Loss of retinal X-cells in cats with neonatal or adult visual cortex damage.

Recordings were made from single retinal ganglion cell somas in cats whose visual cortical areas 17 and 18 were damaged on the day of birth or in adulthood. Neonatal lesions produced a 78 percent loss of X-cells in the retina, while lesions made in adulthood produced a 22 percent loss. Y-cells and W-cells were unaffected. This retinal abnormality needs to be considered when interpreting studies of behavioral deficits and neural mechanisms of recovery after damage to the visual cortex.

Development of rabbit visual cortex: late appearance of a class of receptive fields.

In young rabbits before the age at which the eyes open, only three of the seven receptive field types described in the adult visual cortex are detectable. The remaining four receptive field types-which share the property of having radially asymmetric fields-appear later, coincident with a decline in the percentage of cells that are visually responsive but not classifiable as to receptive field type.

Effects of aging on the densities, numbers, and sizes of retinal ganglion cells in rhesus monkey.

We used sterological procedures that yield unbiased estimates to quantify the densities, numbers, and soma sizes of retinal ganglion cells in seven young adult and six old rhesus monkeys. The retinae were flat mounted so that we could determine whether there are different aging-related losses in different retinal regions. The mean (+/-standard deviation) total number of ganglion cells was 1,529,039 +/- 115,260 in young-adult retinae and 1,556,698 +/- 165,056 in old retinae, a difference that was not statistically significant. There also were no significant differences between young and old retinae in the densities or total numbers of ganglion cells in the four retinal quadrants, in four concentric retinal zones from fovea to peripheral retina, or in smaller hemiretinal regions of the concentric zones. Ganglion-cell soma sizes also did not differ significantly between young and old animals. Moreover, counts of the largest ganglion cells, which probably correspond to P alpha ganglion cells, revealed no selective loss of these cells with aging. These results are consistent with our previous anatomical and physiological studies of the LGN. Together they suggest that the retino-geniculate pathways are relatively unaffected by aging in the rhesus monkey.

Effects of visual deprivation and alterations in binocular competition on responses of striate cortex neurons in the cat.

Electrophysiological recordings were made from single neurons in striate cortex of normally reared kittens (group N), kittens raised with binocular lid-suture (group BD), and kittens raised with one eye lid-sutured and other eye removed (group MD-E). The MD-E group represents a condition in which inputs from the deprived eye have been placed at a competitive advantage over those from the other eye. In agreement with previous studies, fewer cells were responsive to visual stimulation in BD kittens than in N kittens. Among the responsive cells, fewer were direction selective, fewer were orientation selective, and more had inconsistent or fast-adapting responses than in normals. The responsiveness and receptive field properties of striate cortex neurons in the MD-E kittens were less affected by the visual deprivation than in BD kittens; however, they still were abnormal in comparison to normal kittens. Comparison of the ocular dominance distributions for cells in N and BD kittens showed a marked reduction in binocularly driven cells in BD kittens. In addition, in BD kittens, a larger proportion of monocularly driven cells had orientation selective receptive fields than did binocularly driven cells. This difference was not found in normally reared kittens. The results of this study suggest that abnormal binocular interactions contribute to the effects of visual deprivation following binocular lid-suture, probably due to asynchronous light-dark inputs through the closed lids. Removing the other eye and placing inputs from the deprived eye at a competitive advantage during development results in decreased effects on striate cortex neurons. Nevertheless, visual deprivation still produces abnormalities in striate cortex independent of asynchronous or uncorrelated visual stimulation of the two eyes.

Effects of aging on the size, density, and number of rhesus monkey lateral geniculate neurons.

Visual abilities decline during aging, and many of these declines are due to neural changes in the retina or brain. We have begun studies of the monkey visual system to investigate the location and nature of these changes as well as to answer general questions about the effects of aging on neural structure and function. We began with the dorsal lateral geniculate nucleus (LGN) because it is the main structure through which visual information passes on the way to cortex and because the parallel parvicellular and magnocellular pathways are most easily identified and studied in the LGN. In the present experiment, we determined the sizes, densities, and numbers of LGN neurons in young-adult (5 to 12.5 years) and old (23 to 27.5 years) rhesus monkeys. The measures were corrected for tissue shrinkage, and stereological procedures were used that yield unbiased estimates. In young-adult monkeys, neurons densities were lower in the magnocellular layers (about 14,000/mm3) than in the parvicellular layers (23,000/mm3). Neuron density increased about 28% from anterior to posterior in both types of layers. There was an average of approximately 1,267,000 neurons in the parvicellular layers and 148,000 neurons in the magnocellular layers; however, there was substantial variability (1.9-fold) among five brains studied. Aging produced a statistically significant decrease in neuron density in both the magnocellular (29% average decrease) and parvicellular (41% average decrease) layers. However, there was no significant loss of neurons. Rather, the density decrease was due to a small (nonsignificant) decrease in the number of neurons combined with a small (nonsignificant) increase in LGN volume. The increase in LGN volume was due to a significant increase in neuron soma-size and proportional increase in the volume of glial cells, blood vessels, and neuropil. These results, together with those of other studies, suggest that the effects of aging on the primate visual pathway from retina through striate cortex are relatively subtle. It is possible that the major neural changes occur more centrally. Alternatively, individual differences in the effect of aging may require much larger samples or prior screening to observe consistent changes.

