Volume 55 July 1962

Royal Society ofMedicine

President The Rt Hon Lord Adrian OM MD FRS

Meeting April 11 1962

Sherrington Memorial Lecture

The Midbrain and Motor Integrationby D Denny-Brown MD FRCP'

'The motor individual is driven from two sources.The world around it and its own inner world within.Its activity is also partly operated by nervous activityarising spontaneously within the nervous centresthemselves' (Sherrington 1940, p 181).

In a remarkable tribute to Sir Charles Sherrington,his former collaborator and associate, ProfessorE G T Liddell (1960), has drawn an entertainingand vivid picture of the painfully slow develop-ment of knowledge of the physiology of thenervous system up to the time Sherrington beganhis life work. His outstanding contribution wasthe unravelling of the patterns of spinal reflexesand their interaction. Sherrington also maderemarkable contributions to the study of themanagement of spinal reflexes by the brain-stemmechanisms, enough to show the complexity andlability of these in contrast to the more readilyidentifiable spinal level of management. Theclinician must of necessity deal with perturbationof nervous function at all levels and, in hisgropings for better understanding of derange-ment of hemisphere and brain-stem function, isstill in many respects facing the same confusionof ideas of nervous action at these levels thatcharacterized the pre-Sherrington ideas of spinalreflexes.As Professor Liddell has so well portrayed in

his memoir, the development of knowledge offunction had to be preceded by analysis of struc-ture. Often, however, a plausible but wrong idea,derived from some purely anatomical premises,had for a very long time delayed the true ap-preciation of the nature of function. This lastdifficulty has plagued the investigation of cerebralphysiology, where anatomical connexions, ortheir equivalent stimulated effects, have for toolong served as explanation for physiologicalfunction. The physiologist on the other hand

'Putnam Professor of Neurology, Harvard Medical School, andDirector, Harvard Neurological Unit, Boston City Hospital,Boston, Massachusetts, U.S.A.

finds in the brain a plasticity of functionalperformance that makes him unwilling to makeabsolute identification of structure and function.In addition it often appears that at this level thewhole is not necessarily just the sum of all theparts, a haunting doubt that pervades 'Man onHis Nature' (Sherrington 1940).

Decerebrate RigidityThough his few contributions to the physiologyof the brain are outstanding, Sherrington did notbecome embroiled in the controversies overcerebral localization and function: Yet thediscovery of decerebrate rigidity (Sherrington1897a, b) and its systematic analysis is a greatlandmark in the development of neurology, forit is the classic example of release of nervousactivity. Magendie (1823) had in fact describedaccurately the decerebrate phenomenon in therabbit and its abolition by pontine section.Lowenthal & Horsley (1897) had evidentlyobserved the 'cataleptoid state' independently ofSherrington about the same time. But it wasSherrington who performed the seeminglyimpossible feat of demonstrating that the extensorspasm arose in each muscle concerned, with thefinal analysis of the stretch reflex by Liddell &Sherrington (1924). We predict that,just as Bell'sdiscovery of the afferent and efferent roots wasthe most significant single advance in all spinalphysiology, the discovery of decerebrate rigiditywill in future years mark the point at whichunderstanding of the mechanism of the brainreally began. For Sherrington it provided abackground of activity on which to test andenlarge his views of the nature of inhibition, butin addition it was the starting point for theexploration of righting reflexes by Magnus (1924)and by Rademaker (1926) and more recentlyof the nature of cerebellar effects so brilliantlyexploited by Stella et al. (1955), Moruzzi &Pompeiano (1957) and Sprague & Chambers(1953).Of the many remarkable features of decere-

brate rigidity the foremost is its immediate andinvariable appearance following transection be-

527

528 Proceedings ofthe Royal Society ofMedicine

tween the intercollicular level and upper pons.There is no phase of initial shock. Sherrington(1898) observed that the exaggerated posturewas not abolished by section of the dorsal columnsof the spinal cord, or by ablation of the cere-bellum, or by section of the VIII nerves. It wasabolished in any limb by deafferenting that limb.The stretch reflex, however, is a spinal reflex(Denny-Brown & Liddell 1927), and there stillremained the question as to the source of theheightening effect. This question is still not finallysettled. In part there is facilitation of some spinalstretch reflexes by the exclusion of those of othersegments. Thus transection of the spinal cordbehind the cervical enlargement heightens therigidity in the forelimbs (the Schiff-Sherringtonphenomenon). There are also other interactions,for if the cerebellum is removed the rigidityagain appears in the deafferented limb, as Stellaand his associates (1955) have shown. This effectis in part due to a restraining effect of the anteriorvermis on the labyrinthine system through thefastigial nuclei. The extensor spasm then stillseems to follow the pattern of the original decere-brate rigidity, but with the addition of moreneck extension, which in turn frequently sets up aneck reflex causing extension of the forelimbs.The stretch reflex, however, is subdued, and it isnow the alpha neurons that are hyperexcitable.Such observations only inform us that by such

maneuvres it is possible to secure an exaggeratedtonic labyrinthine and neck reflex. Decerebraterigidity itself is not defined. Section of one VIIInerve or destruction of Deiter's nucleus on oneside abolishes decerebrate rigidity on that side,but the effect is from the other labyrinth, destruc-tion of which restores bilateral rigidity. Theessential anti-gravity posture remains in anintercollicular preparation with the VIII nervesand dorsal columns sectioned.

