injected eve. (Received 1 November 1977) with electrolytic lesions. A good correspondence was found between the location of

Transcription

1 J. Physiol. (1978), 281, pp With 6 plates and 3 text-figures Printed in Great Britain OCULAR DOMINANCE IN LAYER IV OF THE CAT'S VISUAL CORTEX AND THE EFFECTS OF MONOCULAR DEPRIVATION By CARLA J. SHATZ* AND MICHAEL P. STRYKERt Fromn the Department of Neurobiology, Harvard Medical School, 25 Shattuck Street, Boston, J1assachusetts 02115, U.S.A. (Received 1 November 1977) SUMMARY 1. The relation between the physiological pattern of ocular dominance and the anatomical distribution of geniculocortical afferents serving each eye was studied in layer IV of the primary visual cortex of normal and monocularly deprived cats. 2. One eye was injected with radioactive label. After allowing sufficient time for transneuronal transport, micro-electrode recordings were made, and the geniculocortical afferents serving the injected eye were located autoradiographically. 3. In laver IV of normal cats, cells were clustered according to eye preference, and fewer cells were binocularly driven than in other layers. Points of transition between groups of cells dominated by one eye and those dominated by the other were marked with electrolytic lesions. A good correspondence was found between the location of cells dominated by the injected eye and the patches of radioactively labelled geniculocortical afferents. 4. Following prolonged early monocular deprivation, the patches of geniculocortical afferents in layer IV serving the deprived eye were smaller, and those serving the non-deprived eye larger, than normal. Again there was a coincidence between the patches of radioactively labelled afferents and the location of cells dominated by the injected eve. 5. The deprived eye was found to dominate a substantial fraction (22 %) of cortical cells in the fourth layer. In other cortical layers, only 7 % of the cells were dominated by the deprived eye. 6. These findings suggest that the thalamocortical projection is physically rearranged as a consequence of monocular deprivation, as has been demonstrated for layer IVc of the monkey's visual cortex (Hubel, Wiesel & LeVay, 1977). INTRODUCTION The development of the mammalian visual cortex can be altered by abnormal visual experience early in life. In the cat most cells of the visual cortex are driven by visual stimulation through either eye, but if one eye is closed during a sensitive * Present address: Department of Neurobiology, Stanford University Medical School, Stanford, California 94305, U.S.A. t Present address: Department of Physiology, University of California, San Francisco, California 94143, U.S.A.

2 268 C. J. SHATZ AND M.P. STRYKER period early in life, that deprived eye loses the ability to influence almost all of the cells in the cortex (Hubel & Wiesel, 1962; Wiesel & Hubel, 1963). In the monkey, this process has been studied anatomically as well as physiologically. Normally, the geniculocortical afferents representing the two eyes are segregated from each other into alternating bands in layer IV of the visual cortex where they terminate (Hubel & Wiesel, 1972; Wiesel, Hubel & Lam, 1974; LeVay, Hubel & Wiesel, 1975). These alternating bands form the anatomical basis of the ocular dominance columns - the physiologically defined grouping of cells according to eye preference - which extend through the full thickness of the cortex (Hubel & Wiesel, 1968; Kennedy, Des Rosiers, Sakurada, Shinohara, Reivich, Jehle & Sokoloff, 1976). After early monocular deprivation, the fourth layer bands representing the deprived eye are considerably diminished in width, while those of the open eye are correspondingly broadened (Hubel, et al. 1977). Exactly parallel changes are found in the pattern of activity from the two eyes seen in micro-electrode penetrations through layer IV (Hubel et al. 1977). In the other cortical layers the deprived eye is even less effective in driving cortical cells (Hubel etal. 1977). Monocular deprivation in the cat has been thought to affect the visual cortex similarly. Under certain conditions, however, the deprived eye has been reported to regain an influence on many cortical cells extremely rapidly (Duffy, Snodgrass, Burchfiel & Conway 1976; Kratz, Spear & Smith, 1976). These findings raise the possibility that inputs from the deprived eye are somehow suppressed (Cass, Sutton & Mark, 1973), rather than physically lost, as has been demonstrated for the monkey. We thought it worth while to examine these possibilities in the cat by studying the physiological and anatomical organization of the two eyes' inputs to layer IV of the visual cortex, first in normal animals and then in animals subjected to prolonged early monocular deprivation. A brief note of these findings has appeared previously (Stryker & Shatz, 1976). METHODS This study used three normal adult cats and six cats monocularly deprived by lid suture from a time prior to eye opening until 81 to 11 months of age, when recordings were made. With the exception of one of the normal cats, the vitreous humour of one eye in each cat was injected with a mixture of mci of 2,3-[3H](N)-proline (specific activity 37-3 Ci/m-mole) and mci of 6_[3H]fucose (specific activity Ci/m-mole) in 100 ul, 0 9 % saline. In cats in which the deprived eye was injected, it was unnecessary to open the eyelid; a small incision made lateral to the canthus was sufficient to allow insertion of a needle. Eleven to twenty-seven days following the eye injection, each animal was prepared for physiological recordings from area 17 of the visual cortex as described previously (Shatz, 1977). Anaesthesia was maintained during the experiment by i.v. infusion of thiopental sodium (2 mg. kg-. hr-1). Eye muscles were paralysed with continuous infusion of a mixture &f gallamine triethiodide (7 mg.kg-'.hr-1) and D-tubocurarine (0 7 mg.kg-1.hr-1) in 5% dextrose in Ringer solution (0'8 ml.kg-'.hr-'). Heart rate and C02 were monitored, and artificial respiration was adjusted to maintain end-tidal C02 between 4 and 4.5 %. Contact lenses of strength appropriate to focus the eyes on a tangent screen at 1-5 m were used to protect the corneas, and the positions of the area centralis and optic disk in each eye were plotted on the screen using a double-beam ophthalmoscope. Techniques of stimulating and recording were similar to those of Hubel & Wiesel (1962). To facilitate recordings from the cells of the fourth layer, care was taken in the preparation of the tungsten micro-electrodes to select only those which were extremely sharp and were insulated to within about jm of the tip.

