HEMISPHERE DECONNECTION AND UNITY INCONSCIOUS AWARENESSJ

R. W. SPERRY

California Institute of Technology

THE following article is a result of studies mycolleagues and I have been conducting withsome neurosurgical patients of Philip J. Vogel

of Los Angeles. These patients were all advancedepileptics in whom an extensive midline section ofthe cerebral commissures had been carried out in aneffort to contain severe epileptic convulsions notcontrolled by medication. In all these people thesurgical sections included division of the corpuscallosum in its entirety, plus division also of thesmaller anterior and hippocampal commissures, plusin some instances the massa intermedia. So far asI know, this is the most radical disconnection ofthe cerebral hemispheres attempted thus far in hu-man surgery. The full array of sections wascarried out in a single operation.

No major collapse of mentality or personalitywas anticipated as a result of this extreme surgery:earlier clinical observations on surgical section ofthe corpus callosum in man, as well as the resultsfrom dozens of monkeys on which I had carried outthis exact same surgery, suggested that the func-tional deficits might very likely be less damagingthan some of the more common forms of cerebralsurgery, such as frontal lobotomy, or even some ofthe unilateral lobotomies performed more routinelyfor epilepsy.

The first patient on whom this surgery was triedhad been having seizures for more than 10 yearswith generalized convulsions that continued toworsen despite treatment that had included asojourn in Bethesda at the National Institutes ofHealth. At the time of the surgery, he had beenaveraging two major attacks per week, each ofwhich left him debilitated for another day or so.

1 Invited address presented to the American PsychologicalAssociation in Washington, D. C., September 1967, and tothe Pan American Congress of Neurology in San Juan,Puerto Rico, October 1967. Original work referred to inthe text by the writer and his co-workers was supportedby Grant MH-03372 from the National Institute ofMental Health, United States Public Health Service, and bythe Hixon Fund of the California Institute of Technology.

Episodes of status cpilcpticus (recurring seizuresthat fail to stop and represent a medical emergencywith a fairly high mortality risk) had also begun tooccur at 2- to 3-month intervals. Since leaving thehospital following his surgery over 51/<> years ago,this man has not had, according to last reports, asingle generalized convulsion. It has further beenpossible to reduce the level of medication and toobtain an overall improvement in his behavior andwell being (see Bogen & Vogel, 1962).

The second patient, a housewife and mother inher 30s, also has been seizure-free since recoveringfrom her surgery, which was more than 4 yearsago (Bogen, Fisher, & Vogel, 1965). Bogen relatedthat even the EEG has regained a normal pattern inthis patient. The excellent outcome in the initial,apparently hopeless, last-resort cases led to furtherapplication of the surgery to some nine more in-dividuals to date, the majority of whom are toorecent for therapeutic evaluation. Although thealleviation of the epilepsy has not held up IQO%throughout the series (two patients are still havingseizures, although their convulsions are much re-duced in severity and frequency and tend to beconfined to one side), the results on the whole con-tinue to be predominantly beneficial, and the overalloutlook at this time remains promising for selectedsevere cases.

The therapeutic success, however, and all othermedical aspects are matters for our medical col-leagues, Philip J. Vogel and Joseph E. Bogen.Our own work has been confined entirely to an ex-amination of the functional outcome, that is, thebehavioral, neurological, and psychological effectsof this surgical disruption of all direct cross-talkbetween the hemispheres. Initially we were con-cerned as to whether we would be able to find inthese patients any of the numerous symptoms ofhemisphere deconnection that had been demon-strated in the so-called "split-brain" animal studiesof the 1950s (Myers, 1961; Sperry, 1967a, 1967b).The outcome in man remained an open question in

723

724 AMERICAN PSYCHOLOGIST

luo. 1. Apparatus for studying laleralization of visual,tactual, lingual, and associated functions in the surgicallyseparated hemispheres.

view of the historic Akelaitis (1944) studies thathad set the prevailing doctrine of the 1940s and1950s. This doctrine maintained that no importantfunctional symptoms are found in man followingeven complete surgical section of the corpus cal-losum and anterior commissure, provided that otherbrain damage is excluded.

These earlier observations on the absence of be-havioral symptoms in man have been confirmed ina general way to the extent that it remains fairto say today that the most remarkable effect ofsectioning the neocortical commissures is the ap-parent lack of effect so far as ordinary behavior isconcerned. This has been true in our animal studiesthroughout, and it seems now to be true for manalso, with certain qualifications that we will cometo later. At the same time, however—and this isin contradiction to the earlier doctrine set by theAkelaitis studies—we know today that with ap-propriate tests one can indeed demonstrate a largenumber of behavioral symptoms that correlate di-rectly with the loss of the neocortical commissuresin man as well as in animals (Gazzaniga, 1967;Sperry, 1967a, 1967b; Sperry, Gazzaniga, & Bogen,1968), Taken collectively, these symptoms maybe referred to as the syndrome of the neocorticalcommissures or the syndrome of the forebrain com-missures or, less specifically, as the syndrome ofhemisphere deconnection.