Hemisection of the Spinal CordFollowing hemisection of the spinal cord in thehigh cervical region the cat and monkey make anastonishing degree of recovery from the paralysisof the limbs on the side of the lesion. As Mettler(1944) has well described in the monkey, it maybe difficult for the casual observer to distinguishthe side of the lesion from the behaviour of theanimal after some months. In unpublished workwith another objectivein 1949 withDrsT Twitchelland L Saenz-Arroyo on cervical hemisection inthe cat,we found in 3 animals that later decerebra-tion, two to ten weeks after hemisection of thespinal cord at C3 level, produced decerebraterigidity of comparable intensity in all four limbs.In two other animals with slightly lower levelsof cervical hemisection the rigidity was not quiteas well developed in the ipsilateral forelimb until

Fig 1 Macaque monkey, 69 days after hemisection ofcervical spinal cord at C4 level on the left side, to showabduction of left arm on tilting in A, and grasp of lefthand on stationary contact in B. In c righting from theleft side and in D its absence from the right side isshown 31 days after left hemisection at CS segment.

an interval of at least three weeks had elapsed.With Dr Twitchell and, later, Dr C W Watson,we found the same phenomenon in two macaquemonkeys with high cervical (C3-C5) hemisection.In both cat and monkey it was also found thatprior to decerebration the animals were unableto right themselves if they lay on the side oppositethe hemisection but could do so rapidly andeffectively if they lay on the side of the lesion(Fig I c). After two to three weeks they learned toroll, or to claw with their uppermost limb if lyingon the opposite side. Primary righting of thehead continued to occur only when lying on theside of the lesion as long as three months afterhemisection. After the first three weeks theipsilateral arm could support the animal when itwas tilted to that side, and with this reaction theredeveloped the ability to use the hand for pre-hension (Fig IA, B).

This finding shows that the descending path-way that mediates decerebrate rigidity is plastic,and bilateral. In the lower limbs it adjusts tothoracic or high lumbar hemisection within tendays. Like the descending pathway for respirationit can cross in the spinal cord below any level ofhemisection and in this respect resembles a net-work rather than a tract. In this phenomenon wesense the operation of a one-ness of fundamentalnervous activity that sets it apart from theunilateral dependence of all the adjustments ofthe organism to the external world. On the otherhand the ascending pathway for most if not allthe body contact righting reaction is crossed, and

Sherrington Memorial Lecture

reaches the other side of the spinal cord in fourto six segments. This is the reason that rightingis preserved on the side of hemisection. It is anacquired modification of the basic mechanism.

The Reactions Released by DecerebrationRademaker's (1926) study of the structuresinvolved in acute midbrain section showed thatin the rabbit and cat the crucial structure lay inthe ventral half of the midbrain, and was vul-nerable to midline section in that region. Heidentified the red nuclei and Forel's decussationas the essential component, destruction of whichabolished phasic labyrinthine and body-on-bodyrighting, releasing decerebrate rigidity.

Subsequent investigations have shown that thesituation is more complex. It has been mostdifficult to keep decerebrate preparations alivefor more than a few days. The experience ofBazett & Penfield (1922), with longest survivalsof 18, 22 and 23 days, has been enlarged by Bard& Macht (1958) with preservation of hypo-thalamus above the lesion, which enabled muchlonger periods of observation. Their pontinetransections were observed for 35 and 39 days,low mesencephalic sections for 30 and 56 days,and high mesencephalic transections for 31 and154 days. Decerebrate rigidity remained intensefor only a few days, then lessened steadily withthe appearance of flexion and crossed extensionreflexes, and stepping movements. Tonic neckand labyrinthine responses appeared. Somerighting of the head occurred in cats with pontinesection, and the uppermost limbs flexed, butthe animals could not right, and if propped up,could not stand. After two weeks the shoulderscould be righted, but not the hindquarters. Catswith low (inferior colliculus) mesencephalic sec-tion could right themselves to a crouching posture,and if stimulated could walk in crouched posture.With high (precollicular) level of midbrainsection the animal could sit, stand, walk spon-taneously, and climb. There were even weakhopping reactions, and coarse proprioceptiveplacing. In that state the rigidity reappeared onlyif the cat was suspended in air, or in any legallowed to hang over the edge of a table. Whenany of the animals, pontine or mesencephalic, layon one side the upper limbs flexed and the under

limbs extended, 'a reflex figure basic to the act ofrighting' (Bard & Macht 1958). I have someobservations of survival of decerebrate monkeysto amplify these conclusions. They also show thatthe first and most important modification ofdecerebrate rigidity that occurs in prolonged sur-vival is the appearance of the positive supportingreaction and related movements of progression.

Chronic Low Decerebration in the MonkeyIn a mature female macaque monkey (lab. no. BG 27)a large lesion of the caudal dorsal two-thirds of themidbrain was produced by direct suction, leavingonly the basis pedunculi intact (Fig 2). The colliculi,aqueduct, central grey matter and caudal mesen-cephalic tegmentum were removed, leaving intact thesubthalamic tegmentum and red nuclei, with thehemispheres above, and the pontine tegmentumbelow. On the day following operation the animal layon one side with extreme extension of all limbs, tailand neck. The fingers and toes were loose, the wristand ankles weakly resistant. On the second and suc-ceeding days the rigidity greatly lessened, and if theanimal was laid on its back the forelimbs flexed,then the hind-limbs.By the sixth day, a positive supporting reaction

appeared, so that passive dorsiflexion of the wrist orankle led to immediate stiffening of the whole of thestimulated limb. When the animal lay on one sidethe upper limbs occasionally showed a flexed posture,and with it the traction response (active flexion) andclosure of the hand or foot in response to a contactstimulus to the palm or sole. No movement of thecontact stimulus was necessary, differentiating thereaction from the grasp reflex (which is also notdependent on posture). Passive extension of digitsstill caused extension of the limb. By the ninth day astrong passive dorsiflexion of any hand or foot led toextension of all the limbs and dorsiflexion of the neck.This spread of reflex effect occurred first tothe diagonally opposite limb, and at times led to areciprocal altemating progressive movement offlexion and extension in all limbs. In the followingdays an extensor thrust and gallop progression couldbe induced in this way. There was never anyrighting, placing or hopping, or lateral tilt reaction,so that the animal had a fixed doll-like appearance.The animal had exhibited periodic respirationthroughout, and could not swallow. It was sacrificedon the sixteenth day.