3 MONOCULAR DEPRIVATION: PHYSIOLOGY-ANATOMY 269 A chief aim of this study was to assess physiologically the spatial organization of responses from the two eyes in layer IV of the primary visual cortex and to relate this to the distribution of thalamic afferents. We therefore chose to make long micro-electrode penetrations in a horizontal and almost parasagittal plane (roughly parallel to the border) in order to improve the chances of encountering many fourth layer cells while traversing several patches of radioactive label representing afferents serving the injected eye (see Shatz, Lindstrom & Wiesel, 1977). With the micro-electrode positioned in this manner, we studied the cortical representation within area 17 of the central 15 degrees of the visual field. An attempt was made to record from a unit every I 00 jm. Each unit encountered was assigned to an ocular dominance group and in most cases was classified as simple, complex or hypercomplex according to the criteria of Hubel & Wiesel (1962; 1965a). Some units were classified as 'lateral geniculate nucleus (LGN)-type' because of the presence of a mutually antagonistic centre-surround receptive field; these units frequently possessed a waveform typical of fibres (Hubel & Wiesel, 1961) (a fast, positive-going spike, frequently notched) and a pattern of spontaneous activity characteristic of LGN afferents. Numerous electrolytic lesions (2-3, A for 2-3 sec) were made along each electrode track in order to mark points of interest. At the conclusion of the experiment, animals were perfused through the heart with 0 9 % saline followed by 10% formol-saline. The visual cortex of each hemisphere was blocked in the plane of the electrode track (almost parasagittally) and sectioned at 30 /Sm. Alternate sections were prepared for cresyl violet staining and autoradiography. Those for autoradiography were coated with Kodak NTB-2 emulsion and exposed for 6-8 weeks before development in Kodak D-19 (Wiesel et al. 1974). Electrode tracks were reconstructed from the Cresyl Violet stained sections, and laminar and cytoarchitectonic boundaries were located according to the criteria of Otsuka & Hassler (1962) and LeVay & Gilbert (1976). Sections containing each track were photographed using brightfield illumination, and darkfield photographs were taken of the immediately adjacent autoradiographs. By superimposing the two sets of photographs, it was possible to mark precisely the position of the lesions on each relevant autoradiograph. The fraction of layer IV occupied by afferents serving the injected eye was estimated in one normal cat, in one cat with the deprived eye injected, and in one monocularly deprived cat in which the non-deprived eye was injected. Darkfield photographs were taken of autoradiographs similar to and including those shown in P1. 2, and high contrast prints were made on Agfa FP3 paper at a total magnification of x. It was verified from adjacent Nissl sections that the borders of layer IV could reliably be determined at this magnification. The outline of layer IV in each section was traced on clear acetate, as were the dense accumulations of label within layer IV. These regions were then cut out with scissors and weighed. Measurements for each hemisphere were based on tracings of 8-16 sections. Regions of layer IV in which the labelling pattern was especially unclear were not included. RESULTS Normal pattern of ocular dominance in layer IV It has long been known that layer IV of the cortex receives the bulk of specific sensory afferents, and modern studies have confirmed this for the two eyes' input to the visual cortex (Wilson & Cragg, 1967; Hubel & Wiesel, 1968, 1972; Garey & Powell, 1971; LeVay & Gilbert, 1976). When isotopically labelled amino acids are injected into the vitreous humour of one eye, some of the label is transported transneuronally to the visual cortex (Grafstein, 1971). The geniculocortical afferents representing the injected eye have been located in layer IV of the monkey's cortex using autoradiography (Wiesel et al. 1974). In the cat's visual cortex as well, this method has shown that the two eyes' inputs are partially segregated from each other (Shatz et al. 1977). We wished to verify that in the cat, as in the monkey, this segregation forms the anatomical basis of the physiological pattern of ocular dominance in layer IV. Two electrode penetrations in area 17 of a normal cat are illustrated in PI. 1.

4 270 C. J. SHATZ AND M. P. STRYKER These penetrations are from the hemisphere ipsilateral to the eye that had previously been injected with radioactive label. In the reconstruction of P1. 1C, cells dominated by the injected eye are represented by filled symbols; lesions made to mark sites along the course of the penetration are indicated by the numbered arrows. P1. 1 A is a montage of two Nissl-stained parasagittal sections which contain most of both penetrations. Such sections were used to make the reconstruction of P1. 1 C. The corresponding region of an adjacent section prepared autoradiographically to reveal the transneuronally transported label is shown in P1. 1 B. Darkfield illumination was used to photograph this and all subsequent autoradiographs, causing the regions containing many silver grains to appear as white patches. The positions of these patches are indicated in P1. 1 C with thin lines. Both penetrations initially passed through the fourth layer near the site marked by the upper open arrow in P1. 1 C and the identical site similarly marked in 1 B. Here the cells encountered in both penetrations were dominated by the injected eye. Here also was a patch of label from thalamic afferents representing the injected eye; both penetrations passed through this patch. Both penetrations then traversed layers V, VI and V again before re-entering and remaining within the lower part of layer IV (IVc). Penetration 1 again encountered cells in layer IV dominated by the injected eye immediately following lesion 5, which coincides with the border of another patch of label. The second penetration encountered a few cells dominated by the injected eye as the electrode re-entered layer IV, where it grazed the border of a patch of label (lower open arrow, P1. 1 B and C). These results and those of similar experiments demonstrate that the patches of radioactive label resulting from transneuronal transport coincide with groups of cells in layer IV dominated by the injected eye Ṫhe coincidence between the anatomy and physiology in layer IV of the cat is similar to that found in the monkey (LeVay et al. 1975). In the monkey, however, the cells in layer IVc are exclusively driven from one eye or the other (Hubel & Wiesel, 1968), whereas in the cat, even in layer IV, most cells are binocularly driven (although between half (Albus, 1975) and 70 % (see Text-fig. 3) are strongly dominated by one eye or the other). Cells driven nearly equally by the two eyes tended to occur between groups of cells strongly dominated by one eye or the other. For the most part the injected eye strongly dominated cells in the centres of the patches of label, where it was rare to encounter cells strongly dominated by the other eye. At the edges of the patches of label binocular cells were encountered along with a mixture of cells monocularly driven from one eye or the other, although occasionally the transitions were quite abrupt. This arrangement makes the ocular dominance columns in layer IV less distinct in the cat than in the monkey. In the other cortical layers, the ocular dominance columns are even less well defined but nevertheless can be detected (Hubel & Wiesel, 1965b; Albus, 1975). This can be seen in P1. 1 C in the portion of the two tracks passing through layer VI. Cells dominated by the injected eye tended to be grouped together near lesions 1 and 2 of penetration 2 and lesion 2 of penetration 1. These groupings correspond roughly to patches of label within layer IV of sections more medial than the one illustrated in P1. 1B. Because of the complex curvature of the cortex in this region, we did not attempt to examine the precision of this correspondence.

