DEAN E. WOOLDRIDGE

MAN'S MYSTERIOUSMEMORY MACHINE

Scientists now think they know whathappens inside the brain when you forget a

phone number you just looked up ...and remember a poem you read years ago.

FE\'\! scientific developments could surpass inimportance a determination of how the brain

achieves its remarkable results. A detailed ex-planation is undoubtedly many years away. How-ever, in the laboratories of life scientists and theoperating rooms of brain surgeons,importantdiscoveries are beginning to penetrate the veil ofmystery that for so many years prevented effectiveapplication of scientific techniques to brain func-tion.

Clearly, memory is one of the most basic ofmental properties. Without the remarkableability of our brains to record current experienceand play it back later, intelligent IiIe as we knowit could not exist. Indeed, the absence of memorywould have profound implications for the ent.ireanimal kingdom. Even single-celled organismscan learn from experience!

Modern research on the memory processes hascast much new light on how and why we learn,and is opening paths to an eventual understand-ing of the more complex phenomena of imagina-tion and conceptual thought. The ultimate sci-entific goal of this work. is a full understanding ofhow the brain works. The practical goals arenew medical techniques to rep vir damaged orabnormal brains, and methods of increasing andimproving the mental output of nurmal people.In this article I shall discuss some of the recentdiscoveries about the memory processes that seemmost pertinent to these long-range goals.

If we think about it, we will probably agree that

© 1963 by McGmw-Hili Book. Com.pany

there must be more than one kind of memory.There is, first, a long-term, permanent memory.Our own names and addresses, the words andgrammatical forms of the language we speak, theaddition and multiplication tables, importantepisodes in our early life-these memories, onceestablished, are with us to stay. To be sure, ourmemory-recall system is less than perfect; all ofus occasionally have mental blocks for particularwords or names. Yet even in these circumstanceswe never doubt that the memory is still theresomeplace; we just can't lay our hands on it. Wemay suspect that our memories fade with thepassing of time, and some probably do, but thisis a slow process. Sometimes the old, long-stand-ing memories even appear to become more vividas a person grows older. In any case, there is alarge inventory of recollections that stays with theaverage person until he dies.

There is also another kind of memory. It isfleeting and short-term. A beginning languagestudent, for instance, in attempting to read aforeign text, may have to look up the same wordtwice when it reappears in the same sentence.The telephone company finds it profitable to payfor television advertisements to exhort its sub-scribers not unly to look up, but also to writedown, the number to be called. And when weare introduced to someone at a party, our audiotory nerves may clearly pass the name along toour brain, but iF we are inauentive ur thinking ofsomething else, it does not stick. For a few sec-onds after the introduction, by deliberate effortwe can retrieve the name from our fast-decayingshort-term memory; but iF we do not immedi-ately focus our attention on the matter, the in-formation vanishes beyond our power of recall.

Such introspective evidence has recently beenbolstered by objective physiological observations.

t>R MAN'S ME M 0 R Y IVrA CHI N E

The study of brain concussion, for example, hasproved pertinent to an understanding of memory.After a true concussion the victim never remem-bers the actual blow that caused him to lose con-sciousness. The neurologist "\T. Ritchie Russellof Oxford has made statistical studies of theseamnesic effects of concussion. * The length of thepnmancnt retrograde O'Inncsia-the period be-(ore the blow that is never again remembered-varies considerably from incident to incident, andin general is greater after the more severe in-juries. However, usually it is several seconds. Ap-parently this is the time it takes the hrain mech-anism to retrieve information from the rapidlydecaying short-term memory and install it in amore permanent storage system. Interruption ofthis process by the blow on the head is believedby Russell to account for the permanent retro-grade amnesia that is associated with concussion.

In addition to a short-term memory lastingonly a few seconds, and a long-term memory last-ing for days or years, there also appears to be amedium-term memory mechanism. This is sug-gested by the peculiar kind of memory deficitsometimes exhibited by patients with damage inthe hippocampal structures that underlie thetemporal lobes of the cortex. (See Illustration I.)Unlike concussion victims, they do not forgetevents that occurred before the damage to theirbrain tissue; they can carryon a normal conver-sation and can briefly remember telephone num-bers or other information; but five or ten minuteslater they not only will have forgotten the num-bers, as most of us would, but will have no recol-lection that there had even been a conversationon the subject! One epileptic patient had hishippocampus removed on both sides. After theoperation, in marketing at a store directly acrossthe street, he could remember quite well whathe had been told to buy. Yet when he went toanother store ten minutes away, by the time hegot there he not only had forgotten what he wasto buy-he couldn't even remember why or howhe had come.