One of the more general and also more interestingand striking features of this syndrome may be

summarized as an apparent doubling in most ofthe realms of conscious awareness. Instead of thenormally unified single stream of consciousness,these patients behave in many ways as if theyhave two independent streams of conscious aware-ness, one in each hemisphere, each of which is cutoff from and out of contact with the mental experi-ences of the other. In other words, each hemisphereseems to have its own separate and private sensa-tions; its own perceptions; its own concepts; andits own impulses to act, with related volitional,cognitive, and learning experiences. Following thesurgery, each hemisphere also has thereafter itsown separate chain of memories that are renderedinaccessible to the recall processes of the other.

This presence of two minds in one body, as itwere, is manifested in a large number and variety oftest responses which, for the present purposes, Iwill try to review very briefly and in a somewhatstreamlined and simplified form. First, however,let me take time to emphasize that the work re-ported here has been very much a team project.The surgery was performed by Vogel at the WhiteMemorial Medical Center in Los Angeles. Hehas been assisted in the surgery and in themedical treatment throughout by Joseph Bogen.Bogen has also been collaborating in our behavioraltesting program, along with a number of graduatestudents and postdoctoral fellows, among whomM. S. Gazzaniga, in particular, worked closelywith us during the first several years and managedmuch of the testing during that period. The pa-tients and their families have been most coopera-tive, and the whole project gets its primary fundingfrom the National Institute of Mental Health.

Most of the main symptoms seen after hemispheredeconnection can be described for convenience withreference to a single testing setup—shown inFigure 1. Principally, it allows for the lateralizedtesting of the right and left halves of the visualfield, separately or together, and the right and lefthands and legs with vision excluded. The tests canbe arranged in different combinations and in as-sociation with visual, auditory, and other input,with provisions for eliminating unwanted stimuli.In testing vision, the subject with one eye coveredcenters his gaze on a designated fixation point onthe upright translucent screen. The visual stimulion 35-millimeter transparencies are arranged in astandard projector equipped with a shutter and arethen back-projected at 1/jo of a second or less—too

HEMISPHERE DECONNECTION 725

fast for eye movements to get the material intothe wrong half of the visual field. Figure 2 ismerely a reminder that everything seen to theleft of the vertical meridian through either eye isprojected to the right, hemisphere and vice versa.The midline division along the vertical meridian isfound to be quite precise without significant gap oroverlap (Sperry, 1968) .

When the visual perception of these patients istested under these conditions the results indicatethat these people have not one inner visual worldany longer, but rather two separate visual innerworlds, one serving the right half of the field ofvision and the other the left half—each, of course,in its respective hemisphere. This doubling in thevisual sphere shows up in many ways: For example,after a projected picture of an object has beenidentified and responded to in one half field, wefind that it is recognized again only if it reappearsin the same half of the field of vision. If the givenvisual stimulus reappears in the opposite half of thevisual field, the subject responds as if he had norecollection of the previous exposure. In otherwords, things seen through the right half of thevisual field (i.e., through the left hemisphere) areregistered in mental experience and rememberedquite separately from things seen in the other halfof the field. Each half of the field of vision in thecommissurotomized patient has its own train ofvisual images and memories.

This separate existence of two visual inner worldsis further illustrated in reference to speech and writ-ing, the cortical mechanisms for which are centeredin the dominant hemisphere. Visual material pro-jected to the right half of the field—left-hemispheresystem of the typical right-handed patient—can bedescribed in speech and writing in an essentiallynormal manner. However, when the same visualmaterial is projected into the left half of the field,and hence to the right hemisphere, the subject con-sistently insists that he did not see anything orthat there was only a flash of light on the left side.The subject acts as if he were blind or agnostic forthe left half of the visual field. If, however, insteadof asking the subject to tell you what he saw, youinstruct him to use his left hand to point to amatching picture or object presented among acollection of other pictures or objects, the subjecthas no trouble as a rule in pointing out consistently(he very item thai he has just insisted he did notsee.

KEY i CASE

FIG. 2. Things seen to the left of a central fixationpoint with either eye are projected to the right hemisphereand vice-versa.

We do not think the subjects are trying to bedifficult or to dupe the examiner in such tests.Everything indicates that the hemisphere that istalking to the examiner did in fact not see theleft-field stimulus and truly had no experiencewith, rior recollection of, the given stimulus. Theother, the right or nonlingual hemisphere, however,did see the projected stimulus in this situation andis able to remember and recognize the object andcan demonstrate this by pointing out selectively thecorresponding or matching item. This other hemi-sphere, like a deaf mute or like some aphasics, can-not talk about the perceived object and, worse still,cannot write about it either.