5 MONOCULAR DEPRIVATION: PHYSIOLOGY-ANATOMY 271 The contralateral eye dominated the majority of cortical cells encountered in the two penetrations, as illustrated in the ocular dominance histogram of P1. 1D. The overall dominance of the contralateral eye in these penetrations was typical of our findings in normal cats (see also Text-fig. 1) and is consistent with previous reports (Hubel & Wiesel, 1962; Blakemore & Pettigrew, 1970; Albus, 1975). The proportion of binocular cells, however, was smaller than usually reported, presumably because most cells studied here were located in layers IV and VI; these layers were generally found to contain fewer binocularly driven cells than the other cortical layers (see Text-fig. 3). An anatomical correlate of the overall physiological dominance of the contralateral eye was examined by comparing the two hemispheres following an injection of radioactive label into one eye. P1. 2 shows a representative autoradiograph from area 17 contralateral (2A) and ipsilateral (2B) to the injected eye in the cat of P1. 1. The densely labelled patches of terminals occupied a greater proportion of layer IV on the contralateral than on the ipsilateral side, consistent with an earlier report (Shatz et al. 1977). This difference was assessed in one animal from photographs of the series of sections through area 17, including those of P1. 2A and B. In both hemispheres, the area occupied by the dense patches of label was measured as a fraction of the area of layer IV. In the hemisphere ipsilateral to the injected eye, 40 % of layer IV was occupied by dense patches of label, while the corresponding measurement was 53 % on the contralateral side (see Table 1). The patches of label, however, are not sharp; grain density generally declines in a gradual fashion from the centre of a patch, indicating a substantial overlap at the borders between afferents representing the two eyes (LeVay, Stryker & Shatz, 1978). Therefore these measurements of area, made from high-contrast photographs, can only be approximate. We nevertheless regard the difference between the two hemispheres as genuine because it was evident in most individual sections. Autoradiographs of the two hemispheres differed in another respect: more label was present in the fourth layer between the dense patches on the side contralateral to the injected eye. This distribution of label may not accurately reflect the arrangement of the afferents representing the contralateral eye because of a technical complication of the transneuronal method - the leakage ('spillover') of radioactivity into the inappropriate laminae of the LGN, and the consequent labelling of the non-injected eye's terminals in the cortex. Spillover has been shown to be much greater within the contralateral than the ipsilateral LGN (LeVay et at. 1978). Effects of monocular deprivation: anatomy Injections were made into one eye of monocularly deprived cats in order to examine whether the difficulty in driving cortical cells through the deprived eye is reflected in the anatomy of the geniculocortical projection, as is the case in the monkey (Hubel et al. 1977). P1. 2C and D shows autoradiographs of representative parasagittal sections in an animal in which the deprived eye was injected, while PI. 2E and F shows sections through the two hemispheres when the non-deprived eye was injected. As in the monkey, when the deprived eye was injected, the patches of label were smaller than normal; this is seen most clearly on the side ipsilateral to the injected eye. When the non-deprived eye was injected (P1. 2F), the patches were much larger

6 272 C. J. SHATZ AND M. P. STRYKER EXPLANATION OF PLATES PLATE 1 Reconstruction of two micro-electrode penetrations made in the visual cortex (area 17) ipsilateral to an injection of 1-5 mci [3H]proline and 1 0 mci [3H]fucose into the eye of a normal cat. The animal survived 16 days after the eye injection. A, montage made from the two parasagittal sections, stained with Cresyl Violet, which contained most of the two electrode tracks. Along each track the electrolytic lesions visible in these sections are indicated by numbered arrows. Anterior is to the left; dorsal is up. B, autoradiograph shown in darkfield of a section immediately adjacent to those in A, showing the patches of radioactive label in layer IV that correspond to the injected eye. Magnification is the same as in A. Radioactive label appears white in this and all subsequent darkfield photos. Positions of open white arrows are identical to those of C, as is position of numbered white arrow which indicates location of lesion 5 in penetration 1. C', histological reconstruction of the two penetrations (PI and P2) based on Cresyl Violet stained sections like those of A. Layers are indicated by Roman numerals or WM (white matter). In this and all subsequent reconstructions, numbered arrows indicate the positions of electrolytic marking lesions made along the electrode track. Symbols refer to each unit recorded: units with simple-type receptive fields are denoted by circles; those with complex-type by squares; triangles indicate other receptive field types (e.g. hypercomplex, etc.) and units that were not classified; diamonds indicate unresolved background activity. Lesion 5 is followed by an asterisk to indicate that its position is also marked on the darkfield autoradiograph. In this instance, filled symbols refer to the units preferring the injected (ipsilateral) eye (ocular dominance groups 5-7). Half-filled symbols indicate units equally driven by the two eyes (group 4). The corresponding arrows in B and C mark identical positions. The dense patches of radioactive label seen in B have been projected onto the reconstruction of C. Thus, in penetration 1, the electrode passed through two patches of label in layer IV, one at the upper right open arrow, the other just following lesion 5. In both cases, cells were dominated by the injected eye. These penetrations were atypical in that simple cells were not isolated in layer IV, probably because the electrode used in this first experiment was too large. D, ocular dominance distribution of all cells recorded in both penetrations (see Text-fig. 1 for explanation of ocular dominance groups 1-7). PLATE 2 Darkfield autoradiographs of parasagittal sections from area 17 showing transneuronally transported label contralateral (A, C and E) and ipsilateral (B, D and F) to the injected eye in one normal (A and B) and two monocularly deprived (MD) cats (C and D; E and F). In one cat (C and D), the deprived eye was injected; in another, the non-deprived eye was injected (E and F). In each case, the patches of radioactive label in layer IV appear more distinct in the hemisphere ipsilateral to the eye injection, for the most part owing to a technical complication of the transneuronal transport method (see text). Intersecting arrows in A indicate orientation of sections in A, C, and E; arrows in B indicate that of B, D and F. D, dorsal; V, ventral; A, anterior; P, posterior. Magnification (B) is identical for A-F. In each section the most dorsal strip of label is located within the cortical regions (on the apex of the lateral gyrus) concerned with the representation of the vertical meridian of the visual field, whereas that located most ventrally is buried within the upper wall of the splenial sulcus and corresponds to peripheral visual field representation. Asterisk in D indicates approximate location of area centralis representation. Dorsal strip of label to left of arrow in D lies within area 18.