This animal showed a complete absence of anyspontaneous movement in spite of integrity of the

Fig 2 Sections froma chronic decerebratemonkey showing theextent of removal ofdorsal midbrain, mostcomplete at the level ofthe posterior colliculus(B) and extendingunder the superiorcolliculus (A),with somepatchy dainage to cere-bellar white matter (c)

529

530 Proceedings ofthe Royal Society ofMedicine

pyramidal system. The state of decerebraterigidity underwent changes similar to thosedescribed in Bard & Macht's cats. The preserva-tion of the corticospinal system avoided the longperiod of depression associated with section ofthis tract, and allowed the rapid development ofelementary contact and traction reactions, andof the elements of righting reactions in the handsand feet in response to asymmetrical bodycontact. These are but details of the 'rightingreflex figure' that Bard & Macht found in theirchronic pontine cats. The primary change is anenhancement of gamma discharge, as Eldredet al. (1953) have demonstrated. It will be notedthat the resulting reactions are not entirelyproprioceptive. It is apparent that lack of activityin the hands and feet of the acutely decerebrateanimal is due to the profound shock that followsthe associated pyramidal section. The tractionreaction, by which the limb flexes in response topassive extension of the fingers, can usually bedemonstrated in the monkey immediately aftercomplete precollicular section of the midbrain,

Fig 3 Representative sections ofmonkey BG 24, showinginfarction of vermis in A, the maximal extent ofremoval of tectum at intercollicular level in B, andsparing of the rostral part of superior colliculus in c.Survival 34 days

but the contact closing reaction is then absent.If a monkey is decerebrated by complete inter-collicular section following removal of theprecentral gyrus some weeks earlier the tractionreaction then also appears, but we have notsucceeded in securing long survivals in this state.

In another chronic experiment I used theavoidance of pyramidal shock by previousablation of precentral cortex to show that thetype of decerebrate rigidity related to loss ofthe cerebellar vermis is essentially the samedisturbance of motor function as that producedby the classical intercollicular lesion. The descrip-tion of recovery of movement in this animalfollows.

Chronic Lesion of Cerebellar Vermis and Tectumafter Bilateral Corticospinal LesionThe precentral gyrus was removed on both sides in ayoung but mature male macaque monkey five monthsbefore induction of a dorsal midbrain lesion includingthe inferior colliculus and the caudal half of thesuperior colliculus, with all the underlying centralgrey and aqueduct, leaving III nerve nuclei, posteriorlongitudinal bundle and central tegmental tractsintact (Fig 3). In addition, clipping the superiorcerebellar arteries resulted in a clear-cut infarctionof all the anterior vermis (Fig 3A) without damage tocerebellar nuclei or superior peduncles. In the fivemonths following the coi Lical operations, the animalhad recovered the ability to grasp clumsily with all

Fig 4 The monkey BG 24 showing, in A the generalextension one day after operation, and in B the effect ofpassive dorsiflexion of the right ankle. In c is shown theextension on being held suspended ten days afteroperation, and in D the absence of abduction in tiltingsideways. In E is shown the grasp of the hand whenthe palm is touched, in F the extension and posi-tive supporting reaction when a flat surface was thestimulus

Sherrington Memorial Lecture

limbs by vision, could use thumb and forefinger topick at objects, had excellent visual placing, slowcontact placing, and very slight spasticity in all limbs.Righting was rapid and effective, and the responsesof the limbs we call the 'tilt reactions', flexion ontilting backward and abduction on tilting sideways,and to hopping, were excellent. There were strongpositive supporting and traction reactions, and arather stiff awkward gait, without the over-adductionthat had been present for the first month.

After the cerebellar and tectal lesion there wasintense decerebrate rigidity interrupted by bursts ofrunning movements in all four limbs. The fingers andtoes were stiffly extended (Fig 4A), but became flexedif the animal lay flat on the back, or if contact wasapplied to the palm of hand or sole of foot (Fig 4B).He showed a strong traction reaction. The animalswallowed well if liquid was placed in the mouth. Theeyes showed roving movements, with reactive pupils,but with no evidence of fixation or visual reaction.By the third day the postural reactions in the limbsbecame very brisk and facile. If the animal was sus-pended in air the limbs extended (Fig 4c). The animalattempted righting by flexing the upper limbs andreaching across the body (Fig 4D). There was at thistime no righting of the head, no compensatory eyemovement, and no lateral tilt or placing reaction.A touch to the palm of the hand induced closure

Fig 5 The monkey BG 24 showing, in A from abovedown, the exaggerated forward tilt reaction whenslowly raised,from a lying posture. In B a passive suddendorsiflexion of the left foot by the examiner results in arepetitive progressive movement of the right arm

(Fig 4E) whereas contact with a flat surface thatextended the fingers passively induced a strongsupporting reaction (Fig 4F). Over-adduction of thelimbs was prominent. The forward tilt reaction nowbegan with only slight change in body contact, andwithout alteration of the posture of the head (Fig 5A),and proceeded all the way from fully flexed to fullyextended posture. A sudden dorsiflexion of wrist orankle could induce progression movement in thediagonally opposite limb (Fig SB). Righting of the headfrom lying on the right side appeared on the fourthday, and at this point the extensor rigidity when lyingdisappeared. There was still no reaction to tilt, andno righting in a free fall. On the sixth day he couldright himself from either side by rotating the trunkand pushing up with opposite arm and leg. He wascompletely expressionless and mute, but now couldrecognize his bowl and reach for it with the left arm,and look to a moving object in the left visual field. Onthe tenth day he could abduct the limbs in order tosupport himselfwhen tilted, and could right by abduct-ing the under arm. A coarse tremor of the trunk ofcerebellar type appeared in sitting posture. By theseventeenth day the ability to abduct the right limbswas greatly improved, and the adductor spasm haddisappeared in all limbs. By the thirtieth day thetilt reactions were extremely brisk. The animal couldnow reach out to objects in either visual field, grasp alarge object, and carry it to his mouth. In doing so hefrequently exhibited a violent cerebellar tremor. Someslow righting in a free fall had now also appeared.On the thirtieth day both cuneate nuclei weredestroyed by direct vision. In the following three daysno change could be detected, and in particular thebody contact righting and hand contact graspingwere unchanged.

