toward the source of the sound, and appearing to be alert and attentive. In Pavlovian terminology this might be called an 'orienting re- sponse' or orienting behavior. Such attention units characteristically would not respond to clicks, tones or noises from a loud speaker on repeated tests, although a new tone or noise might evoke a response the first feu times it was presented. A unit might respond briskly to the appearance of the experimenter and to certain unique stimuli which at- tracted the attention of the cat. For example, the voice of the experimenter, squeaks of a toy mouse, scratching sounds, hissing noises or tapping on a table were effective in causing a unit, otherwise unrespon- sive to repeated noises, to discharge. Clicks from a loud speaker which were ineffective in Bring a unit would do so the moment the experimenter pretended to tap the loud speaker which produced the clicks and thus drew the attention of the animal to the source of the clicks. Hence it was evident that such units, presum- ably 'adapted' or 'habituated' by repetition, could regain functional status by some additional reinforce- ment or 'disinhibition' brought about by an addition or slight change in the stimulating situation. This change would appear to be related to the ARAS or DTPS rather than STPS. Galambos and lellow work- ers conclude "that the neural processes responsible for attention play an important role in determining whether or not a given acoustic stimulus proves ade- quate. Unfortunately attention is an elusive variable that no one has yet been able to quantify. It may be that studies in which cortical unit activity is examined during the course of conditioning and learning will illuminate these matters." Earlier Hernandez-Peon et al. (106) demonstrated how 'attention,' in unanesthetized cats with elec- trodes implanted in the cochlear nucleus, could in- c. ~*s^ — Vf-*~W^ "'">t"*Vv/'v M SEC FIG. 8. Suppression of response in one modality by selective attention in another Indwelling electrodes in the cochlear nucleus of a cat for recording cochlear potentials to click stimuli .1 cat relaxed, cochlear response strong. B: cat re- sponsive and attentive to two white mice in jar, cochlear re- sponse weak. (.': cat relaxed again, cochlear response to click restored. [From Hernandez-Peon et al. 11 06).] fluence the response of the cochlear nucleus to clic ks. Figure 8 shows a regular and uniform response from the cochlear nucleus to clicks when the cat is resting and not apparently attending specifically to the stim- ulus (the obverse of certain cortical units mentioned above which responded only when the cat was at- tentive). When two mice in a glass jar are presented and the cat's orientation and attention is obviously focused upon them, the response of the cochlear nu- cleus to clicks is greatl) suppressed. Olfactory and nociceptive stimuli had similar inhibitory effects upon the auditory response. Finally, when the mice or other distracting stimuli are removed and the cat returns to a resting state, the response of the cochlear nucleus returns to its original magnitude. Here is an apparent example of sensory inhibition imposed upon the sen- sory pathways of a given modality when another com- peting modality becomes the object of attention. Such an inhibitory mechanism imposed at various synaptic sites in thesensorv pathways suggests that the selective exclusion of certain incoming signals mayoccur periph- erally as well as at the cortex or thalamic levels. How- ever, like the centrifugal negative feed-back system mentioned earlier, of which this inhibitory attention mechanism may be a part, the origin of the effect may reside in the cortex or reticular formation of the lower brain stem. The important point would seem to be 1572 IIVXUBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III lhat 'attention' or the process of exclusion of cert. lin sensory messages in favor of others may begin at var- ious levels of the sensory input as well as at the thal- amic and cortical levels. Bv decortication, or by le- sions or transections at critical points in the nervous -wem it should be possible to determine whether the reticular formation, the cortex or both are es- sential to ihis particular form of sensory inhibition. Jasper et ) which succumb to repetition. Using the arousal reaction of the EECi (shift from high-voltage, synchronized slow waves of sleep to low- voltage, desynchronized last waves ol wakefulness a criterion response which disappears upon jo iii [fj repetitions ol an arousal tone (habituation), Sharpless & Jasper studied a great v.nieiv ol characteristics of this phenomenon. Thev conclude that habituation of the arousal reaction is specific to the quality, niodalitv or pattern of .1 given stimulus. Two tvpes ol arous.il reaction were differentiated, a longer Listing one more susceptible to habituation and a shorter lasting one less susceptible. The tunic slow and persistent arousal response is believed to be mediated bv the mesence phalic portion of the ARAS; the taster more differ- entiated arousal response appeared to be dependem upon the DTPS, for specific nonpatterned stimuli the habituation of the activation or arousal response could be detected in the cortex, thalamus .mil mn nlar for- ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS '573 mation, and persisted in the absence of auditory cor- tex. This important experiment, in agreement with other studies, suggests that the tonic, longer-lasting arousal reaction which is subject to habituation by selective but less differentiated stimuli may be a part of a general arousal and attention mechanism sub- served by the ARAS. With its principal locus in the strategically located reticular formation, capable of monitoring input-output functions, it might be thought of as serving a first order protective, pre- paratory and adjustive function in the manner sug- gested by Cannon, Gellhorn, Hess, Ranson and others. The phasic, shorter-lasting arousal reaction, less susceptible to habituation and capable of greater stimulus differentiation and associated less with ARAS than DTPS, would appear to be mediated by a mech- anism better adapted to the special alerting and highly selective attention needed for higher forms of dis- crimination and learning. In the foregoing experiments, as well as in other recent studies (43, 64, 85, 101 103) bearing on arousal, alerting, attention, habituation and condi- tioning, it has become evident that a complex of in- terrelated mechanisms is involved. Specific afferent pathways of communication must be maintained, but it appears necessary also for these to have several points of confluence and convergence within the central nervous system. Specific sensory pathways conduct messages from a variety of specialized and widely dispersed receptors to special destinations in the cortex and elsewhere. Interposed in these path- ways at various strategic locations are relays or syn- apses which are subject to centrifugal influences which regulate and control the sensory influx. Attention, perceptual discrimination, conditioning and learning require, however, more than the de- liverance of specific messages to certain segregated locations, for information from several diverse sources must be brought together if elaboration, correlation and integration are to occur. In conditioning, as we know, one sense mode may come to substitute for another in effecting .1 response, and in order to ac- complish this there must be points of confluence and convergence. The reticular formation of the lower brain stem, the unspecific nuclei of the thalamus and the association zones of the cortex appear to be centers of convergence, as does also the hippocampus. These convergence centers, although interposed at different levels of the central nervous system, appear to be mutually interactive and share some functions in com- How the separate and conjoined sensory influences in these regions of the brain provide the conditions of discrimination, retention, plasticity and modifiability we know to exist will long be the subject of intensive study, but some progress is being made as the results reported here attest. However, neuroanatomicallv, neurophysiologically and behaviorally, we have barely scratched the surface and much remains to be done. So far in this chapter we have considered mainly the neurophysiological and electrophysiological studies which have revealed new mechanisms and new con- cepts of brain function. We shall now turn to another class of electrophysiological events, the electroenceph- alogram 1 1 . 1 .< . Of more global character it represents ongoing activity in the brain recorded at a distance from the surface of the scalp. THE ELECTROENCEPHALOGRAM IN SLEEP AND WAKEFU1 NESS Charat Im slu s of the EEG in Wakefulness Hans Berger, ;i German neuropsychiatrist, is aptly called the father of electroencephalography. In his first lew articles ( 1 <» 22), he described the basic types of electrical activity generated in the brain of human subjects and recordable from the surface of the scalp. These are described by Walter in Chapter XI of this 1 1 11 ml hook. The normal waking EEG is characterized by a prominent alpha rhythm composed of waves of 8 to 1 2 per sec. and about 30 fiY. on the average. The per- sistence, rhythmicity and regularity of the alpha w ,i\ is \ .i! \ hum individual to individual. In a state of quiescence with eyes closed the alpha rhythm is more or less continuous but often shows amplitude modula- tions. Like the ARAS, the alpha rhythm is affected by all types of sensory stimulation. About two fifths (il a second after an unexpected stimulus, the alpha waves block and the EEG may remain in a state of activation with only low-voltage fast or base-line activity for several seconds or longer until the subject has become accustomed to the stimulation. After a lew repetitions of the stimulus the alpha blockade lasts only about 1 sec. before the waves return, and after a number of repetitions habituation may set in with the result that no blocking or suppression of the alpha waves occurs. Visual stimulation is less suscep- tible to habituation than auditory or other types. Berger believed the alpha blockade resulted from the focus of attention upon a specific sense mode with generalized inhibition spreading to other sense zones. '574 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III Adrian & Matthews (6) and more recently Adrian (5) also adhere to the view that attention to a stimulus or even attempts to see or hear a stimulus not actually present is responsible for the disappearance of the alpha waves and their replacement with low-ampli- tude fast and irregular waves. This type of reaction we now call an 'activation' or 'arousal' response and attribute it to increased activity in the reticular sys- tem, including the ARAS and DTPS. However, it iii.n also be a function of selective attention since the waves block with sensory stimulation, mental arith- metic, or attempting to visualize a scene or recall a melody. We need measures of activation and attention, and the EEG offers promise in this regard. Behaviorally a person may feign sleep and deceive an observer but not a trained EEG observer, for his record would show alpha waves or some degree of activation or arousal responses. He could not feign sleep charac- teristics. If a person is supposed to be awake and alert and fixating an area in order to make observations, one would expect the EEG record to show either a condition of activation with low-amplitude fast waves or alpha waves if he has become somewhat habituated to the condition. If either picture shifts to slow and sporadic alpha waves or to even slower delta waves, or mixtures of the two, one can predict quite reliably that the subject is not alert and attentive, but on the contrary is drowsy. Usually the subject will admit dosing or drowsing if questioned at that time; or if a stimulus for reaction-time measurement is given at such a time, it will either be missed entirely or the reaction will be slow. Although the alpha rhythm averages about 10 waves per see. in older children and adults, it has a much lower frequency in very young children. During the waking slate in newborn infants there is no organ- ized or persistent alpha rhythm over the sensor) areas ill the brain. By ;j or 4 months the sensory zones de- velop a persistent alpha rhythm of about 3 or 1 waves per sec I hese increase to about 5 or 6 per sec. at 1 year of age and by to or 1 2 years oi age have reached the adult frequency of to per sec. Figure c| shows the longitudinal development of the occipital alpha rhythm during the waking st.ite in the same child Imiii infant v tn adulthood. Prior to the development ■.I .111 alpha rhythm it is difficult to attract and main- tain the attention of a child of 1 or 2 months of aye. After the child develops an occipital alpha rhythm, the waves can be blinked bv a visual stimulus, indicating that activation via ARAS is possible Before the in- fant develops an alpha rhythm in the occipital area, 50„«I ■ v^^Y^V/ 2iy»«" 'VW^^f\^W^^JVYV\^^v^^^/\^VVv^^ fig. 9. Longitudinal development of occipital alpha rhythm in the same subject from early infancy to -M years of age. Alpha rhythm typically appears in a persistent pattern by the third or fourth month, with a frequency of 3 to 4 waves per sec. By 1 year the frequency is r, to 6 per sec, and by 10 years it may attain the adult average frequency of about 10 per sec. Note that as frequency increases, voltage diminishes. repetitive photic stimulation will induce a rhythmic activitv in synchrony with the stimulus, but only in the immediate range of the natural alpha rhythm which would appear at ■] or 4 months of age. There is some reason to believe that the alpha rhythm is main- tained or synchronized bv unspecific nuclei of the DTPS but is ea|)able of activation bv either ARAS or DTPS. Thus the establishment of a rhythm in a voting infant must mean that the sensory /one involved has taken on a new functional role. The convergence of specific and unspecific influences upon a cortical zone in which rhythmicity and activation properties have been attained indicates that perceptual capacity is also now available. This extension of sensory and perceptual capacities to the cortex implies that a new level of consciousness has been attained as well. At- tention, perceptual discrimination and consciousness are matters of degree, level, temporal relations and perhaps other factors. One wonders about the young infant with slow alpha waves ,\\k\ slow development 11I an evoked potential following stimulation. I o w hat ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS '3/0 extent does the slower time constant of these systems determine the speed of attentive adjustment, of per- ceptual discrimination and so forth? With an alpha frequency of 5 per sec. at 1 year of age, alpha blockade time or latency is twice as long as it is when the alpha frequency has attained 10 per sec. at 10 years or later. At the present time beta waves of 18 to 30 per sec. and under 10 fiv. on the average, and gamma waves of 30 to 50 per sec. and even smaller, hold little prom- ise for use as analytical devices. Both are easily con- fused with muscle potentials and other low-level ar- tifacts. Neither is responsive to sensory stimulation directly as is the alpha rhythm. Beta waves are more easily observed over the frontal half of the head, and over motor and frontal association zones where alpha waves, though present, are not as prominent as pos- teriorly. Beta waves are said to respond to motor movement. Theta waves of 4 to 7 per sec. are seen mainly in temporofrontal regions in children but are often asso- ciated with behavior disorders or other physiopatho- logical instabilities. Delta waves ranging from Less than 1 to 4 per sec. are not seen in the EEG record of normal subjects unless they are drowsy or asleep Slow and large delta waves are usually considered in- dicative of pathology or physiopathology. Characteristic*, of the EEG in Sleep The EEG undergoes striking changes in pattern in the transition from wakefulness to sleep, and has be- come one of the more convenient and reliable ways to assess the state of wakefulness 01 sleep. Berger (22) observed that alpha waves slow and are reduced in amplitude and eventually disappear as sleep develops. Loomis el al. (i66-i6g) studied all-night sleep records and noted not only the abolition of alpha waxes as light drowsiness gave way to deep drowsiness, but that a series of distinct patterns emerged as various stages of sleep ensued. In particular they called at- tention to 12 to 15 per sec. sleep spindles and the ap- pearance of random slow waxes as actual sleep began. Growing out of the work of several groups of investi- gators concerned with the study of stages of sleep (-28. 53. 99' '44. '47. !48> l69> >7°> 2I5> came a fairly consistent and systematic picture of the changing patterns in the transition from wakefulness to sleep. Brieflx' these stages may be described as follows. Stage . 1 (Awake). Alpha waves are present at the start Inn diminish in xoltage and amount as the subject shows slight drowsiness. Stage B {Drowsy). Alpha waxes di- minish still further and vanish leaxing the base line relatively flat or with loxv-voltage fluctuations and sporadic delta waves. Stage C (Light Sleep). Fourteen per sec. spindle bursts develop on a random, low- voltage, delta-wave background. Stage D (Medium Sleep). Delta waves decrease in frequency but increase in voltage and amount; the spindle bursts disappear. Stage E {Deep Sleep). Delta waves increase in duration, voltage and randomness. The transition stage between B and C is one of going from fluctuating awareness or deep drowsiness to complete lack of awareness or loss of consciousness. During Stage C the subject makes no perceptual discriminations and has no memorv for exents or sensory stimulations, unless these are intense enouuh (o wake him up. Each of the aboxe stages can be further subdivided on the basis of certain criteria in relation to the EEG pattern as Simon & Emmons (215) have done. The above classification is based mainly on that of Loomis et al. ( 169), Dax is et al. (53) and Simon & Emmons (21 j 217), each of whom pre- sents illustrations and detailed descriptions. The transition from sleep to wakefulness follows a similar though more variable course in the reverse direction. Changes ma) be more precipitous from one stage to another. With the return of consciousness on waking, alpha waxes are present in reduced amount .Hid xulusje, and max not assume their regular form and character for some time. Henry (99) found that EEG patterns of individuals on waking tend to be more homogeneous and alike than at any other time. As a few hours intervene after waking, the EEG records take on greater individuality. Ontogenetically, as shown in figure to, the EEG sleep pattern emerges with time. These tracings taken at different times from the same child indicate that in light to moderately deep sleep the EEG sleep pattern corresponding to these stages of sleep has not clearly emerged during the first to to 15 days, although there are evidences of incipient delta waves and even spindle bursts in the motor region. B\ [05 days the general pattern of spindle bursts and random sloxx wave ac- tivity is clearly evident, as it also seems to be at about 30 days. The same general type of pattern is also seen at 280 days and thereafter for light to moderate depth of sleep. Others who have studied the EEG of new- born and young infants are Smith (218), Hughes et al. (121, 122), Xekhorocheff (184) and Ellingson (7>). The Sleep-Wakefulness Continuum In attempting to relate the various stages of the EEG pattern to corresponding psychological states '-.:'' II vxiiln « >K ' '1 PHYSIO! "<."> Mi koi'in mi .1 nc;v III II DAYS ik. I.. EEG sleep records in the same , had .11 iiiili -l. hi agea I he firsi io to i g days sliuu little activity characteristic "I sleep beyond the age of i month. Note cln- incipient slow waves and, in the motor and frontal i, , ords in mi the stai i. an ini ipient sleep pindle burst. Records at /'<; and jHd rfayj ari composed "I slow waves and periodii spindle bursts, and differ little from those of a i- in 2-month-old child, 0, occipital; /'. parietal; M, motor; A, frontal. [From l.ind- sley Si Ellingson, unpublished data obtained ,m l in Cradle Society, Evanston, Illinois, 1949. nouv\ 280 DAYS M^A^W^^MiAli^V/^ and their behavioral correlates Lindsle) (159) de- veloped the concepl ol a continuum (see fig. 11 and table 1 I. Figure 1 1 shows some of the stages of the nor- mal EEG extending from an activated or excited state with low-voltage fast activity .is in arousal or alerted states to deep sleep with large, random slow waves. During .1 relaxed state of wakefulness more or less con- tinuous, amplitude-modulated alpha waves are char- acteristic. In drowsiness alpha waves diminish and low-amplitude slow waves begin to appear. In light in moderate sleep spindle bursts and slow waves are conspicuous and in deep sleep only large and random slovt waves are seen [able 1 lists the full range of EEG stages against .1 behavioral continuum, and the corresponding states ,.i awareness and behavioral efficiency. Ii should be noted that the most aroused or excited State hchavior- ,ilK is represented by a low-voltage fa 1 EEG picture, .ind is paralleled by pooi attention and behavioral efficiency On the other hand a slight!) less ictivated state, corresponding to alert attentiveness, favors se- lective and shifting attention, and behavioral 1 1, iency. The optimal alpha rhythm occurs in a re- laxed state of wakefulness where attention is not fixed Iheie are indications thai awareness 01 con- sciousness shows its broadest sweep or scan and per- haps its lowest threshold in the stage of relaxed wake fulness and optimal alpha rhythm. The threshold is elevated and the held restricted as drowsiness and lighl sleep develop. Il is pioh.iliK in ihe Stage ol light sleep that consciousness is lost, since in thai stage there 1- lack ol awareness ol things going on externally and la< k ol ability to make perceptual discriminations. r,. yond thai poinl in the due, tion ol deeper sleep there is no awareness or ineniorv lor events, except tAWIIKH RELAXES OXOW S T DEE' SLEEP in. 11. Normal EEG records characteristic of different stages on the sleep-wakefulness continuum (see table t) ["hi only major omission in the series would be between excited and relaxed, where there should be a low-voltage record resembling excited, but with less marked activation, labeled attentive. [From Jaspei 11.', possibly for dreams which usually are accompanied l>\ a momentary elevation of the pattern of the I'.I'Xi toward wakefulness and alpha rhythms. There are still manv questions about dreams, and recent work on this topic will be taken up subsequently. With respect to consciousness and attention, shifting upward on the continuum from relaxed wakefulness with alpha waves lo alert attentiv eness and an acti- vated or desv nchroni/ed I.I.Ci with low -amplitude fast waves, attention m. iv he heightened or focused ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS '577 but with definite restrictions upon the span of con- sciousness. In other words the threshold to slight change in a restricted field of attention may be low- ered, but the threshold of awareness for events outside that field may be greatly elevated. This is the role of selective attention for which we sought neurophysio- logical mechanisms in the functioning of the ARAS and DTPS. In excited emotion and in situations where strong stimuli arc in abundance (sensory overload), there are