7 The Journal of Physiology, Vol. 281 Plate 1 D C. J. SHATZ AND M. P. STRYKER (Facing p. 272)

8 The Journal of Physiology, Vol. 281 Plate 2 Cont palaterai I ns-, I-I P[-,I N otormai MD. deprived eve injectlon MD non-deprived eye injection C. J. SHATZ AND M. P. STRYKER

9 MONOCULAR DEPRI VATION: PHYSIOLOGY-ANATOMY 273 than normal, fusing to form an almost continuous matrix. In sections from the hemisphere contralateral to the injected eye (P1. 2 C and E), the patches of label were less clearly delineated, at least in part because of the problem of spillover (LeVay et al. 1978). In the cat, then, as in the monkey, there were substantial changes in the geniculocortical projection as a consequence of early monocular deprivation. These changes were assessed, as in the normal cat, by measuring the fraction of layer IV occupied by dense patches of label. When the deprived eye was injected in the animal of P1. 2C and D, this fraction was 25 % on the ipsilateral side and 39 % on the contralateral side (Table 1). On the contralateral side the diffuse appearance of the patches made this measurement particularly difficult. When the non-deprived eye was injected, it was impossible to make this measurement on the contralateral side (P1. 2E), but on the ipsilateral side (P1. 2F) the label occupied 79 % of layer IV (Table 1). TABLE 1. Fraction of layer IV occupied by radioactive label following eye injection % of layer IV Total area Hemisphere with occupied by of layer IV respect to patches of measured Experiment injected eye label (mm2) Normal cat ipsilateral contralateral MD Cat: deprived eye ipsilateral injected contralateral MD Cat: non-deprived eye ipsilateral injected contralateral - - Measurements made from the three animals of P1. 2 (see Methods): one normal and two monocularly deprived (MD). Values must be considered to be approximate because borders of labelled patches are not sharp. Measurements from the hemisphere contralateral to the injected non-deprived eye were not possible due to the indistinct nature of the gaps in the label (see P1. 2E). Although the deprived eye's afferents occupied a smaller, and the non-deprived eye's afferents a larger, than normal proportion of layer IV, there is indirect evidence that these changes might not be strictly complementary. The fraction of layer IV labelled contralateral to a deprived eye injection (39 %) was greater than that remaining unlabelled ipsilateral to an injection of the non-deprived eye in another animal (21 %). This difference suggests that the overlap between afferents representing the two eyes in layer IV may be greater in monocularly deprived cats than in normal cats. Because of possible differences in the degree to which the two animals were affected by deprivation and because of the difficulty in determining the boundaries of the patches of label in the contralateral hemisphere, this suggestion must be regarded as tentative. A peculiarity of the pattern of radioactive label in layer IV associated with injections of the deprived eye was the presence of a more widespread distribution of label along the base of layer IV than superficially within the layer. Such a distribution

10 274 C. J. SHATZ AND M. P. STRYKER gave the patches of label within the fourth layer a pyramidal appearance, as is evident in P1. 2D (see especially the most ventral band of layer IV; note that here the pial surface is ventral). The complementary pattern following a non-deprived eye injection was not observed, a finding also consistent with the possibility of greater than normal overlap between afferents serving the two eyes in deprived animals, especially at the base of layer IV. In this respect, then, the cat may differ from the monkey, in which direct anatomical evidence for a strictly complementary distribution of afferents serving the two eyes has been presented (Fig. 18 of Hubel et al. 1977). Such a difference would be consistent with physiological findings in the two species (see below) but should not be accepted until similar anatomical evidence is obtained for the cat. In the cat, the LGN projects to area 18 as well, and the effects of monocular deprivation were also evident in autoradiographs of that area. Here, as in area 17, the patches of label resulting from an injection of the deprived eye appeared smaller, and those from an injection of the non-deprived eye larger, than normal, although measurements were not made. This is illustrated in P1. 2 D, in which the most anteriordorsal part of layer IV lies within area 18 (upper tier of label to left of arrow). Two additional observations deserve brief note. An injection of the deprived eye resulted in a pattern of radioactive label in the visual cortex that was somewhat less distinct than the pattern produced by injections in a normal cat. It was not possible to determine whether this difference was due to a reduction in the maximum density of the afferents representing the deprived eye, to a diminished ability of these afferents to transport and retain radioactive label, to greater spillover, or to variations in survival times. The area centralis representation of area 17 was less heavily labelled than the peripheral representations. This was most evident in animals in which the deprived eye had been injected (see asterisk in P1. 2D), but was present to a lesser extent when the non-deprived eye or one eye of a normal animal was injected, as has been noted elsewhere (LeVay et al. 1978). The reason for such uneven labelling is unknown. Effects of monocular deprivation: physiology Micro-electrode recordings were made in area 17 of five cats which had been monocularly deprived from birth until 8J to 11 months of age. The overwhelming majority of units were dominated by the non-deprived eye, and few binocularlydriven units were encountered. These findings are consistent with previous reports (Wiesel & Hubel, 1963). Ocular dominance histograms for the cortical cells studied in the hemispheres contralateral and ipsilateral to the deprived eye are shown in Text-fig. 1; similar results from normal cats are included for comparison.* Units with receptive field properties characteristic of cells in the LGN (see Methods) were excluded from all histograms in Text-figs. 1 and 3. In the hemisphere ipsilateral to the deprived eye, 8 % of the cortical cells were dominated by the deprived eye. In the contralateral hemisphere the deprived eye * The sample of cells from normal cats shown in Text-figs. 1, 2 and 3 includes 165 cells studied in this laboratory by Gilbert (1977) using identical criteria for ocular dominance and cell type; his results on laminar differences in these properties were similar to ours.