Single thalamic neurons project to both lateral suprasylvian visual cortex and area 17: a retrograde fluorescent double-labeling study.

Area 17 and the posteromedial lateral suprasylvian (PMLS) visual cortex receive inputs from three thalamic nuclei in common: the lateral division of the lateral posterior nucleus (LPl), the C-laminae of the lateral geniculate nucleus (LGNd), and the medial interlaminar nucleus (MIN). The present study determined whether these projections originate from the same cells via bifurcating axons or from separate populations of cells. Double-label retrograde transport techniques were used to label cells projecting to area 17 with one fluorescent dye and to label cells projecting to PMLS cortex with a different dye. The two dyes used were fast blue and Evans blue. Following injections into the two cortical areas, some cells were double labeled and some were single labeled in all three thalamic nuclei studied. However, the relative number of double- and single-labeled cells, as well as the relative number of cells single-labeled following injections into each cortical area, differed among the three thalamic nuclei. In both MIN and the C-laminae of the LGNd, the number of double-labeled cells was small. Similarly, the number of cells single labeled with the dye placed in PMLS cortex was small in these two nuclei. In contrast, a relatively large number of cells were single labeled with the dye placed in area 17, especially in the C-laminae of the LGNd. These results suggest that in both MIN and the C-laminae of the LGNd, few cells project to both area 17 and the PMLS cortex, few cells project only to PMLS cortex, and a relatively greater number of cells project only to area 17. In LPl, many cells were labeled after the cortical injections. In fact, when the areas of densest labeling for both dyes overlapped, almost every labeled cell in LPl was double labeled. This indicates that almost all LPl cells that project to one cortical area also project to the other via a bifurcating axon.

Thalamic projections to visual areas of the middle suprasylvian sulcus in the cat.

The thalamic afferents to two areas of the lateral suprasylvian visual cortex in the cat were studied by using retrograde transport of horseradish peroxidase (HRP). Injections were localized retinotopically with electrophysiological recording. The posteromedial lateral suprasylvian area (PMLS) of Palmer et al. ('78) receives afferents from the pulvinar (P), the posterior nucleus of Rioch (PN), the C-laminae of the lateral geniculate nucleus (LGNd) and the centrolateral (CL), lateral posterior (LP), medial interlaminar (MIN) nuclei. The anteromedial lateral suprasylvian area (AMLS) receives afferents from CL, P, LP, PN, MIN, and probably from the posterior nuclear group (PO), and the lateral dorsal (LD) and ventral anterior (VA) nuclei. The LP-pulvinar complex has been divided into four zones on the basis of connectivity: geniculate wing, pulvinar, the lateral division of LP, and the interjacent division of LP (Updyke, '77; Graybiel and Berson, '80; Guillery et al., '80). The locations of labeled cells in the present experiments suggest that both AMLS and PMLS receive afferents from each of the four zones, although differences exist in the strength of the projections. While AMLS and PMLS receive afferents from many of the same nuclei (CL, P, LP, PN, and MIN), differences in their afferents also were noted. These differences are of three types. The first is that some nuclei project to only one of the cortical areas. PMLS alone receives input from the C-laminae of the LGNd while AMLS alone receives probable input from PO, LD, and VA. The second difference is in the strength of the projection from some nuclei. AMLS receives a stronger projection from CL and P than does PMLS. The third difference concerns the pattern of distribution of neurons that project to each cortical area. Labeled cells in LP are dispersed after an AMLS injection, but are found in clusters or bands after a PMLS injection. Thus our results indicate that the thalamic afferents to AMLS and PMLS are in general similar: however, differences in input to AMLS and PMLS suggest that inputs to PMLS are predominantly visual while AMLS receives a broader spectrum of afferent information.