If two different figures are flashed simultaneouslyto the right and left visual fields, as for example a"dollar sign" on the left and a "question mark"on the right and the subject is asked to draw whathe saw using the left hand out of sight, he regularlyreproduces the figure seen on (he left half of the

726 AMERICAN PSYCHOLOGIST

field, that is, the dollar sign. If we now ask him whathe has just drawn, he tells us without hesitationthat the figure he drew was the question mark, orwhatever appeared in the right half of the field.In other words, the one hemisphere does not knowwhat the other hemisphere has been doing. Theleft and the right halves of the visual field seemto be perceived quite separately in each hemispherewith little or no cross-influence.

When words are flashed partly in the left fieldand partly in the right, the letters on each side ofthe midline are perceived and responded to sepa-rately. In the "key case" example shown in Fig-ure 2 the subject might first reach for and selectwith the left hand a key from among a collectionof objects indicating perception through the minorhemisphere. With the right hand he might thenspell out (he word "case" or he might speak theword if verbal response is in order. When askedwhat kind of "case" he was thinking of here, theanswer coming from the left hemisphere might besomething like "in case of fire" or "the case of themissing corpse" or "a case of beer," etc., dependingupon the particular mental set of the left hemisphereat the moment. Any reference to "key case" underthese conditions would be purely fortuitous, assum-ing that visual, auditory, and other cues have beenproperly controlled.

A similar separation in mental awareness isevident in tests that deal with stereognostic orother somesthetic discriminations made by the rightand left hands, which arc projected separately tothe left and right hemispheres, respectively. Ob-jects put in the right hand for identification bytouch are readily described or named in speech orwriting, whereas, if the same objects are placed inthe left hand, the subject can only make wildguesses and may often seem unaware that anythingat all is present. As with vision in the left field,however, good perception, comprehension, andmemory can be demonstrated for these objects inthe left hand when the tests are so designed thatthe subject can express himself through nonverbalresponses. For example, if one of these objectswhich the subject tells you he cannot feel or doesnot recognize is taken from the left hand and placedin a grab bag or scrambled among a dozen othertest items, the subject is then able to search outand retrieve the initial object even after a delayof several minutes is deliberately interposed. Un-like (he normal subject, however, these people are

obliged to retrieve such an object with the samehand with which it was initially identified. Theyfail at cross-retrieval. That is, they cannot recog-nize with one hand something identified onlymoments before with the other hand. Again, thesecond hemisphere does not know what the firsthemisphere has been doing.

When the subjects are first asked to use theleft hand for these stereognostic tests they com-monly complain that they cannot "work with thathand," that the hand "is numb," that they "justcan't feel anything or can't do anything with it,"or that they "don't get the message from thathand." If the subjects perform a series of success-ful trials and correctly retrieve a group of objectswhich they previously stated they could not feel,and if this contradiction is then pointed out tothem, we get comments like "Well, T was justguessing," or "Well, I must have done it un-consciously."

With other simple tests a further lack of cross-integration can be demonstrated in the sensory andmotor control of the hands. In a "symmetrichandpose" test the subject holds both hands outof sight symmetrically positioned and not in contact.One hand is then passively placed by the examinerinto a given posture, such as a closed fist, or one,two, or more fingers extended or crossed or foldedinto various positions. The subject is then in-structed verbally or by demonstration to form thesame pose with the other hand, also excluded fromvision. The normal subject does this quite ac-curately, but the commissurotomy patient generallyfails on all but the very simplest hand postures, likethe closed fist or the fully extended hand.

In a test for crossed topognosis in the hands, thesubject holds both hands out of sight, forwardand palm up with the fingers held apart and ex-tended. The examiner then touches lightly a pointon one of the figures or at the base of the fingers.The subject responds by touching the same targetpoint with the tip of the thumb of the same hand.Cross-integration is tested by requiring the patientto use the opposite thumb to find the correspondingmirror point on the opposite hand. The com-missurotomy patients typically perform well withincither hand, but fail when they attempt to cross-locate the corresponding point on the oppositehand. A crude cross-performance with abnormallylong latency may be achieved in some cases afterpractice, depending on the degree of ipsilateral

HEMISPHESK DECONNECTION 727

motor control and the development of certainstrategies. The latter breaks down easily understress and is readily distinguished from the naturalperformance of the normal subject with intactcallosum.

In a related test the target point is presentedvisually as a black spot on an outline drawing ofthe hand. The picture is flashed to the right or lefthalf of the visual field, and the subject then at-tempts as above to touch the target spot with thetip of the thumb. The response again is performedon the same side with normal facility but is impairedin the commissurotomy patient when the left visualfield is paired with a right-hand response and viceversa. Thus the duality of both manual stereognosisand visuognosis is further illustrated; each hemi-sphere perceives as a separate unit unaware of theperceptual experience of the partner.

