B. Dreher

The loss of ganglion cells in the developing retina of the rat.

Horseradish peroxidase (HRP) was injected into both sides of the brain of newborn and adult rats. The number of retinal ganglion cells was estimated by counting cells containing granules of HRP reaction product. The mean number of labelled cells in 2-day-old animals was 169,500 (S.D. +/- 16,000, n = 6). By the tenth postnatal day the mean number of labelled cells had fallen to 113,500 (+/- 2900, n = 3). This value is similar to the mean number of labelled cells in the adult animal (113,000 +/- 2700; n = 4). Thus, during the first few postnatal days the number of retinal cells projecting to the central visual nuclei is reduced by at least 35%.

The morphology, number, distribution and central projections of class I retinal ganglion cells in albino and hooded rats

Class I retinal ganglion cells have been identified in wholemounts of rat retinae following injections of horseradish peroxidase (HRP) into retino-recipient nuclei. Class I cells are characterized by relatively large somata, 3-7 fairly frequently branching large-gauge primary dendrites and relatively thick axons. Cells with a very similar morphology have been visualized in the ganglion cell layer of retinal wholemounts using a neurofibrillar stain. The size of the somata and dendritic trees of Class I cells is affected by the density of all classes of ganglion cells: both somata and dendritic trees of Class I cells located in the region of peak density are smaller than those located in medium- and low-density ganglion cell regions. The mean numbers of Class I ganglion cells labelled following massive injections of HRP into retino-recipient nuclei were 876 (in albino rats) and 944 (in hooded rats), while the mean number of cells stained with the neurofibrillar method in albino retinae was 791. Thus, with the total number of positively identified retinal ganglion cells being 110,000-115,000 [Potts et al., 1982; Perry et al., 1983], Class I cells in both strains of rat constitute less than 1% of all retinal ganglion cells. Nevertheless the dendritic fields of Class I cells cover the entire retina. Although Class I cells are distributed relatively evenly across the retina, the density is slightly greater in the lower temporal retina where the bulk of the ipsilaterally projecting fibres originates. While Class I cells represent up to 10% of ipsilaterally projecting retinal ganglion cells in both strains of rat, fewer Class I cells project ipsilaterally in albinos than in hooded rats. All contralaterally projecting Class I cells appear to send branching axons to the superior colliculus and dorsal lateral geniculate nucleus. Class I cells represent a larger proportion of the ganglion cells projecting to the dorsal lateral geniculate nucleus (4-5%) than that of ganglion cells projecting to the superior colliculus (about 1%). The morphology, numbers, distribution and the pattern of the central projections of Class I retinal ganglion cells in rats suggest that they are likely to be homologues of the alpha-type ganglion cells distinguished in carnivores.

Peak density and distribution of ganglion cells in the retinae of microchiropteran bats: implications for visual acuity.

We have estimated the total number, distribution and peak density of retinal ganglion cells (RGCs) in retinal wholemounts of several species of microchiropteran (echolocating) bats. The estimates are based on counts of Nissl-stained, presumed RGCs. The total number of presumed RGCs varies among the species: from about 4,500 in Rhinolophus rouxi to about 120,000 in Macroderma gigas. In addition, in two species (Nyctophilus gouldi and M. gigas), the estimates are based on counts of positively identified RGCs retrogradely labelled with the enzyme horseradish peroxidase injected into the retinorecipient nuclei. In these two species, the numbers and distributions of retrogradely labelled RGCs and Nissl-stained presumed RGCs are very similar. In all six species studied, the peak-density regions of presumed (or positively identified) RGCs are located in the inferotemporal retinae, and the RGC isodensity lines tend to be horizontally elongated. However, the RGC densities in the high-density regions are only 2-4 times greater than those in the low-density regions in the superior retinae. The somal sizes of RGCs vary from 5 to 16 micron in diameter and are unimodally distributed. There is no indication of the existence of distinct morphological classes of RGCs. The axial lengths of microchiropteran eyes vary from 1.8 mm in R. rouxi to 7.0 mm in M. gigas. For all species the posterior nodal distance (PND) was assumed to be 0.52 of the axial length of the eye. This assumption is based on the analysis of published data concerning schematic eyes of nocturnal vertebrates. These derived values of the PNDs allowed us to calculate the retinal magnification factors and the number of RGCs per degree of visual angle. From these, the upper limits of visual acuity were derived on the basis of the assumptions of the sampling theorem. The estimated upper limits of visual acuity of the six species of echolocating bats vary from about 0.35 cycles/degree in R. rouxi to about 2 cycles/degree in M. gigas. This range is quite similar to the range of visual acuities in murid rodents.

Orientation, axis and direction as stimulus parameters for striate cells

Simple and hypercomplex cells have been recorded from the striate cortex of the cat anaesthetized with N2O/O2 and their responses to stationary flashing and moving bars of light and to moving spots have been studied. The orientation of a bar of light as a stimulus parameter is distinguished from the axis and direction of motion of a spot. The separate effects of these parameters on the responses of striate cells are described. The nature of these responses places clear restrictions on models designed to account for the behaviour of these cells.