Effects of corpus callosum section on functional compensation in the posteromedial lateral suprasylvian visual area after early visual cortex damage in cats.

A visual cortex lesion made in adult cats leads to a loss of direction selectivity and a loss of response to the ipsilateral eye among cells in posteromedial lateral suprasylvian (PMLS) cortex of cats. However, a visual cortex lesion made in young cats results in normal direction selectivity and normal ocular dominance in PMLS cortex. Thus cats with an early lesion demonstrate functional compensation in PMLS cortex. The present experiment determined whether the functional compensation depends upon an intact corpus callosum. Cats received a unilateral visual cortex lesion on the day of birth (day 1) or at 8 weeks of age. When the cats were adult, the corpus callosum was sectioned and 24 hours later recordings were made in PMLS cortex ipsilateral to the visual cortex lesion. Results were compared to cats with a similar lesion and an intact corpus callosum. In cats with a lesion made on day 1, a corpus callosum section did not affect receptive-field properties or ocular dominance in PMLS cortex. Therefore, functional compensation is not dependent on input via the corpus callosum in these animals. However, in cats with a lesion made at 8 weeks. a corpus callosum section resulted in a decrease in the percentage of direction-selective cells and in the percentage of cells driven by the ipsilateral eye. Despite the decrease, the percentage of direction-selective cells still was greater than in cats with an adult unilateral visual cortex lesion. Thus, while partly dependent on callosal inputs, some functional compensation for direction selectivity remains on the basis of ipsilateral inputs.(ABSTRACT TRUNCATED AT 250 WORDS)

Quantitative studies of cell size in the cat's dorsal lateral geniculate nucleus following visual deprivation.

The effects of visual deprivation upon dorsal lateral geniculate (DLG) cell size were compared for seven kittens reared with monocular lid-suture (MD), seven with binocular lid-suture (BD), and six with one eye lid-sutured and the other eye enucleated soon after birth (MD-E). Six additional kittens were reared normally for comparison. For each kitten the cross-sectional areas of 300 cells were measured in one or both nuclei. Measurements were taken from the binocular segment of laminae A and A1 and the monocular segment of lamina A. In agreement with previous studies, cells in the binocular segment of the deprived laminae of MD cats were smaller (33-34%) than those in the non-deprived laminae. Comparisons with normal animals indicated that this difference was due to an increase (10-15%) in size of cells in the non-deprived laminae as well as a decrease (23-25%) in size of cells in the deprived laminae. Cells in the monocular segment also were affected by deprivation in MD cats, and this effect increased with the age (and duration of the deprivation) of the animal. However, it was always smaller than the decrease in cell size in the binocular portion of the DLG. In BD kittens, DLG cells were smaller (7-12%) than normal in all portions of the nucleus, including both the binocular and monocular segments. Direct comparisons between the deprived laminae of MD and BD kittens indicated that the decrease in cell size was greater for MD kittens in the binocular segment, but tended to be greater for BD kittens in the monocular segment. In MD-E kittens, DLG cells in the deprived laminae were smaller (11-17%) than normal in all portions of the nucleus, including both the binocular and monocular segments. Thus, the effects of deprivation were similar to those in BD kittens, even though inputs from the deprived eye had been placed at a competitive advantage in MD-E kittens. These results indicate that two factors may affect cell size in the DLG of visually deprived cats: deprivation per se and abnormal binocular competition. Finally, separate analyses for the ten largest and the ten smallest cells in each lamina of each cat were carried out in an attempt to determine if the changes in cell size were limited to the largest cells. In every case, differences observed for the total sample of cells were paralleled by differences from normal of both the largest cells present and the smallest cells present in the deprived laminae. Since at least two alternative interpretations can account for this finding, the question of whether the large cells are selectively affected by visual deprivation remains unanswered in the cat.

Thalamic projections to the lateral suprasylvian visual area in cats with neonatal or adult visual cortex damage.