This experiment demonstrated the extra-ordinary decerebrate posture related to loss ofanterior vermis if compensation for loss ofcorticospinal function had been first allowed tooccur. As head righting returned, the extensorrigidity of the neck disappeared; as the positivesupporting and contact traction reactions re-turned, decerebrate rigidity disappeared; and asthe lateral tilt reactions returned, adductorrigidity disappeared. The decerebrate state inthis animal is the alpha rigidity described byEldred et al. (1953), long known as the 'aniemic'decerebration of Pollock & Davis (1930).

It is clear that decerebrate rigidity, whether ofthe classical intercollicular type, with hyper-activity of gamma neurons, and consequentlyexaggerated stretch reflexes, or of the cerebellartype with enhanced alpha discharge of laby-rinthine origin, is essentially an unstable condi-tion. Other reactions soon reappear to modifythe extensor rigidity, and in each case thepositive supporting reaction and diagonal progres-sion are the first modifications, soon to befollowed by a negative traction-flexion reaction.At the same time the rigidity becomes plastic,

531

532 Proceedings ofthe Royal Society ofMedicine

suggesting both gamma and alpha activity. Inboth states diffuse contact effects ultimatelyreappear in the form of fragments of body-on-body righting reactions, with local contactreactions in the hand and foot. I therefore findno fundamental difference between the behaviourof the so-called gamma type of decerebrate andthe alpha type, and our problem resolves itselfinto that of the nature of the general reaction thatis released, and the structure that is responsiblefor it.

The Essential StructureMagoun & Rhines (1947) showed that the reti-cular substance of the upper pontine tegmentumhas a predominantly excitatory effect on spinalneurons, chiefly but not exclusively on the gammasystem. This is exerted mainly through the crossedreticulospinal tract. The non-specific nature ofthe pathway is shown by our hemisection experi-ments. The mammalian reticular system isextremely complex, and our best informationcomes from the studies of Coghill (1929) andHerrick (1948) on the less elaborate organizationof salamanders. Coghill showed in these creaturesthat the pattern of progression by diagonal limbmovement appeared as a whole, long beforeisolated fragments of flexion of limbs to localstimulation were obtainable. It also appearsbefore reaction to stimulation in fishes and birds.

Detwiler (1946) found that in young larvalamblystoma swimming movements were per-fectly executed after transection below theauditory vesicle, controlled by the lower medullaand spinal cord. Subsequent to a later stageof development sustained motor activity, includ-ing swimming, was found to fail rapidly if theinfluence of the midbrain was eliminated. 'Themidbrain evidently supplies a factor essential formaintenance of motor efficiency' (Herrick 1948,p 63). Hooker (1944) found that in a comparablestage of development of the human foetus themechanical marionette-like performance changed

to the more graceful fluid movements seen in thenewborn. The comparative anatomists attributesuch co-ordinated behaviour to the effect of theneuropil, the mass of small cells with richlybranched processes that pervade the tegmentum.In this primitive network develop the patterns ofconnexions that reflect the inherited evolutionarymotor sequence of progression and other simplebehavioural patterns. A spinal reflex such as a

flexion reflex is a fragment of a larger context.'Such a reflex movement, taken by itself, isequivocal of purpose; it can enter as a fractioninto various reflexacts ofwholly differentpurpose.It is a fragment of an act, not an act' (Sherrington1940, p 211).The importance of anatomical determinants

of behavioural reflex pattern became apparentwhen the finer analysis of synaptic interaction inthe spinal reflexes by Sherrington revealed thatthe outcome was dependent upon the number ofsynaptic boutons provided for any given motorneuron by a particular pathway. The reflex couldbe graded, modified, inhibited, but the pattern oftransmission was pre-ordained by the anatomicalfactor of bouton density. It therefore should notbe difficult to proceed to the assumption that themost elementary nervous structure, universallypresent throughout the vertebrate series, thereticular formation of the brain-stem and spinalcord, should carry this inbuilt pattern of activity.Ludwig Edinger (1908) not only emphasized theimportance of the reticular system but pointedout its tectal mesencephalic extension and theelementary afferent and efferent pathways con-nected with it. The secondary sensory pathwayof Edinger is the spinothalamic tract ofmammals.Its tactile division, the ventral spinothalamictract, is by no means msigniflcanteveninprimates,and loses most of its fibres as it ascends throughpontine and mesencephalic tegmentum. It pro-vides a substantial contribution of afferentfibres to the colliculi as well as to the mesen-cephalic reticulum. The down-going tecto- andreticulo-spinal pathways are both crossed anduncrossed. OuZihemisection experiments demon-strate that tq control of spinal segments can beexerted by either side alone.Though the results of chronic survival experi-

ments do not substa the particular functionassigned to thered"- b Rademaker (1926),his acute expei demonstrated thatthere is a most im physiological transitionat the level of this"s e. In this transition thepyramidal tract, the substantia nigra, the bra-chium conjunctivwm, the superior longitudinalfasciculus and the- colliculi were specificallyexcluded by Rademaker's experimts. It is moredifficult to implicate any particular part of theall-pervading mesencephalic reticulum, the felt-work of nerve fibres that form the capsule of thered nucleus, and the neuropil of elementaryneurons that form the central grey. Fragments ofprogression and righting reflexes are present inthe spinat segments. Increasing extent of thepontine and mesencephalic tegmentum can besaid only to improve the regularity and initiationof these reactions, for which they provide thetotal pattern. The most critical change occurs inthe immediately precollicular reticulum.