indications that the ARAS or DTPS may be swamped or blocked so that awareness is either se- verely restricted or is so broadened as to be useless. Everyday experience attests to the confusion and the unreliability of observations and testimony made under stressful conditions. Higher-voltage last ac- tivity usually characterizes this slate in contrast to the lower-amplitude activation pattern of alert atten- tiveness. Although sometimes viewed as a kind of epiphe- nomenon, the EEG representing signs of changing cortical activity seems to mirror changes going on in subcortical mechanisms such as ARAS and D I PS. In some respects and for some purposes this ma) be more important to know than what one cortical unit among to billion may be- doing, or low any small sample of units may be responding, vet we need both kinds of information. How docs die EEG reflect changes in subcortical mechanisms? We have seen how it changes from wakefulness t < > sleep and some "I these changes we can account for in terms of ARAS and DTPS activitv when modified l>\ stimulation, lesions or drugs. Alpha waves ma\ be suppressed by natural sensory stimulation or l>\ stimulation of the reticular formation. In the course ol ulavition and quiet alpha waves recede, spindle bursts and slow- waves appear, consciousness is lost and sleep pervades. Blocking the ARAS by lesion or barbiturates shifts the REG pattern from one of activation or alpha waves to one of sleep spindles, slow waves and somno- lence. Sleep spindles have been removed in animal preparations by anterior thalamic lesions, a fact sug- gesting that the transient appearance of sleep spindles during the shift from wakefulness to sleep may repre- sent a shift in the dominance or control manifested by certain thalamic nuclei. Stimulation of medial and intralaminar nuclei of the thalamus by repetitive low-frequency shocks produces recruiting waves in the EEG which resemble the slow waves of sleep, and slightly lower-frequency stimulation actually induces sleep in chronic animal preparations. These are some of the changes reflected in the EEG, and judicious use of it as an indicator will continue to help reveal the effects of even more restricted experimental manipula- tions of subcortical structures and mechanisms. As Adrian (5) so well pointed out, we have thought of the cortex as a screen upon which patterns are thrown by different sense organs, but there are even broader problems of attention, perception, recognition and so forth to contend with. The EEG, when further elucidated by analysis of the subcortical mechanisms whose activity it reflects and by microelectrode studies of the substance from which its potentials arise, 111 a.) well be one of our more important tools in further in- vestigation of higher mental processes We siill eld not know the precise source or origin of the alpha waves, but thev are generally believed i<> be summated dendritic potentials. What determines their time COnstanl and rhythm? Is there .1 cortical or subcortical pacemaker? We have- observed that tin it time 1 onstant is different in young children and adults, but so also is die latency and time constant of evoked potentials, as Ellingson (711 lias shown. Are these- differences clue to structure, chemical constitu- tion, metabolism, niaiiiration.il changes, lack of de- velopment and integration of specific and unspei ifi< sensor) mechanisms, or to a combination of facti I low do these characteristics affect the functions sub- served? The approaches t<> these questions must neces- sarily In- broad and mn-t include- phylogenetic, ontogenetic and longitudinal approaches as well as the detailed experimental manipulations in aeule and chronic- animal preparations. In humans, except for limited observations made during operative- pro- cedures on the brain, we are mainly limited to external measurements such as the- EEG, coupled with psycho logical and behavioral assessments. B\ paralleling these with appropriate animal experiments, certain inferences and deductions may be made. EEG mi,/ Ey* Movement Studies oj Dreaming During Sleep Loomis et nl. (169) Blake- & Gerard (27) and Blake it id. (28) have studied the EEG and general boelv motility during sleep Movement of a part of the body or of the whole body during sleep regularly lightens the level nl sleep as indicated bv the EEG. A move- ment during the- />' stage of sleep or drowsiness in- evitably reinstates a short period of alpha waves, and a major movement during the D stage of sleep has frequently been observed to shift the EEG pattern from delta waves to a C, B or even .1 stage momen- tarily. Thus it appears that kinesthetic, proprioceptive and tactual stimulation excites the ARAS, and it in 1578 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III nun changes the EEG in the direction of wakefulness, though usually very briefly. More recently Aserinsky & KJeitman (13) dis- covered that periods of rapid conjugate eye move- ments during sleep wen- associated with episodes of dreaming. Subsequently Dement and KJeitman (55, 56) combined the EEG and objective measures of eye movements for the study of all-night sleep. They confirmed the fact that rapid eye movements occur periodically during a night of sleep and that if the subject is awakened during or immediately after the series of eye movements, in 80 percent of the instances he is able to report that he had been dreaming and give the content of the dream. When random awaken- ings were introduced in the absence of eye movements only 7 per cent reported dreaming and in the ma- jority of these instances the time of awakening fell within 8 min. of the last period of eye movements which suggested that the dream was recalled from the previous eye movement period. Dement & Wolpert (58) and Dement & Kleitman (57) have greatly extended their observations of eye movements, bodily movement and EEG during sleep, and have determined that there are regular cyclic variations in the depth of sleep as revealed by the EEG and eye movement patterns. Three or four peaks of body and eye movements and sleep light- ened to their stage 1 (comparable to Stage B) occur during a night. Awakening the sleeper during or immediately after one of these periods elicits a report of dreaming. The dream duration appears to parallel the actual eye movement duration which may last from about 10 to 30 min. Contrary to previous belief, the evidence suggests that long dreams are not experienced in a matter of a few near-waking seconds, but rather that the duration of the dream and the duration of the period of lightened sleep and eye movements tend to correspond. Body movements often seem to terminate the dream sequence. The content analysis of the dreams indicates that eye movements participate in the dream sequence but that body movements do not. External and internal stimuli were found to be unimportant in influencing the course of a dream. The 'dream time1 revealed b\ eye movements appears to correspond with the real oi actual time required to experience or partici- pat( in the activity revealed b\ the dream if it were Id be relived in real life The further application of these objective methods "i recording eye movements .mil body movements in association with the EE(J, and I he correlation of these with the report "1 I lie dream upon immediate arousal promises enlightenment about the mechanism of dreams where little existed before. Is Learning During Sleep Possible? Simon & Emmons in a critical review entitled 'Learning During Sleep?' (214) raise serious doubt that any of the few published studies, and several unpublished theses, purporting to show that learning can occur during sleep have met the necessary scien- tific requirements which would justify the conclusion that learning has been induced during sleep. After a critical analysis of the experimental design, statistics, methodology and criteria of sleep employed in these studies, Simon & Emmons conclude that all of them had serious weaknesses in one or more of these areas of criticism. They point out that much of the difficulty in evaluating such studies rests upon a definition of sleep, the establishment of suitable criteria of sleep and, perhaps most important of all, whether the period of sleep training was properly monitored throughout to insure that the criteria of sleep were met. According to their own apparently carefully controlled experiments, Simon & Emmons (216, 217) report that learning did not occur during actual sleep (comparable to Stages C and D discussed above). Even when using a less rigid criterion of sleep based on the transition stage from wakefulness to sleep (comparable to Stage B), but with the restriction that alpha waves be absent from the record at least 30 sec. before a stimulus and for at least 10 sec. after- ward, they still found no significant evidence of learning. These results would suggest that the best answer that tan In' given at present to this question is that learning dining actual sleep is not possible. Despite this conclusion main commercial enterprises across the country advertise that sleep learning is possible, and oiler phonograph records or tape recordings oi Foreign languages and other dillicult-to-le.irn and time-consuming subjects [o be plaved during the ionise of a night's sleep or some portion of it. The important question is whether the user oi such a program of sleep learning is being exposed during waking or near-waking slates in which case the result if positive would not be sleep learning but conscious or semiconscious waking learning. Even though the sleep-learning program were not turned on until an hour or so alter going to bed, we know hum the results discussed above in relation to die. mi periods (-,!>, 58) thai there .ne cycles of oscillation ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS 1 579 between deeper sleep and near-waking states through- out the night. In some of the lighter periods (Stage B) when alpha rhythms return and eye movements per- sist for 10 to 30 min. according to Dement & Wolpert (58) there would be considerable opportunity for drowsy, semiconscious or waking learning to occur. Unless the entire period of presumed sleep were moni- tored by EEG and perhaps eye movement recording as well, there would be no way of knowing whether any accretion of learning which occurred as a result of the night's stimulation was due to learning during periods of actual sleep or of relative wakefulness. The experiments of Simon & Emmons, though well controlled with respect to sleep and wakefulness by continuous EEG recording, are subject to the criticism that they involved only one night of sleep-learning trials for each subject. It might be argued that a longer period of exposure would be required; however, the nature of their material was such that it might easily have been learned in one night. In one of the experiments 96 items of information pretested before the night of sleep and read with answers at 5-min. intervals during the EEG monitored sleep, were posttested the next morning for recall and recognition. There was no learning of material which had been presented during the B to E stages as shown by the EEG, performance of the experimental group so exposed being no better than that of a control group given no training. The amount of correct recall increased markedly for those information items presented during periods when alpha waves were present. Since alpha waves when present correlate highly with wakefulness and consciousness, learninu during their presence might well be expected. Since they selected their subjects on the basis of their domi- nant alpha rhythms, the absence of alpha waves for prolonged periods undoubtedly meant deep drowsi- ness and loss of awareness or light sleep, and as a consequence no significant learning. What if any are the reasons one might have to expect that learning would be possible during sleep? Let us examine first the implications and conditions of learning during normal wakefulness, then the conditions which exist during sleep which might be presumed to make learning possible or impossible. In normal waking learning the association of two events simultaneously, or successively in reasonable conti- guity, one or more times under the proper conditions of psychological set or attention and with the proper motivation, generally results in the development of a bond of relationship between the events which upon the appearance of one tends to recall the other. Rep- etition strengthens this relationship. In a conditioned reflex situation, a conditioned stimulus is paired with an unconditioned stimulus until through repetition it gains the potential it did not have originally of releasing the response of the unconditioned stimulus. In both of these instances of learning there is implied perceptual discrimination of the stimuli, which re- quires selective attention, a certain degree of general alertness, and some degree of motivation. Although these and other conditions may be necessary for optimal learning in the waking state, there is no certainty that they would have to obtain during sleep if learning and conditioning were to occur then, since electrical recordings show the cortex, reticular formation, thalamus and hippocampus to be exhibit- ing very different patterns of activity during sleep than during waking. Since specific sensory pathways remain open during sleep, impulses giving rise to evoked responses still reach specific receiving zones of the cortex, and also reach and affect the thalamus and hippocampus. The reticular formation on the other hand is less responsive and the entire unspecific sensory system including the brain stem, thalamus and cortex has a higher threshold of excitability and exhibits a different pattern of electrical activity. Sensory messages reaching the cortex do not result in perceptual discrimination in the absence of unspecific influences. The reason for this is not known, nor is it known precisely where perceptual discrimination occurs. As Galambos & Morgan point out in Chapter LXI of this Handbook it is not known how or where learning takes place in the brain, although it appears from recent studies that the reticular and limbic systems may have as much to do with this as the cortex. Therefore, on the basis of present neurophysiological knowledge about learning, one cannot prejudge nor preclude the possibility of learning during sleep. However, the burden of proof would seem to reside with those who maintain that it can. With respect to the lack of consciousness during sleep, and therefore the lack of awareness of and memory for events which occur during sleep, one might wonder how learning could be expected to occur. Lacey et al. (150, 151), as well as others have been able to produce what might be called 'uncon- scious' conditioning. In studying autonomic responses in two groups of college students, they presented a list of words to which association responses were to be given. A shock accompanied 1 of 40 words each time it was presented. For one group that word was one with a rural connotation; for the other, an urban i 580 HANDBOOK OF PHYSIOLOGY "NEUROPHYSIOLOGY III connotation. Subsequent testing of the words revealed that a conditioned anxiety had spread to words in the list with a rural connotation for the one group and with an urban connotation for the other. Many of the subjects were unaware that they had been shocked on a word belonging to a particular category and that they were responding with autonomic anxiety responses in other words of that category. In this experiment however the subjects were awake and not asleep. They were quite aware of the stimuli as words with meaning, but were 'unconscious' or unaware of the categorization or abstraction process which devel- oped beyond the simple perception of a word. Such extended cognitive conditioning goes on in a brain awake and alert due to shock reinforcement and exists without awareness as a relational process. But what of sleep where the simple perceptual process appears not to develop due to lack of ARAS influence? Can a lower order of perception occur and leave a useful residual in memory? Can dissociation of levels or hierarchies of perception occur during sleep and be redintegrated during subsequent wakefulness? Consciousness, Attention, Hypnosis and the EEG consciousness. The term consciousness has suffered both from the broadness and narrowness of its conception. This is illustrated by a review entitled 'Consciousness Reconsidered' by Schiller (208) and by five volumes cm ering the annual Macy conferences from 1950-54 entitled Problems of Consciousness, edited by Abramson (1). The broad framework of these treatments of consciousness ranges from the anatom- ical, concerned with the seat or locus of consciousness, to the zoological or phylogenetic approach. Also included are philosophical, psychological, psycho- analytic, anthropological, sociological, biological, biochemical, neurological, neurophysiological and oilier view-points. It is perhaps significant however that the neurophysiologie.il approach because of its recent advances and contributions to new concep- lions ol brain ortiani/alion and function has decidedly influenced thinking, and has given new impetus to discussion ol an old topic. 1 [ere we shall be concerned mainly with certain -.elected aspects of the neuro- physiological, electroencephalograph ic, psychological and behavioral manifestations Oi consciousness. Iii the Macy conferences .\\»\ in Schiller's review then- is an evident failure to resolve .1 definition of consciousness which would satisfy the various inter- disciplinary approaches to this problem. Because ol this the Macy ferences which began with the somewhat restrictive title 'Levels of Consciousness' shortly shifted to Problems of Consciousness. In his summary Schiller states: "Exclusively physiological and exclusively introspective accounts are incompre- hensive and give rise to artifacts. Although they are complementary, integration of knowledge is hard to achieve because their points of reference and scales of observation are wide apart." Lindsley (160) has argued for some kind of opera- tional definition of consciousness in terms of which we can observe, measure and evaluate. He proposed that consciousness is a state of awareness, and that under- lying this is sensory or perceptual discrimination which might serve as an anchoring point in terms of observation and measurement. Thus to determine the state of awareness of consciousness at some point on the sleep-waking continuum (see table 1) we can measure the intensity and other characteristics of the stimulus necessary to arouse the subject to a point where he can discriminate the stimulus presented. The act of discrimination may be an overt or covert response, the latter being measurable as an autonomic, somatic (electromyographic) or central-nervous- system (electroencephalographic ) response, as well as by subjective judgment and verbal report. Piaget in the 1954 volume of Abramson 1 1 I stated that consciousness might be analyzed in two ways, one, by studying the earliest or most elementary forms of awareness, or by concentrating upon states oi consciousness which were in process ol disappearing and returning, and two, by Studying the develop- mental changes in awareness as these are revealed by objective criteria of language, judgment and so forth. He discussed the second of these approaches under the heading of consciousness of necessity in which awareness of logical necessity was dealt wiih as it develops in children, lor example, 7-vear-olds clearly see the logical necessity of the number of beads in a short broad vial equaling those in a long thin one. This is a cognitive consciousness or a con- sciousness of relationships, which Piagei attempts to trace developmentallv in terms ol total operational structures. He believes these structures not only help to explain changes in consciousness, but are eventually isomorphic with corresponding neurological struc- tures. He states that consciousness is essentially a sv stem of meanings that mav be cognitive or affective. We shall now revert to the first approach mentioned bv Piaget, namely what are some of the earliest forms of awareness, or when does .1 newborn infant lirsi manifest signs of awareness. Kleitman in the 1950 and 1954 volumes of Abramson (1 1 emphasized ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS the point of view that there is a vertical stratification of consciousness ranging from near zero in the new- born baby asleep (or an adult in deep dreamless sleep) to an alert attentive adult after two cups of coffee. Lindsley in the 1950 volume and Monnier in the 1952 volume also stress the fact that consciousness is a matter of degree and level of complexity of perceptual discrimination. The newborn infant with a relatively nonfunctional cortex or the anencephalic monster without the cortex are rather alike in being essentially brain-stem creatures in whom reflexes and arousal responses exist, but very little differentiation of discrimination and response. Discrimination if present is at its lowest ebb, and attention, except for the crudest form of orienting reflex, is literally absent. As Lindsley (160) has indicated: "Although a new- born or young infant may show reaction to stimula- tion it is undifferentiated and not at all selective. Sensory discrimination, if present at all, is most elemental), and in this sense one must conceive of consciousness as being very restricted." The onset of the occipital alpha rhythm in the young human infant (see fig. 9) at 3 or 4 months of age appears to be correlated with the first consistent behavioral mani- festations of integrated attention. Docs the absence of persistent alpha rhythms in sensory /ones prior to that indicate a lack of awareness and consciousness in the field of these sense modalities? It probably docs so far as perceptual discrimination and integration are concerned. As we observed earlier, the return of alpha waves signifies the emergence from sleep and also the onset of awareness or consciousness of the environment which can only be discriminated as this stage of the EEG develops. There is a correspondence in the development and maturation of the structure and function (EEG) of the brain and the perceptually oriented behavior it manifests. Consciousness and EEG Characteristics Let us turn now to the EEG in relation to some reversible changes of consciousness. Gibbs et al. (90) were among the early workers to study the EEG changes in epilepsy and conditions of impaired consciousness. Davis & Davis (52) surveyed this topic in a very thorough fashion in 1939, and very little can be added except in respect to interpretation in terms of recent neurophysiological concepts. With regard to determining the state of consciousness and its electrical correlates in the early stages of sleep, they draw upon the description of experiments con- ducted by Davis et al. (53). "The subject lies down to sleep with a rubber bulb in one hand, and is instructed to squeeze it once whenever he feels that he has just 'drifted or floated off for a moment and twice if he feels that he has awakened from 'real sleep.'