Previous transneuronal anterograde tracing studies have shown that the retino-thalamic pathway to the posteromedial lateral suprasylvian (PMLS) visual area of cortex is heavier than normal in adult cats that received neonatal damage to visual cortical areas 17, 18, and 19. In contrast, the strength of this projection does not appear to differ from that in normal animals in cats that experienced visual cortex damage as adults. In the present study, we used retrograde tracing methods to identify the thalamic cells that project to the PMLS cortex in adult cats that had received a lesion of visual cortex during infancy or adulthood. In five kittens, a unilateral visual cortex lesion was made on the day of birth, and horseradish peroxidase (HRP) was injected into the PMLS cortex of both hemispheres when the animals were 10.5 to 13 months old. For comparison, HRP was injected bilaterally into the PMLS cortex of three cats 6.5 to 13.5 months after they received a similar unilateral visual cortex lesion as adults. In cats with a neonatal lesion, retrograde labeling was found in the large neurons that survive in the otherwise degenerated layers A and A1 of the lateral geniculate nucleus (LGN) ipsilateral to the lesion. Retrograde labeling of A-layer neurons was not seen in the undamaged hemisphere of these animals or in either hemisphere of animals that had received a lesion as adults. As in normal adult cats, retrograde labeling also was present in the C layers of the LGN, the medial interlaminar nucleus, the posterior nucleus of Rioch, the lateral posterior nucleus, and the pulvinar nucleus ipsilateral to a neonatal or adult lesion. Quantitative estimates indicate that the number of labeled cells is much larger than normal in the C layers of the LGN ipsilateral to a neonatal visual cortex lesion. Thus the results indicate that the heavier than normal projection from the thalamus to PMLS cortex that exists in adult cats after neonatal visual cortex damage arises, at least in part, from surviving LGN neurons in the A and C layers of the LGN. Although several thalamic nuclei, as well as the C layers of the LGN, continue to project to PMLS cortex after an adult visual cortex lesion, these projections appear not to be affected significantly by the lesion.

Development of the projections from the dorsal lateral geniculate nucleus to the lateral suprasylvian visual area of cortex in the cat.

In the study reported in the preceding paper, we used retrograde labeling methods to show that the enhanced projection from the thalamus to the posteromedial lateral suprasylvian (PMLS) visual area of cortex that is present in adult cats following neonatal visual cortex damage arises at least partly from surviving neurons in the dorsal lateral geniculate nucleus (LGN). In the C layers of the LGN, many more cells than normal are retrogradely labeled by horseradish peroxidase (HRP) injected into PMLS cortex ipsilateral to a visual cortex lesion. In addition, retrogradely labeled cells are found in the A layers, which normally have no projection to PMLS cortex in adult cats. The purpose of the present study was to investigate the mechanisms of this enhanced projection by examining the normal development of projections from the thalamus, especially the LGN, to PMLS cortex. Injections of HRP were made into PMLS cortex on the day of birth or at 1, 2, 4, or 8 weeks of age. Retrogradely labeled neurons were present in the lateral posterior nucleus, posterior nucleus of Rioch, pulvinar, and medial interlaminar nucleus, as well as in the LGN, at all ages studied. Within the LGN of the youngest kittens, a small number of retrogradely labeled cells was present in the interlaminar zones and among the cells in the A layers that border these zones. Such labeled cells were virtually absent by 8 weeks of age, and they are not found in normal adult cats. Sparse retrograde labeling of C-layer neurons also was present in newborn kittens.(ABSTRACT TRUNCATED AT 250 WORDS)

Do blue-eyed white cats have normal or abnormal retinofugal pathways?

Three white cats that had blue eyes and no tapetum were studied by behavioral, electrophysiological, and anatomical methods in order to determine whether they showed evidence of abnormal retinofugal pathways comparable to those found in Siamese cats and in other mammalian forms having melanin deficits. The three cats were normal in every respect. However, several other white cats, obtained subsequently, do show an abnormality of the retinogeniculate pathway identical to the abnormality of Siamese cats. Cats of the second type are thought to be homozygous for the Siamese gene and also the express the White gene. Because the characteristics Siamese pigmented "points" fail to develop in the presence of the white gene, cats of the second type are not distinguishable from other white cats on the basis of eye color or coat color. In terms of their central visual pathways and of their patterns, however, they are recognizably Siamese. It is not known how common "crypto-Siamese" cats are in the white cat population, but the possibility of their occurrence suggests that, in general, white cats should not be used for studied of the central visual pathways.

Effects of lateral suprasylvian visual cortex lesions on visual localization, discrimination, and attention in cats.