Wirh these patients, memories seem to getstarted in the normal way and to continue prop-erly for a few minutes, but then they disappearentirely. Apparently, there is some kind of mech-anism in the deep structures on the sides of thebrain that normally converts a medium-termrecollection, which (an persist unaided for a Iewminutes, into a truly long-term memory trace.

Thus it appears that the nervous activity pro-duced in our brains by external events does not

"Brain, Memory, Learning (Oxford UniversityPress, 1959).

go directly into a permanent record. Instead itis held for a time in two suspense files-a fewseconds in the first, a few minutes in the second.This delay has at least one value. It gives timefor the important "attention" mechanism of thebrain to select some small part of what is goingon around us for admission to our consciousawareness. Thus at a noisy party, we can selectthe host's introductory remarks out of the confus-ing babble of surrounding conversation, forconscious attention and transmittal to our me-dium- and long-term memory mechanisms.

Once it is well installed in the permanentstorage system the memory trace has additionalinteresting properties. One of the most significantis its tendency to strengthen with time.

OLD MEMORIES;SELF-STARTING NEURONS

INC A S E S of severe concussion, in additionto the bricE interval just before the blow that

is never again recalled by the victim, there isoften a much longer interval that is temporarilyerased (rom the memory. Depending on theseverity of the blow, the events of the precedinghours, nays, or weeks may be blank. On first re-covering consciousness after a very severe blow,the victim may even give the date as several yearspreviously, with no recollection of the interven-ing period. As he recovers, the period of retro-grade amnesia shrinks, but always from the pasttoward the present. The older memories are thehardest to eliminate, and they are the first toreturn. With continued recovery, the amnesiadiminishes until it finally reaches the minimumof a few seconds beyond which no further recov-ery of memory is possible.

The effect of electroshock treatments of men-tally ill patients also demonstrates the strengthof older memories. A series of such treatmentsgradually induces amnesia for past events-pro-gressively from the present working back into thepast. When the shock treatments are terminated,the patient gradually recovers the lost memories,from the past working up to the present. Simi-

A fter It distinguished career in industry, DeanE. Wooldridge resigned as president of ThompsonRamo Wooldridge Inc. early this year to specializein scientific research and writing. He is now Re-search Associate of the California Institute of Tech-nology, where he took his Ph.D. in physics in 1936.Thc material in this article is adapted from hisforthcoming book, "The Machinery of the Brain"(McGraw-Hill).

SENSORIMOTOR CORTEX

FRONTAL LOBE

CALLOSUMCORPUS

OPTIC CHIASMA

59

VISUAL CORTEX

CEREBELLUM

6RAINSTEM

Illustration I: Diagrammatic view of the right half of the brain, seen from the left, showing its threeprincipal parts: the cerebral cortex, the cerebellum, and the brainstem, The cerebral cortex is the partexposed when the skull is cut away. It consists of several square feet of sheet material, one-tenth toone-fourth of an inch thick, which is wrapped and squeezed around the brainstem and the cerebellum.Fissures divide the cortex into lobes. On the inner surface of the temporal lobe is the hippocampus.

larly an old person may have a vivid recollectionof early events but be unable to remember recentoccurrences. Again, when a young child with atotal vocabulary of, say, fifty words suffers asevere brain concussion, he may lose half hisvocabulary, and the words he loses are the oneslast learned.

The peculiar strength and invulnerability ofold memories has led Russell to put forth someinteresting speculation. When a memory traceis being laid down or recalled, electric currentspass among the particular neurons (nerve cells)that are activated by the event or recollectionthat is being experienced. The passage of suchcurrent from neuron to neuron is generally be-lieved to strengthen the connections betweenthem, thereby making it easier for the sameneuronal pattern to be activated in the future.This idea partially explains why repetitionstrengthens memories, such as those underlyingthe learned muscular processes involved in a golfswing, [or example. But Russell points out thatthe neurons in the brain are not just passive con-ductors of electrical nerve impulses. Without anyapparent reason, nearly all neurons independ-ently send occasional pulses of electricity into theneighboring nerve cells. These randomly gen-erated electric currents, Russell suggests, tend tofollow the well-established neuronal paths; thisautomatically strengthens the memory traces thathave already been laid down. In this way, thelonger the memory has been in the storage mech-anism of the brain, the stronger its trace will be.