. . . The accuracy of the signalling is remarkable, con- sidering how unfavorable drowsiness is for intro- spection and signalling. . . . The common denomina- tor in the subjective reports of the experience of 'floating" is a depression of sensory perception. Some identify the state by suddenly realizing that they have ceased to hear noises or that they have lost their awareness of the bed clothes or the position of their body. Others stress the appearance of visual phantasies or interruptions in the train of logical thought, but in all cases there is loss of awareness, particularly of immediate external stimuli. This transient clouding of consciousness appears to be correlated with definiU objectivt n> m i; it activity at the brain." The 'floating or drifting off experiences occur in the B stage when alpha waves are greatly diminished or absent lor 5 see. or more. As these experiences become more persistent and longer last- ing, low-voltage delta waves appear and merge into the C Stage ol real sleep with spindle bursts of 14- pei -sec. w .iv es. induced physiologii vi. changes. I he breathing of an 8 per cent oxygen mixture was carried out in three subjects to the point of unconsciousness by Davis et at. i",4>, with EEG and handwriting as indicators of change of state of consciousness. After 4 iniii. of low oxygen a subject reported "'fullness in the head, ringing in the ears and a desir [sic] to breathe more deeply." Alter 17 min. he felt "fuzzy," dizzy and experienced a spell of remoteness. His writing deteriorated in quality and composition and finallv stopped entirely (hand "froze"'). In each of the three subjects with the onset of unconsciousness large slow waves became persistent in a 'locked' pattern. Although none realized they had lost consciousness, they were unaware when the mouthpiece of the breathing apparatus had been removed. The slow waves became 'unlocked' after breathing room air for 5 sec. and after 5 sec. more the subject was able to write in a distorted fashion. At this point slow waves were receding. After 2 min. on room air all slow waves had disappeared and the record showed low-voltage fast activity, and not until 4 min. or more did alpha waves and control picture return. Thus we see that the deterioration of consciousness and writing be- havior began with the loss of alpha waves, but uncon- sciousness did not become complete, nor behavior I582 IIVMHiOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III blocked, until very large slow waves developed a locked pattern. The return of consciousness was correlated with the breaking up of the slow wave pattern, but did not become clear until some alpha waves returned. Voluntary overbreathing or hyperventilation car- ried on from 2 to 4 min. will often induce delta waves of increasing magnitude in normal subjects. In 15 experiments the Davises reported consistent results, with modification of consciousness as the high-voltage delta waxes became persistent. Subjects failed to respond to commands, although one reported hearing the command and said afterward he was unable to comply. If hyperventilation were carried further, it is quite likely that a complete loss of con- sciousness would occur. In connection with Metrazol seizures which had been induced in a schizophrenic patient who had a normal control EEC), it was found that the patient be- came unconscious when convulsions associated with high-voltage fast waves set in. Lack of consciousness persisted through the subsequent low-voltage, rela- tively isoelectric period following the cessation of clonic convulsions, and through the high-voltage delta wave period in which waves gradually became better organized and more regular. Consciousness was re- Stored when the electrical activity finally returned to normal frequency ranges. Thus unconsciousness oc- curred in this instance during high-voltage fast waves, but continued through two or more patterns of slow waves of diverse type. There is always the possibility, however, that unconsciousness which occurs during the fast high-voltage stage in a major seizure shifts into a prolonged period of sleep with slow waves from which the patient eventually arouses. Thus, the slow w.imn following a convulsion may literally be deep sleep waves which are also accompanied by uncon- sciousness. But some patients cannot be as readily aroused from such ,1 stage as can a normal person with similar patterns in .1 stage of deep sleep. following coma-producing injections of insulin in a schizophrenic patient, it required better than an hour before all alpha waves disappeared and low - voltage delta waves replaced them Ai this point the patient was asleep with obvious loss of const lonsness In about >, hr. verv large I -to-j-per-sec. slow waves in a lot ked pattern appeared and continued dining the remainder ol the coma, which was cleared in 1 1 min. bv an injection of glucose. Forty min. following glucose the EEG was approaching normal and con- st iousness w as restored. |)eep ether and deep alcoholic intoxication both produce large slow delta waves in their unconscious anesthetic and stuporous phases, respectively. Ether initially produces a shift from alpha waves to a fast activity before giving way to delta wave activity of larger and slower nature. Alcohol initially enhances alpha activity, then diminishes and slows it; but no marked changes occur until deep intoxication is reached when the pattern becomes a more or less continuous moderate to high-voltage delta wave activity. SEIZURE PATTERNS Willi MODIFICATION OF CONSCIOUS- NESS. Figure 12 illustrates a spontaneous grand mal or major seizure pattern in a young man 19 years of age. The entire seizure is shown except for the pro- longed terminal sleep phase. The attack began with an increase in the voltage of alpha and other activity (at arrow, fig. 12). At the point marked tomi phase the patient's limbs and body assumed a rigid extensor position, and his EEG showed a further increase in voltage and amount of fast activity. About 5 sec. after the onset of the tonic phase, when high-voltage fast activity was still marked, consciousness was lost. The tonic phase gradually shifted to a clonic jerking phase, and as this subsided there was a period of flat, isoelectric record which ushered in the complete relaxation of the comatose phase. Alter some seconds random slow waves began to return and this picture persisted for some time, gradually forming more regular delta waves which eventually disappeared in about 30 min., the EEG returning to near normal. Figure [3 shows a major seizure induced bv electro- shock in which essentiallv the same pattern is exhibited as in the naturally occurring grand mal seizure shown above Again consciousness was lost at or near the start of the tonic phase. The sequence of tonic extensor phase with high-voltage fast activity, followed bv clonic convulsions with slowing of waves 10 correspond to rate of jerking, isoelectric record corresponding to the relaxed comatose phase, and finally continued relaxation with development of random slow waves, is characteristic of both the induced and natural major attacks. Unconsciousness merges with sleep at the end, ami the arousal and preseizure EEG pattern ina\ not return for 1 ;, min. to an hour or more. In contrast to the major seizures, figure 1 1 shows a petit mal or minor attack, with characteristic spike and slow wave pattern lasting u to [5 sec. with no behavioral change other than opening of the eves and staring straight ahead. Such an attack usually indicates an 'absence' or momentarv 'blank-out,' ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS >58.3 flOtripiM w*--~v>~^*~n/~wws~ ^^-~>-^WKWkn>A/wV^^ R. Parietal CLONIC PHASE \J\A^IWJ\F%^M — — — flOK- /Sjtj I0OM" fig. 1; EEG record of a major grand mal seizure showing the onset with hypersynchrony about 6 sec. before the tonic phase was ushered in with high-voltage fast activity which gradually shifted to slow waves at ;{ per sec. and the clonic phase Consciousness was lost near the onset of the tonic phase during fast high-voltage waves. [From l.indslev, unpublished observations.] but without falling or necessarily very notice. 1 lilt- change in behavior. If speaking, the patient in.i\ stop or continue with poor integration. A movement already started may be fixed or frozen, or may be continued with poor control. Our major concern here is with the state of consciousness and that is variable, ranging from a complete "blank-out" to partial awareness and ability to continue counting, or to stop counting momentarily and continue again as if nothing had happened. These and other varia- tions to receive, integrate and respond to stimuli and commands are apparently due to the variations in the magnitude and extent of the seizure discharge and perhaps also bear some relationship to its origin, whether in the thalamus, cortex or elsewhere. (See also Chapter XIY by Gastaut & Fischer-Williams in this Handbook. I The loss of consciousness in syncope, and especially carotid sinus syncope, has been attributed to a fall in arterial pressure and decreased oxygen tension of blood in the brain, but Lennox et al. (156) and Ferris it al. (73) concluded that some other factor than cerebral anoxia must produce unconsciousness in carotid sinus syncope oi the central type. Forster et al. (76) found unconsciousness associated with slow waves in the circulatory type, but with fast waves in the central type. In recent years Bonvallet et al. (29) have brought forth evidence which mav have a bearing upon unconsciousness and syncope. They have shown that spontaneous fluctuations of electro- cortical activity arc related to sympathetic tone, and that visceral and nociceptive stimuli produce marked activation of the cortex and parallel sympathetic changes. Two mechanisms are operating, a direct influx of such stimuli to the bulbar reticular forma- tion with immediate cortical activation, and a delayed humoral process which acts upon the pontomesen- cephalic reticular activating system and then the cortex. The sympathetic tone mav be just as impor- tant to the maintenance of wakefulness and conscious- ness as the influx of proprioceptive and exteroceptive stimuli. 1 ,84 II Willi! M IK OF PHYSIOLOGY NEUROPIIYSKH .(ICY III fic. 13. EEG record (it an electroshock convulsion similar to the grand mal attack in tig. 1 j. The record was started 8 sec. alter the shock to avoid shock artifact. The record begins with the tonii phase and high-voltagi fast activity; consciousness was lost immediately. The clonic phase follows quickly with first, 3-per-sec. waves, then slowing to 1 per sec., and subsequently to the flat isoelectric period during which the patient was Completely relaxed. The end of the record shows large random slow waves gradually organizing into a sequence of large I-per-sec, waves characteristic of deep sleep. J From Lindsley, unpub- lished observations. I *f*rrW**vvWW^^ ^-^-w, — - — s~~vsSK^«^YW^n'Wvw^^ TONIC PHASE CLONIC PHASE i/^^vvvwlAl,'^^%o^^J^^*^v^^ — -— — -—^^ ,V\^Vl/YVVYV\/VVVVVVV'V',^^ — ■v- jijuu»-a^a^juxJUau1^xajuju4^Ji~_A~-jiAj^ — uj~jujluuJU-UX-L_ ui. — * a^_ji. U^Jl^-A^-A^X^-L^L-J A — (-_X— i- — t — L— A~ — Aft_-A/s- -L> h> JL— — A^— Ls— Aa— Kr M K I '*« I XXI*" ^nAnAAnAA^jmniWJWV\fWV, AAAj . ... ^ajw^ — -I ^mmmmmmw^m?, i ' imam vwv — -i ooened eyes- s'ored 300 M" 11.. 1 1 Mi, record oi .1 minor petit mal seizure showing characteristic 'spike-and-slow-wave' sequence. The attack lasted about 14 sec. Consciousness was in abeyance, and there was amnesia foi the attack. From Lindsley, unpublished observations. I hese workers have also demonstrated that disten- iKin of the carotid sinus, with experimental control hi arterial pressure and cerebral circulation, is capable (if inhibiting liy nervous transmission the brain-stem activation ol the cortex. This is the only known met nanism oi afferent influx to the reticular formation which inhibits it, and ihus reduces cortical activation and promotes slow-wave development comparable to thai in .1 si. iic (il sleep. The influence ol sympathetic tone as an activator ol the ARAS, oi the sudden drop in tone as a deai tivator, may under certain emotional in fatigue conditions be able to induce either accen- tuated hypersynchrony and fast activation waves such .is appear to trigger the grand mal attack and produce unconsciousness, or else slow waves which also are frequently associated with unconsciousness. The carotid sinus mechanism and its inhibitor) effect on the ARAS may well be a participant in uncon- sciousness produced in seizures, syncope and narco- lepsy Gellhorn and collaborators (23, »<>, 88) have emphasized the relation of the unspecific to the specifii sensory systems in relation to perception, attention and consciousness, Gellhorn (86) feels that interaction ol impulses from the diffuse, hypothalamic-corticaJ projection system with (hose of the specific projection ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS '585 systems makes possible the various degrees of aware- ness that may be distinguished by physiological and psychological criteria. Gellhorn (86) has demonstrated that nociceptive stimuli in lightly anesthetized cats, by interaction with acoustic and optic stimuli, lead to increased reactivity in the projection areas of each of these modalities. A similar interaction between sense modalities has been demonstrated at the reticu- lar formation by Hernandez-Peon & Hagbarth (104). These investigators have also shown that afferent and corticifugal influences interact at the reticular forma- tion and result in either interference or facilitation. Both 'occlusion' and "subliminal fringe' phenomena have been demonstrated. They believe it likely that reticular unresponsiveness, attributable to interaction upon reticular neurons from two or more sources, may well influence sensory perception and conscious- ness. Presumably moderate interaction could lead to facilitation, whereas excessive bombardment could result in complete blocking of reticular activation of the cortex causing disturbance in attention and awareness, and even complete loss of consciousness as observed in some seizure states. The kind of sudden and intense barrage from afferent and corticifugal sources in strong emotions could be responsible for the confusion and immobilization often attendant upon such circumstances. The EEG in these condi- tions (see fig. 11; table 1) usually shows a picture of complete and prolonged flattening or increased high-frequency activity. Such records have been observed in acute and chronic anxiety patients by Cohn (50) and Lindsley (158). temporal course of consciousness. Fessard (74) has written a very challenging paper on nervous integra- tion and conscious experience. He takes from Cobb (49) the notion: "It is the integration itself, the relationship of one functioning part to another, which is mind and which causes the phenomenon of consciousness." Fessard calls this phenomenon of consciousness an "experienced integration' or EL With regard to the temporal organization of con- sciousness he makes three points: a) "each new EI involves a process of reorganization that cannot be instantaneous"; b) time and EI are intimately related in that states of consciousness succeed one another; and c) in order to know the nature of the integrative processes resulting in EI we must attempt to deter- mine the neural mechanisms corresponding to memory. In relation to Fessard's first point, it can be shown as in figure 7 that two brief flashes of light 50 msec, apart register as one evoked potential in the visual cortex of the cat, until the reticular formation has been stimulated, and then for a short time afterwards the pair of flashes elicit a pair of evoked potentials. The same two flashes, as Lindsley (162) has found, would be seen by a human subject as one when 50 msec, apart, but as two when 100 msec, separate them. Thus in this instance, as well as in the case of fusion of repetitive flashes above flicker level, succes- sive El's require time for reorganization and integra- tion. The second point emphasizes Lashley's (155) concept of "serial order of behavior' and stresses the importance of coherence of temporal sequences, for meaning depends upon order. Travis (226) has made an attempt to investigate the temporal course of consciousness by recording the EEG from subjects while they were asked to rest and let their minds wander, without attempting to direct the stream of consciousness. Periodically the experimenter said, 'Now,' and the subject reported whatever conscious state w.i^ interrupted. A variety ol types of imager) was revealed, but mainly visual, auditor) and kinesthetic. Reports classified as abstract thinking or mental blankness also were encountered. Alpha blocking or an activation pattern tended to be associated with visual images, kinesthetic sensations and mental effort, while mental blankness and abstract thinking appeared to be accompanied by Strong or well-developed alpha waves. Travis con- cluded that large and regular alpha waves are indica- tive of a state of cortical equilibrium and represent a generalized psychic activity, whereas the breaking up of this collective action into more rapid and ir- regular oscillations of much smaller amplitude represents a relatively high degree of specificity in psychic activity. In present-day terms this would be equivalent to saying cortical activation is associated with attention. Penfield (192-194) has developed an interesting collection of unique observations and subjective reports by the artificial induction of awareness through stimulation of the brain of conscious patients at time of operation. He stresses that sensation cannot be located in the cortex but instead the sensory material is reorganized in a higher center which he identifies as the "centrencephalon" or "higher brain stem.' He states that the centrencephalon '"must be in certain portions of the diencephalon, mid-brain and pons." Elsewhere (193) he includes "those parts of the higher brain stem, which have symmetrical connections with both hemispheres," pointing out i586 H \M)BOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III that the intralaminar systems of the thalamus and the reticular formation of the brain stem and the nonspecific projection systems satisfy the definition. Penficld (194) points out that the "'stream of con- sciousness" as well as "man's experiential record" or memorv involves the temporal lobe where direct electric. il stimulation in conscious man evokes flash- back memories. The hippocampus or a more central portion of the centrencephalic system may be the actual storehouse, but these bear sufficiently close functional relationship to the temporal lobes so that the latter constitute the effective points of activation of such memories and experiences. Penfield (192) adds that not only do sensory impulses arriving at the cortex descend to central integrating centers where "highesl level" final integration is presumed to take place but in their final form are impressed upon the premotor cortex for outflow to effectors. However, he believes that there is an alternative pathway for directional voluntary impulses via subcortical motor centers. Sperry (220) has put forth an interesting point of view indicating that "all brain excitation has ulti- matclv one end, to aid in the regulation of motor coordination. Its patterning throughout is determined on this principle. It follows that efforts to discover the neural correlates of consciousness will be more successful when directed on this basis than when guided by arbitrary correlations with psychic ex- perience, stimulus patterns, or outside reality, or by analogies with various types of thinking machines." He would approach the problem of perception, thinking and even consciousness by attempting to understand motor integration and adjustment. He believes that we have been preoccupied with "sensor) avenues to the stuck of mental processes" and these "will need to be supplemented by increased attention to the motor patterns, and especially to what can be inferred from these regarding the nature of the asso- ciative and sensor) functions." viiisii'ix \\n mi ml Jouvel (135) has investi- gated in in. in at the time of operation the effect of 1 li . tive visual a 1 ten 1 ion and the intrusion of distract- ing stimuli Hi other modalities upon the surface EEG over the occipital cortex and the subcortical responses m the optic radiations He has found that repeau-d single flashes of lighl produce .1 marked augmentation ol the responses in the optic radiations when the mbjei 1 was asked to attend to and count the flashes The surfaci EEG hows an increase in rapid low- volt igi H n\ it) at the start ol the pei iod oi attention and then a decrease. When stimuli of a distracting nature are introduced to other sense modalities than vision, there is a reduction of the subcortical visual responses almost to total disappearance. This would indicate interaction and suppression among the alternative sense modes competing for the field of attention. Olfactory, auditory and nociceptive stimuli were very effective in producing this suppression, but tactile stimulation from objects placed in the hands seemed less effective. Mental calculation and problem solving also caused reductions in the mag- nitude of the optic radiation responses. In a patient with damage in the brain stem of 6-mo. duration who had in addition to other neurological indications difficulty in maintaining vigilance, it was found that nociceptive stimuli did not affect the visual responses. Jouvet feels that the subcortical reduction of response observed at the time of attention to another sense mode calls for a distinction between the neural mecha- nisms put in play at the time of the arrest reaction (EEG arousal) and those exhibited at the time of attention. He states that his results confirm those already obtained by others (100, 106) in animals and previously mentioned in this chapter. However, he is uncertain whether the inhibition he has demon- strated is at the level of the lateral geniculate body or the retina, since both have been shown to be affected negatively by reticular stimulation, as well as positively. He did not find habituation of the subcortical visual responses to light flashes presented once a second over a period of several minutes. Oswald (1881 has studied the human EEG under a variety of experimental circumstances relative to vigilance and habituation. During intent listening to brief tones of near-threshold intensity to which the subject was to respond bv pressing a kev , two tenden- cies were found, one, for sleep and alertness to alter- nate regularly and rapidly at the rate of the signals, and two, a slower downward drift toward sleep to the point of failure to respond. The tones were presented at intervals of IO, 5 or ;j sec. during sessions lasting 1 -, to 30 min. First there was ,1 blocking "I alpha to each tone with return for a few seconds between tones. Later this tendency was reversed with alpha hlockage appearing between tones and alpha bursts being triggered bv the tones. The ques- tion arises of course as to whether this is merely a shift from vigilance to a drows) state, with reversed tendencies for the response, or whether this is a unique reaction of the reticular system to repeated near-threshold stimuli representing a form of habitua- tion. Similar alternations of sleep signs and alertness ATTENTION, CONSCIOUSNESS, SLEEP AND WAKE FULNESS 158/ appeared while listening to rhythmic music and maintaining apparent wakefulness while tapping to the rhythms. Because of the special interest of the subjects in the music and some tendency to concen- trate upon particular instruments or passages, there is possible reason to see in these results the influence of selective attention, with suppression of activity in some periods and not in others, and the alternation or variation of these patterns of attention with sleep- like suppressions of alpha rhythm. It is conceivable also that different areas of the cortex may react differently to selective attention, although this has not been the particular subject of extensive study. The effect of alerting upon reaction time has been studied by