If two objects are placed simultaneously, one ineach hand, and then are removed and hidden forretrieval in a scrambled pile of test items, each handwill hunt through the pile and search out selectivelyits own object. In the process each hand may ex-plore, identify, and reject the item for which theother hand is searching. It is like two separateindividuals working over the collection of test itemswith no cooperation between them. We find theinterpretation of this and of many similar perform-ances to be less confusing if we do not try to thinkof the behavior of the commissurotomy patient asthat of a single individual, but try to think insteadin terms of the mental faculties and performancecapacities of the left and the right hemispheresseparately. Most of the time it appears that themajor, that is, the left, hemisphere is in control.But in some tasks, particularly when these areforced in testing procedures, the minor hemisphereseems able to take over temporarily.

It is worth remembering that when you splitthe brain in half anatomically you do not dividein half, in quite the same sense, its functionalproperties. In some respects cerebral functionsmay be doubled as much as they are halved becauseof the extensive bilateral redundancy in brain organ-ization, wherein most functions, particularly in sub-human species, are separately and ra.ther fullyorganized on both sides. Consider for example thevisual inner world of either of the disconnectedhemispheres in these patients. Probably neither ofthe separated visual syslems senses or perceivesitself to be cut in half or even incomplete. One

may compare it lo the visual sphere of the hcmi-anopic patient who, following accidental destructionof an entire visual cortex of one hemisphere, maynot even notice the loss of the whole half sphereof vision until Ihis has been pointed out to him inspecific optometric tests. These commissurotomypatients continue to watch television and to readthe paper and books with no complaints aboutpeculiarities in the perceptual appearance of thevisual field.

At the same time, I want to caution against anyimpression that these patients are better off mentallywithout their cerebral commissures. It is true thatif you carefully select two simple tasks, each ofwhich is easily handled by a single hemisphere,and then have the two performed simultaneously,there is a good chance of getting better than normalscores. The normal interference effects that cornefrom trying to attend to two separate right and lefttasks at the same time are largely eliminated inthe commissurotomized patient. However, in mostactivities that are at all complex the normally uni-fied cooperating hemispheres still appear to dobetter than the two disconnected hemispheres. Al-though it is true that (he intelligence, as measuredon IQ tests, is not much affected and that Ihepersonality comes through with little change, onegets the impression in working with these peoplethat their intellect is nevertheless handicapped inways that are probably not revealed in the ordinarytests. All the patients have marked short-terrnmemory deficits, which are especially pronouncedduring the first year, and it is open to questionwhether this memory impairment ever clears corn-pletery. They also have orientation problems,fatigue more quickly in reading and in other tasksrequiring mental concentration, and presumablyhave various other impairments that reduce theupper limits of performance in functions that haveyet to be investigated. The patient that has shownthe best recovery, a boy of 14, was able to return topublic school and was doing passing work with B toI) grades, except for an F in math, which he had torepeat. He was, however, a D student before thesurgery, in part, it would seem, for lack of motiva-tion. In general, our tests to date have been con-cerned mostly with basic cross-integrational deficitsin these patients and the kind of mental capacitiespreserved in the subordinate hemisphere. Studiedcomparisons of the upper limits of performancebefore and after surgery arc still needed.

728 AMERICAN PSYCHOLOGIST

SPACIALCONSTRUCTION

SIMPLE LANGUAGECOMPREHENSION

NONVERBALIDEATION

FIG. A. Schematic outline of the functional lateralizationevident in behavioral tests of patients with forebraincominissurotoray.

Much of the foregoing is summarized sche-matically in Figure 3. The left hemisphere in theright-handed patients is equipped with the ex-pressive mechanisms for speech and writing andwith the main centers for the comprehension andorganization of language. This "major" hemispherecan communicate its experiences verbally and inan essentially normal manner. It can communicate,that is, about the visual experiences of the righthalf of the optic field and about the somestheticand volitional experiences of the right hand andleg and right half of the body generally. In addi-tion, and not indicated in the figure, the majorhemisphere also communicates, of course, about allof the more general, less lateralized cerebral activitythat is bilaterally represented and common to bothhemispheres. On the other side we have the muteaphasic and agraphic right hemisphere, which can-not express itself verbally, but which through theuse of nonverbal responses can show that it is notagnostic; that mental processes are indeed present

centered around the left visual field, left hand,left leg, and left half of the body; along with theauditory, vestibular, axial somatic, and all othercerebral activities that are less lateralized and forwhich the mental experiences of the right and lefthemispheres may be characterized as being similarbut separate.