This hypothesis explains the curious persist-ence of old recollections. It also appears to pro-vide a physiological basis for other subjectiveobservations. For example, college students cram-ming for examinations often find that what wasdifficult and confusing the preceding night seemsclearer and simpler the next morning, after agood night's sleep. The performance of athletesis frequently improved when they return to prac-tice after laying off training for a while. Russell'sideas make these observations understandable.Once the correct memory patterns are establishedin the brain, the passage of time will auto-matically strengthen them because the randomlygenerated electric currents of the neurons tendto traverse, and thereby strengthen, the est.ab-Iished connections. Therefore, if the student hascrammed correct information into his memorystorage system, he may make more progress instrengthening that learning by laying off thesubject entirely for hours or clays than by furthercramming. This is especially true if fatigue iscausing so many mistakes that further study isactually likely to impair the quality of the storedmemory trace.

TO this point, the clues we have consideredas to the nature of the memory mechanisms havelargely come from statistical surveys of the effectsof brain concussion and clinical observations ofmemory defects after surgical operations. Suchinformation is valid when properly interpreted.It would be more convincing, however, if it were

60 MAN'S MEMORY MACHINE

possible to "do something to" the brain of aparticular patient and imrnecliately observe aneffect on his memory. Unlikely though it mayseem, exactly this result was stumbled ontotwenty-five years ago by the brain surgeon WilderPenfield, director of the Montreal NeurologicalInstitute.

ELECTRICAL EVOCATION

OF THE PAST

PEN FIE L D had pioneered in the develop-ment of surgical techniques for the allevia-

tion and cure of epilepsy. Before the surgicalexcision of the defective brain tissue responsiblefor a patient's epileptic attacks, it was his practiceto employ an electrical exploratory technique.First he would remove the section of the skullcovering the damaged area. Then he would re-store the patient to consciousness and ask him todescribe his sensations as electric current was in-jected into various spots of the exposed cortex.Such electrical stimulation, which is completelypainless, had been found capable of providingclues to the precise limits of abnormal brain tis-sue. But in 1936, when electrical stimulation wasthus applied to a spot of one patient's exposedcortex, she suddenly reported that 'she felt trans-ported back to her early childhood. In the op-erating room she essentially relived an episodeout of her remote past, even feeling- again thesame fear that had accompanied the originalevent.

Another of Penfield's early patients seemed tosee herself as she was while giving birth to herchild. A young man saw himself-years earlier-with his cousins at their home in South Africa;he could hear them laughing and talking. Onewoman heard the voice of her small son in theyard and the neighborhood sounds of honkingautos, barking dogs, and shouting youngsters.Another patient seemed to hear an orchestra play-ing a number that she did not herself know howto sing or play, and that she only vaguely re-called having heard before. Still another patientwas moved again by the beauty of a Christmasceremony performed years earlier in her churchat home in Holland.

Since 1936 many other brain surgeons haveobserved the same triggering of memory by cor-tical stimulation. These electrically elicited ex-periences always appear to be real happeningsout of the past although the patient usually hasnot been consciously carrying them in his mem-ory. The episodes recalled are frequently incon-sequential, but they are never vague. Their vivid-

ness, in fact, differentiates them from ordinarymemory. Instead of a remembering, according toPenfield, "it is a hearing-again and seeing-again-a living-through moments of past time ." • Kever-the less, the patient does not lose contact with thepresent. He seems to have concurrent existences-one in the operating room, and one in the partof the past that he is reliving. The term "doubleconsciousness" is employed by brain surgeons todescribe these peculiar sensations of their pa-tients.

Experience induced by cortical stimulationalso differs in other ways from ordinary memory.The induced recollection cannot be speeded lipor slowed down. It unfolds at its own pace, withevents in their original time sequence. When thestimulating current is turned off, the recollectionabruptly stops. If the current is turned on again,the same remembered episode may return to thepatient's mind, but it never simply continueswhere it left off. Instead, the recalled episodestarts again from the beginning-as though itwere stored on a film or tape which automaticallyrewinds each time it is interrupted.