Lansing et al. (154)- It was found that ordinary visual reaction times to the onset of a visual stimulus without special alerting and when alpha waves were absent or present spontaneously averaged 280 msec. When a brief auditory signal was introduced as a forewarning up to 1 sec. before the visual stimulus, the reaction time was markedly reduced to about 206 msec, if the auditory stimulus preceded the visual stimulus by at least three tenths of a second. Reaction time was reduced if the forewarning stimulus was less than three tenths of a second ahead of the visual stimulus, but much less, and the alpha waves had not been blocked by the time the visual stimulus was presented. In other words unless the interval between the auditory and visual stimuli is greal enough, activation or alpha blocking does not occur in time to facilitate the response. Figure 15 shows the curve of reduction of reaction time as a function of the length of the forewarning period and also the curve for the degree of alpha blocking. The two curves show remarkable similarity, suggesting that alpha blocking or activation is indeed related to reduction in reaction time. The set to respond after an alerting signal is presumed to be triggered by the auditory signal acting upon the ARAS, the influence of which is known to be alpha blockade or activation. This process in turn facilitates the speed of processing the over-all reaction. Since input and output time in a visual reaction time situation are relatively constant and fixed, the reduction in reaction time is probably mainly a reduction in central cortical processing time. A similar type of influence has been demonstrated by Fuster (83) in monkeys trained to make perceptual discriminations between two objects exposed tachisto- scopically (see fig. 16). The monkey has been trained to select the correct object in order to get a food reward from under it. After thorough training, he is placed before the one-way screen and can only see i'T 1 1 111 1 1 1 1 i_ 0 200 400 600 800 1000 F0REPERI0D IN MILLISECONDS in. 1 -,. Reaction time and alpha blocking plotted as a function of the fbrepe 1 interval. Note that the reaction time is reduced to a minimum when the foreperiod ranges from 300 to iooo msec, and that the alpha-blocking curve follows a similai time course. Once activation occurs there is no further reduc- tion of the reaction time, Relaxed unalerted reaction time, 280 msec , alerted reaction time, minimum -;oii msec. From Lansing et nl. 1 154).] the objects when the tachistoscopic flash of varying durations is presented. As he makes his correct choice, he reaches through a door in front of the object selected which stops the cluck and provides an over-all reaction time for the process The monkey has electrodes implanted in the mesencephalic reticular formation and can be thus stimulated. Figure 17 shows the curves for per cent correct response and reaction time for trials without (control I and with reticular stimulation. It will be noted that reticular stimulation improved performance and reduced reaction time. Here again is .m example of facilitation through ARAS influence. Activation or alerting induced in this experiment by direct stimulation of the reticular formation reduced perceptual discrimi- nation reaction time. In the human experiment just cited indirect activation was produced by a fore- warning signal and reduced over-all reaction time. In both experiments there is further indication of the facilitation, and thereby a reduction, of central processing time. eeg in hypnosis. The question arises, since hypnosis has some characteristics in common with sleep, such as trance-like states with limited awareness, and some in common with selective attention, whether the EEG is one of wakefulness or sleep, and whether any indications of specific arousal or alerting are evident. 1.88 II VNUBOOK Oh 1'IIVsIl il ( K.Y NKl'ROPHYSIOLOGY III SLIDING DOOR PARTIALLY RAISED TACHISTOSCOPE -V- \ — J ACTIVATES \ \ TIMER — o -o CONTROL ■ ■ ST M SOOpA 600 !00 1 90 ' °^~- ^"""^ LJ 5 1- BO - ^o s N N g 70 " \\ I - \ \ (T GO - \ \ \ \ Z \ UJ ■snn 5(1 \ 2 CHANCE 0 , I 1 1 „' 40 30 20 10 40 30 DURATION OF EXPOSURES (msec) fig. 17. Percentage of correct response and mean reaction times as a function of tachistoscopic exposure duration. Cor- rect responses increased and reaction times decreased during reticular stimulation (solid lines). Each point represents 100 trials. [From Fustcr (83).] fig. 16. Visual-discrimination reaction-time apparatus. Monkey is trained to discriminate objects for a reward, then allowed to see the objects only during a brief tachistoscopic flash. The light starts the clock timer; the monkey's choice and reaching through the door stops the clock, so giving the reac- tion time. [From Fuster (83).] Although comparatively few studies in the EEG field deal with hypnosis, those that have been pub- lished (15, 27, g8, 168, 171, 232) are in general agreement for the most part. In a waking trance, the EEG in hypnosis docs not differ significantly from that of the same person in a normal waking slate. The EEG does nol resemble the KEG in sleep unless a subject in a trance has been allowed to sjo to sleep or deliberately put to sleep by suggestion. With the general relaxation which occurs during hypnotic episodes, there is sometimes an increase in alpha activity, or ,1 falling oil if slighl drowsiness supcrv cues. Unless .1 special suggestion is made requiring efforl mi tension, there is no tendency for the EEG to show activation. In general then it can be said thai the EEG in hypnosis is similar to that during waking. Under hv pilosis the KE( • lends 10 parallel that during normal stales and, if the hypnotic subject is required to sleep, he has a sleep EEG; if he is required to soke a problem, his EEG resembles thai when he is not under h\ pnotic influence. The efficacy of suppressing pain reactions under hypnosis, or of obliterating cognizance to one scum- mode in deference to another, suggests a similarity ol this aspect of hypnosis to selective attention under normal conditions. Loomis et al. (168), however, found that inserting a pin into the arm of a hypnotized subject, after repeated suggestions that he would feel nothing, produced effective alpha blocking lasting 28 sec. Repeated trials produced less effective blocking. The threat to insert the pin into the arm but without actually doing so had no effect on the alpha waves, even when the suggestion was made that the pin was in the arm and painful. Thev also observed that alpha waves were present during hypnosis as before, and that opening and closing the eyes caused alpha blockade as prev iouslv . 1 lowcv er, w hen they suggested that the subject was blind in the presence of Light, the alpha waves were not blocked. This result differs from that obtained by others (27, 171, 232), all of whom found that negative suggestion was ineffective in reversing the usual alpha-blocking reaction to Light. Comparatively little progress has been made electroencephalographically or neurophysiologically in coming to grips with the mechanism which might underlie hypnosis and hypnotic phenomena. Ac- cording lo Barber (141 hvpnosis is not a "slate of consciousness" Il is nol a •'thin',;" or an "entity." "ilvpnosis is a descriptive abstraction referring 10 an interpersonal relationship which i^ characterized by a number of overlapping processes." 1 Le sees \w pnotic phenomena as extensions of normal modes and man- ners of behaving and reacting, but under the influence of strong "belief" in the hypnotist's influence. Al- though he strongly emphasizes that hvpnosis is not ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS !589 a state, he recognizes that in normal sleep or awaken- ing from sleep a suggestion is not reacted upon unless the proper 'set' has been established. He finds a similar thing in hypnosis, that a readiness and will- ingness to respond as the hypnotist directs is a factor in the effectiveness of suggestion. This would seem to imply some kind of plane of reduction, since the sub- ject would not necessarily assume this set under nor- mal conditions. Further careful study is needed of hypnotic influences if the neural mechanism of this remarkable condition is to be understood. Presumably much of this will have to be done with the EEG as the principal liaison between modern neurophysiological conception and the hypothetical states of hypnosis. SUMMARY A cursory review of the history of concepts con- cerning sleep and wakefulness centers in the brain has been attempted with impressions of the steps which have led to the modern concept of wakefulness as dependent upon the reticular system. Several sources of stimulation which influence the reticular formation in the lower brain stem and, through it, the ascending reticular activating sv stein (ARAS) art- discussed. The ARAS influences the cerebral cortex and higher brain centers diffusely, lint may also have more differentiated effects, some of which may act through the diffuse thalamic projection astern (DTPS) in such a manner as to regulate cortical activity and excitability, especially in associational fields. In addition to this nonspecific influence the ARAS and DTPS may act differentially with respect to the fields of reception served by the specific tha- lamic projection systems (STPS). Wakefulness is maintained by excitation of the reticular formation and the ARAS through collaterals from all sensory pathways, by corticifugal impulses originating in various regions of the cortex and by humoral factors which affect particularly the rostral portions of the reticular formation. Increased activity in the ARAS through .my of these sources of excita- tion acts upon the cortex by changing the pattern of its electrical activity from the slow waves and spindle bursts of sleep, or the alpha waves of relaxed wake- fulness, to a pattern of low-voltage fast waves, com- monly referred to as 'activation.' Electrocortical ac- tivation is accompanied by behavioral arousal and by alertness and attention. The elusive term 'consciousness' has been consid- ered as a graded form of awareness, ranging from the simplest perceptual discriminations to the more com- plex cognitive forms of abstraction and thinking. It has no precise locus on the sleep-wakefulness contin- uum described in terms of EEG patterns, behav- ioral characteristics and states of awareness. Uncon- sciousness in which perceptual contact with the environment is lost can be identified roughly with the onset of sleep in Stage C of the EEG in which delta waves and 14-per-sec. spindle bursts predominate. Other forms of unconsciousness induced by breathing low oxygen mixtures, by hyperventilation, by insulin coma, by deep alcoholic intoxication and by some seizure states are maink associated with the onset and persistence of large ami slow waves. In grand mal sei- zures and in some physiological, drug and anesthetic conditions, loss of consciousness is associated with high-voltage fast activity. Attention is closely allied to arousal and wakeful- ness and, like wakefulness and consciousness, appears to be a graded phenomenon extending from general alerting, .is in the orienting reflex, to specific alerting, as when attention is focused upon a given sense mode and dominates sensory input to the point of exclusion of other sense modes. Still higher or more finely focused attention m.iv lie restricted to a limited aspect of a given sense mode. Recent neurophysiological findings have been con- sidered which not only broaden the scope of modern concepts of brain organization and function, but bear specifically upon the mechanisms which may underlie and subserve the processes of attention, perception and learning. REFERENCES 1. Abramson, H. A. (editor). Problems of Consciousness. New York: Macy, 1950-54. 2. Adey, W. R., J. P. Segundo and R. B. Livingston. J. Neurophysiol. 20: 1, 1957. 3. Adey, W. R., S. Sunderland and C. W. Dunlop. Electro- encephalog. & Clin. Neurophysiol. 9: 309, 1957. 4. Adrian, E. D. J. Physiol. 88: 127, 1936. 5. Adrian, E. D. The Physical Background of Perception. Oxford: Clarendon Press, 1947. i> Adrian, 1 D. and B. H. C. Matthews. Brain 57: 355, 1934. 7. Akimoto, H. and O. Creutzfeldt. Arch. Psychiat. 196: 494. '957- 8. Akimoto, H., N. Yamaguchi, K. Okabe, T. Nakagawa, K. Abe, H. Torii and K. Masahashi. Folia psychiat. neurol. Japonica 10: 117, 1956. 9. Albe-Fessard, D., C. Rocha-Miranda and E. Oswaldo- Cruz. Electroenccphalov. &■ Clin. Neurophysiol. 1 1 : 777, 1959. 10. Albe-Fessard, D. and A. Rogeul. Eleclroencephalog. & Clin. Neurophysiol. 10: 131, 1958. '59o II Willi* ll ik I I] 1'IIYMl il l il. Y NEUROPHYSIOl HOY III 11. AmaSSI.an, V. I. am j K D1.V110. ./. Xeurophysiol. 17: 575, 46. '954- 47 12. Amassian, V. I. smi H H. Waller. In Reticular Forma- 48. lion of the Bruin, edited by H. H. Jasper, L. D. Procter. R S. Knighton, W. C. Noshay and R. T. Costello. Boston: 49. Little, 1958, p. 69. 50. 13. ASERINSKY, I. AND N. Kimimvv J. Appl. Physiol. 8: 1, 51. '955- 14. Barber, I X J. Psychol. 45: 115, 1958. 52. 15. Barker, W. and S. Burgwin. A.M. A. Arch. Xcurol. Psy- chiat. 62: 41 2, 1949. 53. 16. Bartley, S. H. and G. H. Bishop Am. ./. Physiol. 103 : 1 73, '933- 54 17. Bartley, S. H. and G H. Bishop. Am. J. Physiol. 103: '59- '933- 55- 18. Beattie, J. and D. Sheehan. J. Physiol. 81 : 218, 1934. 56. 19. Berger, H. Arch. Psychiat. 87:527, 1929. 20. Berger, H. ,/. Psychol, u. Neurol. 40: 160, 1930. 57. 21. Berger, H. Arch. Psychiat. 94 16, 1931. 22. Berger, H. Anh. Psychiat. 97: 6, 1932. 58. 23. Bernhaut, M., E. Gellhorn vmi A. T. Rasmussen. ./. Xeurophysiol. 16: 21, 1 953. 59. 24. Bishop, G. H. Am. ./. Physiol. 103: 213, 1933. 25. Bishop, G. H. Cold Spring Harbor Symp. Quant. Biol. 4: 60. 3°5. '936- _'i>. Bishop, G. H. In: Reticular Formation of the Brain, edited 61. by H. H Jasper, L. D. Proctor, R S. Knighton, W. C. Noshay and R. T. Costello. Boston: Little, 1958, p. 413. 62. 27. Blake, H. and R. W. Gerard. Am. ./. Physiol. 119: 692, ■937- 63. 28 Blake, H., R. W. Gerard and N. Kleitman. ./. Neuro- physiol. 2: 48, 1939. 64. 29. BONVALLET, M., P. Dell and G. Hiebel. Electroencephalog. & Clin. Neurophysiol. 6: 119, 1954. 65. ■jo. Bradley, P. B. In: Reticulat Formation of the Brain, edited by H H.Jasper, I. D Proctor, R. S. Knighton, V\ C 66. Noshay and R. 1 Costello. Boston: Little, 1958, p. 123. 67. 31. Bradley, P. B uid J Llkes. J. Physiol. 120: 14P, 1953. (2. Bradley, P. B. andJ. Elkes. Brain 80: 77, 1957. 68. 33. Bri mi k. I' Compt. rend. Soc. de hint. 1 18: 1235, 1935. 34. Brimi r, F. Compt. rend. Soc. de biol. 122: 460, 1936. 35. Bremer, F. Compt. rend. Soc. de biol. 1 22 : 464, 1936. {Ii BREMER, F. and N StOUPEI Acta neurol. et psychiat. belg. 69. 58: 401, 1958. 37. Bremer, F. and N. Stoupel. Arch, internal, physiol. 67: 70. !" '959- 38. Bremer, F. and ( lin/i .In/; internal. Physiol. 62: '954- 7'- ig Bi 11 1 1 1 - I B. and E. Beck In Neurological Basis of Be- haviour, edited by G I W Wolstenholme and C M. 72. O'Connor. Bosion: Little, iq',8, p. go. )o Brodal, A I In / rmation of the Brain Stem: Ana- 73. tomical Aspects and Functional Correlations. Edinburgh: Oliver, 1957. 74. |i BUSER, P \m> P BoRENSTEIN li^tmencephalog. ef Clin. Neurophysiol. 1 1 28-,, 1959. I .■ I'.i ii- P P BoRENSTEIN sm>| BrUNER Electroencephalog. yc ' \ physiol. 1 1 : 305, 1959. I; Buser, P., M b'lsii \mi R Hernandez-Peon. I la yS. neurol. lalinoam. \ .'lift, 1958. II Chano, II I / Neurophysiol. 13:235, 1950 -- 1 1 h \-< II I / Vi tropl 1 tiol 1 1 95, 1951. Chang, H.-T, ./. Xeurophysiol. 15: 5, ig52. Chow, K. L. A.M. A. Arch. Neurol. & Psychiat. 71 : 7^2, 1954. Chow, K 1. . W. C. Dement and S. A. Mitchell, Jr. Electroencephalog. & Clin. Neurophysiol. 11: 107, 1959. Cobb, S. A.M. A. Arch. Neurol. & Psychiat. 67: 172, 1952. Cohn, R. ./. Nero, ii Merit. l)a. 104: 351, 1946 I II 1 l/ll LDI, ().. (i. B.MMCAK INIR AND R. .|l-"N(i. EleC- troencephalog. & Clm. Xeurophysiol. 8: 163, 1956. Dams, II and P. A. Davis. A. Res. Nero. J. Stoll. A.M I Arch Neurol. v; Psychiat. 67: 155, 1952 Jasper, H. H. and J. Droogleever-Fortuyn .1. Res. Nerv & Menl. Dis., Proc. 26: 272, 1947. Jasper, H. H., R. Naquet and E. E. King. Electroencepha- log. S Clin. Neurophysiol. 7: 99, 1955. Jasper, H., G. F. Ricci and B. Doane. In: Neurological Basis of Behaviour, edited by G. E. W. Wolstenholmc and C. M. O'Connor. Boston: Little, 1958, p. 277. Jouvet, M. Rev. Psychol. Francaise 2, No. 4: 1, 1957. Jouvet, M. and R. Hernandez-Peon. Electroencephalog. & < 'lin. Neurophysiol. Suppl. 6: 39, 1956. Jung, R., O. Creutzfeldt and O.-J. Grusser. Deutsche med. H'chnschr. 82: 1050, 1957. (English translation: Ger- man M . Month 3: 1, 1958.) Kabat, H., H. W. Magoun and S. VV. Ranson. J. Comp. Neurol. 63 : 2 I I , I 936. Karplus, J. P. and A. Kreidl. Arch. ges. Physiol. 129: 138, '9°9- Kayser, C. J. physiol., Paris 41 : i-A, 1949. >59* llWDIIllUK OF PHYSIOLOGY \l I ROPHYSIOLOGY II! 141. Kerr, D. I. B. and K.-E. Hagbarth. J. Neurophysiol. 18: 362, 1955. ; (j KrRR, F. W. L. and H. L. O'Learv. Electroencephalog. & Clin. Neurophysiol. 9: 461, 1957. 1 4 j. Killam, K. F. and E. K. Killam. In: Reticular Formation of the Brain, edited by H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay and R. T. Costello. Boston: Little, 1958, p. ill. 144. Ki.aue, R. ./. Psychol, u. Xeurol. 47: 510, 1937. 145. Kleitman, N. Sleep and Wakefulness. Chicago: Univ. Chi- cago Press, 1939. i4t). Kleitman, N. and N. Camtlle. Am. J. Physiol, too: 474, 1932. 147. Knott, J. R., F. A. Gibbs and C. E. Henry. J. Exper. Psychol. 31 : 465, 194 J. 148. Knott, J. R., C. E. Hf.nry and J. M. Hadlev. J. Exper. Psychol. 24: 157, 1939. 149. Kovacs, L. Gior psichiat. e neuropat. 60: 344, 1932. 150. Lacey, J. I. and R. L. Smith. Science 120: 1045, 1954. 151. Lacey, J. I., R. L. Smith and A. Green. Psychosom. Med. 17: 208, 1955. 152. Landis, C. and W. A. Hunt. The Startle Pattern. New York: Farrar & Rinehart, 10,39. 153. Lansing, R. \V. Electroencephalog. & Clin. Neurophysiol. 9: 497. '957- 154. Lansinc, R. VV., E. Schwartz and D. B. Lindsley. J. Exper. Psychol. 58: 1, 1959. 155. Lashley, K. S. In: Cerebral Mechanisms in Behavior, edited by L. Jeffreys N™ York: Wiley, 1951, p. 112. 156. Lennox, W. G., F. A. Gibbs and E. L. Gibbs A.M. A. Arch. Neurol S Psychiat. 24: 1001, 1935. 157. Li, C.-L., C. Cullen and H. H. Jasper. J. Neurophysiol. '9 '3'. '956- 158. Lindsley, D. B. In: Second International Symposium on Feel- ingi and Emotions, edited by M. L. Rcymcrt. New York: McGraw-Hill, 1950, p. 238. 159. Lindsley, D. B. Electroencephalog. & Clin. Neurophysiol. 4: 443. '952- 160. Lindsley, I) B In; Clinical Neurosurgery, edited by I.J. Jackson and R. K Thompson. Baltimore: Williams & W'ilkins, 1957, p. 175. 161. LlNDSLEY, D. B. In: Reticular Formation of the Brain, ed- ited by H H.Jasper, L. I). Proctor, R S. Knighton, W. (: Noshay and R. T. Costello. Boston Little, 1958, p. 513. id.;. Lindsley, I). B. In: Description mid Analysis of Behavior, edited by R. Patton. Pittsburgh: Univ. Pittsburgh Press, 1958, p. 48. 163. I.IMiM I •. . 1) B, J. BOWDEN AND H. W. MaCOUN. EleC- bhalog Clin Neurophysiol. 1 : 473. 1949. I64 I.1M1SIIY, I) B. I. II. SCHREINER, W. H. KnOWLES AND H. W. Magoun. Electroencephalog. & Clin. Neurophysiol. 2: l!f(. '95°- 165. Lindsley, I) H, I.. H. Schreiner and H. W. Macoun. / Veurophysuil 1 2 107, 1949. 166. LOOMIS, A. I.. I. V ll\kv>\ \\ii(;. Homari .Science 61: 239, 1935 1 * > 7 . Looms, A I. . I. N, Harvi y and G. Hobart. Science 82: <»H- 168. LOOMB, A. I.., E. N. Hakvin wii ti Hobari J Expei Psychol, mi 249, ig |6 169. Looms, A. I., E. N. Harvey and G. Hobart, J. Exper. Psychol. 21 127, 1937. 170. Loomis, A. L., E. N. Harvey and G. A. Hobart. J. Neurophysiol. 1 : 413, 1939. 171. Lundhoi.m, H. and H. Lowenbach. Character & Person- ality 1 1 : 145, 1942. 172. Macoun, H. W. A. Res. Net S Went. Dis., Proc. 30:480, ■952- 173. Magoun, H. W. and R. Rhines. J. Neurophysiol. 9: 165, 1946. 174. Marinesco, G., O. Sager and A Kreindler. /.tschr. ges. Neurol. Psychiat. 122: 23, 1929. 175. Mauthner, L. W'ein. med. Wchnschr. 40: 962, 1001, 1049, 1890. 176. Monnier, M., M. Kai.berer and P. Krupp. XXI In- tel nat. Physiol. Congress, Buenos Aires. 1959. 177. Monnier, M. and R. Tissot. In: Neurological Basis of Be- haviour, edited by G. E. W. W'olstenliolme and C. M. O'Connor. Boston: Little, 1958, p. 105. 178. Morgan, C. T. and E. Stellar. Physiological Psychology. New York: McGraw-Hill, 1950, p. 357. 179. Morison, R. S. and E. W. Dempsey. Am. J. Physiol. 135: 281, 1942. 180. Morison, R. S. and E. W. Dempsi •> Am. / Physiol. 138: 297. '942- 181. Moruzzi, G. and H. W. Magoun. Electroencephalog. & Clin. Neurophysiol. 1 : 455, 1949. 182. Nauta, W.J. H. J. Neurophysiol. 9: 285, 1946. 183. Nauta, W. J H. and D. G. W'hitlock. In: Brain Mech- anisms and Consciousness, edited by I. I). Adrian. F. Bremer, H. H. Jasper and J. 1'. Delafresnaye. Oxford Blackwell, '954. P- 81. 184. Nekhorocheff, I. Rev. neural. 82: 487, 1950. 185. NlEMER, VV. T. AND J. JlMENEZ-C\s 1 I 1 INOS J. Cornp. Xeurol. 93 : l o 1 , 1 950. 186. O'Leary, J. L. and L. A. Coben. Physiol. Re;. 38: 243. 1958. 187. Olszewski, J. In: Brain Mechanisms and Consciousness, ed- ited by E. D. Adrian, F. Bremer, H. H. Jasper and J. F. Delafresnaye. Oxford: Blackwell, 1954. p. 54. 188. Oswald, I. ./. Meat. Sc. 105: 269, 1959. 189. Pavlov, l.Scient. Month. 17:603, 1923. 190. Pavlov, I. Conditioned Reflexes. New York: Oxford, 1927. 191. Pechtel, C, J. H. Masserman, L. Schreiner and M. Levitt J \ Went. Dis. 121:26, 1955. 1 < I _■ I'lMlllli, W. Brain 77: 1, 1954. 193. Penfield, VV. In: Brain Mechanism, and C tsness, ed- ited by E. I). Adrian, F. Bremer. H. II Jasper and J. F. Delafresnaye. Oxford: Blackwell, 1954, p. 284. 194. Penfield. VV. XIV Internal. Congr. Psychol,, Proc. 47, 1955. 195. Pieron, H, /<■ problcmc physiologique tl. Phtiinnil.nl. 41 ,1 201. Rose, I I I Res Ven S Went. Dis., Proc. 30 454, 1952. ATTENTION, CONSCIOUSNESS, SLEEP AND WAKEFULNESS [593 202. Rosenblueth, A. and W. B. Cannon. Am. J. Physiol. 135: 690, 1942. 203. Rossi, G. F. and A. Brodal. A. M. A. Arch. Neurol. & Psychiat. 78:439, 1957. 204. Rossi, G. F. and A. Zanchetti. Arch. ilal. biol. 95: 199, >957- 205. Rothballer, A. B. Electroencephalog. & Clin. Neurophysiol. 8: 603, 1956. 206. Scheibel, M. E. and A. B. Scheibel. In: Reticular For- mation of the Brain, edited by H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay and R. T. Costello. Bos- ton: Little, 1958, p. 31. 207 . Scheibel, M., A. Scheibel, A. Mollica and G. Moruzzi. J. Neurophysiol. 18: 309, 1955. 208. Schiller, F. A.M. A. Arch. Neurol. & Psychiat. 67: 199, ■95^ 209. Schreiner, L. H., D. B. Lindslev and H. VV. Macoun. J. Neurophysiol. 12: 207, 1949. 210. Schreiner, L. H., D. M. Rioch, C. PeChtel and J. H. Masserman. J. Neurophysiol. TO: 234, 1953. 21 1. Segundo, J. P. Ada neurol. lalinoam. 3: 245, 1957. 212. Segundo, J. P., R. Arana andJ. I) French. ./. Nrurosurg. 12: 601, 1955. 213. Siiarpless, S. and H. Jasper. Brain 79: 655, 1956. 214. Simon, G. W. and W. H. Emmons. Psychol. Bull. 52: 328, ■955- 215. Simon, G. W. and W. H. Emmons. .Science 1 -•.] mlili, m,<> 216. Simon, C. W. and W. H. Emmons. ./. Exper. Psychol. 51: 89. >956- 217. 218. 219. 220. 221. 223. 224. 225. 226 227. 228. 229 230. *3l- 232. Simon, C. W. and W. H. Emmons. Am. J. Psychol. 69: 76, I956. Smith, J. R. J. genet. Psychol. 53: 471, 1938. Speigel, E. A. Zlschr. ges. exper. Med. 55: 183, 1927. Sperrv, R. \V. Am. Scientist 40: 2gi, 1952. Sprague, J. M., L. H Schreiner, D. B. Lindsley and H. W. Magoun. J. Neurophysiol. 1 1 : 501, 1948. Starzl, T. E. and H. \V. Magoun. J. Neurophysiol. 14: ■33. '95'- Starzl, T. E. and D. G. Whitlock. J. Neurophysiol. 