Projections from the lateral geniculate nucleus in the cat and monkey

J. Anat. (1967), 101, 4, pp. 677-692 677With 7 figuresPrinted in Great BritainProjections from the lateral geniculate nucleusin the cat and monkeyM. E. WILSON AND B. G. CRAGGM.R.C. Cerebral Functions Research Group, Department of Anatomy,University College LondonThe surprising claim of Glickstein, Miller & Smith (1964) that the lateral geniculatenucleus (LGN) in the cat emits a crossed projection through the corpus callosum tothe contra-lateral visual cortex has important implications for work on split-brainedcats (Myers, 1956) and monkeys (Downer, 1959). This claim was based on lesionsmade by a particular stereotaxic approach to the LGN, and we thought it worthwhileto see whether a different stereotaxic approach, which avoided the corpus callosum,would give the same result.We found no evidence of a crossed geniculo-striate projection in either cat ormonkey, but in the course of this work made the unexpected finding that the LGNin the cat projects to three separate ipsi-lateral areas of neo-cortex. These comprisevisual areas I and II and the lateral part of the middle suprasylvian gyrus. Visualareas I and II have been shown to correspond to areas 17 and 18 of Otsuka & Hassler(1962), by Bilge, Seneviratne & Whitteridge (1963) and Hubel & Wiesel (1965). Wehave not found these cytoarchitectonic boundaries sufficiently well defined to beuseful for localizing the degenerated projections. Various control lesions have beenmade to determine whether the projections to visual area IL and the suprasylviangyrus arise in the LGN or in neighbouring structures.MATERIAL AND METHODSTwenty-three cats and two monkeys were anaesthetized with pentobarbital sodium,and given penicillin. Brain lesions were made under sterile conditions with a coagulatingcurrent passed through a needle electrode held in a stereotaxic machine. In thecats, a burr hole was made in the cranium just anterior to the lateral edge of thetentorium. The needle was then advanced through the postero-lateral neo-cortex andhippocampus to a known co-ordinate in the diencephalon, on a track that was nearlyhorizontal. In the monkeys, the head holder was altered to allow the neck to beflexed so that Reid's baseline was vertical. The electrode was then lowered througha burr hole in the occipital bone to pass through the cerebellum, entorhinal cortexand hippocampus and so reach the LGN from the ventral side. After survivalperiods of 1-3 weeks, the brains were perfused and subsequently stained by themethod of Nauta & Gygax (1954), and a mounted series of sections stained withcresyl violet.A multi-compartmented sieve was designed to process the sections for the Nautamethod. This had the advantage that the loose frozen sections were kept in strictserial order and were not damaged by tedious individual handling. The sieve was

678M. E. WILSON AND B. G. CRAGGmade by drilling twenty-two holes each i in in diameter through a disc of Perspex1 in thick and 51 in in diameter. A handle was attached so that the horizontal disccould be moved from one polyethylene dish to another. The lower face of the discwas softened by soaking in acetone, and then pressed flat on a sheet of nylon nettingtill dry. The Perspex that embedded the nylon netting was protected by a coat ofAraldite or DePex (Gurr). Successive sections were placed in the twenty-two compartmentswhile the sieve was in I in of distilled water, and this batch of sectionswas then processed as a unit by moving the sieve from one dish to the next. Thedishes were filled to a depth of about I in with an appropriate solution (i.e. about100 ml), and were agitated throughout the process by a mechanical rocker. A singlesection was taken from the sieve at the end of the treatment in 15 %0 AgNO3 andprocessed in the ammoniacal-silver bath. The latter was adjusted if necessary to givethe required result, as explained previously (Cragg, 1965), and the rest of the batchwere then stained together. After reduction, the batch was treated with two baths ofsodium thiosulphate, washed and dehydrated still in the sieve. The sections werethen taken individually into creosote beechwood oil and mounted on slides. Thesieve is now available commercially from Electrophysiological Instruments Ltd., at21 Marshall Street, Edinburgh, Scotland.RESULTSProjections from the lateral geniculate nucleus in the monkeyIn the first monkey, there was a thin-needle track through the cerebellum, entorhinalcortex, subiculum and fimbria, ending in a lesion 1-5 mm in diameter whichaffected the postero-dorsal pole of the LGN and the inferior part of the pulvinarnucleus. The occipital lobes were studied in parasagittal section. In the cortex therewas dense fibre degeneration in the antero-ventro-lateral part of the striate area thatabuts the lunate sulcus and lies immediately dorsal to the inferior occipital sulcus(i.e. in the cortical representation of the macula). This patch of degeneration wasentirely confined to the cortex containing the stria of Gennari, and did not extendforwards into the peristriate cortex. The cortical degeneration was ipsi-lateral to thelesion, all parts of the contra-lateral striate and peristriate cortex being clear ofdegeneration.In the second monkey, there was a similar needle track but the lesion was fartherforwards in the middle of the antero-posterior extent of the LGN. The lesion was1 mm in diameter and entirely confined to the lateral limb of the LGN. In the occipitallobes, which were cut frontally, there was dense fibre degeneration in the ventralwall of the calcarine sulcus on the medial side of the hemisphere at the posterior endof the brain (i.e. the degeneration was in the representation of the peripheral visualfield). The degeneration was again confined to the striate area, and none was seen inthe peristriate cortex, or the occipital lobe contra-lateral to the lesion. The degenerationwithin the striate area in these two monkey brains was dense in the deeper layersand became gradually less dense superficially. Most of the fine preterminal fibreswere seen in layer 4, some reached layer 3, but there were none in layer 2.