11 MONOCULAR DEPRIVATION: PH YSIOLOG Y-ANATOJM Y 275 dominated 13 %. This difference between the hemispheres in the effect of deprivation, though small, is consistent with our finding that the patches of radioactive label transported from the deprived eye occupied a smaller proportion of layer IV on the ipsilateral than the contralateral side (see Table 1). For the analyses by cortical Normal Deprived z Group contralateral hemisphere 0~ 100 II I~ ~ ~ ~ LI c ipsilateral hemisphere Text-fig. 1. The ocular dominance distribution of cortical cells in area 17 of normal (left) and monocularly deprived (right) cats. Units with LGN-type receptive field properties were excluded from this sample. Each cell was assigned to one of seven ocular dominance groups after Hubel & Wiesel, Cells driven exclusively by the contralateral eye belong to group 1; those preferring the contralateral eye strongly belong to group 2 and weakly to group 3; cells driven only by the ipsilateral eye belong to group 7; those preferring the ipsilateral eye belong to groups 5 and 6. Cells in group 4 are equally driven by the two eyes. Cells from monocularly deprived animals are plotted in one or the other histogram according to their location in the hemisphere contralateral (top) or ipsilateral (bottom) to the deprived eye. Filled bins indicate cells dominated by the deprived eye. layer which follow, however, we have combined the samples from the two hemispheres by assigning each cell to an ocular dominance group as if it had been recorded from the contralateral hemisphere (for example, in Text-figs. 2 and 3, cells belonging to group 1 were monocularly driven from the deprived eye). Combined in this fashion, the results of Text-fig. 1 show that 12 % of all cortical cells studied here were dominated by the deprived eye. This fraction is slightly greater than those of previous

12 276 C. J. SHATZ AND M. P. STRYKER reports (Wiesel & Hubel, 1963, 1965; Hubel & Wiesel, 1970; Blakemore & Van Sluyters, 1974; Olson & Freeman, 1975; Kratz et al. 1976). The difference may perhaps be attributed to a somewhat greater contribution of layer IV to our sample than to others'. A similar bias in our sample from normal cats may also account for the somewhat smaller proportion of binocularly driven cells (70 %) than has been reported previously (see Text-fig. 3) ~~~~~~~~~~~~~~ Normal Deprived LGN-type 20[!-m Text-fig. 2. Ocular dominance histograms of cortical units in area 17 of normal (left) and monocularly deprived (right) cats according to receptive field properties. Hypercomplex and unclassified cells were omitted from this sample. In these histograms, results from the hemispheres ipsilateral and contralateral to the deprived eye were combined so that groups 1 to 3 indicate cells preferring the deprived eye. All other conventions are as in Text-fig. 1. Of units dominated by the deprived eye, very few were complex cells. See text for further details. The receptive field properties of units dominated by the deprived eye in all layers of the visual cortex showed that most of these units were cortical cells rather than afferent fibres. This finding is illustrated in Text-fig. 2. Many (38 %) of the units dominated by the deprived eye (i.e., ocular dominance groups 1, 2, 3 and half of group 4) had simple-type receptive fields. A few units (6 %) dominated by the deprived eye had complex-type receptive fields. These simple and complex units are very likely cortical cells because cells with such receptive field properties have never been

13 MONOC ULAR DEPRIVATION: PHYSIOLOGY-ANATOMY 277 found in the lateral geniculate nucleus (Hubel & Wiesel, 1961). Thirty-three per cent of the units driven by the deprived eye were classified as LGN-type (see Methods). The remaining units (23 %) were classified as 'abnormal'. Most of these units fatigued easily, had no clear orientation preference, lacked the centre-surround organization and spontaneous activity typical of LGN cells, and none had a fibre waveform. We therefore think that these were cortical cells rather than afferent fibres. It is also interesting to note that 5 % of the cells dominated by the non-deprived eye also had abnormal receptive fields. These cells were frequently encountered at points where eye preference was changing. Most (66 %) of the cortical cells dominated by the deprived eye were located within layer IV. Text-fig. 3 shows the ocular dominance distributions of cells from each cortical layer in monocularly deprived animals; here normal data are provided for comparison. Within layer IV of the deprived animals 22 % of the cells studied were dominated by the deprived eye, while in layer IV of normal cats each eye had exclusive influence on about 20 % of the cells. The deprived eye dominated 26 % of the cortical cells recorded in layer IVc, and 20 % of those within layer IVab, hinting that monocular deprivation may differ in its effects on the sublaminae of layer IV, although a larger sample is required to assess the significance of this difference. This observation is noted because of the finding, described above, that radioactive label is more widely distributed at the base of layer IV than superficially, following an injection of the deprived eye (see PI. 2 D). Results in the cat, then, are similar to those in the monkey (Hubel et al. 1977) in that while the deprived eye dominates a substantial fraction of cortical cells in layer IV, its access to the other cortical layers is severely restricted. In layer V, for example, normally the most binocular cortical layer (see Text-fig. 3), no cells were dominated, and only five out of fifty-two cells were driven at all, by the deprived eye. Similarly within layer VI, only 4 % of the cells were dominated by the deprived eye. The difference between layers IV and VI in the effect of monocular deprivation was present when the simple cells alone were considered: twenty-one of the eighty-nine layer IV simple cells (24 %) were dominated by the deprived eye, while for the simple cells of layer VI, this figure was only two out of forty-three (5 %). This result suggests that the severity of the effect of monocular deprivation is determined more by cortical layer than by cell type. Within layer IV, however, the complex cells were affected by deprivation more severely than the simple cells (complex: 2/45 (4 %) dominated by deprived eye; simple: 21/89 (24 %)). Pattern of ocular dominance in layer IV after monocular deprivation The relation between cells driven by the deprived eye and the regions of layer IV which appeared to receive terminals from the deprived eye was examined (as it was in normal cats; P1. 1) by recording from layer IV in monocularly deprived cats in which one eye had been previously injected. An example from such an animal whose deprived eye was injected is shown in P1. 3. In this and all subsequent Plates the filled symbols represent units dominated by the deprived eye. A long tangential penetration was made in area 17 contralateral to the deprived eye. During the final third of this penetration (P1. 3C), units influenced by the deprived eye were found in two

14 X on _ C. J. SHATZ AND M. P. STRYKER Normal Deprived - 20_ = _ t, 0 E 40 _ + t! 40_ ID -n --vr, g0 0 '_-> 0-6 z _ = r IJnfiLimEJE -1 I r-1 I 40 - > 20- ol -[no l l rin > I 0 I I I Group Text-fig. 3. The ocular dominance distribution of cortical cells in area 17 of normal (left) and monocularly deprived (right) cats according to the cortical layer in which they resided. Because the border between layers II and III is indistinct, cells located within the upper half of the combined thickness of the two layers were assigned to the histogram labelled II + III top, while those located within the lower half were placed in the histogram marked II + III bottom. Histograms were compiled from the same sample of cortical units as that shown in Text-fig. 1. Filled bins in right histograms indicate those cells dominated by the deprived eye. In these histograms, results from the hemispheres ipsilateral and contralateral to the deprived eye were combined so that groups 1 to 3 indicate cells preferring the deprived eye. Other conventions as in Textfig.1.