15: 449. '952- Taussig, L. Rei: neurol. Psychiat., Prague 28: 387, 1931. Tissot, R. and M. Monnier. Helvet. physiol. et Pharmacol. Acta 16: C33, 1958. Travis, L. E. J. Exper. Psychol. 21 : 302, 1937. Tromntr. I. Deutschi Ztschr. Nerocnh. 105: 191, 1928. Verzeano, M , I) B. Lindsley and H \V Macoun. J. Neurophysiol. 16: 193, 1953. von Economo, C. In: Handbuch der normalen und patho- logischen Physiologic, edited by A. Bethe, G. von Bergmann, G. Embdcn and A. 1 llingei Berlin Springer, 1926, vol. >7, P 59'- VON I.conomo, C. Ergebn. Physiol. 28: 312, 1929. W \i Kir, A 1 Congr. lalinoam. neurocir., Montevideo 6: 926, '955- Walter, W. G. J. \ Rr) Mai 1 3 -,9, 1938. WiGCERS, (' I Physiology m Health and Disease. Philadel- phia Lea, 1937, P- 250. Wikler, A. Proc. Soc. Exper. Biol. ~J Med. 79 jbi, 1952. CHAPTER LXV Perception HANS-LUKAS TEUBER1 Psychophysiological Laboratory, New York Universily-Bellevue Medical Center, New York City CHAPTER CONTENTS Psychophysics Reduction of Sensory Qualities Attempted Restitution of Qualities in Psychophysics Multidimensional Nature of Sensory 'Attributes' Implications for Neurophysiology' Studies of Perceptual Processes Impact of Gestalt Psychology and Operational Behaviorism The Need for Converging Operations Some Central Problems for a Theory of Perception Patterning Selectivity Reaction to Relations Similarity Serial (Temporal) Order Equivalence of Certain Temporal and Spatial Patterns Perception of Shape Basic Aspects of Figure Processes Minimum articulation, assimilation and contrast Figure and ground Principles of grouping Ontogenetic Considerations Phylogenetic Considerations Vertebrates Higher invertebrates: cephalopods Salticidae Invertebrates with compound eyes Changes in Perception of Shape after Cerebral Lesions Effects of total removal of 'projection systems' Effects of subtotal removal in man Completion and extinction of patterns Coexistence of specific and general perceptual changes Effects of subtotal lesions in animals below man Isolation Studies Pattern vision after removal of congenital cataracts Effects of early visual deprivation on shape perception in subhuman species Alternative interpretations 1 The preparation of this chapter has been aided by the Commonwealth Fund of New York, and by the U. S. Public Health Service under their program grant M-3347 to the Psychophysiological Laboratory, Depth, Distance and Other Aspects of Spatial Localization The Traditional Approach (Depth from Clues) and Its Alternative 1 Depth from Gradients) Monocular clues Depth from gradients Kinetic depth ellri Is Binocular parallax rVcuit) of binocular depth perception I In l'» iis ill I ill, 11 lusiun \iiiliicu v Spat r Perception Binaural parallax Auditory localization during head and body movements Interaction Between Posture and Distance Receptors in Spatial Localization Effects of body tilt: the Aubert phenomenon and its variants Oculogravic effects and related phenomena Abnormalities of Space Perception after Cerebral Lesions Distortions in the tridimensional structure of visual space Abnormality of visuopostural interaction Spatial disorientation' after parietal lesions 1 )epri\ ation Studies Depth perception after early \ isu.il deprivation Tactile deprivation: effects on body scheme Recombination 1 Rearrangement and Disarrangement) Studies Persistent spatial disorientation in lower species Adaptation in man to prolonged visuospatial inversion or distortion Partial spatial reorganization during short-term experi- ments •Reafferent' stimulation as prerequisite for adaptation Double localization: monocular diplopia and diplophonic effects Perception of Apparent Motion Afterimages of Motion Induced Motion Autokinetic Effects Stroboscopic Motion Tactile and auditory apparent motion Stroboscopic effects in subhuman species A Physiologic Hypothesis 1595 396 ll\M)BOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III Perception of 'Real' Motion Characteristics in Normal Subjects Minimal rates Differential thresholds Phenomenal stages Role of size and surround Velocity transposition 'Paradoxes' of seen motion \ I 'normalities of Perception of Motion Altered motion perception after cerebral lesions Isolation studies Recombination (disarrangement) studies Constancies, Illusions and Figural Aftereffects Constancies : Examples and Measurement Interpretations ((instancy in animals and children The need for parametric studies Recent work on constancy of color and brightness Role of instruction and experimental setting Effects of 'reduction' Loss of Constancies after Cerebral Lesions Deprivation and Recombination Studies Illusions: Phenomena and Interpretations Illusions as misapplied constancy effects Perceptual habituation; decrement of the Miillcr-Lyer illusion on repeated trials Intermodal transfer of Miiller-Lyer decrement; 'haptic' illusions Figural aftereffects Conclusion it may seem strange to find a chapter on perception in a handbook of neurophysiology, doubly strange if the chapter begins with the claim that there is no adequate definition of perception and ends with the admission thai we lack a neurophysiology theory. The two difficulties have a common source: physiolo- gists distinguish sensation from perception and deal with sensor) processes, in different modalities, as dependent upon receptor mechanisms and subsequent neural events. As a result, perception assumes (he role ui sume supplementary higher process, super- imposed upon these sensory capacities and devoid of an) obvious neural correlate. After excluding per- ceptual phenomena from sensory physiology, we are thus hard pul to explain how different sense modalities inn i. M i in perceiving, how we apprehend shapes or sizes, distance or depth, or, more broadly put, how things manage to look, or feel or sound the way they do (284). In the absence <>l an accepted definitionand theory ui perception, any selection from iis vasl literature will be arbitrary. Yci we shall try to stress those phei una which reveal (lie inadequacy i>l tradi- tional distinctions between sensation and perception, and those problems which seem most in need of physiologic interpretation. We shall deal, first, with subjective intensity and psychophysics in order to show that, even here, where simple one-to-one rela- tions between stimulus and sensation are assumed to exist, there actually arise complex issues that require a new approach to our search for physiologic corre- lates. The same issues will be raised with regard to the more traditional topics for students of perception : the problems of perceived shape, depth and motion, the constancies and the illusions. None of these topics can lie covered completely, but we shall attempt to concentrate throughout on those perceptual phenomena that can be demon- strated not only in man, but in lower forms. Corre- spondingly, we shall try to stress methods for studying perception that might be applicable to a wide variety of vertebrate and even invertebrate species. This is done in spite of the risk of seeming capricious in our choice of experimental evidence. The comparative approach permits one to utilize naturally occurring differences of neural structure in formulating guesses about tin' way in which neural structures might determine perceptual performance. Moreover, the problems of pattern vision in monkey, octopus, or bee will seem less remote if one realizes that (he nervous systems ol infrahuman forms are available for experimental intervention not permissible in man and that the greater simplicity, especially of invertebrate neural organization, may let us discover correlations between structure and function (hat might go undetected in the study of human subjects. In contrast to the usual picture drawn by surveys ul comparative sensory physiology, most ol the phe- nomena we shall note cannot be interpreted in terms of known physiologic mechanisms. These unexplained effects are presented, nevertheless, because they m.iv suggest directions in which the search for neural bases ul perception should proceed. Our effort .it reviewing these phenomena might be justified il we remember thai the facts observed in any field are largely a function of the questions that are being asked. Understanding of neural activity "is likelv to progress more rapidly il we keep before us the facts ol behavior which we hope eventually to explain" (308). If perceptual phenomena continue to be refractory io neurophy siology . one mighl wonder whether current approaches (o 'basic' sensory proc- esses are correctlv conceived. The difficulty may be analogous to the one raised by phenomena of complex coordinated movements (see Chapter LXVI1 by Paillard in this volume). As PERCEPTION r597 long as one's efforts are concentrated on reflex mecha- nisms of the spinal cord, the problems of patterning of skilled movements seem residual and nearly in- soluble questions, relegated to some 'highest inte- grating activity' of the central nervous system. The very fact, however, that patterning should thus re- main unexplained casts serious doubts on the ade- quacy of current theories of coordination on spinal as well as supraspinal levels (531 ). The distinction between sensory and perceptual processes, and the preoccupation with the former at the expense of the latter has historical rather than logical reasons. We can trace the origin of this dis- tinction by considering, first, a problem which cuts across individual modalities of sensation, the problem of measurement of sensory intensity and quality (psy- chophysics). PSYCHOPHYSICS Reduction of Sensory Qualities The traditional distinction between sensation and perception and the resulting difficulties have a long history [admirably treated by Boring (53)]. Empiri- cism has always been sensationism; experience could not be acquired except through the senses, and the senses could not err except for erroneous interpreta- tion of their unequivocal message. Sensations were elementary conditions of learning, while perceptions were complex and learned; they were considered as compounds in which present sensations gained their meaning through related residues, 'images,' 'ideas' of previous sensory events [see Berkeley (38)]. Yet the sensations, so conceived, were actually postulates of classical physics. From its renaissance beginnings, physics has sys- tematically restricted its data by reducing the quali- tative richness of everyday sensory experience to certain quantifiable aspects of matter and of matter- in-motion. Galileo established physical acoustics by suggesting that perceived pitch might be reducible to frequencies of vibration in a medium such as air or water. [See Galileo Galilei (141); the Galilean discovery was communicated two years earlier by his correspondent Mersenne (344)] Newton analo- gously referred perceived differences in hue to corre- sponding differences in vibrations of the ether, likening these to vibrations of air which "according to their several bignesses, make(s) several tones in sound." He did so in spite of his increasing prefer- ence for a corpuscular theory of light; in fact, his curious insistence on seven primary colors was based on an explicit analogy to the seven tones within the octave (361). The progressive elimination of sensory qualities in physics culminated in the nineteenth century,2 and it was at that time that 'psychophysics' arose as a systematic effort to reintroduce the 'lost' qualities through a special form of experimentation. The very form of these experiments at first reinforced the distinction of sensation and perception. Attempted Restitution 0/ Qualities in Psychopfq In the typical psychophysical procedures of the nineteenth century, as formulated by Weber (523), Fechner (122), von Helmholtz (501) and Wundt (551), attributes of sensation, such as pitch, were ex- plored by systematic \ariation of a single physical dimension, such as frequency , as if pitch were the per- ception of the frequency of a tone. Thus, a one-to-one correspondent!' between physical and sensory dimen- sions was assumed, although Weber I 523) had already established that such a correspondence was not linear: for anv given intensity of stimulation, /, the just noticeable difference, A/, was known to vary with the intensity of stimulation, where / was small, SI tended to be small, where / was larije, A/ tended to be lame, in fact, Weber believed that one might formulate a law : A/ / = constant (Weber's law I. For instance, in judging differences of weights placed successively on the finger, the just-discrimi- nable difference (or, as we now say, the Weber frac- tion) was thought to be approximately constant at 1 30. From Weber's law, Fechner derived, by inte- grating, the expression .V = A' log / (Fechner's law) where S is the magnitude of sensation, measured in some appropriate unit, and / the stimulus measured in terms of physical units defining the absolute threshold. This simple logarithmic relation between stimulus and sensation cannot be valid if just-discriminable differences of sensation are unequal, i.e. Fechner's law cannot hold if Weber's law is incorrect. The evidence, available for many sensory dimensions, - It could be argued that classic physics, by eliminating most aspects of everyday experience from its primary data, elevated somatic sensation to the status of prototype of all sensation; mechanics implies action and reaction through contact, especially if forces acting at a distance are denied. The much greater abstractness of modern physics is accomplished by abandoning any attempt at reconciling the structure of physical reality with that of perceptual phenomena. '598 IIWDBOOK OF PHYSIOLOGY M I ROPHYSIOLOGY III tends to show thai Weber fractions arc only very approximately equal; over a given ranee of stimulus values, the Weber and Fechner expressions are usually leasl adequate at the extremes of the range. Moreover, recent developments in psychophysical scaling meth- see Stevens (451)] suggest that Fechner's loga- rithmic expression might have to be replaced by a power function. V Stevens points out, the systematic measurement ol sensation in relation to a given stimulus dimension can he performed without recourse to the concept of differential thresholds and without assuming the constancy of such differences. When a picture in black and white is viewed, first in the sunlight and then in the shade, we ordinarily perceive little, if any, change in the relative brightness of its features. This could mean that the differences in brightness have remained the same under changing illumina- tion (a> Fechner might have said) or, alternatively, that the ratios of light and dark portions have re- mained unaltered. The assumption of constancy of subjective differences would lead to the formulation \f/ = kilogip (a restatement of Fechner's law) in which the psychological magnitude \p is related to the logarithm of a physical dimension ip, and a constant ki. If we assume, instead, that the proper relation is based on equal ratios, we obtain \p = k?s Plati iu ■;•'■ 1 who jectured thi 'powei law" after he had asked eight aitists to paint, independently, a hi ay halfway I «>-i \s ■ i. in.- white and hl.uk The grays turned out 'presqui identiques and remained so undei different illumina- tions I ..it. 1 in his career, Plateau retracted his views, 1 aril) mi. 1. 1 the influence ..1 certain psychophysical experiments bj 1 1 Ibo tl 99 s- . Stevi us 1 power. The values for the exponent n cover a con- siderable range. For loudness, the exponent is 0.3; for the subjective intensity of electric shock to the fingers, it is 3.5 (451 ). The apparent subjective magni- tude of an artificial 'star' grows roughly as the square root of the photometric level. These bare figures do tell us a good deal about the differences between various sensory dimensions. For apparent loudness, where the exponent is around 0.3, there is enormous compression of the scale of magni- tudes, since in order to double the apparent loud- ness, we must multiplv the physical energy by 10 (or the sound pressure by the square root of 10). For the 'unphysiologic' mode of stimulation by direct electric shock to a subject's finger, the converse is true; here, the subjective intensity shoots up as the 3.5 power of the current applied — a fact to bear in mind when applying direct electrical stimulation to peripheral nerves But whether the organism com- presses or expands a given stimulus dimension, the basic psychophysical relation would be simple: equal stimulus ratios produce equal subjective ratios (451).5 Multidimensional Nature nj Sensory 'Attributes' Whether logarithmic or exponential, these psycho- physical relations suggesl a simplicity of sensor) processes which would seem to set them apart from the complexity- of everyday perception of objects. This impression of simplicity, however, is deceptive. The classical attempt at restitution of sensory qualities, as we termed it, presupposes the existence of inde- pendent one-to-one relations between sensory and stimulus dimensions. But il pitch is nothing else than perception oi frequency, or hue of wavelength, it would be difficult to understand in what sense Galileo or Newton made discoveries about these physical correlates of sensation. The fact is that psychophysie.il relations need not be one-to-one, or isomorphic in the sense that to each definable physical dimension there corresponds one and onl) one sensory dimension, litis lack oi parallelism has such far-reaching impli- cations that it needs to be considered here in some detail. For Wunclt, each sensation had two 'attributes,' vi/. its specific qualitv and intensity (551). To this meagei inventory, extension in space and duration in 'Attempts at identifying electrophysiologic correlates ol sensor) scales e.g Granit (168) must be reconsidered in the light ni psy< hophysical evidence which shows the inadequai ies oi the Weber-Fechner formulation. PERCEPTION '599 Invariable points "i n 1 1 1 1 r 675 650 625 600 575 550 525 500 475 450 Millimicrons fig. I. Contours for constant hue. A given contour describe' the combinations of wavelength and intensity of a mono" chromatic light which yield impressions of equal hue. Only three points within the visible spectrum are 'invariable,' i.e. show no shift in apparent hue with changing intensity. Except for these invariable points (a yellow at 572 m/i, a green at 503 itui and a blue at 478 m/il , all other points vary in a predictable fashion according to the Bezold-Briicke effect. There is a fourth invariable point which is not shown, since it lies outside the spectrum, being constituted by a specific mixture of long and short waves, in the purple. [From Purdy (391 I. time were added b) Kiilpe (296), and his four at- tributes of sensory quality, intensity, extensity and protensity (duration) have guided the search for neural correlates of perception for more than half a century. There is, however, no reason why this should be so. Consider the case for vision. Here, a basic sensory quality would be color (perceived hue), ordinarily thought .is corresponding to wavelength (or mixture of wavelengths) ; the intensity dimension would be represented as perceived brightness. Yet, with few exceptions, hue is not only dependent upon wavelength, but upon intensity; except for certain invariable points in the spectrum, all colors shift in hue towards either yellow or blue as the intensity is increased. This phenomenon of shift in hue with changing intensity — the Bezold-Briicke effect — can be quantified by plotting contours of equal hue (see fig, 1) which describe the combinations of intensity (in photons) and wavelength (in millimicrons) which will yield equal-appearing; hues [see Purdv (391)]. Perceived hue, thus, is a joint function of wavelength and intensity. Correspondingly, perceived brightness is a joint function (although a different function) of the same physical dimensions; even a casual glance at a spectrum will reveal that different wavelengths, although energy is constant, have different-appear- ing brightness within the spectrum. Moreover, when over-all energy is reduced, the relative brightness of the different colors shifts in accordance with the Purkinje phenomenon. (Characteristically, this Pur- kinje shift fails to appear in the rod-free part of the retina.) We thus have ample evidence for the dependence of hue on both wavelength and energy (as the Bezold- Briicke effect shows) and for brightness, again, on energy and wavelength (as the Purkinje effect shows), although two different functional relations establish these two psychological dimensions, hue and bright- ness. Lest there be a suspicion of some lingering pre-established harmony (two physical for two psycho- logical dimensions), we might consider a third psycho- logical dimension of colors, saturation. Saturation, too, varies with wavelength (being minimal in the spectral yellow and violet) and with encrgv (being maximal at intermediate energy levels I, so that we can get three systematically discriminable aspects of color sensations out of the appropriate combinations of only two physical 1 stimulus 1 dimensions [see Boring (-,ti . I In- situation has considerable generality. In audi- tion, perceived pitch is easilv identified with fre- quency, and loudness with intensity of sound waves. Yet lor pure tones, one finds an analogue of the Bezold-Briicke effect since perceived pitch changes as intensity is changed, even though frequency is held constant. Low-frequenc) tones, when raised in in- tensity, sound lower in pitch, while high-frequencv tones sound higher in pitch when their intensity increases I iv>, V" !• There even is an 'invariable point' which falls roughlv into the region of maxim. il sensitivity along the frequency spectrum. As a result, we can plot contours of equal pitch (fig. 2) which indicate the changes in frequency that need to be made in order to counteract the alterations in pitch produced by a given change in intensity.8 But loudness, too, is ,1 joint product of frequency 6 Happily for instrumental music, the shifts depicted in fig. 2 are found in this pronounced fashion only for pure tones, but much less for the complex 'timbered' tones produced by musical instruments It is not clear whether this relative constancy of pitch for complex tones is due to the presence of harmonics or whether other factors play a role. However, the phenomenon of greater pitch constancy of these complex tones belongs to the large group of effects called 'perceptual constancies' discussed below. It should be remembered, in this context, that the statements about color sensations made in the preceding sections are similarly restricted in scope; they apply primarily to 'film' colors, i.e. colors which do not in any obvious way belong to the surface of seen objects. [See Katz (249. 