Experiments were carried out to begin to define the behavioral functions of the lateral suprasylvian (LS) visual area of the cat's cortex. Behavioral tasks were chosen for analysis on the basis of previous suggestions in the literature concerning possible functions of LS cortex and its afferent pathways. These tasks included the ability of cats to orient the head and eyes to a stimulus presented in particular locations in the visual field, the ability to learn successive reversals of a two-choice visual pattern discrimination, and the ability to maintain or shift attention between relevant or irrelevant visual form and brightness cues. Eight cats were trained on each of these tasks. Four of the cats then received bilateral lesions of LS cortex, including the AMLS and PMLS regions, and the remaining 4 cats were used to assess normal retention. The LS cortex lesions had no significant effect upon performance of any of the behaviors tested. Thus, this region of cortex appears to play no essential role in simple brightness, form, and pattern discrimination performance, visual reversal learning, maintaining and shifting visual attention, or orienting the head and eyes to stimuli in the visual field. These results are discussed in relation to previous lesion studies involving large regions of the cat's extrastriate cortex and studies in other species. Possible functions of LS cortex, based upon recent electrophysiological studies, are suggested.

Recovery of pattern discrimination ability in rats receiving serial or one-stage visual cortex lesions.

Two groups of 10 hooded rats were trained on a pattern discrimination between horizontal and vertical striped stimuli which were equated for contour-length and total luminous flux, and in which consistent local luminous flux cues were eliminated. In one group of rats, visual cortex removals were performed in two stages with training between the operations. Nine out of the 10 rats were able to relearn the pattern discrimination (median of 344 trials) after the completed bilateral visual cortex lesions in one stage. In agreement with previous studies, none of these animals were able to relearn the discrimination after more than 10 times (550 trial limit) the trials required for original learning. However, several rats with total one-stage lesions could relearn the pattern discrimination if very extended periods of training were given.

Retinal ganglion-cell densities and soma sizes are unaffected by long-term monocular deprivation in the cat.

Three cats were raised with monocular deprivation for 5.2-7.2 years, and ganglion-cell densities and soma sizes were measured in their flat-mounted retinae. The retinae were Nissl-stained so that ganglion cells could be measured whether or not they maintained normal central projections. Measurements were made in the area centralis, peripheral binocular segment, and monocular segment of the retinae. There were no significant differences between the deprived and non-deprived retinae in the densities or soma-sizes of alpha cells or other (non-alpha) ganglion cells at any of these retinal locations. These results support the view that the most distal effects of monocular deprivation occur at the retino-geniculate contact, and they suggest that even after long-term monocular deprivation, effects in the lateral geniculate nucleus do not produce secondary, retrograde changes in the retina.

Critical period for the marked loss of retinal X-cells following visual cortex damage in cats.

Visual cortex damage in newborn kittens produces a 78% loss of retinal X-cells whereas damage in adult cats produces only a 22% loss. Retinal Y- and W-cells are unaffected. The present experiment showed that the critical period for the severe loss of retinal X-cells ends between birth and 2 weeks of age. These results have implications for understanding the neural mechanisms of recovery from early visual cortex damage.

Development of neuronal responses in cat posteromedial lateral suprasylvian visual cortex.

We studied the normal development of responses to visual stimulation among neurons in the posteromedial lateral suprasylvian (PMLS) visual cortex, an extrastriate visual cortical area in cats. Recordings were made from 495 single neurons in 19 kittens that were 2, 3, 4, 8, or 12 weeks of age, and the results were compared with those from normal adult cats. The percentage of neurons that respond to light increased from 57% in 2-week-old kittens to approximately adult values in 8-week-old kittens (81%). The strength and consistency of neuronal responses also increased with age. Nearly all of the responsive cells had well-defined excitatory receptive-field centers, and the receptive-field center sizes were similar to adults at all ages studied. However, few cells (5%) had inhibitory receptive-field surrounds in 2-week-old kittens. The incidence of surround inhibition increased to adult levels (about 40% of the cells) by 8 weeks of age, and the strength of surround inhibition also increased with age. Most cells responded best to moving stimuli in 2-week-old kittens, just as in adults. However, only about 20% of the responsive cells were direction sensitive at 2 weeks of age. The percentage of direction-sensitive cells increased gradually with age and reached approximately adult values by 8 weeks of age (74%). Once cells developed complete direction selectivity, with no response in the null direction, directional tuning width was similar to that in adults. When tested with slits of light flashed at various orientations or with spots and slits moving in various directions, few cells (8% or less) showed orientation selectivity at any age, just as in adults. Most of the cells were binocularly driven, and the ocular dominance distribution was similar to adults at all ages studied. These results indicate that many response properties of PMLS neurons are similar to those of adults as early as 2 weeks of age, soon after the time of eye opening. However, some properties show marked developmental changes. The mechanisms and sources of these changes are considered. In addition, the relevance of these results to mechanisms of compensation following early damage to visual cortical areas 17, 18 and 19 is discussed.