Penfield's work raised a hope of determiningjust what parts of the brain are involved inmemory. And some progress was indeed made insuch localization. It was found, for example, thatthe temporal lobes are the only areas in whichelectrical stimulation elicits experimental recall.This suggests that the records of past events arestored in those regions of the cortex.

However, another explanation is equally pos-sible. Current injec.ted into a temporal lobemight be conducted to some other region wherethe actual memory traces are stored. In fact, itis certain that some memory is stored outside thecortex. A decorticate cat or dog-one which hashad its cortex removed-is still capable of learn-ing simple tasks. Learning involves memory, andthe memory trace established by a decorticateanimal obviously cannot be in the cortex.

\t\'e know too that the stored programs of be-havior that determine reflexes and guide auto-matic bodily control processes operate properlywhen the cortex has been removed. Apparently,the brainstem has its own memory storage sys-tem. This is consistent with other evidence thatthe brainstem is a more or less self-sufficient organthat controls the reflexes, the visceral processes.and other primitive aspects of behavior that pre-serve the physical health and weU-being of theanimal. The same evidence ascribes to the cortex,a much later evolutionary development and a

*StJeech and Brain Mechanisms (Princeton Univer-sity Press, (959).

particular specialty of the higher apes and man,the role of elaboration and refinement-of help-ing the primitive brainstem perform its controlfunctions in a more precise and effective manner.(This, of course, is one way of defining intelli-geuce.)

Thus the proper question about the location ofmemory traces is not whether nIL memory isstored in the cortex, but whether any memory isstored in the cortex. As the cortex expanded fromspecies to higher species, did it bring along itsown storage facilities ~or taking care of the in-creased memory needs of higher intelligence, orinstead did it simply draw on the recording serv-ices of the older and deeper regions of the brain? .An answer to this important question appears tohave been provided by an ingenious series of ex-periments on cats, that has been under way since19r14 in the biological research laboratories ofR. W. Sperry, at the California Institute of Tech-nology.

CORTICAL STORAGE: TIlE CASEOF THE ONE-EYED CATS

ON the working hypothesis that some mem-ories at least can be stored in the cortex,

Sperry and his associates set out to find whetherdifferent memory traces could be established inthe two cortical hemispheres. They knew that thiswould be impossible to discover in the normalbrain, for the two hemispheres are ordinarilyconnected by a gigantic intercommunicatingcable called the corpus callosum) that contains ahundred million or more separate nerve conduc-tors busily cross-feeding information [rom onehemisphere to the other. Therefore the first stepin the Caltech experiments was to cut the corpuscallosum of the test animal. It was known thatsuch an operation would not destroy the animal.I n fact, investigators had long been mystified bythe small effects on behavior and intelligence thisoperation produced. As we shall see, cutting thecorpus callosum does cause major changes in theindividual, but it took sophisticated experimentsto determine what they were.

The surgical preparation of each test animalalso included cutting its optic chiasma. This isa center where the optic nerves from the twoeyes come together. Normally about half thenerve fibers of each eye cross over, in the chiasma,so as to proceed to the other side of the brain.These crossing fibers were cut so that the left eyeof the animal remained connected only to theleft side of its cortex and the right eye to theright side. The animal's vision was impaired

BY DEAN E. WOOLDRIDGE 61

somewhat by the operation, but not enough tobe significant in the experiments. (See Illustra-linn 11.)

Thus the starting material for the Caltechwork consisted of "split-brain" animals-cats thathad been modified by having their cortical hemi-spheres disconnected from one another and theiroptical wiring rearranged to connect each hemi-sphere cleanly to a corresponding eye. Thesemodified cats were the experimental "apparatus."The experiments themselves involved a trainingprocedure. O~e end of the <tat's cage was pro-vided with two swinging doors on which cardswith symbols could be placed. One card wouldcarry a circle, the other a cross. When the cat, byaccident or design, pushed open the door withthe circular symbol, it was rewarded with a mor-sel of food; when it opened the other door, itreceived an annoying puff of air in its face. Afterthirty or forty such trials per day, the animalgenerally learned in a week or so to open thedoor with the circle and avoid the one with thecross.