It may be noted that nearly all of the symp-toms of cross-integrational impairment that I havebeen describing are easily hidden or compensatedunder the conditions or ordinary behavior. Forexample, the visual material has to be flashed at1/]0 of a second or less to one half of the field inorder to prevent compensation by eye movements.The defects in manual stereognosis are not apparentunless vision is excluded; nor is doubling inolfactory perception evident without sequential oc-clusion of right and left nostril and elimination ofvisual cues. In many tests the major hemispheremust be prevented from talking to the minor hemi-sphere and thus giving away the answer throughauditory channels. And, similarly, the minor hemi-sphere must be prevented from giving nonverbalsignals of various sorts to the major hemisphere.There is a great diversity of indirect strategies andresponse signals, implicit as well as overt, by whichthe informed hemisphere can be used to cue-in theuninformed hemisphere (Levy-Agresti, 1968).

Normal behavior under ordinary conditions isfavored also by many other unifying factors. Someof these are very obvious, like the fact that thesetwo separate mental spheres have only one body,so they always get dragged to the same places, meetthe same people, and see and do the same thingsall the time and thus are bound to have a greatoverlap of common, almost identical, experience.Just the unity of the optic image—and even afterchiasm section in animal experiments, the conjugatemovements of the eyes—means that both hemi-spheres automatically center on, focus on, andhence probably attend to, the same items in thevisual field all the time. Through sensory feed-back a unifying body schema is imposed in eachhemisphere with common components that similarlycondition in parallel many processes of perceptionand motor action onto a common base. To getdifferent activities going and different experiencesand different memory chains built up in the sepa-rated hemispheres of the bisected mammalian brain,as we do in the animal work, requires a con-

HEMISPHERE DECONNECTION 729

FIG. 4. In tests for ipsilateral motor control, different hand postures inoutline drawing arc projected one at a time to left or right visual field (seeFigure 1). Subject attempts to copy the sample hand pose with thehomolateral and the contralateral hand.

siderable amount of experimental planning andeffort.

In motor control we have another importantunifying factor, in that either hemisphere can di-rect the movement of both sides of the body, includ-ing to some extent the movements of the ipsilateralhand (Hamilton, 1967). Insofar as a response in-volves mainly the axial parts and proximal limbsegments, these patients have little problem indirecting overall response from sensory informationrestricted to either single hemisphere. Control ofthe distal limb segments and especially of thefiner finger movements of the hand ipsilateral tothe governing hemisphere, however, are borderlinefunctions and subject to considerable variation.Impairments are most conspicuous when the sub-ject is given a verbal command to respond with thefingers of the left hand. The absence of the cal-losum, which normally would connect the languageprocessing centers in the left hemisphere to themain left-hand motor controls in the oppositehemisphere, is clearly a handicap, especially in theearly months after surgery. Cursive writing withthe left hand presents a similar problem. It maybe accomplished in time by some patients usingshoulder and elbow rather than finger movement.At best, however, writing with the left hand is not

as good after as before the surgery. The problemis not in motor coordination per se, because thesubject can often copy with the left hand a wordalready written by the examiner when the sameword cannot be written to verbal command.

In a test used for more direct determination ofthe upper limits of this ipsilateral motor control, asimple outline sketch of a finger posture (seeFigure 4) is flashed to a single hemisphere, andthe subject then tries to mimic the posture with thesame or the opposite hand. The sample posturecan usually be copied on the same side (i.e., throughthe main, contralateral control system) withoutdifficulty, but the performance does not go so easilyand often breaks down completely when the subjectis obliged to use the opposite hand. The closedfist and the open hand with all fingers extended seemto be the two simplest responses, in that thesecan most often be copied with the ipsilateral handby the more adept patients.

The results are in accord with the thesis(Gazzaniga, Bogen, & Sperry, 1967) that theipsilateral control systems are delicate and marginaland easily disrupted by associated cerebral damageand other complicating factors. Preservation of theipsilateral control system in varying degree in somepatients and not in others would appear to account

730 AMERICAN PSYCHOLOGIST

for many of the discrepancies that exist in theliterature on the symptoms of hemisphere dccon-nection, and also for a number of changes betweenthe present picture and that described until 2 yearsago. Those acquainted with the literature willnotice that the present findings on dyspraxia comemuch closer to the earlier Akelaitis observations thanthey do to those of Licpmann or of others ex-pounded more recently (see Geschwind, 1965).

To try to find out what goes on in that speechlessagraphic minor hemisphere has always been one ofthe main challenges in our testing program. Doesthe minor hemisphere really possess a true stream ofconscious awareness or is it just an agnostic autom-aton that is carried along in a reflex or trancclikestate? What is the nature, the quality, and thelevel of the mental life of this isolated subordinateunknown half of the human brain—which, like theanimal mind, cannot communicate its experiences?Closely tied in here are many problems that relateto lateral dominance and specialization in the hu-man brain, to the functional roles mediated bythe neocortical commissures, and to related aspectsof cerebral organization.