Projections from lateral geniculate nucleus679Projections from the lateral geniculate nucleus in the catThe cat brains were cut into coronal blocks by cuts 8 and 23 mm in front of theposterior limit of the occipital lobe. The posterior block contained the point of entryof the needle electrode, the middle block had most of the cortical-fibre degeneration,and the anterior block was free from degeneration in the two brains in which it wasstudied. The blocks were reassembled and photographed before sectioning, and theprints used to reconstruct the distribution of the cortical fibre degeneration. Sincein the cat the needle electrode entered the neo-cortex and passed through the whiteA B tA>iit; A.t. i s > } We X ;4P 4 P 2 P 0 ~ .41 ~ ~4OF~~~~~~~~(Fr3 Fr5 Fr6Fig. 1. A The position of a needle track aimed at the LGN showing the structures traversedat six levels in coronal sections at the indicated distances in mm anterior (Fr) or posterior (P)to the interaural plane. B The size of a typical needle track in a Nauta preparation. Degeneratedfibres can be seen under higher magnification. Scale = 100jtm.matter before reaching the diencephalon, it was important to know what corticalfibredegeneration would be produced by the needle track alone if no thalamic lesionwere made. Figure 1 shows a typical track through the brain. It should be noted thatthe first thalamic structure encountered by the needle is the posterior end of the LGN.Control lesionsIn two cats (C 1, C 2) the needle was inserted just short of the LGN and no lesionmade, while in three other cats (C3-C 5) the needle was in the third ventricle abovethe thalamus when the current was passed. In the posterior block (as defined above)of these five brains degenerated fibres were found spreading from the needle trackto the inferior half or three-quarters of the postero-lateral gyrus on both its lateraland medial surfaces, and to the posterior suprasylvian gyrus. In C 1 there was a

680M. E. WILSON AND B. G. CRAGGmoderate amount of fibre degeneration in the crown only of the middle suprasylviangyrus extending forwards into the middle block. The lateral gyrus was however clearof degeneration in the middle block, and in the other four brains (C2-C 5) the wholeof the middle block was clear of degeneration apart from an occasional degeneratedfibre. No reason has been found for the degeneration in C 1 being more extensive thanin C2-C5.In two other cats (C 10, C 16) the needle track ended in a thalamic lesion (describedin a later section), but there was no fibre degeneration in the cortex of the lateral ormiddle suprasylvian gyri more than 12 mm in front of the occipital pole. In additionLGLS.PLGFig. 2. A composite representation of the distribution of cortical-fibre degeneration in thecontrol brains C 2-C 7. The dotted lines show where the brains were cut into posterior, middleand anterior blocks. The needle electrode entered the cortex within the region of the black circle.ESG, ectosylvian gyrus; LG, lateral gyrus; LS, lateral sulcus; PLG, postero-lateral gyrus;SSG, suprasylvian gyrus; SSS, suprasylvian sulcus.to the inadvertent damage done by the needle track in the white matter, the cortexat the point of entry was sometimes damaged in the superficial layers. In two furthercats (C 6, C 7) the relevant superficial layers of cortex were damaged deliberatelywithout inserting a needle. One of these small lesions was confined to the posterolateralgyrus, and the other to the posterior suprasylvian gyrus, but in both brainsthe only degenerated fibres found were localized to the immediate vicinity of thelesion.Most of the information to be presented in the rest of this paper was derived fromthe study of the cortex in the middle block, 8-23 mm anterior to the occipital pole.The results in the control brains may be summarized (Fig. 2) by saying that thelateral gyrus from 12 mm ahead of the occipital pole was almost entirely free fromdegeneration in all nine of the control lesions, while in all seven of the control lesionswithout thalamic involvement the lateral gyrus was free throughout the middleblock. The suprasylvian gyrus was similarly free from degeneration except in onebrain (C 1), in which the crown but not the walls of the gyrus contained degeneration.The posterior block however contained degeneration caused by the needle trackwhich must be allowed for in drawing conclusions from the results of thalamic lesions.