The Caltech procedure differed from the nor-mal animal-training routine, not only becausesplit-brain animals were used, but also in anotherimportant respect. Each cat was provided with aneye-patch, so that in all of its training it used onlyone selected eye. The crucial point in the ex-periment came after the animal had been trainedto near perfection in choosing the correct visualpattern, when the patch was then shifted to theother eye and the cat. was again presented withthe problem it had learned. The result was spec-tacular. Changing the eye-patch was exactlyequivalent to changing cats! When employing theeye connected to the untrained half of its brain,the animal appeared to have not the slightestrecollection of ever having been in the problembox. Training had to start all over again. And theeat's rate of learning, using the untrained hemi-sphere connecter! to the second eye, was exactlythe same as that of an entirely fresh, untrainedcat. The second eye / hemisphere combinationcould be taught equally well the same or theopposite discrimination from that learned by thefirst. In other words, the animal could be taughtto open the door with the circle when it wasusing the right half of its brain, and to open theone with the cross when its left half was in use.Such an animal would shift its performance auto-matically, without confusion, when the eye-patchwas shifted. Sperry and his co-workers thus estab-lished that each half of the brain possesses itsown memory mechanism. Each can learn its ownhabits and can control the entire organism, if the

ments of the left side of the body. Withonly this isolated remnant of cortex, theanimal could still perform the learnedroutine and learn new discriminationswith the same left paw. Sperry's conclu-sion was that he had cornered the memorytraces for these new habits of tactile dis-crimination in the sensorimotor cortex ofthe corresponding cerebral hemisphere.

Experiments elsewhere have yielded re-sults consistent with the Caltech findings.In rats, as well as cats, the memory trace(or the left paw has been localized in ornear the right sensorimotor cortex. On theother hand, if a visual trick has beentaught, the memory record appears to bestored near the visual cortex, in quite a

CEREBELLUM different part of the brain.Thus it appears that memory records

are stored both in the cortex and in theVISUAL CORTEX brainstem. If the learned task is simple

enough [or the brainstem to handle alone,cortical memory may not be involved. Acat deprived of its auditory cortex canlearn to discriminate between two simpletones. On the other hand, it can be trainedto discriminate between complex musical

patterns involving simultaneous or preciselytimed sequences of tones only if its auditory cor-tex is intact. Probably it is thus in general-thememory trace for a complex discrimination re-sides in the cortex, whereas the trace for a simplerroutine, and perhaps even for the simpler aspectsof the same routine, resides in the deeper and"older" structures of the brain.

Are the split-brain results applicable to man?Comparable experiments are not done on hu-mans, and accident or disease rarely requirescutting of the corpus callosum of a human pa-tient. Thus the prospects of direct confirmationare not good. Nevertheless, the results of otheranimal brain research have usually proved per-tinent to the brain of man. It is quite likely thatwe allocate difficult memory chores to the cortexand e<lsyones to the brainstem in much the sameway as do rats, cats, and monkeys.

62

FRONT AL LOBE

OPTICCHIASMA

TEMPORALLOBE

BRAINSTEM

Illustration II: Diagrammatic view of the brain, seenfrom below. In the split-brain operation, experimenterscut the corpus callosum. They also sever the opticchiasma through which the optic nerves normally connectboth eyes with both sides of the brain.

lines of cortical intercommunication are cut anddifferent training experiences are provided tothe two hemispheres.

The Caltech work has not been confined tovisual experiments. For example, cats with dis-connected cortical hemispheres have been trainedto press one of two levers in accordance with thesmoothness or roughness of the surface. As withvisual discrimination, there was no transfer oflearning from the left to the right paw; the twopaws could learn different or conflicting responses.The split-brain experiments have also been per-formed on monkeys, with the same results.

In Sperry's experiments, the memory mech-anisms were separated into two independent sys-tems by cutting the corpus callosum, whichcommunicates between the two halves of the cor-tex. These results strongly suggested that thememory storage systems involved are located inthe cortical hemispheres, rather than in the brain-stem. Other experiments have indicated justwhere in the cortex the memories are stored. Forexample, one of the split-brain cats was taughta pedal-pressing routine with the left front paw.Then all of its cortex on the corresponding(right) hemisphere was removed, except for thesensorimotor cortex and the surrounding region.This is the area which registers fine sensationsfrom and controls the precise muscular move-

CHANGING ONE INTO TWO

THE Cal tech experiments on cats havea rather spectacular split-personality impli-

cation. This was demonstrated in even moredirect fashion in later work with monkeys. Firsta monkey was subjected to a split-brain opera-tion, in which its corpus callosum and optic chi-asma were cut, just as in the cat experiments.