With such in mind, I will try to review brieflysome of the evidence obtained to date that pertainsto the level and nature of the inner mental lifeof the disconnected minor hemisphere. First, it isclear that the minor hemisphere can perform inter-modal or cross-modal transfer of perceptual andmnemonic information at a characteristically hu-

man level. For example, after a picture of someobject, such as a cigarette, has been flashed to theminor hemisphere through the left visual field, thesubject can retrieve the item pictured from acollection of objects using blind touch with the lefthand, which is mediated through the right hemi-sphere. Unlike the normal person, however, thecommissurotomy patient is obliged to use the cor-responding hand (i.e., the left hand, in this case)for retrieval and fails when he is required to searchout the same object with the right hand (see Figure5). Using the right hand the subject recognizesand can call off the names of each object that hecomes to if he is allowed to do so, but the righthand or its hemisphere does not know what it islooking for, and the hemisphere that can recognizethe correct answer gets no feedback from the righthand. Hence, the two never get together, and theperformance fails. Speech and other auditory cuesmust be controlled.

Jt also works the other way around: that is,if the subject is holding an object in the left hand,he can then point out a picture of this object orthe printed name of the object when these appearin a series presented visually. But again, theselatter must be seen through the corresponding halfof the visual field; an object identified by the lefthand is not recognized when seen in the right halfof the visual field. Intermodal associations of thissort have been found to work between vision, hear-ing and touch, and, more recently, olfaction in

FIG. 5. Visuo-tactilc associations succeed between each half of the visualfield and the corresponding hand. They fail with crossed combinations inwhich visual and tactual stimuli are projected into opposite hemispheres.

HEMISPHERE DECONNECTION 731

various combinations within either hemisphere butnot across from one hemisphere to the other. Thisperceptual or mnemonic transfer from one sensemodality to another has special theoretical interestin that it is something that is extremely difficult orimpossible for the monkey brain. The right hemi-sphere, in other words, may be animallike in notbeing able to talk or write, but in performances likethe foregoing and in a number of other respects itshows mental capacities that are definitely human.

Other responses from the minor hemisphere inthis same testing situation suggest the presenceof ideas and a capacity for mental association andat least some simple logic and reasoning. In thesame visuo-tactual test described above, the minorhemisphere, instead of selecting objects that matchexactly the pictured item, seems able also to selectrelated items or items that "go with" the particularvisual stimulus, if the subject is so instructed. Forexample, if we flash a picture of a wall clock to theminor side and the nearest item (hat can be foundfactually by the left hand is a toy wrist watch, thesubjects significantly select the watch. It is as ifthe minor hemisphere has an idea of a timepiecehere and is not just matching sensory outlines. Or,if the picture of a dollar sign is flashed to the minorside, the subject searches through the list of itemswith the left hand and finally selects a coin suchas a quarter or a 50^ piece. If a picture of ahammer is presented, the subject may come up witha nail or a spike after checking out and rejectingall other items.

The capacity to think abstractly with symbols isfurther indicated in the ability of the minor hemi-sphere to perform simple arithmetical problems.When confronted with two numerals each less than10, the minor hemisphere was able in four of sixsubjects so tested to respond with the correct sumor product up to 20 or so. The numbers wereflashed to the left half of the visual field or pre-sented as plastic block numerals to the left handfor identification. The answer was expressed bypointing to the correct number in columns of seenfigures, or by left-hand signals in which the fingerswere extended out of the subject's sight, or bywriting the numerals with the left hand out of sight.After a correct left-hand response had been madeby pointing or by writing the numeral, the majorhemisphere could then report the same answerverbally, but the verbal report could not be madeprior to the left-hand response. If an error was

made with the left hand, the verbal report ovi-tained the same error. Two different pairs of nu-merals may be flashed to right and left fields simul-taneously and the correct sum or products signaledseparately by right and left hands. When verbalconfirmation of correct left-hand signals is requiredunder these conditions, the speaking hemispherecan only guess fortuitously, showing again that,the answer must have been obtained from the minorand not from the major hemisphere. This hasbeen demonstrated recently in a study still in prog-ress by Biersncr and the present writer. The find-ings correct an earlier impression (Gazzaniga &Sperry. 1967) in which we underestimated thecapacity for calculation on the minor side. Normalsubjects and also a subject with agenesis of thecallosum (Saul & Sperry, 1968) were able to addor to multiply numerals shown one in the left andone in the right field tinder these conditions. ThecomnnVsurotomy subjects, however, were able toperform such calculations only when both nu-merals appeared in the same half of the visual field.