Projections from lateral geniculate nucleus681Fr5O Fr65 Fr80 Fr9O0Fig 3. The distribution of cortical-fibre degeneration following a lesion in the LGN in C 8. Thelesion is shown at four levels of the atlas of the cat's brain published by Jasper & Ajmone-Marsan(1955). In the cortical-diagram note the wide separation of the two medial projections at theposterior end of the middle block and their convergence at the representation of the verticalmeridian anteriorly. The suprasylvian degeneration in the wall of the gyrus is only seen in thecross-sections. SS, splenial sulcus.

682M. E. WILSON AND B. G. CRAGGLGN lesionsThree cats (C 8-C 10) had thalamic damage entirely confined to the LGN, while inothers there was additional involvement of medially placed structures. The threelesions to be described were small enough to produce an interesting topographicaldistribution of degeneration in the visual cortex. To understand this it is necessaryto remember that the vertical meridian of the visual field is represented on the medialedge of the LGN (Bishop, Kozak, Levick & Vakkur, 1962; Seneviratne & Whitteridge,1962). On the cortex, the vertical meridian is represented at the boundary between theareas designated as Visual I and Visual II (Talbot & Marshall, 1941 ; Bilge, Seneviratne& Whitteridge, 1963). Hubel & Weisel (1965) and Whitteridge (1966) showed thatthis boundary coincided with that between the area striata (17) and the area occipitalis(18) as defined by Otsuka & Hassler (1962). Furthermore, the anterior part ofthe LGN projects to the anterior visual cortex, and the posterior part of the LGNposteriorly (Minkowski, 1913).In the first cat (C 8), the lesion extends from postero-lateral to antero-medial LGN(Fig. 3), and thus affects the representation of the peripheral visual field at the posteriorend of the LGN, and the field near the vertical meridian at the anterior end ofthe LGN. Since the lesion extends to the antero-medial tip of the LGN, the mostanterior degeneration seen on the cortex (Fig. 3) corresponds to the anterior limitof the visual area. In the cortex, two areas of degeneration are found, one on themedial wall of the hemisphere in area 17 or Visual I, and the other on the top of thelateral gyrus in area 18 or Visual II. In the posterior part of the middle block, adegeneration-free zone intervenes between these two areas of degeneration whichare widely separated, as are the representations of the peripheral field (away fromthe vertical meridian) in Visual I and Visual II. Anteriorly, the two areas of degenerationcome together at the representation of the vertical meridian on the lateral side ofVisual I and the adjoining medial side of Visual II (see Fig. 3). It will be argued belowthat this topographical correspondence between the distribution of the cortical-fibredegeneration and the position of the lesion in the LGN is a strong reason for thinkingthat the cells of the LGN project to the two areas Visual I and Visual II, and that theresult is not due to damage to fibres of other origin which happen to be passing nearthe lesion in the LGN.In C8 no degeneration was found on the medial lip of the lateral sulcus correspondingto area praeoccipitalis (19), the Visual III of the electrophysiologists.However, a sparse but definite projection was found to the lateral wall of the middlesuprasylvian gyrus, a region that was clear of degeneration in all the control brains.In the three areas of degeneration described above, the broken fibres were of similarmedium calibre, and were dense in the deeper layers of the cortex, but scarcelyextended above layer 4.In the second cat, C9, the lesion was on the lateral edge of the LGN, and passedforwards to the anterior extremity of the nucleus, with some possibility of damageto the reticular nucleus just beyond, and to one small bundle of fibres from thenucleus lateralis posterior. This then was a lesion in the representation of the farperipheral visual field. In the striate cortex the fibre degeneration was ventro-medialnear the splenial sulcus (Fig. 4). A second patch of degeneration appeared on the

Projections from lateral geniculate nucleus683lateral half of the lateral gyrus. Since Visual IL and Visual III abut in this region,it is difficult to say whether the degeneration observed contained a separate projectionto Visual III in this brain. Definite fibre degeneration was again seen in thelateral wall of the middle suprasylvian gyrus, and scanty degeneration also occurredin the adjoining medial wall of the ectosylvian gyrus. The last finding and the possibilityof damage to the reticular nucleus and nucleus lateralis posterior in this brainare the only respects in which these results differ from the pattern set by C8.P1 Fr8 Fr17ffAi~~~~~~~APu Pu LPLFig. 4D The distribution LP~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~P~of cortical-fib degeneration inIn theGtird at,10,theesioL a that three levels LPCL MdvPL~~~~M'~GM "CTONRPed~~~~~~~~~~~~~~~~~~~~~~~~~~~eFr5O0 Fr6-5 Fr8O0produced N deeerto \NR7/.ind