The Effect ofPrecollicular SectionIn the clinic, as a result of displacements of thebrain at the opening of the tentorium, states ofdecerebrate rigidity of all four limbs occur inwhich the eyes turn upward, the head retracts

Sherrington Memorial Lecture

and all limbs are not only stiffly extended but alsointernally rotated, with strongly clasped hands.This condition does not occur as a result oftransection of the midbrain in animals. In manit is occasionally seen as a one-sided reaction,also related to tumour of the brain-stem as in thecase reported by Cathala (1922). The internalrotation and extension of the limbs on one side isthen associated with deviation of the head andeyes up and back to the opposite side. The originof this curious posture has been a mystery, thoughthe rotation of the limbs is of importance inrelation to better understanding of the rotarytremors of the limbs that occur in so-called'midbrain tremor' and hemiballismus. Hess(1949) found that if one interstitial nucleus isdestroyed the overaction of the other is such as tolead to a torsion of the head and eyes. In themonkey we have succeeded in producing thisdystonic torsion of the head and limbs only by alesion that lay between the interstitial nucleusand red nucleus (Fig 6A). When the condition isunilateral and severe the animal rolls over andover in a corkscrew spiral of progression (Fig6B). The reaction is most complex and I shall notdiscuss it further here except to say that the levelof disordered function in this type of decerebratestate is pretectal and not mesencephalic. In manthe bilateral decerebrate attitude of head retrac-tion with internal rotation of limbs arises becausethe pretectal region is displaced downwards by

Fig 6 The pretectal lesion in monkey BG 37 is shown inA, and in B the resulting tonic torsion ofright limbs andhead

large brain lesions into the tentorial opening,so that the level of impaction is actually justabove the midbrain. It is a disorder of labyrin-thine co-ordination of righting of head and eyeswith secondary limb effects.

Overaction of labyrinthine and body contactrighting of the head, in addition to decerebraterigidity, thus requires a level of section rostralto the red nucleus in the monkey, severing itsconnexion with the remainder of subthalamus.The tonic labyrinthine reactions mediated by thepontine reticulum are orientated to gravity,whereas the phasic and tonic pretectal responsesare orientated to the longitudinal axis of thebody. It has not yet been possible to separate thebody-on-head and labyrinth-on-head mechanismsat this level. Both control the distribution of bodyposture by connexions with the reticulum in theregion of the red nucleus. It is the destruction ofthis mechanism that results in the release of thegamma type rigidity that is characteristic ofintercollicular decerebration.

Sherrington found that the dorsal columnswere not of importance in decerebrate rigidity,though one might have expected this contributionof the proprioceptive system to play some veryessential part. He did not discuss this strikingdiscrepancy. It will be remembered, however,that Brouwer (1927) showed that the dorsalcolumns were a relatively late development inthe mammalia. They are important for the instinc-tive grasp reaction and other types of orientatedcortical contact reactions that affect the motorsystem at a higher level, which we have recentlydiscussed in relation to the function of- the basalganglia (Denny-Brown 1960).

Decerebrate rigidity is thus an unstabledisequilibrium of the integration of the simpleorganization of progression and righting resultingfrom the loss of all those factors that normallymodify them. The innate patterns of motororganization that are laid down in the pontineand mesencephalic reticulum are primarilyconcerned with the regulation of the spinalstretch reflex. This resistance to shock (the 'isola-tion dystrophy' of spinal segments) reflects afunctional independence that probably is derivedfrom their primal position in the whole neuralorganization. The most simple reactions ofprogression and righting that rapidly emergedfrom the rigid state share this autonomy. Such-reactions, however, require afferent mechanismsthat can discriminate adequacy of stimulus, forwhich we have given some remarkable examplesrelating to opening and closing of the hand orfoot. Limited lesions of the tectum did notabolish such afferent organization in the twochronic 'decerebrate' monkeys we have described,but it is likely that the afferent mechanism is'-as

5~33

534 Proceedings ofthe Royal SocietySfMedicine

extensive as is the reticular substance. It wasapparent that the complex motor automatismsthat eventuallyrecovered in thesecond (cerebellar)animal nevertheless failed to restore the highlyintegrated individual existing before the cerebellarand tectal lesion. What was missing was evidenceof reaction to any total environmental situation,and the normal 'drive' that activates behaviour.The animal was reduced to the status of anautomaton. Since cerebellar vermis lesions alonehave no such effect our attention was turned tothe function of the quadrigeminal plate.

The Mesencephalic TectumStimulation of the tectum opticum leads to a widevariety ofmovements, including movements oftheeyes and of the limbs. Dilation of the pupils, anddeviation of the eyes to the opposite side arecommon effects from stimulation of the superiorcolliculus, pricking of the ears and vocalizationfrom the inferior pair. In the monkey, Ferrier(1876) first described the wide opening of the eyeswith elevation of the eyebrows and turning of thehead and eyes away from the side of stimulationthat is characteristic of the full effect. Thisis possibly only a general reaction, and not adefinitive mechanism for these movements.Ablation of the corpora quadrigemina has beenreported many times since Flourens (1824), whoconsidered total blindness to be the sole result inbirds and rabbits. Bechterew (1883) and Munzer& Wiener (1898) and others found that blindnesswas incomplete. Many of the early experimentersreported forced movements and immobility ofthe pupils, but these were later thought to be dueto damage to neighbouring structures, the pos-terior longitudinal fasciculus and posteriorcommissure (Stefani 1881, Sgobbo 1900). Stefanidescribed the disturbance of vision in the pigeonas closely resembling that which follows removalof the cerebral hemispheres. The bird could flyand avoid obstacles, drop down and perch onobjects, but would not be able to find corn evenwhen hungry and would not attempt to escapewhen the observer reached out to pick it up.Sgobbo (1900) destroyed the colliculi in dogs anddescribed a similar amblyopia, present in theopposite visual field from a unilateral lesion.Removal of the inferior colliculus caused somedefect in hearing and also some visual impair-ment. Movements of the eyes and pupillaryreactions were unaffected. It was supposed thatthe absence of this type of visual disturbance inman is due to the increased importance of thegeniculocalcarine pathway. In the monkey andin man the colliculus nevertheless receives asizable complement of fibres from the lateralgeniculate body, and these are distinct from thosethat pass to the pretectal nucleus and posterior

commissure, thought to serve the pupillary lightreflex (Keller & Stewart 1932, Magoun 1935).Removal of the superior colliculus in the

monkey without damage to the calcarine cortex,radiation or pulvinar is a most difficult procedure.Pasik et al. (1961) reported such removal in fourmonkeys, and with removal of striate cortex intwo others. Optokinetic responses were stillpresent except in one animal lacking striate cortexas well as colliculi. Ocular movements and vesti-bular nystagmus were unaffected. They found noevidence that the colliculi were critical for anyvision remaining after removal of striate cortex.No amblyopia was reported in animals withintact geniculostriate system.We ourselves have accomplished removal of