15 MONOCULAR DEPRIVATION: PHYSIOLOGY-ANATOMY 279 groups within layer IV. Lesion 9 marks one group, and the other group follows lesion 10. In each case, units driven by the deprived eye were located within a patch of radioactive label in the fourth layer. This coincidence is shown in P1. 3B, the autoradiograph of the section containing the final portion of the penetration. Between the densely labelled patches in layer IV in this Plate, no physiological influence of the deprived eye was detected although some radioactive label above background levels is present. We think that this label does not indicate a widespread distribution of deprived-eye afferents, but is probably a consequence of spillover within the LGN which, as mentioned above, produces such a pattern in the contralateral hemisphere of normal cats (see P1. 2A and LeVay et al. 1978). On the other hand, the non-deprived eye was seen to influence the responses of both isolated units and unresolved background activity within the dense patches of label from the deprived eye (see P1. 3B). An intermingling of responses from the two eyes was particularly evident in the penetration shown in P1. 4, for which, unfortunately, the autoradiographic procedure failed. Here, even within the two groups of cells dominated by the deprived eye, cells dominated by the non-deprived eye were interspersed, and background responses could be elicited from both eyes. Such intermingling of responses is consistent with the possibility suggested from the anatomy (Table 1) that terminals representing the non-deprived eye may overlap part of the territory occupied by terminals representing the deprived eye, although in the case of P1. 4, the electrode may have merely traversed the border of each patch of deprived-eye terminals. In another animal in which the deprived eye was injected, we failed to encounter any cells in layer IV dominated by the deprived eye (P1. 5 C). From PI. 5B, the autoradiograph of the section containing the penetration, it is apparent that the electrode passed between two patches of label as it first crossed layer IV, neatly avoiding both (see lesion 2). When the electrode re-entered layer IV at the end of the penetration, it stopped just short of another patch of label (open arrow). These and the previous results suggest for layer IV of the cortex that within patches of deprived-eye afferents identified autoradiographically, and only within them, cortical cells may be dominated by the deprived eye. A similar question may be asked about the non-deprived eye: how is its widespread physiological influence related to the distribution of its afferents? This question was investigated by recording from monocularly-deprived cats in which the nondeprived eye had been injected. P1. 6 shows an example of one penetration from such an experiment. Here we encountered only one group of cells driven by the deprived eye in layer IV, marked by lesions 2 and 3. This group was found within the gap in the almost continuous band of label representing the non-deprived eye's territory in the fourth layer, as shown in the autoradiograph of P1. 6B. These results suggest that within layer IV the non-deprived eye influences cells only within the regions in which its afferents terminate. In layer IV of monocularly deprived cats, most cells were driven exclusively from one eye or the other and were grouped according to eye preference. Some regions of layer IV appeared to receive input from both eyes, in that cells driven monocularly from one eye were intermingled with those driven from the other, and some binocularly driven cells were also encountered (Pls. 3 and 4). In regions not receiving an

16 280 C. J. SHAPZ AND M. P. STRYKER PLATE 3 Reconstruction of the final third of a tangential penetration through area 17 in an animal monocularly deprived for 9 months. In this animal, the deprived eye was injected with 2*0 mci [3H]proline and 1P8 mci [j3h]fucose, and survived 13 days after the injection; recordings were made from the contralateral hemisphere. A, Cresyl Violet stained section containing a portion of the electrode track and lesions 8-11 (numbered arrows). B, darkfield autoradiograph of the same region from the immediately adjacent section. In the fourth layer, lesion 9 marks one patch of label corresponding to the deprived eye, while another patch follows lesion 10. Scale for A and B (shown in A) = 1 mm. C, reconstruction of the electrode track; conventions same as in P1. 1 except that in this and Pls. 4-6 filled symbols denote units dominated by deprived eye. As the electrode passed through layer IV, two groups of units dominated by the deprived eye were encountered: the first was located near lesion 9; the second after lesion 10. Histogram inset in this Plate and in Pls. 4, 5 and 6 show the ocular dominance distribution of all cells recorded along the portion of the electrode track illustrated. Ocular dominance groups are numbered conventionally, as in Text-fig. 1. Filled bins denote cells dominated by the deprived eye. Sections oriented with anterior to left; dorsal, up. PLATE 4 Reconstruction of a micro-electrode penetration through area 17 in an animal monocularly deprived for 11 months. Injection of radioactive label in this animal failed. Recordings were made from the hemisphere contralateral to deprived eye. A, Cresyl Violet stained parasagittal section containing the entire electrode track and all seven lesions (arrows); sections oriented anterior to left; dorsal, up. B, reconstruction of the track; conventions as in P1. 3; within layer IV, units dominated by the deprived eye were found between lesions 3 and 4, and 5 and 6. C, ocular dominance distribution of the units which were identified as cortical cells on the basis of their receptive field properties. Filled bins (left) indicate cells dominated by deprived eye. PLATE 5 Reconstruction of a micro-electrode penetration through area 17 in a cat monocularly deprived for 81 months. In this penetration we failed to record any units driven by the deprived eye within layer IV. Animal survived 12 days after 1 mci [3H]proline and 1-5 mci [3H]fucose were injected into the deprived eye. Recordings were made from hemisphere ipsilateral to deprived eye. Conventions are as in Pls. 3 and 4. Scale for B same as A. In B, an accumulation of silver grains running diagonally upwards to the right from lesion 2 (filled arrow) reveals a portion of the electrode track, which missed the two nearby patches of label corresponding to the deprived eye. The track ends just short of another patch of label located to the left of the open arrow in B. Open arrows in B and C at identical locations. PLATE 6 Reconstruction of a penetration in area 17 of a cat in which the non-deprived eye was injected. Animal survived 27 days after the eye injection (1.5 mci [3H]proline and 1-5 mci [3H]fucose). Period of monocular deprivation was 9 months. Recordings were made from the hemisphere contralateral to deprived eye. Conventions are as in Pls As the electrode entered layer IV, a group of cells driven by the deprived eye was encountered between lesions 2 and 3. Positions of these lesions were found to coincide with a gap in the radioactive label corresponding to the non.deprived eye in the darkfield autoradiograph of B. Scale for A-C is shown in C.