684M. E. WILSON AND B. G. CRAGGlimited value because the posterior placement of the lesion caused the corticaldegeneration to occur in the posterior block. The surprising finding that the LGNappears to project to Visual II as well as to Visual I makes it desirable to determinewhether these degenerated projections could have arisen from structures borderingthe LGN. Examination of Nissl-stained sections of the thalamus did not reveal anyretrograde reaction in the nuclei medial to the LGN in any of these experiments.It nevertheless seemed desirable to place lesions in these more medial structures tosee whether the resultant pattern of degeneration would fit that seen after lesions ofthe LGN.Lesions of structures medial to the lateral geniculate nucleusOne cat (C 11) had a lesion that was centrally placed at the posterior end of theLGN, but that moved medially farther forwards to involve the medial interlaminarnucleus of the LGN. In the cortex degenerated fibres were distributed as an almostcontinuous band in the dorso-medial part of the lateral gyrus, as would be expectedof a lesion close to the representation of the vertical meridian. However, degeneratedfibres were also present in the lateral wall of the lateral gyrus, in a region that maybe either area 18 or 19. Some degeneration was again present in the lateral wall ofthe suprasylvian gyrus in the middle part of its antero-posterior extent. These resultsare similar to those obtained in C8 except for the added degeneration in the lateralwall of the lateral gyrus, and the added involvement of the medial interlaminarnucleus in the lesion.In cat C 12 a lesion was made within the posterior half of the pulvinar nucleus withinvolvement of the medial interlaminar nucleus of the LGN. No retrograde cellreaction was seen in the LGN in Nissl-stained preparations. No fibre degenerationoccurred in the dorsal or medial parts of the lateral gyrus (area 17), but degenerationin the lateral half of the lateral gyrus extended down into the depths of the lateralsulcus. In the suprasylvian gyrus, degeneration was moderate on the crown, butbecame dense in the lateral wall. There was also moderate degeneration in the medialwall of the ectosylvian gyrus, and a small patch at the bottom of the splenial sulcus.The last-named region had been free of degeneration in the brains described above,with the possible exception of C 11: no degeneration was seen in this brain, but themedial interlaminar nucleus was damaged anteriorly, and there was a gap in theseries of sections of this brain at an anterior level.In C 13 a small lesion damaged the medial interlaminar nucleus of the LGN andpart of the N. posterior and N. lateralis posterior. Cortical fibre degeneration wasfound in the lateral two-thirds of the lateral gyrus extending down the lateral wallof the gyrus, in the lateral half of the middle suprasylvian gyrus again extendingdown the wall, and in a small patch at the bottom of the splenial sulcus. Scantscattered degeneration occurred elsewhere in the medial wall of the hemisphere(area 17) and at the tip of the medial wall of the ectosylvian gyrus. The degenerationin area 17 was probably due to the passage of the needle track through the centralpart of the LGN. Degeneration found on the posterior ectosylvian gyrus is probablydue to track damage in the white matter.In C 14 the lesion penetrated the medial interlaminar nucleus and extended as farforwards as the N. ventralis anterior, causing also some damage in the pulvinar

Projections from lateral geniculate nucleus 685nucleus (see Fig. 5). The pattern of degeneration was similar to that in C 13, exceptthat degeneration in the medial wall of the hemisphere was confined to a small patchat the bottom of the splenial sulcus. These four brains (C 11-C 14) had in commondamage to the medial interlaminar nucleus of the LGN, and all possessed a de-Pul . >LYLP4%. .\ MD /7'\-~~~~~ ~~~~LP L','' 1/ DL , Hb /'GL, .'~3 ( 2/MD'hP'a0TNGM9 t _ *NCP -a . GI.to' GM V%R.4-:~~~~ AL4Fig. SnN siSN SN1 ~ N :- j~~~~~~~~~~~~~~~~~~~~~~~~~~4 .~~~~~~~~~~~~~~~~~~~~~~~~~~~~generatedproecto to th lateral \atoAh aea.gr hc a otpeeeaeK'3,Q..5.the Figlesion affecing (C8C1)wthemda intrelamionarncesof the LGN.Fr5-0 Fr6-5 -0 ~~~~~~~~~~~~~Fr6-.4~-inrCe14rshown(Cdiagramalsoshowed degeneration in a small region at the bottom of the splenial sulcus. Thislatter region and the lateral part of the lateral gyrus have in common that they may