75% to 100% of the superior colliculus in fivemacaques with survivals of 27 to 91 days. Theoperation was done from behind by gentledisplacement of the occipital pole laterally, andintroduction of a sucker by direct vision, oneside at a time. In two animals there was patchysuperficial damage to one occipital lobe on itsouter surface, but any degeneration in the lateralgeniculate bodies was difficult to find. In anotherthere was no such damage and in two microscopicsections are not yet available, though no grossdamage could be seen. The effect of these largelesions of the superior colliculus in the monkeyhas in our hands been most dramatic. In thedays immediately following unilateral removalthe animal behaved as if there was a completehemianopia on the side opposite the lesion. Hisgeneral behaviour was greatly altered, and he wasnow mute with totally expressionless facies. Allthe mimetic responses characteristic of the socialhabits of monkeys were absent. He showed nofear of the examiner. As some object moved intothe visual field on the side of the lesion he showeda violent start and moved away, usually circlinground to the side of the lesion. Within three tofive days he reacted to the appearance of theexaminer in the intact field of vision by retreatto the back of the cage, and by watching intentlyin the same general direction from which theexaminer had come. He seemed to lose theobject of his interest quickly, and would go onfor some minutes looking expectantly at theplace where the stimulus had entered his remain-ing visual field, sometimes making reachingmovements with one or other arm. He had nodifficulty in reaching out to a moving object witheither hand, and had little difficulty taking holdof it if it was within reach. Nevertheless the eyesdid not fix and looked only approximately in thedirection of the object. The ear opposite the sideof the lesion was no longer pricked to a suddensound. The pupils were always equal. There was aslight divergent strabismus, but ability to look

Sherrington Memorial Lecture

Fig 7 Monkey BG 25 following removal of posterior three-quartersof superior colliculus and oral half of inferior colliculus. The usualimmobile posture is shown in A, the presence of visual placing in B,reaching for a piece of white cotton on a wire moving silently acrossthe visual field in c, and in D the ultimate recovery of reachingfor a still object.up, down, or to either side was not impaired. Theoptokinetic response was present only when thestripes moved into the visual field from theoperated side.The removal of the remaining superior col-

liculus (and in one monkey where both sides were Fig 8 The tecremoved at one operation) resulted in a profound the animal shdisturbance of visual and general behaviour(Fig 7). In three animals (BG 25, 33, 36) a movingobject larger than 2 cm square could secure theattention of the animal, and he would reach to- Iwards the object if it came within 50cm ofhim. Hisreaching was approximately accurate, and wasnot associated with visual fixation (Fig 7c). Inone animal (BG 25) this reaction to moving ob- Ajects was present in both visual fields from theday after operation and was associated withaccurate visual placing in both fields (Fig 7B). C.

After five weeks this animal was able to locateand reach for a still object 2 cm X 3 cm for thefirst time (Fig 7D). This animal had the leastextensive removal of the superior colliculus,sparing its anterior one-quarter on the right side,and rather less on the left (Fig 8). The oralone-third of the posterior colliculus was removedin all animals. In three other animals vision formoving objects returned on one side after periods I-of one, three and 34 days respectively. This lastanimal (BG 32) appeared to be -completely Fig 9 Monkey BG 31, witunreactive to visual stimuli for 34 days, and after period of five weeks. Thothe fifth week reacted only to very large moving shown in A. In B the aninobjects in one visual field. feeling his way around a st

T down is shown the absenceThe last of the five animals (BG 31) remained of the observer's hand ancompletely unreactive to visual stimuli (except for when he is touched

ctal lesion in BG 25,Wown in Fig 7

th no visual reaction for aie convergent strabismus is,nal falls from a shelf whiletrange cage. In c from aboveof reaction to the approach

id the violent flight reaction:

535

536 Proceedings ofthe Royal Society ofMedicine

Fig 11 Photic driving in BG 31 with stroboscopicflashes at approximately 2 per second

Fig 10 The lesions ofthe animalBG 31 showing completeremoval of both superior colliculi, forward to wherethese lie lateral to the posterior commissure (uppersection). The lateral part of both posterior colliculi wasalso removed (lowest section). The central grey was leftintact

a blink to a bright sudden flash) for five weeks.He collided with walls and other objects, fell offedges of tables and shelves (Fig 9B) and wouldreact to an approaching stimulus only by aviolent start when it touched him (Fig 9c). Inthis animal there was found to have been aremarkably complete removal of the superiorcolliculi (Fig 10) including the extension underthe pulvinar on each side up to the level of theposterior commissure. In this animal there wasalso no reaction of the pupils to light. The eyeswere slightly convergent, but conjugate rovingmovements of moderate range in all directionsstill occurred in response to tactile stimulation.The electroencephalogram, recorded by electrodesimplanted between skull and dura five weeks afteroperation, and recorded three days later withoutanesthesia, showed normal rhythms. With photicdriving, spikes were recorded over both occipitalregions, particularly evident during sleep (Fig 11).In this animal, and in others in the period whenthere was no response to visual objects, there wasno visual placing. Tactile placing continued to beaccurate and rapid.