Functional influence of areas 17, 18, and 19 on lateral suprasylvian cortex in kittens and adult cats: implications for compensation following early visual cortex damage.

The aim of the present study was to investigate the mechanisms of physiological compensation that is seen in the posteromedial lateral suprasylvian (PMLS) cortex of cats that received visual cortex (areas 17, 18, and 19) damage early in life. The strategy was to compare the response properties of PMLS neurons just after visual cortex damage (before any compensation has occurred) with the properties of PMLS neurons in normal cats and cats with long-standing visual cortex damage. Fourteen animals (aged 8 weeks, 18 weeks, 26 weeks, or adult) received a unilateral visual cortex lesion and recordings were made from ipsilateral PMLS cortex within about 24 h. An additional 4 adult cats were studied within either 24 or 3 h of a bilateral visual cortex lesion. Results from these animals were compared with results from normal cats and cats with long-standing visual cortex damage studied previously in this laboratory. At all ages studied, an acute visual cortex lesion reduced the percentage of direction-sensitive cells in PMLS cortex from nearly 80% in normal cats to about 20% after the lesion. In 8- and 18-week-old kittens, nearly all of the remaining PMLS cells responded best to stimulus movement but were not direction sensitive. In 26-week-old and adult cats, the remaining cells were divided between those that responded to movement without a directional preference and those that responded as well to stationary flashed stimuli as to moving stimuli. The presence of receptive-field surround inhibition was not affected significantly by an acute lesion at any age. In addition, few PMLS cells were orientation selective to elongated slits of light in cats with an acute lesion, just as in normal cats. The ocular dominance distributions of PMLS neurons also were normal following an acute visual cortex lesion at all ages studied. These results suggest that the influences of areas 17, 18, and 19 on the response properties of PMLS neurons are the same when the properties first reach maturity as in adult cats. The results also suggest that the mechanisms of physiological compensation for an early visual cortex lesion differ for different response properties. Compensation of direction sensitivity and orientation selectivity (an anomalous property) develops de novo after the early lesion. In contrast, compensation of ocular dominance appears to be due to the maintenance of a preexisting property that is present immediately after the lesion. Thus, plasticity after early visual cortex damage represents multiple developmental changes in the remaining visual pathways.

Influence of the cortico-geniculate pathway on response properties of cat lateral geniculate neurons.

The influence of the cortico-geniculate pathway on identified X and Y lateral geniculate cells was studied by reversibly cooling visual cortical areas 17 and 18. The majority (86.5%) of cells changed their response to visual stimulation when cortex was inactivated, and both X- and Y-cells were modulated by the cortical input. The influence of the visual cortex was complex, with both excitatory and inhibitory actions. Furthermore, the mechanism underlying the basic center-surround receptive field organization was influenced.

Reduction in numbers of large ganglion cells in cat retina following intravitreous injection of antibodies.

Antibodies were prepared against large ganglion cells isolated from bovine retina and injected into the vitreous chamber of 1 eye in 6 adult cats. The other eye of each cat received either a control pre-immune gamma-globulin injection or was untreated. After a survival time of 9-86 days, ganglion cell density was assessed from Nissl-stained retinal whole-mounts. In each cat, there were fewer large ganglion cells (alpha-cells) in the immunoglobulin-injected retina than in the control retina. The reduction in large ganglion cells occurred in patches adjacent to areas of approximately normal large ganglion cell density. Counts of the number of large ganglion cells in both eyes of the 6 cats indicated that the immunoglobulin injected eyes had from 8% to 61% (mean 32%) fewer large ganglion cells than the paired control eyes. This was significantly greater than the difference in the number of large ganglion cells between pairs of normal or control-injected eyes. The magnitude of the effect was not related to the survival time following the immunoglobulin injection. Cell size measures of all ganglion cells in selected areas of retina indicated that the small ganglion cells were unaffected by the antibodies. However, there was a suggestion that the largest of the medium size ganglion cells were affected in addition to the large ganglion cells. Counts of total ganglion cells per unit area in affected regions of retina revealed a reduced overall density, suggesting that the ganglion cells were lost rather than decreased in size. These results indicate that antibodies to the large ganglion cells can be used to reduce the number of large ganglion cells (alpha-cells) in the cat retina. Since these cells correspond to the Y-cell functional class of ganglion cells in the cat retina, the antibodies may provide a useful tool for studying Y-cell function in the visual pathways.