According to a doctrine of long standing in theclincial writings on aphasia, it is believed that theminor hemisphere, when it has been disconnectedby commissural or other lesions from the languagecenters on the opposite side, becomes then "wordblind," "word deaf," and "tactually alexic." In con-tradiction to this, we find that the disconnectedminor hemisphere in these commissurotomy pa-tients is able to comprehend both written andspoken words to some extent, although this com-prehension cannot be expressed verbally (Gazzaniga& Sperry, 1967; Sperry, 1966; Sperry & Gazzaniga,1967). If the name of some object is flashed to theleft visual field, like the word "eraser," for example,the subject is able then to search out an eraser fromamong a. collection of objects using only touch withthe left hand. Jf the subject is then asked what theitem is after it has been selected correctly, hisreplies show that he does not know what he isholding in his left hand-—as is the general rule forleft-hand stereognosis. This means of course thatthe talking hemisphere does not know the correctanswer, and we concluded accordingly that theminor hemisphere must, in this situation, have readand understood the test world.

These patients also demonstrate comprehensionof language in the minor hemisphere by being ableto find by blind touch with the left hand an objectthat has been named aloud by the examiner. For

732 AMERICAN PSYCHOLOGIST

example, if asked to find a "piece of silverware,"the subject may explore the array of test itemsand pick up a fork. If the subject is then askedwhat it is that he has chosen, he is just as likelyin this case to reply "spoon" or "knife" as fork.Both hemispheres have heard and understood theword "silverware," but only the minor hemisphereknows what the left hand has actually found andpicked up. ]n similar tests for comprehension ofthe spoken word, we find that the minor hemisphereseems able to understand even moderately advanceddefinitions like "shaving instrument" for razor or"dirt remover" for soap and "inserted in slotmachines" for quarter.

Work in progress shows that the minor hemi-sphere can also sort objects into groups by touchon the basis of shape, size, and texture. In sometests the minor hemisphere is found to be superiorto the major, for example, in tasks that involvedrawing spatial relationships and performing blockdesign tests. Perceptive mental performance in theminor hemisphere is also indicated in other situa-tions in which the two hemispheres function con-currently in parallel at different tasks. It has beenfound, for example, that the divided hemispheres arecapable of perceiving different things occupying thesame position in space at the same time, and oflearning mutually conflicting discrimination habits,something of which the normal brain is not capable.This was shown in the monkey work done someyears ago by Trevarthen (1962) using a system ofpolarized light filters. It also required section ofthe optic chiasm, which of course is not included inthe human surgery. The human patients, unlikenormal subjects, are able to carry out a doublevoluntary reaction-time task as fast as they carryout a single task (Gazzaniga & Sperry, 1966).Each hemisphere in this situation has to perform aseparate and different visual discrimination in orderto push with the corresponding hand the correct oneof a right and left pair of panels. Whereas inter-ference and extra delay are seen in normal subjectswith the introduction of the second task, these pa-tients with the two hemispheres working in parallelsimultaneously perform the double task as rapidlyas the single task.

The minor hemisphere is also observed to demon-strate appropriate emotional reactions as, for ex-ample, when a pinup shot of a nude is interjectedby surprise among a series of neutral geometricfigures being flashed to the right and left fields at

random. When the surprise nude appears on theleft side the subject characteristically says that heor she saw nothing or just a flash of light. However,the appearance of a sneaky grin and perhaps blush-ing and giggling on the next couple of trials or sobelies the verbal contention of the speaking hemi-sphere. If asked what all the grinning is about, thesubject's replies indicate that the conversant hemi-sphere has no idea at this stage what it was that hadturned him on. Apparently, only the emotionaleffect gets across, as if the cognitive componentof the process cannot be articulated through thebrainstem.

Emotion is also evident on the minor side in a cur-rent study by Gordon and Sperry (1968) involvingolfaction. When odors are presented through theright nostril to the minor hemisphere the subject isunable to name the odor but can frequently tellwhether it is pleasant or unpleasant. The subjectmay even grunt, make aversive reactions or ex-clamations like "phew!" to a strong unpleasantsmell, but not be able to state verbally whether itis garlic, cheese, or some decayed matter. Again itappears that the affective component gets across tothe speaking hemisphere, but not the more specificinformation. The presence of the specific informa-tion within the minor hemisphere is demonstratedby the subject's correct selection through left-handstereognosis of corresponding objects associated withthe given odor. The minor hemisphere also com-monly triggers emotional reactions of displeasure inthe course of ordinary testing. This is evidencedin the frowning, wincing, and negative head shakingin test situations where the minor hemisphere,knowing the correct answer but unable to speak,hears the major hemisphere making obvious verbalmistakes. The minor hemisphere seems to expressgenuine annoyance at the erroneous vocal responsesof its better half.