Even more extraordinary than the loss of visionfollowing bilateral ablation of the colliculi wasthe change in general behaviour of these monkeys.Whether some vision for movement and placingremained or not the animals appeared to betotally unaware of events in their environment.They would explore for food when hungry, andshowed the usual periods of sleep in a naturalposture. The remainder of the time was spent instaring aimlessly into space (Fig 7A). They tookno care of their fur or toilet. After an initial startwhen picked up they would sit on the observer'sarm or nestle in a corner and stare. If stimulatedthey would run in a straight line until theyencountered an obstacle, then would sit still indazed uncertainty. If suddenly disturbed theyshowed a pilomotor response but their burst ofactivity soon evaporated, with relapse intoapathy. Only once did one of these animals utterany sound; when he was suddenly grasped by anexaminer, after not being handled for a week, heuttered a sharp cry before submitting in silence.

If recovery of an optokinetic reaction occurredon one side, it was associated with circling to thatside whenever the animal walked. The circlingcould be stopped by surrounding the animal witha striped drum rotating in the same direction at afaster rate and could be started by rotation of thedrum in the opposite direction. If improvementcontinued, visual placing appeared several days totwo weeks after the return of optokinetic eye andhead movement, and later, vision for movingobjects. The reaction of the pupil to light wasreduced in amplitude in proportion to the loss ofvisual function, being lost when no visualfunction returned (BG 31). The lesions thennecessarily also damaged the closely relatedpretectal nucleus. The pupils remained large andequal in all animals. The dorsal grey matter ofthe aqueduct of Sylvius was removed in three ofthe five animals, but without damage to thenucleus of the III nerve, or to the ventral partof the central grey, or to the brachium conjunc-tivum. There was no difference in behaviour inthose two where the central grey matter remainedintact. The lesion involved the rostral quarteror third of the inferior colliculus in all animals.

Sherrington Memorial Lecture

The Pattern-setter of the Nervous SystemIt is extraordinary that the removal of a relativelysmall area of the neuraxis, the tectum of themidbrain, should inactivate the elaborate hemis-pheric organization for reaction to events in theexternal world. It is even more remarkable thatthis should occur in relation to a nucleus withoutdirect cortical representation. The evoked corticalpotential is still intact, with the structural ap-paratus thought to be essential to mammalianvision. The defect is in the reactivity of theorganism to the environment, evident in a grossreduction in all types of externally directedbehaviour. Thetectum appears to be the primarydriver of the mesencephalic reticulum of whichStarzl et al. (1951) demonstrated the cerebral'alerting' activity. The reaction is clearly alliedto the loss of 'facio-vocal' activity reported byKelly et al. (1946) in cats, from peri-aqueductaland tegmental lesions, but our own experimentsindicate that behaviour is defective in a moregeneral sense, related to the loss of the perceptualfacilitating apparatus of the tectum. Sprague etal. (1961), observing similar mutism and indif-ference following lateral tegmental lesions in thecat, have attributed these symptoms to sensorydeprivation due to damage to the lemniscus. Therewas no such damage in our animals. To theSherringtonian physiologist the concept of an'alerting' system without relation to some co-herent integrating mechanism has been difficultto comprehend. Our observations on the colliculiindicate that, just as the innate patterns of move-ment and righting are laid down in the structureof the elementary mesencephalic tegmentum, thecomparable templates of the most instinctiveinborn patterns for afferent perception are laiddown in the tectum.The lack of reaction of the decerebrate state

has always been regarded as amotorphenomenon,as if that explained all its aspects. The problemhas long been known to the neurologist in adifferent form, namely as 'akinetic mutism', or'coma vigil'. In other circumstances this reactionis associated with dystonia. The peculiarity ofcollicular lesions is the absence of motor defectexcept for visual fixation. Such a condition is rarein the clinic, except in relation to diffuse tumours.We therefore present an example, that of a patientwhose brain was sent to our laboratory someyears ago by Dr James Poppen, who kindlyallows me to cite the case:

A 17-year-old schoolgirl sustained a severe headinjury as a result of being thrown from a horse. Shewas found deeply unconscious with left pupil dilatedand fixed, right pupil miotic.The lower extremitieswere rigid in extension, the upper in flexion. A lumbarpuncture showed grossly bloody spinal fluid. X-raysof the skull revealed DO fracture. The coma continued,

Fig 12 The midbrain of the patient cited in the text

with intermittent opisthotonic spasms. A subduralclot was removed from over the left hemisphere.None was found on the right, though the brainappeared cedematous. The decorticate posture lessenedin severity and became intermittent, with aimlessrestless movements of the left arm and chewingmovements of the jaw in the intervals. Pneumo-encephalography showed only a general dilation of theventricular system. The spasms and involuntary move-ments gradually lessened in frequency and degree,and had entirely disappeared by the third month afterthe accident. Throughout her illness the patientshowed no expression and stared into space for hoursat a time. She showed no evidence of recognition ofpeople or of any event in her surroundings. She diedin this state eight months after the injury.Autopsy disclosed no significant cerebral atrophy

or deformity. Small superficial yellow traumaticexcavations 3 5 x 1 5 cm were found on two temporal-gyri on the right side, but were very shallow. Severalsmall yellow or grey lacunar lesions were found in thecentral white matter of all lobes of the brain, but therewas no disintegration of white matter. There was asmall lesion in the right habenular and pretectalnucleus, and the left cerebral peduncle was damagedwhere the tentorial edge had been once impactedagainst it. The left lateral tegmentum of the midbrainand the superior colliculus, chiefly on the left side,showed severe scarring and loss of substance (Fig 12).

This case demonstrates the importance of thecolliculi to perception in the human brain.Though such relatively isolated lesions are rarewe would draw attention to the common affectionof this region in tentorial herniation, and theprobable implication of parts of it in conditionssuch as peduncular hallucinosis and the moresevere manifestations of Wernicke's encephalo-pathy.