686M. E. WILSON AND B. G. CRAGGcontain part of area 18 or 19 or both. It is, however, difficult to say whether theregions receiving the degenerated projection should be designated as area 18 or 19(see Discussion).In another cat, C 15, the lesion damaged mainly the pulvinar nucleus, with someinvolvement of the N. lateralis posterior and N. lateralis dorsalis, but did not touchthe LGN or its medial interlaminar nucleus. The cortical degeneration was confinedto the crown and lateral wall of the suprasylvian gyrus in the middle part of itsantero-lateral extent. The lateral gyrus was entirely free from degeneration (seeFig. 6).A lesion that extended most of the length of the N. lateralis posterior was made inanother cat (C 16) without involvement of other nuclei. No retrograde reaction wasseen in Nissl preparations in the LGN or pulvinar nucleus. Fibre degeneration wasfound only on the posterior ectosylvian and posterior suprasylvian gyri, and wasprobably due to track damage. Waller & Barris (1937) suggested that N. lateralisposterior projects to the anterior end of the suprasylvian gyrus. Unfortunately, thefrontal block was not stained. The results in C 16 do, however, exclude a projectionto the lateral gyrus from this source.Combined lesions of the LGN and more medial structuresIn the course of making the lesions described above, seven other brains (C 17-C23)were studied in which a lesion in the LGN had spread to involve more medial structuresincluding the medial interlaminar nucleus and the pulvinar nucleus. In onebrain (C 17) the lesion was just beneath the LGN, damaging the latter medially andinvolving the medial interlaminar nucleus, and part of the posterior nucleus, N.lateralis posterior and perhaps part of the pulvinar nucleus. Dense degeneration wasdistributed laterally across the lateral gyrus from its medial edge, but did not descendfar into the lateral wall of the gyrus. There was also some slight degeneration on themedial wall of the hemisphere especially at the bottom of the splenial sulcus, and inthe middle suprasylvian gyrus. In the other six cats the lesion involved the LGNdorso-medially, as well as the medial interlaminar nucleus and other medial structures.Two of these lesions were so far posterior that the degeneration in the cortexwas confined to the posterior block where it was mixed with degeneration due to theneedle track through the white matter. In the other four brains (C 18-C21) there wasdegeneration in the middle suprasylvian gyrus and in the lateral part of the lateralgyrus, and at the bottom of the splenial sulcus. Besides this constant finding, therewas degeneration in the crown of parts of the lateral gyrus corresponding to thevariable antero-posterior placement of the lesion in the LGN.Contra-lateral degenerationAll the fibre degeneration described above was ipsi-lateral with the lesion, and nocontra-lateral degeneration in the lateral gyrus was seen in any of the eleven brainswith lesions of the LGN (C 8-11, 17-23). The lesions confined to more medial structuresdid not cause contra-lateral degeneration either. The needle track was fine andavoided the corpus callosum (Fig. 1). The range of survival times covered the periodof 1-3 weeks specified by Glickstein, Miller & Smith (1964).

El ~ ~0JO-Projections from lateral geniciilate nucleus 687LLE~~~~~~~~~~oe~~~~~~~~~~~~~.oLiC)/- 0.CL~ ~ ~ ~ ~ ~ ~ ~ ~ ~C)~~~~~~~~~~~~~~~~~~~~~~~~~~~EC~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~LLLL~~~~~~~~~~~~~~~43Anat. IoI