Observations like the foregoing lead us to favorthe view that in the minor hemisphere we deal witha second conscious entity that is characteristicallyhuman and runs along in parallel with the moredominant stream of consciousness in the majorhemisphere (Sperry, 1966). The quality of mentalawareness present in the minor hemisphere may becomparable perhaps to that which survives in sometypes of aphasic patients following losses in themotor and main language centers. There is no in-dication that the dominant mental system of theleft hemisphere is concerned about or even aware of

HEMTSPHEKK DECONNECTION

the presence of the minor system under most ordin-ary conditions except quite indirectly as, for ex-ample, through occasional responses triggered fromthe minor side. As one patient remarked im-mediately after seeing herself make a left-hand re-sponse of this kind, "Now 1 know it wasn't medid that!"

Let me emphasize again in closing that the fore-going represents a somewhat abbreviated andstreamlined account of the syndrome of hemispheredeconnection as we understand it at the presenttime. The more we see of these patients and themore of these patients we see, the more we be-come impressed with their individual differences,and with the consequent qualifications that must betaken into account. Although the general picturehas continued to hold up in the main as described,it is important to note that, with respect to manyof the deconnection symptoms mentioned, strikingmodifications and even outright exceptions can befound among the small group of patients examinedto date. Where the accumulating evidence willsettle out with respect to the extreme limits ofsuch individual variations and with respect to apossible average "type" syndrome remains to beseen.

REFERENCES

AKELAITIS, A. J. A study of gnosis, praxis, and languagefollowing section of the corpus callosum and anteriorcommissure. Journal oj Neurosurgery, 1944, 1, 94-102.

BOGEN, J. E., FISHER, E. D., & VOGEL, P. J. Cerebralcommissurotomy: A second case report. Journal oj theAmerican Medical Association, 1965, 194, 1328-1329.

BOGEN, J. E., & VOGEL, P. J. Cerebral commissurelomy: Acase report. Bulletin oj the Los Angeles NeurologicalSociety, 1962, 27, 169.

GAZZANIGA, M. S. The split brain in man. ScientificAmerican, 1967, 217, 24-29.

GAZZANIGA, M. S., BOOEN, J. E., & SPERRY, R. W.Dyspraxia following division of the cerebral commissures.Archives oj Neurology, 1967, 16, 606-612.

GAZZANIGA, M. S., & SPERRY, R. W. Simultaneous doublediscrimination following brain bisection. PsychonomicScience, 1966, 4, 262-263.

GAZZANIGA, M. S., & SPERKY, R. W. Language after sec-tion of the cerebral commissures. Brain, 1967, 90, 131-148.

GuscirwiND, N. Disconnexion syndromes in animals andman. Brain, 1965, 88, 237-294, 584-644.

GOKDON, H. W., & Si'KKRV, R. W. Olfaclioii following sur-gical disconnection of the hemispheres in man. In ,Proceedings oj the Psychonomic Society, 1968, in press.

HAMILTON, C. R. Effects of brain bisection on eye-bandcoordination in monkeys wearing prisms. Journal ojComparative and Physiological Psychology, 1967, 64,434-443.

LEVY-AGRESTI, J. Ipsilaleral projection systems and minorhemisphere function in man after neocommissurotomy.Anatomical Record, 1968, 160, 384.

MYERS, R. E. Corpus callosum and visual gnosis. InJ. F. Delafresnayc (Ed.), Brain mechanisms and learn-ing. Oxford: Blackwcll, 1961.

SAUL, R., & Si'ERRY, R. W. Absence of commissurotomysymptoms with agenesis of the corpus callosum. Neu-rology, 1968, 17, in press.

SPERRY, R. W. Brain bisection and mechanisms of con-sciousness. In J. C. Eccles (Ed.), Brain and consciousexperience. New York: Springer-Verlag, 1966.

Si'ERitY, R. W. Mental unity following surgical discon-nection of the hemispheres. The Harvey lectures.Series 62. New York: Academic Press, 1967. (a)

SPERRY, R. W. Split-brain approach to learning problems.In G. C. Quarton, T. Melnechuk, & F. O. Schmitt (Eds.),The neurosciences: A study program. New York:Rockefeller University Press, 1967. (b)

SPERRY, R. W. Apposition of visual half-fields after sec-tion of ncocortical commissures. Anatomical Record,1968, 160, 498-499.

SPERRY, R. W., & GAZZANIGA, M. S. Language followingsurgical disconnection of the hemispheres. In C. H.Milikan (Ed.), Brain mechanisms underlying speech andlanguage. New York: Grunc & Stratton, 1967.

SFKRRY, R. W., GAZZANIGA, M. S., & BOGEN, J. E. Functionof neocorlical commissures: Syndrome of hemispheredeconnection. In P. J. Vinken & G. W. Bruyn (Eds.),Handbook of neurology. Amsterdam: North Holland,1968, in press.

TREVARTHEN, C. B. Double visual learning in split-brainmonkeys. Science,, 1962, 136, 258-259.