537

538 Proceedings ofthe Royal Society ofMedicine

ConclusionI hope the observations reported here will serveto illuminate the importance of the discovery ofdecerebrate rigidity. Attention has for too longbeen focused on the transient rigid component.The total phenomenon illuminates the vital partthe mechanism of the mesencephalic and tectalreticulum plays in all cerebral function. Thisprimitive organization of small neurons isevidently the vital facilitator of thalamo-cortical-function. Just as cortical activity is dependent onthalamic integrity, so both also require thecolliculi. Some surgeons have long been impressedby the importance of the peri-aqueductal greymatter to the general manifestations of con-sciousness (Bailey & Davis 1942). Skultety et al.(1954) showed that lesions of the peri-aqueductalgrey matter alone in cats did not disturb reactivity.The physiological evidence has been difficult toestablish, and our own observations show nodirect relationship to mechanisms for sleep.Consciousness is a term that implies a subjectiveelement which the physiologist tries to avoid.The mesencephalic tectum is, however, essentialfor the reactions we call general awareness, forwhich it has an initiator function, just as itsventral component, the mesencephalic reticularsubstance, is vital to organization of movement.Further, though there is reason to believe that thereactions concerned are at the primitive instinc-tive level, our experiments indicate that they arethe essential substrate for all the more highlydeveloped behavioural reactions.We hesitate to use the word 'primitive', but

perhaps I may quote from C J Herrick (1921)regarding the comparable region of the trulyprimitive tubular nervous system of cyclostomes.The grey matter in these animals borders thecentral canal and the processes of the cells radiateoutwards. He says: 'The nervous organization issuch as to make possible a relatively smallnumber of reactions for all sorts of sensorystimulations, and these reactions are for the mostpart simple total movements of the whole body. .. rather than complex adjustments involvingprecise co-ordination of many separate organs.'The pericanalicular grey matter of the mid-

brain has long been regarded as the essential areafor brain function in fishes but its role had beenthought to have been transferred to the hemis-pheres in all higher vertebrates (Monakow 1895).In the primate nervous system this is still the vitalcentre of the brain. Its more differentiatedperipheral layers, the reticulum and tectum, aremore essential than the immediately peri-aqueductal core. Though physiologists haveresisted Henri Bergson's notion of an elan vital as

being too imprecise for scientific proof, this smallarea deserves consideration as the most vital forunitary function of the organism. It is thephysiological 'ego', the mainspring of la vie enrelation.

REFERENCESBailey P & Davis E N

(1942) Proc. Soc. exp. Biol., A. y. 51, 305Bard P & Macht M B

(1958) In: Ciba Foundation Symposium on the NeurologicalBasis of Behavior. Eds. G E W Wolstenholme & C MO'Connor. Boston; p 55

Bazett H C & Penfield W G (1922) Brain 45, 185Bechterew W (1883) Pflfig. Arch. ges. Physiol. 31, 60Bergson H (1909) L'Evolution Cr6atrice. ParisBrouwer B (1927) Anatomical, Phylogenetical and Clinical Studieson the Central Nervous System. Baltimore

Cathala J (1922) Rev. Neurol. 38, 1504Coghill G E (1929) Arch. Neurol. Psychiat., Chicago 21, 989Denny-Brown D (1960) Lancet ii, 1099, 1155Denny-Brown D & Liddell E G T (1927) J. Physiol. 63, 144Detwiler S R (1946) J. exp. Zool. 100, 103Edinger L (1908) Vorlesungen uiber den Bau der nervosen Zent-

ralorgane des Menschen und der Thiere. 7th ed. Leipzig; Vol. 2Eldred E, Granit R & Merton P

(1953) J. Physiol. 122, 498Ferrier D (1876) The Functions of the Brain. LondonFlourens P (1824) Recherches Exp6rimentales sur les Propri6t6s

et les Fonctions du Syst6me Nerveux, dans les AnimauxVert6br6s. Paris

Herrick C J(1921) J. comp. Neurol. 32, 4(1948) The Brain of the Tiger Salamander. Chicago

Hess W R (1949) Das Zwischenhirn, Syndrome, Lokalisationen,Funktionen. Basel

Hooker D (1944) The Origin of Overt Behavior. Ann Arbor,Mich.

Keller A D & Stewart L (1932) Amer. J. Physiol. 101, 64Kelly A H, Beaton L E & Magoun H W

(1946) J. Neurophysiol. 9, 181Liddell E G T (1960) The Discovery of Reflexes. OxfordLiddell E G T & Sherrington C S

(1924) Proc. roy. Soc. ser. B 96, 212Lowenthal M & Horsley V

(1897) Proc. roy. Soc. 61, 20Magendie F (1823) J. Physiol. exp. path. 3, 376Magnus R (1924) Korperstellung. BerlinMagoun H W (1935) Amer. J. Physiol. 111, 91Magoun H W & Rhines R (1947) Spasticity. The Stretch-reflexand Extrapyramidal Systems. Springfield, Ill.

Mettler, F A (1944) J. comp. Neurol. 81, 339Monakow C von (1895) Arch. Psychiat. Nervenkr. 27, 386Moruzzi G & Pompeiano 0 (1957) Arch. ital. Biol. 95, 31Manzer E & Wiener H

(1898) Mschr. Psychiat. Neurol. 3, 379Pasik, T Pasik P & Bender M

(1961) Trans. Amer. neurol. Ass. 86, 229Pollock L J & Davis L (1930) J. comp. Neurol. 50, 377Rademaker G G J (1926) Die Bedeutung der roten Kerne und

des ubrigen Mittelhirns fair Muskeltonus, Korperstellung undLabyrinthreflexe. Berlin

Sgobbo F P (1900) Manicomio mod. Nocera 15, 153Sherrington C S

(1897a) Proc. roy. Soc. 60, 411(1897b) Lancet i, 373(1898) J. Physiol. 22, 319(1940) Man on His Nature. Cambridge

Skultety F M, Meyers R & Brown G(1954) Trans. Amer. neurol. Ass. 79, 121

Sprague J M & Chambers W W(1953) J. Neurophysiol. 16, 451

Sprague J M, Chambers W W & Stellar E(1961) Science 133, 165

Starzl T E, Taylor C A & Magoun H W(1951) J. Neurophyslol. 14, 479

Stefani (1881) Atti Accad. med.-chir. Ferrara, qusted by Sgobbo(1900)

Stella G, Zatti P & Sperti L'1955) Amer. J. Physiol. 181, 230