25" ; I hi id IIWDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III -40 •30 -20 -10 0 10 20 30 Intensity in db (zero db = 1 dyne/cm7) 40 no. a < mitours for constant pitch. A given contour describes the combinations of frequency and intensity of a pure tone which yield impressions of equal pitch. The ordinate was arranged so that a contour with positive slope shows how pitch increases with intensity (for high frequency tones). [Based on data from Stevens '450).] and intensity. The equal-loudncss contour (449) indi- cates, for a standard tone of low frequency, how intensity has to be changed to balance changes in frequency in a comparison tone so that the compari- son tone appears to equal the standard in loudness (see fig. 5). In addition, figure j shows that at least two other psychological dimensions, 'volume' and density,' can lie obtained by appropriate covaria- tion of frequency and intensity. It is apparent from these sets of contours that two dimensions on the physical (stimulus) side can yield more than two perceptual dimensions. Moreover, a perceptual dimension is not established by independ- ent variation, as used to be believed in the classical psychophysics "i the nineteenth century i Yii). It is actually impossible to vary any of the dimensions de- picted in figure ; in isolation, that is without varying the others What we have, inste.nl, is .1 ease of 'de- pendent constancy,' or in variance; we can maintain a percepl l>\ appropriate variation of all of the other relevam perceptual dimensions. The concept of in- 64 62 - 60 58 - bb V 1 Z~ — X -c < 1 400 450 500 550 Frequency in cycles per second 600 fig. 3. Isophonic contours for pitch, loudness, volume and density. F.ach contour defines the combinations of frequency and intensity at which a comparison tone will he perceived as equal in pitch or loudness, or volume, etc., to the standard tone of 500 cps and 60 db. [From Stevens (449).] variance will assume increasing significance as we progress to other problems of perception in this chapter. Implications jo Neurophysiologic Studies of Pin, jitual Processes There are at least three lessons that may be de- rived from this brief review of nineteenth- and twen- tieth-century psychophysics. They concern a) the stage in the nervous system where sensors differentia- tion occurs; b) the complcxitv of sensory processes which makes their separation from perception prob- lematic; and 1 ) the need for uniform methods — an enlarged psvehophysics — to attack, together what used to be treated separately as sensory and per- ceptual phenomena. We have seen that the organism can respond dif- ferentially to sensory stimulation in such a way that the responses can have more dimensions than the stimulus itself. Different functions relating frequency and inteusitv of sound yield at least four sensory attributes, and perhaps more. If this is true, then our search for neural liases of such differentiation should proceed with yreat caution; clearly, the differentia- tions iiiav occur at auv stage, from the periphery (the receptors) to the final paths antecedent to the discriminating motor responses. Thus, it should not be surprising to find an organism making discrimina- tions which appear to go beyond the differentiation in its peripheral receptors. Nor need one assume that PERCEPTION l60I all differentiations of which a receptor structure is capable ought to be preserved in later stages within the nervous system; differentiation made possible at the periphery may be lost at some subsequent stage. The absence of color discrimination in the cat may- be a case in point, since the cat retina appears to make color discriminations even though behavioral evidence for such discriminations on the part of the 'whole cat' is lacking (see Chapter LX by Neff in this volume). If sensory discriminations, according to the views developed here, no longer bear a simple one-to-one relation to the stimulus dimensions, then one of the stronger nineteenth-century arguments for separating sensation and perception is lost. In the new 'dimen- sional9 psychophysics, sensations have a complexity not unlike that of perception, even though the earlier attempts at separation have not been entirely aban- doned. Consider the classical psychophysical pro- cedures: the matching of a monochromatic color with a dichromatic mixture, the determination of just- discriminable differences along a particular stimulus dimension, the discovery of absolute thresholds. In all these procedures, the hoped-for simplicity ol re- sults was marred by 'errors'; matches between stimuli differed, depending on whether the standard stimulus was placed to the left or right of the variable ("space errors'); stimuli were over- or underestimated, de- pending on whether they were given first or second in a successive comparison ('time errors'); judgments of weights differed markedly, depending on the presence or absence of much lighter or much heavier weights in the series submitted for comparison Cintra- serial effects'). All these phenomena, though interest- ing in their own right, tended to be treated as sources of 'error' in strict psychophysical experiments; the) were to be canceled by balancing the stimulus presen- tation so that effects of spatial position, or temporal order, or of constitution of the series could be elimi- nated. There is, of course, nothing illegitimate about such attempts at controlled experimentation; but the distinction between the intended results as 'primary,' or 'sensory,' and the space or time effects as 'errors,' produces rather than proves the elementary nature of sensation. At the same time the study of percep- tion becomes an exclusive concern with these 'errors' which are left over, as it were, as mere residuals after the more serious business of sensorv psvehophysics has been accomplished. Actually, intraserial effects and related phenomena can be dealt with just as rigorously as the supposedly more elementary threshold phenomena. Helson's work on 'adaptation levels' (197) shows that the subjective magnitude of a singly presented stimulus depends upon the weighted geometric mean of the series of stimuli that the subject has worked with; but, as in all psychophysical procedures, explicit or implicit verbal instructions are crucial since they determine which stimuli are perceived by the sub- ject as forming part of the series. [However, Stevens' protest (452) should be noted.] That this is so, can be demonstrated very simply. Ask a subject to move a particular weight out of the way (with the excuse that it is cluttering up the table), and this particular weight will not enter into the formation of the series; it will have no effect on the subjective magnitude of the other weights within the test [see Brown (65); also Bruner (71)]. The purity of the psychophysical experiment thus depends largely on the perceptual set (71) adopted by the subject, and the distinction between sensory and perceptual aspects of the task is spurious. Beyond the consideration of space and time errors, and of intraserial effects, perception studies in the later nineteenth century were primarily concerned with two sets of phenomena, the so-called "illusions' and 'constancies.' To this day, these have remained central problems in perception. Every textbook shows samples oi some geometrical optical illusions — the famous Muller-Lyer pattern, for instance (see fig. I 1 I .uul most texts list the speri.il theories developed to account for these effects [cf. Boring (53, pp. 238- 2 1 y, Woodworth & Schlosberg ( J40J, pp. 417-423)]. Yet to speak of illusions .is speei.il cases — curiosa of perception, as it were is tendentious. As soon as one admits that perception lacks simple one-to-one corre- spondence to physical stimuli, the explanation for perceptual illusion will be sought among the general laws of perception. Once these laws are known, the illusions themselves will be understood (53). The same may be said lor the so-called constancies, e.g. the relative constancy of perceived brightness, size or shape, with varying illumination, tilt or dis- tance. The classical treatment of these effects as re- sults of "unconscious inference' by von Helmholtz (501) implies that we ought to be able to perceive the "image' on our retina, and then elaborate our perception by recourse to some judgmental factors into a corrected or interpreted view of the distal ob- ject to which the image refers. But to assume that we ought to 'see' the retinal image may be just as naive an assumption as the belief that in hearing a given pitch, we ought to perceive a frequency. The very notion of a retinal 'image' is a curiously anthropo- i6oi IIWIIBOOK OF PHYSIOLOGY NELROI'HYSIOLOGY III morphic way of describing the receptor processes which antedate perception Although the 'unconscious inference' theory of von rlelmholtz has been revived in some of the cur- rent studies of perception from the 'tr ansae tionalist' viewpoint [see Ames (7), Ittelson & Cantril (236)], the prevailing approach to the 'constancies' and to the 'illusions' is based on the belief that nearly all perception is perception of objects, and that the con- stancies pervade perception on all levels, from the most elementary to the most complex. Impact "i (,< stalt Psychology and Operational />'< haviorism This unitary approach to perception can be credited largelv to two sources: first, to Gestalt psychology (and phenomenology) and second, to certain trends in operational behaviorism. The Gestalt psychologists, notably Wertheimer (536), Kohlcr (269) and KotTka (284), from the outset rejected all notions of elemen- tary sensory processes. To hear a pure tone or to see a single contour meant to them to perceive organized structures, and they considered it their task to search for the laws underlying all perceptual organization. In this respect, the) continued the tradition of phenom- enology the careful description of sensory experi- ence. In such descriptions, errors of perception were not features to be eliminated; they were to be studied as clues to the basis of normal function. Thus, Pur- kinje (393) was convinced that "illusions of the senses tell us the truth about perception" ("dass Sinnes- tauschungen Gesichtswahrheiten sincl "); and he re- corded the manifold phenomena of entoptic imagery — the illusions produced by inadequate stimulation of tin- retina, tiering (2011, 202), the main antagonist of von rlelmholtz, surveyed phenomena of simultane- ous and successive contrast, of afterimages and mem- ory colors, demonstrating that a patient study of sensor) phenomena leads to the discovery of many problems for sensors physiology which a more rigid psvchophvsical approach would unavoidablv pass by. It is clear lh.it such phenomenal distinctions are I1.1 id on consensual validation, an appeal to intro- spections of the other man. In modern behaviorism appeal- ol this son ,uc proscribed; die tendency has been to reduce all studies of perception to that of sensory discrimination, As Boring ( ~, () has pointed out, this tendency, too, results m ,, denial oi any dichotomy between sensation and perception; in behav torism, sensation (in the form of discrimination) has absorbed perception, while in phcnomcnologv and Gestalt psychology, perception has absorbed sensa- tion. The Need for Converging Operations The reduction of perception to discrimination has not gone without contradiction within the framework of modern behaviorism itself. Thus, Garner et al. (144) have pointed out that to reduce perceptions to the overt discriminatory responses by which they are observed would seriously restrict the scope of experi- mentation. The reduction of perception to discrimina- tion is usually defended by operational bchaviorists, such as Graham (166), by referring to Bridgman's critique of concepts in physics. For Bridgman (61), "a concept is synonymous with the corresponding set of operations," i.e. the experimental operations by which the concept is established. However, this does not mean that any operation will generate a meaningful concept. If one thinks of perception as intervening between stimulus and response, and thus distinct from either of these two terms, it is obvious that a particular kind of operation is needed to estab- lish whether a series of events is perceptual in nature. This can be done, as Garner el al. (1441 suggest, by means of converging operations. These are any sets of two or more independent experimental procedures which lead to the same terminus, i e, establish .1 con- cept by ruling out alternative interpretations. To take an example from a currently popular area of experimentation consider the supposed difference in perception of words, or of drawings, or symbols with different emotional content [e.g. Met Jinnies (342 I . In a hypothetical experiment of this type, lour words are exposed at high speeds (i.e. tachistOSCOpically) and the subject is required to recognize them. If two of these words are neutral, and two vulgar, the result will be that the neutral words are read .11 high speeds, while the vulgar words are not, or at least not until their exposure has been made quite long. Such a result is eouunonlv interpreted .is perceptual defense, since failure to read is considered identical with failure to perceive However, the failure to report vulgar words may be .1 characteristic of the subject's responses rather than his perceptions. A converging operation would be 10 pair vulgar responses wilh the neutral stimulus words, and vice versa l 144). Only under these conditions could one decide whether the 'defense' was, in fact, perceptual, ll should lie noted ih. 11 both operations are needed il one cares to estab- PERCEPTION 1603 lish whether the defense was a mere withholding of unpopular responses or a truly perceptual effect. The need for converging operations is particularly pressing in work on perception in young children or in animals below man. In the absence of verbal re- port we have to depend entirely on differential motor reactions on the subject's part, but this does not mean that any given set of these reactions is sufficient to reach conclusions regarding the subject's perceptual repertoire. Earthworms will pull pine needles into their burrows by grasping them at the base (where the pair of needles hang together). Conversely, they pull leaves into their burrows by grasping the tip, never the stem, so that the leaf rolls itself into a tube as it disappears into the burrow. This highly adaptive behavior looks like form perception, but the basis of discrimination is gustation [cf. Mangold (338)]. The worms avoid powdered extracts of the base of leaves and ingest powdered extracts of the tips, and con- versely for pine needles. Thus, converging operations are needed to define the sensory basis of discrimination; moreover, any tendency to reduce all perception to discriminatory responses would reintroduce the classical sensory bias and force us to omit all those problems from con- sideration which are crucial in developing a physio- logic theory of perception. SOME CENTRAL PROBLEMS FOR A THEORY OF PERCEPTION What are those perceptual phenomena which the traditional physiology of the senses has left out? It is difficult to make a complete list, but we can enumerate half a dozen aspects of perception which anv physio- logic theory would base to take into account. These aspects are, briefly: a) patterning, b) selectivity, r) reaction to relations or ratios of stimulation, d) reac- tion to similarity, e) apprehension of serial (temporal) order, and f) equivalent reactions to certain spatial and temporal sequences [cf. Lashley (308)]. Patterning Perception is patterned, since some parts of a stimulus array are always perceived as belonging together while others are not [cf. Schumann (415)]. We tend to see (or feel) 'things,' and not the holes between them [cf. Koffka (284)]. Temporal sequences are analogously structured into events [cf. Johanssen (241)]. Such patterning would imply, as its neural counterpart, an interaction of concomitant and suc- cessive processes in afferent systems. Interaction be- tween different sense modalities is perhaps a special form of this general patterning of perception. The difficulty here is not the interaction as such, but its organization — the how and why of interaction — in direct analogy to the problems of coordinated move- ment. Selectivity Perception is selective. Although the structuring of what we perceive is largely determined by the actual distribution of stimuli in an array, we can selectively attend to one part as against others (see fig. 14). Such selectivity, in neural terms, would probably amount to some preliminary priming or sensitization (308), favoring one part of a neural system over others and hence one way of structuring a complex sensory input over its alternatives. Another, and currently attractive, view would assign a special role in this selection or filtering to extralemniscal afferents (245, 324), or to efferent pathways to sense organs (139, 140, 168, 203, 204 Rem lion in Relations Perception is relational, that is the organism reacts to ratios of excitation rather than absolute amounts. A classic instance of this is the case of 'transposition' of learned reaction to size differences [cf. Kohler (266)]. Once trained to select the larger of two stimuli (e.g. two squares, a, 10 cm and b, 20 cm in height), the subject readily selects the larger member of another pair of stimuli in which all dimensions have been doubled (thus, he will choose b', when a' = 20 cm and b' = 40 cm). There is little difficulty in con- structing models of central nervous activity that would incorporate this feature, but most speculations about the neural basis of discrimination (see Chapter LX by Neff in this volume) either have ignored it, or have tried to reinterpret these phenomena in Pav- lovian terms by invoking principles of conditioning and of "primary stimulus generalization' (374). Con- siderable ingenuity has been displayed in these re- interpretations [cf. Spence (434, 435) and Hull (226)], but the attempts have not been convincing [for cri- tique, see Kliiver (259), Lashley (304), Hunter (227, 228), Rudel (407, 408) and others] and are contra- dicted by some experimental evidence (162, 464). Thus, a child or animal (chick, monkey or chim- panzee) can be trained to select an object of inter- 1604 HANDHIKIK OF rlt\slc .Y NEl'ROI'HYSIOI.OCV III mediate size among three objects differing in size; having learned to do so, the subject can, under some conditions, 'transpose' this middle-size reaction to a new set of objects of greater or lesser absolute size (162, 407). Apparently, most discriminations involve responses to relationships between the dbcriminanda. Similarity Closely related is the central problem of similarity in perception. As Mach (333) pointed out, geometric similarity is not necessarily identical with optic (or, more generally, perceived I similarity, since we can- not predict, on the basis of purely physical characteris- tics of a stimulus, which patterns will be perceived as similar.7 Stimuli are similar in some respects and not in others so that the problems oi similar- it\ involve all those listed up to now, patterning, selection and transposition. Perception of similarity is basic to our capacity- for recognition, and thus for inclusion of what we see or hear or feel into classes oi stimuli. What the neural basis of such classificatory activity might be has not even been a matter of much speculation [but see Semon (419), Craik (94), I hi ib (188) and MacKay (335)]. Some guesses about it would be needed in understanding agnosia, a condition claimed to consist of a specific loss or impairment of this classificatory activity in the presence of particular brain lesions. As we shall see, however, there is considerable controversy as to whether agnosia, in this sense of the term, could ever occur. Serial I / emporal) Order A yie.it deal ol work on perceptual phenomena has been restricted to the studv of --t.it it" patterns, usually visual displays in two dimensions. While these pat- terns serve to underscore the importance oi grouping and perceptual selectivity, undue reliance on these static demonstrations leads to a neglect of temporal sequences in perception, the problems ol serial order. ■ 1 . 1 k« the riddle "I octave similarity, To must listeners, tones one 01 tave apart are more similai than are tones within the octave, and tones differing l>\ an octave are readily con- fused Perhaps Newton's preoccupation with the musical u.is noi so strange afta all, although ii remains odd thai he ga 1 octavi itructure to the visible spectrum ["he anthropomorphii charactei ol ilus attribution is lerscored by the difficulty ol showing octave similarity in animals below < 1 although then ar< data suggesting that octave similarity iSlsIs I., I I .lis |it In most languages, the sequence of words in a phrase conveys specific meaning, and the sequence of pho- nemes within a word has crucial importance in all forms of human speech (see Chapter LXYI11 1>\ Zangwill in this volume). The apprehension of such patterns of successive stimulations clearly requires some temporary storage of the information received at any one moment (62, 64), and capacity to deal with the completed sequence as if it were a simultaneous pattern. Equivalence of Certain Temporal and Spatial Pattt 1 n 1 Lashley (307) has pointed out that within limits we can apprehend a visual (or tactile) configuration, irrespective of whether we explore it by scanning, part for part, or whether we see it "at one glance' (or have it impressed upon the skin). The neural basis for this transposition of serial into simultaneous patterns is unknown but would seem to require some central mapping of temporal into (simultaneous) spatial orders. If we look back upon this list of problems, it will be apparent that they are interdependent. Nor is it likely that the listing is exhaustive (5, 21, 489, 490). Yet enumeration of these half dozen aspects of per- ception should remind us of the existence of problems which traditional physiology has rarclv considered. The list may also help us to indicate a striking simi- larity (in logical structure) between two seemingly. unrelated fields; experimental embryology and the psychophysiolosy of form perception. In both areas, the origin of form is a central problem. In both, clues for analysis can be derived from essentially three sources: studies involving delect, isolation, or recom- bination (530). In defect experiments, the develop- ment of organic form (or perception of patterns) is investigated in the presence of a lesion in the struc- ture. In isolation experiments, the behavior of a part of the structure is studied alter it has been cut off from tin' remainder of the living svstem. or the entire svstem is studied under conditions of artificially diminished input. In recombination studies, finally, one inquires into effects ol altered arrangement in the spatial relationships ol parts, as in embryologic ex- periments involving transplantation I ( ; ;, Vi"1. or in studies of vision after inversion of eves in tlv or fish or new 1 1 [48, 1 ;6, 1 (9), or alter the prolonged wear- ing of distorting or inverting spectacles in man (285, 455 \'<~ ■ I It-Id, unpublished observations). We shall therefore turn next to a review ol normal pattern perception, especially the problems ol shape, selec- PERCEPTION 1605 tivity and transposition, and their implications for studying motion and depth perception, illusions and constancies. To each of these sections, we shall add the corresponding discussions of altered perception following defect, isolation or recombination of parts. PERCEPTION OF SHAPE Basic Aspects of Figure Processes MINIMUM ARTICULATION, ASSIMILATION AND CONTRAST. Perhaps the best way to begin a survey of the role of patterning in perception is to consider those rare situations where patterning is minimal. The classical demonstration for this in the visual modality is the homogeneous field, or Ganzfeld [see Metzger (345 1 and KofTka (284)]. The subject looks into a half-sphere which has been painted a uniform gray or while. 