688M. E. WILSON AND B. G. CRAGGDISCUSSIONIn the monkey, our first lesion destroyed the postero-dorsal pole of the LGN, andproduced cortical fibre degeneration in the antero-ventro-lateral corner of the striatecortex in the angle between the lunate and inferior occipital sulci. This result isconsistent with previous work, for Clark & Penman (1934) found that a lesion in themacular part of the retina produced transynaptic neuronal atrophy in the posterodorsalpole of the LGN, and Talbot & Marshall (1941) recorded electrical responsesin the antero-ventro-lateral part of the striate cortex to photic stimulation of themacula. Degenerated fibres in the striate cortex after lesions in the LGN were notconcentrated in the stria of Gennari, and this is consistent with the preservation ofthe latter after undercutting the striate cortex, as found by Clark & Sunderland(1939). In both our monkeys with LGN lesions, fibre degeneration in the striate areaextended to the boundary of areas 17 and 18. Myers (1965) showed that this part ofthe striate area is connected to the adjacent part of area 18. The abrupt cessation ofthe degeneration at the boundary of area 17 with area 18 was thus convincingevidence that the LGN in the monkey does not project into the appropriate part ofarea 18. Fibre degeneration was, moreover, absent from the rest of areas 18 and 19.Our results in the cat show that the striate area receives a topographically organizedprojection from the LGN, e.g. medial lesions in the latter produce fibre degenerationin the lateral edge of area 17, while a lateral lesion (C9) caused fibre degenerationdeep in the medial wall of the lateral gyrus (see Fig. 7). The rostro-caudal arrangementis consistent with the results of earlier workers who studied the location of retrogradedegeneration in the LGN after making lesions in parts of the lateral gyrus (Minkowski,1913; Waller & Barris, 1937). The mapping of the visual field by electrophysiologicalmethods in the LGN (Seneviratne & Whitteridge, 1962; Bishop, Kozak, Levick &Vakkur, 1962) and on the visual cortex (Talbot & Marshall 1941; Bilge, Seneviratne &Whitteridge, 1963) again implies the same topographical relationship between theLGN and the visual cortex.Our finding in the cat of a second projection from the LGN that supplies afferentsto area 18 was unexpected, although this area was known to contain a second topographicalrepresentation of the visual field (Visual II) as shown by Talbot (1942)and Bilge, Seneviratne & Whitteridge (1963). Indeed Talbot (1942) suggested thatVisual IIreceived an independent projection from the LGN, for the responses therewere not 'depressed by narcosis or cautery of the lateral gyrus', and were 'independentof convulsants applied at the medial locus'. Moreover, Doty (1958) was able torecord photic responses from the lateral half of the lateral gyrus after the moremedial area had been removed. However, removal of the middle of the lateral gyrus(containing VisualII) produced only a small area of retrograde degeneration in themedial edge of the LGN. This degeneration was attributed to the lesion extendinginto the optic radiation running to the striate cortex (Doty, 1958).There is a possibility that the projection to Visual II degenerates after lesions in theLGN because of damage to fibres of extraneous origin which might pass through theLGN. There are however, two facts that make this explanation improbable: first,the disposition of the degeneration in Visual II is topographically organized in amanner complementary to that in Visual I (see Fig. 7), being thus compatible with

Projections from lateral geniculate nucleus689the mapping of the visual field in Visual II (Whitteridge, 1966). Fibres arising in moremedial structures and passing through the LGN towards the internal capsule wouldbe interrupted equally by medial or lateral lesions in the LGN, and would not thereforebe expected to show a topographical distribution related in the correct mannerto the position of the lesion in the LGN. Secondly, lesions have been made in thestructures surrounding the LGN medially, and the resulting cortical degenerationin the lateral part of the lateral gyrus only slightly overlapped that found in Visual IIafter LGN lesions, and never extended to the medial boundary of Visual II wherecentral vision is represented and where degeneration was found after lesions in themedial part of the LGN.These two arguments do not apply to the cortical degeneration seen in the suprasylviangyrus after LGN lesions, for this was also present after lesions in more medialstructures and did not show any marked topographical distribution. There is, however,independent electrophysiological evidence of a direct projection from the LGNto the suprasylvian gyrus (Vastola, 1961) which has been corroborated by the evidenceof Thompson, Smith & Bliss (1963). Buser, Borenstein & Bruner (1959) have shownthat a pathway via the thalamus medial to the LGN may be important for photicallyevokedresponses on the crown of the middle suprasylvian gyrus. They did not showthat this pathway is essential, and in any case they may have been dealing with anarea of the suprasylvian gyrus more medial than that implicated in our experimentswith the LGN lesions. Bruner (1965) has reduced but not abolished the photicallyevokedresponse in the lateral part of the middle suprasylvian gyrus by injectingpotassium chloride into the thalamus medial to the LGN. No studies of retrogradedegeneration in which only the middle suprasylvian gyrus was removed have beenreported. Marshall, Talbot & Ades (1943) in one cat removed the lateral gyrus onone side and, after allowing time for retrograde degeneration to occur, were unableto obtain the short latency response from the middle suprasylvian gyrus. It is unfortunatethat no histology was reported, as this result may imply that the middlesuprasylvian gyrus is supplied by collaterals of fibres to the lateral gyrus and thatthese collaterals are insufficient to sustain the parent-cell body when the main branchto the lateral gyrus is destroyed.In addition to these three projections from the LGN discussed above, we founddegenerated fibres in the lateral half of the lateral gyrus after lesions of structuresmedial to the LGN (Fig. 7). This cortical degeneration seemed to be mainly in area 19,or Visual III of Whitteridge (1966) and Hubel & Weisel (1965), with perhaps someextension into the lateral part of area 18 or Visual II. When degeneration was presentin the lateral part of the lateral gyrus, it also appeared at the bottom of the splenialsulcus. Although this degeneration extended some way along the upper wall of thesulcus, it was still more deeply situated than that caused by the most lateral lesionin the LGN. By analogy with the situation on the dorsal surface it would appearthat this is a projection to either area 18 or 19 related to the extreme periphery of thevisual field. Cytoarchitectonic studies of this region are conflicting: Winkler & Potter(1914) and Gurewitsch & Chatschaturian (1928) described an area 18 here, butOtsuka & Hassler (1962) distinguish area 18 only on the dorsal surface of the brain,and describe the region in the depth of the splenial sulcus as area 19. However, it isdifficult to imagine a Visual II or Visual III (area 18 or 19) in the splenial sulcus in43-2