17 The Journal of Physiology, Vol. 281 A Plate 3 11~~~~~~~~~~~~~~~~~~~~~~~~~~1+1 I TZ H= =; cr :I= * 8*76 1 mm ov~~~~~~~~~~~~~~~~~.0 = so WmA T I C. J. SHATZ AND M. P. STRYKER (Facing p. 280)

18 - - s w X x I The Journal of Physiology, Vol (DC0) EZZZ'j- I I [co _ LO - MX _ N Plate 4 E ( C. J. SHATZ AND M. P. SThYKER

19 The Journal of Physiology, Vol. 281 Plate 5 + > > 5 A*' ~~~~~~~~~~. CE)~~~~~~~~ C. J. SHATZ AND M. P. STRYKER

20 The Journal of Physiology, Vol. 281 Plate 6 A C. J. SHATZ AND M. P. STRYKER 2 4 6

21 MONOCULAR DEPRIVATION: PHYSIOLOGY-ANATOMY 281 autoradiographically demonstrable input from the injected eye, cells were driven exclusively from the other eye (Pls. 5 and 6). These results allow us to elaborate a picture of the organization of the fourth cortical layer following monocular deprivation. The deprived eye's territory is smaller than normal; the non-deprived eye's territory is larger; and the two territories are roughly complementary but may overlap considerably. DISCUSSION This study has shown that in the cat, as in the monkey (Hubel et al. 1977), prolonged early monocular deprivation causes a rearrangement of the geniculocortical projection to layer IV, with corresponding physiological changes. The study focused on the representation of the central part of the visual field within cortical area 17. Anatomically, area 18 and the representation within area 17 of the peripheral visual field appeared also to be affected by monocular deprivation in a similar fashion, but physiological and quantitative anatomical studies of these regions were not made. Two kinds of explanations have been proposed to account for the loss in effectiveness of the deprived eye in the eat's visual cortex. According to one hypothesis, the deprived eye physically loses many of its connexions to cortical cells. This has been demonstrated to be the case for layer IVc of the monkey (Hubel et al. 1976, 1977). An alternative hypothesis proposes that the deprived eye's connexions to the cells of the visual cortex are somehow functionally suppressed so that they remain intact but cease to drive the cortical cells effectively (Cass et al. 1973; Kratz et al. 1976; Duffy et al. 1976). The correspondence observed in the present experiments between the reduced patches of geniculocortical afferents representing the deprived eye and the eye-preference of the cortical cells within these patches demonstrates that the first hypothesis is valid for layer IV of the cat's visual cortex. Where patches of the deprived eye's afferents were present, the deprived eye was effective in driving cortical cells. The present results deal only with the effects of prolonged deprivation and therefore do not exclude the possibility that a functional suppression may precede the physical loss of deprived-eye afferents (Blakemore, Van Sluyters & Movshon, 1976). Even with prolonged deprivation the possibility that some degree of functional suppression occurs cannot be excluded by the present experiments. Our interpretation of these experiments relies on the assumption that in autoradiographs of layer IV the dense patches of silver grains reflect the distribution of geniculocortical afferents representing the injected eye. In normal cats, the diffuse label found between the dense patches can be attributed to spillover (see Results and LeVay et al. 1978), and we assume that in deprived animals the situation is similar. This technical consideration, then, makes it unnecessary to attribute the radioactive label found in layer IV between the dense patches to afferents from the deprived eye which are somehow functionally suppressed. It was intriguing to find that about 20 % ofthe fourth layer, assessed either anatomically or physiologically, appeared to be resistant to the effects of monocular deprivation. In the normal adult cat, about 20 % of the layer IV cells are monocularly driven by the contralateral eye, and a similar fraction are exclusively driven by the ipsilateral eye (Text-fig. 3). The concidence between these two fractions suggests the

22 282 C. J.SHATZ AND M.P.STRYKER possibility that cells which are destined to be monocularly driven may be resistant to the effects of deprivation, while those nearly equally driven by the two eyes may be taken over by the open eye through a competitive interaction (Hebb, 1949; Wiesel & Hubel, 1965; Guillery, 1972; Sherman, Guillery, Kaas & Sanderson, 1974; Hubel et al. 1977). A model for the way in which monocular deprivation might exert its effects must take account of recent findings on the development of ocular dominance columns in layer IV. In the kitten (LeVay et al. 1978), as in the monkey (Rakic, 1976; Hubel et al. 1977), geniculocortical afferents serving the two eyes are initially intermingled within layer IV; the adult pattern emerges by a progressive segregation of the two sets of afferents. Segregation begins during the third week of life in the kitten and does not attain its adult extent until after the sixth week (see LeVay et al for further discussion). The onset of the critical period for monocular deprivation occurs during the fourth week (Hubel & Wiesel, 1970) at a time suitable for the competitive processes envisioned - after the onset of segregation but well before it stops. The more dramatic effects of monocular deprivation on the other layers of the cortex, seen physiologically in both cat and monkey, may be accounted for by either of the mechanisms which were proposed for layer IV: functional suppression, or a further loss of intrinsic cortical connexions from layer IV to the other layers (Hubel et al. 1977). In view of the findings concerning layer IV, a further loss of connexions to cells in the other cortical layers does not seem unlikely. Anatomical studies of the intracortical connexions serving the deprived eye may provide evidence on this point. We wish to thank Gail Grogan for preparing the autoradiographs, Carolyn Yoshikami for helping with the photography, Drs Simon LeVay and William Harris for discussions and critical readings of the manuscript and Dr Charles Gilbert for sharing his results with us. We are especially grateful to Drs David Hubel and Torsten Wiesel for making this study possible and enjoyable, and for criticizing the final product. This work was supported by grant EY to T. N. Wiesel from the National Institutes of Health. C.J.S. held a fellowship from the Harvard University Society of Fellows. REFERENCES ALBUS, K. (1975). Predominance of monocularly driven cells in the projection area of the central visual field in the cat's striate cortex. Brain Res. 89, BLAKEMORE, C. & PETTIGREW, J. D. (1970). Eye dominance in the visual cortex. Nature, Lond. 225, BiAKEMoptz, C. & VAN SLUYTERS, R. C. (1974). Reversal of the physiological effects of monocular deprivation in kittens: further evidence for a sensitive period. J. Physiol. 237, BLAKEMORE, C., VAN SLUYTERS, R. C. & MOVSHON, J. A. (1976). Synaptic competition in the kitten's visual cortex. Cold Spring Harb. Symp. 40, CAss, D. T., SrTTON, T. J. & MARK, R. F. (1973). Competition between nerves for functional connexions with axolotl muscles. Nature, Lond. 243, DUFFY, F. H., SNoDGRAss, S. R., BURCEarIL, J. L. & CONWAY, J. L. (1976). Bicuculline reversal of deprivation amblyopia in the cat. Nature, Lond. 260, GAnsY, L. J. & PoWEiLL, T. P. S. (1971). An experimental study of the termination of the lateral geniculo-cortical pathway in the cat and monkey. Proc. B. Soc. B 179, GILBERT, C. D. (1977). Laminar differences in receptive field properties of cells in cat primary visual cortex. J. Phy8iol. 268, GRAFsTEIN, B. (1971). Transneuronal transfer of radioactivity in the central nervous system. Science, N.Y. 172,