11 illumination is lowered to the point where the subject no longer sees the texture of the walls of this sphere (its microstructure), lie docs not perceive any surface bul a mist of undefined depth. Similar effects can be obtained by an ingenious variation of the method devised by Hochberg ei G"S NEUROPHYSIOLOGV III fig. 4 A configuration illustrating assimilation, contrast and the role of boundaries in structuring .1 visual field. Bisecl the gray ring by placing a pencil across it, vertically, and note the abrupt darkening of the right half of the ring, and the increase in brightness of the left half, by moving the pencil to the left or right, contrast can be "drawn" from one side to the other. [Based on Koffka (284). Rubin himself and others Mich as Wertheimer (537) and Musatti (356) to define the stimulus determi- nants which decide what pan of a pattern will become figure and whal part ground. It is often, but not al- ways, the enclosed portion that assumes figure character, but ,1 complete enumeration of factors has never heen achieved. Ii is possible to make the stimulus situation am- biguous so that what is figure can also be seen as ground and conversely fig . . ii is under these con- 1 greater or lesser ambiguity that subjective determinants (attitudes, past history) on the part of the perceivcr can pla\ their role in determining whal is seen Although the figure-ground principle was in 1 developed foi the somewhat artificial situation ol bidimensional patterns on paper, ii probably holds for iridiinension.il visual objects and seems to have its analogues in other sensor} modalities (250, 553). The era! importance of the figure-ground principle is underscored I >y the observation that ambiguous figures are apparently not rccouni/cd by most subjects if on second presentation their figure-ground articulation has become reversed 1 41 >l i principles of grouping. What determines figure formation in a stimulus array? Wertheimei (537) has derived a scries of principles of grouping or phe- nomenal laws, the most important being those ol proximity, similarity and 'good figure.' These 'laws' of grouping have heen severely criticized. For one. PERCEPTION 1607 fig. 5. Figure and ground. Left, white goblet or black faces (Rubin's vase-figure), right, black claw or white protrusions. [From Rubin 14m, their universality lias been questioned; their validity might seem to be limited to line drawings or similar visual patterns. There is, however, some evidence that these principles transcend the situations and sense modality from which they were derived. A fundamental aspect of figure perception relative ease of 'tranposition' — was first described lor auditory patterns; a melody can be recognized whether sung by a bass or soprano, or transposed from one ke\ in another (496). Similarly, a shape remains recogniz- able, within limits, in spite of variations in overall dimensions, color and background, or whether pre- sented to the sense of sight or touch. The early work of the Gestaltists has stressed these facts ol transposition rather than their limit--, but these limits merit more systematic study. Melodies become unrecognizable when played backwards, and there seem to be serious restrictions in intennodal transfer, e.g. in identification of visual with tactile patterns (292). The most urgent objection, however, to applica- tion of Gestalt principles in perception is their non- objective and nonquantitative character which makes specific predictions of perceptual events so difficult. There is, however, no obvious reason why this dif- ficulty should not be overcome. Rigorous psvcho- physical methods can be applied, as has been done by Bobbin (50) in dealing with the principle of closure, or by Heise & Miller (191) in dealing with auditory patterning. It is here that information theory may be profitably applied, particularly in specifying the structure of the stimulus array. Thus, Attneave (11, 1 2 ) has given an ingenious method of serial guessing of the component parts of visual patterns in analog) to Shannon's technique (422) for estimating the re- dundancy in message sequences A figure is the "bet- ter," the smaller the number of sequential guesses needed to specify it. A final objection to Gestalt principles in percep- tion, as they are usually stated, comes from those, such as Hebb (188), who believe that form percep- tion is learned. Gestalt psychologists, by stressing the ubiquity of the principles of grouping on different ontogenetic and phylogenetic levels, have tried to show that these principles antedate learning (560); in fact, Kohler (267) rejected their designation as 'innate,' because he believed that these principles reflect basic physical aspects of the brain processes which correspond to perception. If the regularities revealed by these principles are simple physical laws, then they are prior to, and independent of, the de- velopment of organic form Bv contrast, modern empiricists have tried to show- that perceptual patterns have to be acquired labori- ously during normal ontogenetic development. For Hebb (188), perceived forms have some "primitive unity' (his term for figure-ground articulation) prior to learning, but this unit formation does not, he believes, suffice in mediating shapes. Shape is acquired in earlv infancy through successive scanning of con- tours by eye movements which are guided reflexly along predominant brightness gradients in the visual t6o8 HANDBOOK OF PHYSIOLOGY ^-.NEUROPHYSIOLOGY III o • h a, o «5 II II fig. 6. Forms used in experiments on metacontrast (successive interaction of contours). Figures marked a are not perceived if they precede figures marked b within critical intervals, and if their outer contour coincides with the subsequent position of the inner contour of the b figures. [Modified from Werner (534).] field. Repeated scanning movements lead to organ- ized neural traces, or 'cell assemblies* [due to some process similar to Rappers' neurobiotaxis (247)]; eventually, the scanning movements as such can drop out, since parts (it the figure will now activate the trace and the corresponding perceptual habit of seeing (or feeling) a given form. Transposition of forms (the ability to recognize them independently of size or of orientation on receptor surfaces) would likewise be acquired in early development. An in- direct argument in favor of these views is sometimes seen in the difficulties of pattern perception under various conditions of •reduced' stimulation, under low illuminati with small visual angles or with brid periods of time (as on tachistoscopic presenta- tion I achisloscopic exposures can impede the pel don 11I patterns to such an extent that one might doubi whether patterns are ever perceived under 1.1I conditions so that apprehension ol different parts is stiicilv simultaneous. Apparently some scan- ning is net c--.11 v even for 'good' figures. In fact, ease ol perception under reduced conditions, such as tachistoscopy, can be used as still another objective index for goodness of a figure. This rank order of difficulty of shapes under abnormal exposure condi- tions has been explored, and the results lilted to a physical diffusion model of shape perception by Bit- terman et al. (46) and Krauskopf el al. I J<|i I. That shape perception takes time can be demon- strated even more strikingly by successive tachisto- scopic presentation. If the two patterns shown in figure 6, disk and annulus, are exposed each for ta to 20 msec., and the annulus follows the disk at a critical interval of about 150 msec, the disk will not lie per- ceived. For the effect to appear, it is necessary that the inner contour of the annulus occupy the same region of the visual field as the outer contour of the disk [f the order of succession is reversed (annulus shown before diski, both patterns are seen. Ap- parently, the outer contour ol the disk is still in the process ol lot in.ition when the surrounding annulus begins to form. The phenomenon has been known for .1 long time under the name of •metacontrast' see Stigler (454) and later Wet tier (534)]; it has been reported to occur when the first pattern is presented to one eye, and the second to the other eve, but there PERCEPTION 1609 is no adequate quantitative comparison of the mo- nocular with the binocular conditions.9 Possibly a variant of this procedure is the wipe-out phenomenon which results when an information- bearing visual stimulus (e.g. a triangle) is flashed tachistoscopically, and followed within critical time by a white flash, with appropriate time-relation. The first (figured) flash cannot be reported; it has been 'wiped out' by the bright field which followed (324). Unfortunately, this effect seems to appear only if both exposures are given to the same eye; it is thus possible that the second (unfigured) flash serves merely to obliterate an afterimage that would nor- mally appear and assist in the perception of the briefly exposed pattern. Ontogenetic Considerations Our knowledge of shape perception in children is surprisingly scant, and a good deal of the work is contradictory. It has been asserted that children shovt the effect of the 'laws' of grouping more than adults; or, differently put, that they can overcome these ef- fects less (e.g. by selective attention to detail, set or instructions). Thus, hidden figures (like those illus- trated in fig. 14) are difficult to unscramble for chil- dren; in fact, most of the patterns developed by Gottschaldt (164, 165) present insoluble perceptual tasks for children below the age of six, according to Witkin (542) and Ghent (148). There also are marked individual differences in the performance of adults in this respect (542). As Ghent (148) has pointed out, the crucial source of difficulty may be the sharing of contours; the young child is disproportionately handi- capped whenever the contours of the embedded figure also form integral contours of the embedding configuration. By contrast, the partial concealment ol figures by intersecting lines ("mixed figures') presents less difficulty to normal children (fig. 7). According to Stern (447, 448), children below the age of six are less disturbed than adults in recognizing patterns shown to them in unusual orientation (<■ g. upside down). Such an observation would weigh 9 The phenomenon gains in interest in view of recent dem- onstrations that resolution of successive light flashes at cat or monkey optic cortex (where it is measured by discrete evoked potentials) can be improved by intercurrent electrical stimulation of the midbrain reticular formation 1 3^4 1. Micro- electrode recordings for individual neurons of the optic cortex of the cat have revealed a similar effect, trains of light flashes are followed by discrete neural discharges at higher than normal rates if reticular formation or nonspecific thalamic nuclei are intercurrently stimulated (245)- FIG. 7. "Mixed figures' employed in testing normal child- ren and those with brain injury. Bach of the nine composite drawings was presented individually and without the num- bers shown See Teuhei (t)-, illustrated subsequently in Ghent 1 148).] against the virus n| Hebb 1188), since it would imply greater rather than less transposability of shape at younger ages. Actually, the evidence is unclear. Children who cannot read often look at drawings in picture books which they hold upside down; their own drawings may show orientations on the page which are peculiar From the adult standpoint (5771. Yet more recent studies (22g, 230; Ghent, unpub- lished observations) indicate that children may have more (not less) trouble than adults in identifying in- serted pictures. Their preference for holding picture books in peculiar ways may be related to consistent positional preferences which they exhibit just as strongly when presented with unfamiliar geometric patterns. A weakness of many of the available studies of child perception is their failure to employ converging opera- tions. The indices of a particular perception are often drawings made by the children [cf. Eng (113), Yolkelt (493) and Osterrieth (372)], which preclude decision as to whether the peculiarities lie in the child's perceiving, in his drawings, or in both. Less ambiguous results can be expected from systematic application of identical methods for the study of form perception in children and in nonverbalizing sub- It. in HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III ; f * V ••^v ^ -fY ^ \ *•••• % fic. 8. Drawings' produced by a chimpanzee. The experimenter provided the animal with the regular outlines shown, and the animal scribbled over them, filling interspaces and 'completing' designs which the experimenter had deliberately left unfinished. [From von Schiller (510).] human forms, as has been done by Riissel (409) and Gellcrman (147). An obvious, but again, little explored difference between children and adults is the marked difficulty children have with apprehension of peripherally ex- posed figures and with tachistoscopy. That children require a much longer exposure time for the percep- tion of patterns has been adduced in favor of the view that pattern perception is learned (188), but alterna- tive interpretations are possible. Apparently, the two sources of difficulty, limited peripheral span and in- creased time requirements, interact in some complex fashion. It is known thai even with unlimited exposure children take much longer than adults in surveying an array of patterns, and brain lesions in children pro- duce disproportionate slowing in searching time, according to Tcuber et al. (468). Piagct has sii'-mested that children may have virtually a tubular field, and some of the peculiarities of their perceptions can perhaps lie understood in these terms (148). - netu Considerations Discover) of neural correlates of pattern perception will be made e likcK if we can define some of the essential similarities and differences that exist between pattern perception in man and various other species. We ate verv far from such a goal. VERTEBRATES. Work on subhuman primates bv Kohler (j;oi and KJiiver (259) has disclosed puzzling resemblances among the perceptions of monkeys, anthropoid apes and man Most revealing in this respect is Kluver's systematic snrvev (259) of equiva- lent responses in monkeys. Kluver's method involves presenting the animal with a pair of discriminanda (e.g. a triangle vs. a circle), and then varying these stimuli until the learned choice breaks down. This method thus defines ranges of similarity or equiva- lence of stimuli; it reveals that all of the patterns which are functionally equivalent (i.e. elicit the same response) in the monkey look similar to man. Analogous conclusions are suggested by the analy- sis of figural preferences in the scribbling of chimpan- zees studied by von Schiller (510). These scribblings, to be sure, are never representational drawings, but they can be influenced by the figural properties of visual patterns placed on the paper by the experi- menter prior to letting the animal scribble over it (see fig. 8). Lashley (1501) has extended Kluver's technique of 'equivalent and nonequivalcnt stimuli' to the study of pattern perception in the rat. Apparently, rodents as well as primates tend to group stimulus arrays into figure and ground. Figures that are easilv discrimi- nated by man are also 'easy3 for monkey and rat, and stimuli lacking identiiiabilitv lor one species are also readily confused by the other, as shown in figure 9 (301). These similarities (of perceived figural simi- laritv ) were si. surprising that Lashle) considered but rejected the possibility that results might have been different if rat and not man had constructed the test patterns. Comparable data on lower vertebrates are scant [sec, however, Pache I ;; 3) for frogs and Herter (jot.) for Itsh1 but oiler no support lor assuming any abrupt changes in the evolution of vertebrate pattern vision. PERCEPTION I 6l I Training figures Equivalent stimuli rain BESS [ojiyj F3P I nan mil fig. g. Equivalcnt-nonequivalent stimuli, designed according to Kliiver (259) and used by Lashley (301) to establish range of perceived similarity in rats Trained to select, say, the solid upright triangle shown in the In/) row of the training ligures, the animal will transfer its choice to the corresponding outlines of an upright triangle and even to incomplete outlines. Thus it will select, on successive presentations of pairs of test stimuli (in lop row), those shown to the left in each pair. The same applies to the training ligures shown in the second, third and fourth rows, and to the corre- sponding arrays of test stimuli ( )n actual presentation of training or test ligures, lateral positions of rewarded and unrewarded patterns are randomized, so that the rat cannot solve these problems by acquiring simple left- or right-going tendencies. There must be some differences. Although positive evidence is hard to find, one might search for these differences in two directions, a) There are some indi- cations that lower animals may show narrower limits in transfer from one learned discrimination to other discriminanda, and b) conversely, under certain natural conditions lower animals mav show nondiffer- ential reactions to stimuli that for man are obviously different. To take the first possibility first: Lashley's rats failed to transfer reaction from a white triangle (on a black ground) to a black triangle (on a white ground), a transfer possible for monkey, ape and man. (Note, however, our difficulties with the identification of faces in photographic negatives. ) Similarly, rats are less tolerant of rotation of figures — an isosceles triangle is not equivalent to the same triangle rotated 90°; there are similar though lesser difficulties on such transfer tests for apes and for human children (147). The other possible source of data for species differ- ences lies in observations such as Buytendijk's (83) who showed for toads how the range of perceptual equivalence (perceived similarity) expands or con- tracts in particular directions depending on the ani- mal's internal state. A toad is placed into a cage with inanimate objects, such as matches and strands of dried moss. As long as the hungry animal has not re- ceived any food, it does not react to these objects. After it has eaten a worm, it begins seeking the broken i6ia HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III matches, and after it has eaten a spider, it begins picking up the moss. In ethology, the systematic study of species-specific behavior [cf. Tinbergen (486)], attempts are made to assess internal states or "moods' in animals by survey- ing the range of equivalent reactions to varied dummies as stimuli (Attrappenversuche). Thus, mobbing reactions in small birds can be elicited by balls of feathers on a stick. The resulting observations are sometimes interpreted as if they suggested the exist- ence of particular schemata or kernel perceptions which act as releasers of particular behavior sequences [IRM, "innate releasing mechanisms'], for example those of courtship, especially in birds [see Huxley (232)] or of predator recognition (486). There is con- troversy about the generality of some of the observa- tions [cf. HirschW al. (216) and Tinbergen (487)], and particularly about the innateness of the reactions. The most radical claim in this respect is implied in other recent descriptions of bird orientation. For the warbler, a night migrator, Sauer (412) has reported accurate orienting reactions to stellar constellations (including the artificial constellations in a plane- tarium), and this in birds which had been reared in isolation and without any previous exposure to the starry skv. Phenomena of animal camouflage provide strong indications for similarity in pattern vision among dif- ferent vertebrate species [cf. Thayer & Thayer (480)]. There, the principles of perceptual grouping (272, 537) turn into rules for concealment, as in the con- struction of hidden figures (142, 164, 165). Experi- mental proof for the efficacy of protective coloration, or cryptic attitudes (93) is available in a few species, for camouflaged insects as prey (85) and fish and birds as predators (4")8, 554). The evidence suffices to establish that these concealing features act at least as strongly for lower vertebrates as they do for man, indicating that perceptual principles involved in grouping and in camouflage antedate the evolution of the human nervous system. Some species differences, however, might lie in the relative ease with which concealment can be overcome, conceivably, lower forms arc less able to override t h<-si ■ factors I >v selective attention, but evidence on this point is sparse. This review (it pattern perception in children and subhuman forms reveals the fragmentary state of our information. The available methods have not been used sufficiently to establish valid comparisons be- tween species. It is often forgotten that the same species that can be shown to react t < > broadlv sche- matic stimuli in some situation (as stressed b) the ethologists) can also be trained, in the laboratory, to make refined perceptual distinctions.1" Even within a given species we usually lack systematic studies de- fining both the extent of equivalence (generalization) and of discrimination. These limitations will become still more obvious when we turn to a review of pattern perception in the higher invertebrates. Does percep- tion in these forms differ as much from that in the vertebrates, as the differences in the nervous systems of the phyla would suggest? higher invertebrates: cephalopods. Studies of pattern vision in higher invertebrates with large image-forming eyes (such as the octopus) again sug- gest a rather puzzling similarity in visual organization between these forms and vertebrates. Experiments in- volving successive discrimination [cf. Boycott & Young (57, 58)] and transfer to equivalent patterns (461 ) have revealed few differences, except that an octopus trained to go to an upright triangle will transfer this reaction to a rotated triangle. This is not found in rats (301) nor in pigeons (4881, although it does appear in monkeys (3581, chimpanzees (14b) and children (146). Furthermore, Sutherland (459, 460) showed that the octopus seems virtually incapa- ble of discriminating oblique lines (\ vs. 1, even though it learns promptly to tell a vertical (j) from a horizontal ( — ) line. This specific difficulty in discriminating mirror images prompted Sutherland to postulate a neural scanning mechanism in the octopus, whereby shapes are classified as follows. The visual impulses are led into an array of cells arranged in rows and columns, and the total excitation is counted separately in columns and rows. Reaction to shape is then deter- mined by ratios of horizontal to vertical excitation, a mode of determination which would make horizontal and vertical maximally discriminable, and oblique lines not at all (4,1) l'1' \s Sutherland himsell pointsout, Lashlev (301) had noted a similar tendency to confuse mirror images in rats. Quite recently, Rudel (unpublished observation- has shown that normal children below age -i\ have 10 That these discriminations are not only the product of artificial settings is indicated by the role of individual recogni- tion among birds in the maintenance of pecking orders I fis- turbances in the pecking order ol chicks (Leghorn fowl an most readily induced by changes around the head 01 by abrupt changes in body coloi If the changes are gradually introduced, they have no effect see Guhl >v Ortman (174 "The same mechanism would account lor equivalent re- actions to identical shapes of different sizes, .1 1 eaction found for the octopus 1 jfii 1, pigeon |88 rat |0i and monkey (1159). PERCEPTION 1613 A/ XO"»Y FIG. 10. Figures used in testing form perception in honey bees. The bees could be trained to distinguish each of the figures in the upper row from each in the lower row, but failed to distinguish among those in the upper row, or among any of those in the lower row. [Modified from Hertz (J07).] analogous difficulties with mirror images; thus, they failed to learn a distinction between \ and /. (How- ever, Rudel's finding that C and 3 are practically