690M. E. WILSON AND B. G. CRAGGthe topographical scheme proposed by Whitteridge (1966). The alternative hypothesisthat the degeneration in the splenial sulcus lies in Visual I cannot be supportedeither for topographical reasons, even allowing that there may be no representationof central vision in the medial interlaminar nucleus (Stone & Hansen, 1966). A thirdpossibility is that the degeneration in the splenial sulcus is neither in Visual I, IInor III, but in some other cytoarchitectonic area, perhaps cingulate cortex.The exact origin of this projection from structures medial to the LGN is uncertain.The relevant lesions always involved the medial interlaminar nucleus, but also extendedinto adjacent parts of the pulvinar nucleus. One medially placed lesion (C 15)was restricted to the pulvinar, lateralis dorsalis and lateralis posterior nuclei, andcortical-fibre degeneration was confined to the middle suprasylvian gyrus (see Fig. 7).Garey (1965) has reported retrograde degeneration in the medial interlaminar nucleusfollowing cortical lesions placed lateral to area 17 of Otsuka & Hassler (1962). Theseresults suggest that the medial interlaminar nucleus may be the origin of the projectionto the lateral part of the lateral gyrus (Visual III), but we cannot exclude thepossibility that the most medial part of the pulvinar nucleus untouched by the lesionin C 15 might also project to Visual III. Whether the medial interlaminar nucleushas an independent projection to the suprasylvian gyrus could not be determined.The brief report by Garey (1965) also states that the LGN projects ipsi-laterally toareas 17-19, but it is not clear whether the medial interlaminar nucleus is includedin this statement.We found no evidence of a contra-lateral cortical projection from the LGN ineleven cats and two monkeys. The contrary result by Glickstein, Miller & Smith(1964) is perhaps attributable to their approach through the corpus callosumdamaging transcallosal fibres interconnecting the visual areas of cortex. As thesingle control lesion reported by these authors was ventral to the LGN, it is probablethat the track was anterior to that used in the other experiments, and might thushave missed the forward edge of the relevant transcallosal fibres.SUMMARY1. A stereotaxic approach to the thalamus was designed to avoid the corpuscallosum. An adaptation of the Nauta-Gygax method to batch-staining that preservedserial order and avoided the handling of individual sections is described.2. Stereotaxic lesions were placed in the LGN of two monkeys. Fibre degenerationwas confined to the ipsi-lateral striate cortex and ceased abruptly at the end of thestria of Gennari.3. Stereotaxic lesions were made wholly within the LGN in three cats. There weretwo areas of fibre degeneration within the lateral gyrus corresponding to Visual I andVisual II (areas 17 and 18) and one on the lateral wall of the middle suprasylviangyrus. The degeneration in the first two areas was topographically organized.4. Stereotaxic lesions were made in thalamic structures medial to the LGN inthirteen cats. The medial interlaminar nucleus of the LGN was damaged in elevencats, three of which had no damage in the main body of the LGN. Fibre degenerationwas found on the lateral edge of the lateral gyrus, probably in area 19 (Visual III)and also in the depth of the splenial sulcus. There was no indication that accidental