23 MONOCULAR DEPRI VATION. PHYSIOLOGY-ANATOMY 283 GUILLERY, R. W. (1972). Binocular competition in the control of geniculate cell growth. J. comp. Neurol. 144, HEBB, D. 0. (1949). Organization of Behavior. New York: John Wiley. HUBEL, D. H. & WIESEL, T. N. (1961). Integrative action in the cat's lateral geniculate body. J. Physiol. 155, HUBEL, D. H. & WIESEL, T. N. (1962). Receptive fields, binocular interaction, and functional architecture in the cat's visual cortex. J. Physiol. 160, HUBEL, D. H. & WIESEL, T. N. (1965a). Receptive fields and functional architecture in two nonstriate visual areas (18 and 19) of the cat. J. Neurophysiol. 28, HUBEL, D. H. & WIESEL, T. N. (1965 b). Binocular interaction in striate cortex of kittens reared with artificial squint. J. Neurophysiol. 28, HUBEL, D. H. & WIESEL, T. N. (1968). Receptive fields and functional architecture of monkey striate cortex. J. Physiol. 195, HUBEL, D. H. & WIESEL, T. N. (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. J. Physiol. 206, HUBEL, D. H. & WIESEL, T. N. (1972). Laminar and columnar distribution of geniculocortical fibers in the macaque monkey. J. comp. Neurol. 146, HUBEL, D. H., WIESEL, T. N. & LEVAY, S. (1976). Functional architecture of area 17 in normal and monocularly deprived macaque monkeys. Cold Spring Harb. Symp. 40, HUBEL, D. H., WIESEL, T. N. & LEVAY, S. (1977). Plasticity of ocular dominance columns in monkey striate cortex. Phil. Trans. R. Soc. Ser. B 278, KENNEDY, C., DES ROSIERS, M. H., SAKURADA, O., SHINOHARA, M., REIVICH, M., JEHLE, H. W., & SOKOLOFF, L. (1976). Metabolic mapping of the primary visual system of the monkey by means of the autoradiographic 14C-deoxyglucose technique. Proc. natn. Acad. Sci. U.S.A. 73, KRATZ, K. E., SPEAR, P. D. & SMITH, D. C. (1976). Postcritical period reversal of effects of monocular deprivation on striate cortex cells in the cat. J. Neurophysiol. 39, LEVAY, S. & GILBERT, C. D. (1976). Laminar patterns of geniculocortical projection in the cat. Brain Res. 112, LEVAY, S., HUBEL, D. H. & WIESEL, T. N. (1975). The pattern of ocular dominance columns in macaque visual cortex revealed by a reduced silver stain. J. comp. Neurol. 159, LEVAY, S., STRYKER, M. P. & SHATZ, C. J. (1978). Ocular dominance columns and their development in layer IV of the cat's visual cortex: a quantitative study. J. comp. Neurol. 179, OLSON, C. R. & FREEMAN, R. D. (1975). Progressive changes in kitten striate cortex during monocular vision. J. Neurophysiol. 38, OTSUKA, R. & HASSLER, R. (1962). ftber aufbau und gliederung der corticalen sehsphare bei der katze. Arch. Psychiat. NervenKrankh. 203, RAKIc, P. (1976). Prenatal genesis of connections subserving ocular dominance in the rhesus monkey. Nature, Lond. 261, SHATZ, C. (1977). A comparison of visual pathways in Boston and Midwestern Siamese cats. J. comp. Neurol. 171, SHATZ, C. J., LINDSTROM, S. & WIESEL, T. N. (1977). The distribution of afferents representing the right and left eyes in the cat's visual cortex. Brain Res. 131, SHERMAN, S. M., GUILLERY, R. W., KASS, J. H. & SANDERSON, K. J. (1974). Behavioral, electrophysiological and morphological evidence of binocular competition in the development of the geniculocortical pathways of cats. J. comp. Neurol. 158, STRYKER, M. P. & SHATZ, C. J. (1976). Ocular dominance in layer IV of the normal and deprived cat's visual cortex. Neurosci. Abstr. 2, WIESEL, T. N. & HUBEL, D. H. (1963). Single-cell responses in striate cortex of kittens deprived of vision in one eye. J. Neurophysiol. 26, WrIESEL, T. N. & HUBEL, D. H. (1965). Comparison of the effects of unilateral and bilateral eye closure on cortical unit responses in kittens. J. Neurophysiol. 28, WIESEL, T. N., HUBEL, D. H. & LAM, D. M. K. (1974). Autoradiographic demonstration of ocular dominance columns in the monkey striate cortex by means of transneuronal transport. Brain Res. 79, WILSON, M. E. & CRAGG, B. G. (1967). Projections from the lateral geniculate nucleus in the cat and monkey. J. Anat. 101,