indistinguishable for small children, while u and n are easily discriminated, does not seem to conform to Sutherland's hypothetical scanning mechanism, nor would ii seem to be predictable from an earlier theor) of shape recognition advanced by Dcutsch (103) (More recent observations by Sutherland himself indicate that for octopus, too, there is much less diffi- culty with u and n than with C and 3, although the difference between the two tasks is not as great as in the child. ) An incidental outcome of Sutherland's work is a new and more convincing interpretation of Boycott & Young's well-known report of altered memory for visual form in the octopus following extirpation of the vertical lobe system from the animal's brain (58). Such animals tend to lose discrimination, and attack both positive and negative shapes.1- However, some of the animals can be retrained by presenting shapes at very short intervals (a situation which interferes with learning in normal animals, since they stop attacking shapes to which they are frequently exposed). Young and Boycott have interpreted these observa- tions by assuming that the negative memory in ani- mals with verticalis lesions becomes very short-lived, and that it can be maintained only by frequent expo- sures (every 5 to 10 mini to the negative shape, even '- Although this effect is often quoted as being specific for visual perception, it exists just as well for tactile discrimination in a blinded octopus lacking the vertical lobe (533). if the shape is not attacked. Since Sutherland (460) has shown that "exposure to a shape tends to reduce the tendency to attack a shape" in the normal octopus, it is likely that '"verticalis removal may raise general tendency to attack and that discrimination can only- show up when this tendency is reduced, and that this reduction may be brought about by very frequent presentation of shapes." salticidae. The octopus and other cephalopods are not the only invertebrates whose eyes are capable of forming single images. The jumping spiders (Salticidae) possess such eyes, and apparently depend on them in courtship and in their recognition of prev (190, 221, 222). Heil's Held studies (190) describe visual recognition of dead (and hence, immobile 1 prey. Such an achievement is unexpected, since it has been assumed that most invertebrates require mining targets for adequate perception. Laboratory studies similar to those in the octopus would be desirable. INVERTEBRATES Willi COMPOUND EYES. By far the IllOSt pressing need, however, is for further elucidation of pattern vision in higher invertebrates with compound eyes. That visual orientation is important in bees has been made clear b\ von Frisch (499, 500) but the wa\ in which visual patterns are recognized by the bees remains elusive Hertz, in numerous studies (207- 2119), demonstrated that bees distinguish star-shaped from closed flower patterns, with preference for those with radiating (and hence, with broken) contours (see tig. 101. Zerrahn (557 1 showed further that the bees' preference increased directly with increasing numbers of contours within a given test pattern. Wolf (546) found that rotating patterns were chosen in preference to identical stationary ones; flickering sources attracted more bees in direct proportion to the rate of flicker. All these observations seemed to suggest that "pattern" might reduce to "flicker,' i.e. the num- ber of ommatidia stimulated in rapid succession as the bee flies over a pattern rich in black-white contrast. This cannot be the whole story, however, since von Frisch had shown as early as 191 4 [see also Knoll (264) and Friedlander (133)] that bees can be trained to distinguish bipartite disks (e.g. one lateral half blue, the other half yellow), depending on the right- left pattern (e.g. yellow-blue vs. blue-yellow). Such an accomplishment in an animal with compound eyes is all the more remarkable if one recalls the inability of the octopus to distinguish mirror-image patterns. Even more puzzling are the demonstrations by Hertz (210) that with some patterns, bees can be trained to [614 II VNDHOOK OF I'HVSIOLOGV NEUROPHYSIOLOGY III choose the one with fewer contours. Some most un- expected features of visual perception in bees, finally, are implied in certain open-field experiments on the 'clustering' of bees at food sources [see Kalmus (246)]. It is well known that bees congregate at flowers or food dishes wherever there are other bees. This cluster- ing can be enhanced by placing a mirror below the food source, and even more by placing 'supernormal' (i.e. outsi/ed) cardboard dummies of bees around the food dish. Clearly, the analysis of pattern vision in bees and in other invertebrates with compound eyes is in need of much further study [see also Use (234)]. Changes in Perception of Shape After Cerebral Lesions In comparing perception across species one hopes to find differences in function that might correspond to known differences in structure. As we have seen, the differences in perceptual repertoire of different species are on the whole less marked than the simi- larities. Studies of perception in the presence of defects in neural structures aim at finding characteristic changes as the result of ablation, injury or disease. Yet, here too, lack of change or resiliency of percep- tion is a major finding. Nevertheless, there are certain alterations in perception after cerebral lesions that may cast light on the normal bases of perception. As in the phylogenctic studies, however, evidence for these changes can be no better than the methods em- ployed in analyzing their nature. In this respect, methods developed in testing nonverbalizing organ- isms turn out to be useful in assessing performance in man with brain injury; conversely, studies of animals with experimental ablations profit from tasks derived in clinical settings (466). In die following, we shall briefly consider effects ol total removal of primary (cortical) projection s\ stems, then effects of subtotal lesions within these s\ stems, and finally changes found alter various lesions encroaching on neocortical struc- tures outside the primary projection fields. EFFECTS OF To I \1 REMOVA1 OF 'PROJECTION SYSTEMS.' 1 01 the visual system, it is generally agreed thai total destruction of the geniculostriatc sector results in total and irreversible loss oi pattern vision (see Chapter LX by Neff in this volume). This is believed to take place, whether the lesions eliminate the lateral geniculate ies, or the optic radiation, or the visit. il cortex. It is not clear whether in man some reactions to light in.iv eventually return Nee, again, Chapter LX) In the experimental monkey, bilateral occipital lobec- tomy abolishes pattern vision, but reactions to lumi- nous flux are preserved (261). The question of re- sidual vision in man after complete destruction of striate cortex will remain unanswered until testing methods such as Kluver's are applied in suitable clini- cal cases Likewise, we may need to reinvestigate the tradi- tional belief that effects of total striate cortex re- moval on pattern vision are progressivelv less severe as one moves down the phylogenctic scale (339). The condition of rodents and carnivores following such an operation may not be as different from that of pri- mates as has been claimed. In still lower forms, dif- ferences undoubtedly appear. Removal of the entire forebrain in fish is said to have no effect on visual dis- criminations whether established pre- or postopera- tively [see the review by Herter (206)]. The status of visual capacity in birds following forebrain removal is puzzling and badly needs further examination. Visser & Rademaker (491, 492) report that pigeons without forebrains avoid vertical but not horizontal strings in their path; they alight on the back of a cat while they are in flight, but avoid cats while walking, etc. What is needed is an extension of available testing tech- niques to the study of decerebrate birds (and of birds with lesions of the optic tectum). If there are phyletic differences in visual organization, then homologous removals of central nervous system tissue in different species would not be expected to lead to homologous changes in perception. The loss of shape perception in primates following visual cortex resection may have parallels in the dis- orders of auditory patterning, in primates and carnivores, after certain bilateral removals of 'audi- tory cortex' [sec Diamond & Neff (104) and Chapter LX by Neff in this Handbook], Definition of the mini- mal effective lesion in the auditory svstetn, however, is difficult because ol the uncertainty regarding the full extent of auditory projection fields in carnivores and primates. This difficulty is even greater in the somatosensory svsiem [cf. Cole & Clees (92)] where total loss of reactions to tactile patterns .titer restricted cortical removal has not yet been demonstrated ex- perimentally, irrespective of whether the subjects were primates, carnivores or rodents. effects 01 subtota] removai in vivs. Partial de- struction of the visual cortex in man is followed by a v.uietv of symptoms which may cast lighl on neural correlates ol pattern v ision. There are two outstanding features of residual vision in such defective visit. il PERCEPTION l6l5 fig. 11. Complete loss of peripheral field (bilateral hemi- anopia) with sparing of central vision, following penetrating mis- sile wound of the brain. The course of the projectile is indi- cated in the diagrams below the visual fields in order to show the similarity between this case of "peephole vision' and an earlier one reported by Holmes (218). This patient showed no obvious difficulties with form perception in the remnant of his visual field. [From Teuber ltd. (469).] fields, the resiliency of pattern vision as such (which can be mediated by small remnants of the visual cortex) and the subtle but significant changes in patterning after lesions even of moderate m/<- Visual patterns can be perceived in cases of very extensive field defect. The two complementary exam- ples shown in figures 1 1 and 12 are taken from a study of visual performance after penetrating gunshot wounds of the brain (469). In the first case, a tunnel field, following a through-and-through bullet wound of the posterior lobe substance, was compatible with grossly normal pattern perception in the centrally located remnant of the field. This resiliency is par- ticularly remarkable since the anatomical arrange- ments in the optic radiations and cortex make it likely that the lesion producing this tunnel field implicated vascular supply to both occipital lobes sparing only the pole. Return of normal pattern vision in the isolated remnant is not easily reconciled with the theory that patterns are recognized by some cortico- cortical interaction between a primary projection field (e.g. here, the visual cortex), and the surrounding 'associative' region (e.g. the prestriate cortex). Equally remarkable is the capacity, in the second case, illustrated in figure 12, to discriminate large visual patterns (triangles, squares, circles) that are circumscribed around the central island of blindness. Such a capacity is specifically denied by certain theories <>l shape perception, such as that of Dculscli (103). The resiliency oi shape perception in remnants of defective visual fields should not detract from the ob- servation that there are lasting deficits which merit experimental studs. These deficits can be grouped according to their relative specificity (467), i.e. the degree to which they depend on focal lesions in the primary projection system. Thus the scotoma repre- sents a specific and circumscribed loss of shape per- ception, a gap in the substrate corresponding to a gap in the visual field. After the first year following injury, these areas of acquired blindness remain unchanged in outline, position and density (469). The defects are homonymous and similar in the two monocular fields. In spite of their similarity in the two eyes, they turn out to be incongruent on careful plotting, suggesting that the cell populations which represent each retina beyond the thalamus overlap without being identical. This lack of congruence aside, the scotomata reflect, in shape and position, the specific anatomical arrange- ments within the visual pathways. It is difficult to con- ceive of any other consequence of cerebral lesions which would be as static, localizable and cir- cumscribed. In the presence of these scotomata, however, there i6i6 IIWDIIlli IK OK I'HYSIO] ' M.'i Ml k( il'U\sloI.OGY III pig 12. ( Sentral scotoma, sur- rounded by spared peripheral vision, following penetrating mis- sile wound of tlie occipitopolar region of the human brain. I he course of the missile suggested involvement of upper and lower lips of both calcarine fissures. This patient was capable ol distinguishing large forms i tri- angle, circles, squares) presented so that they surrounded the central area of blindness. [From Tcuber tl al. 1469).] arc other changes of a more subtle nature which do not conform to any simple point-for-point correlation of structure and function. If one tests those areas of the visual field which lie outside the scotoma, one finds that shape perception is disproportionate!) handicapped by tachistoscopic presentation (33, 386). In the same regions of the field, there are other asso- ciated disturbances: contours may fade more rapidly; fusion thresholds for flickering light are reduced; and there is impairment of dark adaptation, and of per- ception of real and apparent motion. These changes involve the visual field as a whole (23, 294, 470-473); in this respect, these changes are less specific than the scotoma since they appear even in those parts of a visual field which seem intact on routine perimetric testing. However, the delects are restricted to cases with circumscribed visual field delects, a fad suggest- ing that occipital lesions produce twofold effects: those (hat are focal (scotoma), and those that are less focal, involving visual functions over the entire field and main aspects of visual performance. COMPLETION V\|) EXTINCTION OK PATTERNS. These subtle but persisteni alterations can be manifested l>> additional changes in the mode of functioning of the injured si 1 list raie. Under a v arieiv of conditions, figure processes .in- 'completed1 across scotomatous regions so that there i- no gap in perception although part of the stimulus figure falls into areas of blindness 1 30, ; ;, 70, 157, 386, 469). A characteristic instance ol this 1 completion or tilling-in process is shown in figure 13. Sin li .1 filling-in is well known for the area ol the normal blind spot i 501 1, bul in thai case, completion is usii.iiu attributed to the fad that there is discon- tinuity in the retina bul urn in the cortex (20). I In argument 1 .1 1 be advanced lor completion of pat- terns across scotomata acquired after cerebral lesions. Moreover, even in the transient scotomata of mi- graine (303) or of 'visual fits' (469), completion is regularly observed. The phenomenon is thus a basic feature of perception; it is difficult to reconcile with most perceptual theories which invoke a scanning of the cortical projection areas [e.g. Pitts & McCulloch (381 )] or with those versions of 'field tfieorv' which assume a) that all of the perceptual process takes place in the primary projection field and b) that the percep- tion as such is based on an 'isomorphic process' which takes place within the projection field (271, 281). Either a scanning or a field theory would have to be modified in view of these completion phenomena. As the result of completion, defective visual fields may be functionally more extensive than the dis- tribution of plotted scotomata would indicate. A converse process, however, restricts perception in seemingly functioning regions (29). This is the phenomenon of extinction of patterns in certain areas of a defective field. 'Extinction' occurs in mildly impaired areas of the visual field when other rela- tively intact areas are stimulated simultaneously, Thus, a pattern may be seen on exposure to an amblyopic half of a field, as long as this exposure takes place against a relatively homogeneous back- ground. However, as soon as a second pattern of any soil is simultaneously exposed in the other (less impaired) half of the field, the first pattern either becomes less distinct ('obscuration') or vanishes altogether ('extinction') (29). As soon as the pattern in the less impaired pari of the field is removed, the one in the affected field becomes visible. I In- phenomenon his man) variants, described by Bender (28). It can occur within a lateral half of a visual field, eg. between its upper and lower quad- PERCEPTION 1617 A • • • • A, • • B LJ B, □ c LJ c, D Zl H "■ H FIG. 13. 'Completion' of forms projected partly into the blind half of the visual field in a case of left homonymous hemianopia. A-E are patterns presented on a translucent screen (at 100 msec.) by a tachistoscope, A, fixation point. Patterns A—C subtended 10° to either side of the fixation point along the horizontal meridian. Patterns I) and E, 7° and 8°, respectively. The patient reported what he saw I I / Inst verbally, then by drawings, then by selection from a series "I comparison patterns. [From Bender & Tcubcr (30); see also Teuber el al. (469).] rants. It occurs in the tactile and kinesthetic modality Thus, a patient with a right parietal lesion may be able to report a contact on his left hand and on his right hand, as long as each hand is stimulated singly. On 'double simultaneous stimulation' of both hands he may report only the contact on his right (the less affected side). One and the same patient may exhibit extinction for touch and vision, or for one or the other modality alone. Unfortunately, it is still not clear why extinction occurs when it does, and why it fails to occur when it does not (295). Right parietal or parietooccipital lesions are frequently found in cases of extinction, but they are not obligatory.13 It 13 In fact, the earliest demonstration of one-sided 'neglect' on bilateral stimulation was made in experiments on dogs with is conceivable that the phenomenon of extinction on multiple simultaneous stimulation represents a special and exaggerated form of normal lateral interactions between contour processes in a sensory field [see Fry & Bartley (135)]. The famous syndrome of Balint (17), in which the patient can perceive whatever he fixates at any time but nothing else, may be another form of extinction, and similar disturbances may play a role in those forms of so- called agnosia in which a patient cannot organize a large array with multiple patterns (121, 547; see also 331). It is quite obscure where and how these enhanced lateral interactions within a sensory sphere take place. They need not occur within the primary projection field itself; it is tempting to assign some role in these abnormal processes to nonspecific projection systems (324). A possible analogue in normal states (beyond the short-ran^e interaction of contours) is the inter- ference effect on pain l>v countcrirritation, i.e. an intercurrent painful process set off elsewhere in the both. In the auditory sphere, a possible normal analogue would he the inability to attend simul- taneously to dichotic messages (two sets of information presented each to one ear) (63, 87, 352; Chapter I. Will by Zangwill in this volume). COEXISTENCE OF SPECIFIC VND GENERAL PERCEPTUAL changes. The review of perceptual alterations after subtotal lesions of the visual projection system in in. in lias revealed an obligatory association of specific symptoms (scotoma) with less specific symptoms found elsewhere in the defective field. In addition, one can construct perceptual tasks which are affected in an entirely nonspecific way, i.e. irrespective of whether there are scotomata or not and irrespective of the site of the cerebral lesion. These tasks employ "hidden' figures m which line drawings are concealed by embedding them in interlacing contours (see above and fig. 14). Cerebral lesions which result in visual field defects produce disproportionate difficulties in cortical removals from either the occipital or the frontal lobes. Jacques Loeb (328) showed that such dogs would invariably turn toward meat on their operated side and neglect another piece exhibited simultaneously to the other side of the fixation point. Loeb's method of double stimulation was applied after that in neurologic patients to demonstrate minimal sensory or visual impairment (72, 369, 386). The curious 1 transient) disregard of stimuli opposite an acute unilateral frontal lesion in dog and monkey has been investigated further by Bianchi 143), Kennard & Ectors (254) and, quite recentlv, by Welch & Stuteville 1532). i6i8 HANDBOOK OF PHYSIOLOGY NEUROPHYSIOLOGY III / / ^~ \I 01 17 fk;. 14. Hidden-figure test (sample page), modified after Gottschaldt (164, 165). The subject is required to find the figure .it the /"/' within each of the lower {embedding) figures. [From Teuber & Weinstein ( 4 7 9 1 - the perceptual analysis of such patterns (161 , 386, I7<|i. However, while field defects are sufficient to produce such a deficit, they are not necessary. In- juries in any lobe of the brain, in either or both hemispheres, lead to significant loss on this task (fit;. 15; 479). Neither does it seem necessary that there be other symptoms of cerebral lesion, such as somatosensory or motor changes (see fig. 16), since following brain injury subjects with or without such symptoms perform equally poorly as compared with normal controls. Only aphasic patients, as a group, can be shown to fall significantly below the others with brain injury (see, again, fig. 16), who in turn arc surpassed significantly b\ the controls. M Perceptual changes after cerebral lesions in man thus range from those that arc most specific (scoto- mata) to those thai are general, or nonlocalizablc. Which one of these alterations appears seems to depend on the nature and level of the task employed. Such findings recall the belief of Flourens (127) who assigned to all major sectors of the forebrain an action commune, in addition to their action propre. If one considers the wide scope of perceptual selectivity in higher forms, one can perhaps understand its de- pendence on such common action of the hemispheres and its nonspecific decline alter brain injury in man. Analogous hierarchical findings can be obtained for man's somatosensory systems. Here, alterations in basic sensory thresholds after lesion of the somato- sensory projection system correspond to the scotomata in the visual sphere. By their very nature, these deficits might be expected to interfere with the per- ception of what are traditionally known as 'higher' or complex aspects of objects presented through the sense of touch. However, the classical concept of 'astereognosis' (i.e. agnosia for touch) suggests that recognition of object qualities can be impaired in the absence of any "primary' or 'elementary' deficit in that modality. Difficulties in tactile perception, then, could reflect either a mere consequence of sensory alterations (such as changes in thresholds) or a separate higher- 35 uj 30+ o oc UJ CD 3 25-- 20- ■ R L BIL NF F UNI NP P NT T NO 0 fig. 15. Average number of hidden figures (see fig. 14) correctly traced by normal adults (controls, C) and by subjects with brain injury, grouped according to location of lesion : /., left unilateral lesion, R, right unilateral lesion; /•", /', '/', (>, frontal, parietal, temporal and occipital, respectively; A7-, NP, NT, NO, nonfrontal, nonparietal, etc. [From Teuber & Weinstein 1.479).] 35T 30-- cr 25+ O o 20" CO 5 3 z 15- • 1 n