Projections from lateral geniculate nucleus 691damage to fibres arising medial to the LGN could account for the projection foundafter a lesion confined to the LGN.5. There was no evidence of a crossed geniculo-striate projection in eleven catswith lesions in the LGN.We are grateful to Miss M. J. Toyne for histological assistance.REFERENCESBILGE, M., SENEVIRATNE, K. N. & WHITTERIDGE, D. (1963). The primary visual receptive area of thecerebral cortex in the cat. J. Physiol., Lond. 169, 36 P.BISHOP, P. O., KOZAK, W., LEVICK, W. R. & VAKKUR, G. J. (1962). The determination of the projectionof the visual field on to the lateral geniculate nucleus in the cat. J. Physiol., Lond. 163, 503-539.BRUNER, J. (1965). Afferences visuelles non-primaire vers le cortex cerebral chez le chat. J. Physiol., Paris56, Suppl. 12, 1-120.BUSER, P., BORENSTEIN, P. & BRUNER, J. (1959). Etude de systemes 'associatifs' visuels et auditifs chez lechat anesthesia au chloralose. Electroenceph. clin. Neurophysiol. 11, 305-324.CLARK, W. E. LE GROS & PENMAN, G. G. (1934). The projection of the retina in the lateral geniculatebody. Proc. R. Soc. B, 114, 291-313.CLARK, W. E. LE GROS & SUNDERLAND, S. (1939). Structural changes in isolated visual cortex. J. Anat. 73,563-574.CRAGG, B. G. (1965). Afferent connexions of the allocortex. J. Anat. 99, 339-357.DOTY, R. W. (1958). Potentials evoked in cat cerebral cortex by diffuse and by punctiform photic stimuli.J. Neurophysiol. 21, 437-464.DOWNER, J. L. DE C. (1959). Changes in visually guided behaviour following mid-sagittal division ofoptic chiasm and corpus callosum in monkey (Macaca mulatta). Brain 82, 251-259.GAREY, L. J. (1965). Interrelationships of the visual cortex and superior colliculus in the cat. Nature, Lond.207, 1410-1411.GLICKSTEIN, M., MILLER, J. & SMITH, 0. A. (1964). Lateral geniculate nucleus and cerebral cortex.Evidence for a crossed pathway. Science, N.Y. 145, 159-161.GUREWITCH, M. & CHATSCSHATURIAN, A. (1928). Zur Cytoarchitektonik der Grosshimrinde der Feliden.Z. Anat. Entwgesch. 87, 100-138.HUBEL, D. & WIESEL, T. N. (1965). Receptive fields and functional architecture in two non-striate visualareas (18 and 19) of the cat. J, Neurophysiol. 28, 229-289.JASPER, H. H. & AJMONE-MARSAN, C. (1955). A Stereotaxic Atlas of the Diencephalon of the Cat. Ottawa:National Research Council of Canada.MARSHALL, W. H., TALBOT, S. A. & ADES, H. W. (1943). Cortical responses of the anaesthetized cat togross photic and electrical afferent stimulation. J. Neurophysiol. 6, 1-15.MINKOWSKI, M. (1913). Experimentelle Untersuchungen uber die Beziehungen der grosshirnrinde undder Netzhaut zu den primaren optischen zentren, Besonders zum corpus geniculatum externum. Arb.hirnanat. Inst. Zurich 7, 255-362.MYERS, R. E. (1956). Function of corpus callosum in interocular transfer. Brain 79, 358-363.MYERS, R. E. (1965). Organization of visual pathways. In Functions of the Corpus Callosum. Eds. E. G.Ettlinger, A. V. S. de Reuck and R. Porter, pp. 133-138. London: Churchill.NAUTA, W. J. H. & GYGAX, P. A. (1954). Silver impregnation of degenerating axons in the central nervoussystem: a modified technique. Stain Technol. 29, 91-93.OTSUKA, R. & HASSLER, R. (1962). Ober Aufbau und Gliederung de corticalen Sehsphare bei der Katze.Arch. Psychiat. NervKrankh. 203, 212-234.SENEVIRATNE, K. N. & WHITTERIDGE, D. (1962). Visual evoked responses in the lateral geniculate nucleus.Electroenceph. clin. Neurophysiol. 14, 785.STONE, J. & HANSEN, S. M. (1966). The projection of the cat's retina on the lateral geniculate nucleus.J. comp. Neurol. 126, 601-624.TALBOT, S. A. (1942). A lateral localization in cat's visual cortex. Fedn Proc. Fedn Am. Socs exp. Biol. 1,84.TALBOT, S. A. & MARSHALL, W. H. (1941). Physiological studies on neural mechanisms of visual localizationand discrimination. Am. J. Ophthal. 24, 1255-1263.THOMPSON, R. F., SMITH, H. E. & BLISS, D. (1963). Auditory, somatic sensory, and visual response interactionsand interrelations in association and primary cortical fields of the cat. J. Neurophysiol. 26,365-378.

692 M. E. WILSON AND B. G. CRAGGVASTOLA, E. F. (1961). A direct pathway from the lateral geniculate body to association cortex. J. Neurophysiol.24, 469-487.WALLER, W. H. & BARRIS, R. W. (1937). Relationships of thalamic nuclei to the cerebral cortex in the cat.J. comp. Neurol. 67, 317-339.WHITrERIDGE, D. (1966). Personal communication.WINKLER, C. & POTER, A. (1914). An Anatomical Guide to Experimental Researches on the Cat's Brain.Amsterdam: Versluys.