But in addition, a frontal lobotomy was per-formed on one side of the monkey's brain. In thisoperation the nerve fibers running from thebrainstem to the forward part of the cortex arecut. 'When this operation is done on both sidesof the brain, it is known to produce a relaxed,"I don't care" sort of animal. It has been per-formed on some psychotic human patients to re-lieve unbearable tensions and emotional over-activity.

After the surgery on the monkey, the equiva-lent of an eye-patch was employed to force theanimal to use one eye or the other in viewing itsenvironment. With the monkey employing theeye that was connected to the unmodified corticalhemisphere, a snake was displayed to the animal.Monkeys are normally deathly afraid of snakesand the split-brain monkey was no exception. Itshowed the usua I fright and escape reactions.Then the conditions were changed so that themonkey had to employ the eye connected withthe hemisphere with nerves cut in the lobotomy.Again the snake was displayed. This time, themonkey could not have cared less; the snake heldno terrors for it. It was as though two differentanimal personalities now inhabited its body!

Sperry is working to extend his techniques topermit cutting most of the brainstern as well asthe cortex into disconnected symmetrical halves.It is fascinating to contemplate the possibilityof preparing an animal with essentially two sepa-rate brains, each capable of receiving its ownsensations, recording its own memories, learningits own behavior patterns, developing its ownemotional habits and personality attributes, andperhaps even sleeping and waking independentlyof the other. Fantastic though it seems, this maywell be a consequence of the exciting work nowunder way in biological research laboratories.

MEMORY AND METAPHYSICS

M U C H remains to be learned about mem-ory. We are still in the dark as to the

microscopic neuronal mechanism of the memorytrace. \Ve have not even identified any specificpart of the brain as the repository of all storedinformation. And no surgeon's scalpel has yetbeen able to remove the recollection of a singleevent or of an isolated habit. Nevertheless, re-cent developments, a few of which have been out-lined here, justify an optimistic view of the fu-ture. Brain-research scientists are now supple-menting traditional psychological methods of in-vestigation with powerful new physiologicallyoriented techniques. They are learning how to

BY DE A N E. WOO L n RID G E 63

modify and manipulate the machinery they arestudying and thereby establish definite relationsbetween brain structure and mental phenomena.In all science real progress toward the solutionof a complex problem commences when meansare found to break the problem down into anumber of simpler component parts. The dis-covery of the multiple-memory mechanisms, theelicitation of p<tst recollections by electrical stim-ulation, the evidence from the Caltech split-brainexperiments-each is a small island of knowledgein what has heretofore been a vast sea of ignor-ance.

The emphasis now, and for some time to come,must he on understanding rather than applica-tion. Ultimately, however, new knowledge aboutthe memory mechanisms and other aspects ofbrain function seems certain to lead to new waysto alleviate the effects of diseased or damagedbrain tissue. It is also not too much to hope forimprovement in our ability to educate and getmaximum service from our mental equipment,as we learn more about the workings of the ma-chinery of the brain.

In addition to these scientific and practical im-plications of current research, there is an im-portant philosophic by-product. Scientists arebeginning to come to grips with the phenomenaof consciousness. The split-brain work, for ex-ample, may already have provided a surgical tech-nique for causing two distinct personalities tooccupy a single human body. Separability of theindividual into two identities is also suggestedby the peculiar sensation of "double conscious-ness" experienced by some of Wilder Penfield'spatients. Other research has shown that whetheror not we are conscious depends solely onwhether a suitable pattern of electrical activityexists in a specific center of the brainstern. Andit has also been established that the peculiarlypersonal sensations of fear, horror, rage, pleasure,and ecstasy can be artificially induced by elec-trical stimulation in appropriate small regions ofthe hrain.

Surely these growing indications that our sub-jective sensations are determined in a regularand predictable way by the physical conditionof specific parts of the brain are of the greatesthuman signifiGlIlce. Eventually such knowledgemay well transfer the phenomena of conscious-ness out of metaphysics and into the realm de-scribed by the physical laws of nature. It wouldbe hard to imagine a development of more far-reaching importance to science and philosophy.Yet it could come as a consequence of researchon memory and the brain.

Harper's Magazine, June 1963