A correlation of receptive field properties with conduction velocity of cells in the rat's retino-geniculo-cortical pathway.

Summary1.The receptive field properties and responses to electrical stimulation of 126 P-cells recorded from the dorsal lateral geniculate nucleus (LGNd) were studied in the hooded rat.2.Eighty-five cells had a concentric (Kuffler, 1953) receptive field organisation (46 off-centre on-surround; 39 on-centre off-surround). Of the remaining cells 29 had co-extensive on/off excitatory discharge regions, nine had on-centres with suppressive surrounds and two cells gave on-responses but had no suppressive surround. One cell was identified as suppressed-by-contrast.3.On the basis of the battery of tests developed for the identification of cell types in the cats retina and LGNd, 35 of the cells with a Kuffler-type receptive field organisation were identified as Y-like. The majority of the remaining cells, both concentric and others, reminded us of the different subclasses of W-cells of the cat. Nine concentric cells in most of the tests exhibited X-like properties.4.All of the Y-like cells were driven by relatively fast conducting retinal ganglion cell axons, comprising the t1 conduction velocity group. The majority of the remaining cells were driven by slower axons comprising t2 or t3 conduction velocity groups.5.Thus, in the rat, as in other mammalian species studied so far, there is a correlation between the conduction velocity groups in the retino-geniculo-cortical pathway and the functional groups based on the cells’ receptive field properties. There seem to be functional equivalents of the cats Y- and W-cell classes but evidence for a distinct X-like class of cells is lacking.

Evidence that the early postnatal reduction in the number of rat retinal ganglion cells is due to a wave of ganglion cell death

Horseradish peroxidase (HRP) was injected into the retino-recipient nuclei of each hemisphere in newborn rats. The animals were perfused 3--9 days after the injections; the number of retinal ganglion cells in retinal wholemounts was estimated by counting cells containing granules of HRP reaction product. The mean number (150,500) of labelled cells in 3-day-old rats was significantly higher than those in older animals (117,000, 121,000, 113,000 respectively on 6th, 8th and 9th postnatal days). However, in animals of any of the ages studied, the estimated numbers of ganglion cells were virtually the same as those in the animals of the same age but injected with HRP only 15--20 h before the perfusion. Thus, the reduction in the number of retinal cells projecting to the central visual nuclei observed during the first few postnatal days is due to a wave of retinal ganglion cell death; ganglion cell death induced by the neonatal removal of the contralateral superior colliculus has a similar time course.

The survival of neonatal rat retinal ganglion cells in vitro is enhanced in the presence of appropriate parts of the brain.

SummaryThe enzyme horseradish peroxidase (HRP) was injected into the visual centres of the brains of neonatal rats. Following dissociation of retinae into tissue culture, the ganglion cells could be identified by appropriate histochemical staining for HRP reaction product. Cultures were prepared of dissociated retinae from rats aged 2–6 days postnatal. After 3 h the cultures were fixed, and HRP-labelled cells visualized and counted. Estimates were made of the number of ganglion cells per retina at each age. Results indicated a loss of ganglion cells during the first few postnatal days. This loss paralleled that observed in vivo. It was further found the retinal ganglion cells died rapidly in vitro when cultured in a minimal medium. Only 50% of ganglion cells originally plated remained viable after 24 h. However, the survival rate could be increased to 100% by coculturing the cells with diencephalon and mesencephalon; these contain the retinorecipient nuclei. Coculturing with cerebellum did not result in such an enhanced survival rate. Ganglion cells could be maintained over longer periods of time by reinoculating the cultures with additional tissue containing diencephalon and mesencephalon. These results support the hypothesis that developing neurons require trophic factors from their target tissues in order to survive.

Convergence of parvocellular and magnocellular information channels in the primary visual cortex of the macaque.

We investigated whether responses of single cells in the striate cortex of anaesthetized macaque monkeys exhibit signatures of both parvocellular (P) and magnocellular (M) inputs from the dorsal lateral geniculate nucleus (dLGN). We used a palette of 128 isoluminant hues at four different saturation levels to test responses to chromatic stimuli against a white background. Spectral selectivity with these isoluminant stimuli was taken as an indication of P inputs. The presence of magnocellular inputs to a given cortical cell was deduced from its responses to a battery of tests, including assessment of achromatic contrast sensitivity, relative strengths of chromatic and luminance borders in driving the cell at different velocities and conduction velocity of their retino‐geniculo‐cortical afferents. At least a quarter of the cells in our cortical sample appear to receive convergent P and M inputs. We cannot however, exclude the possibility that some of these cells could be receiving a convergent input from the third parallel channel from the dLGN, namely the koniocellular (K) rather than the P channel. The neurons with convergent P and M inputs were recorded not only from supragranular and infragranular layers but also from the principal geniculate input recipient layer 4. Thus, our results challenge classical ideas of strict parallelism between different information streams at the level of the primate striate cortex.

High precision systems require high precision “blueprints”: A new view regarding the formation of connections in the mammalian visual system

It is well established that early in development interconnections within the mammalian visual system are often more widespread and less precise than at maturity. The literature dealing with the formation of visual connections has largely ignored differences in developmental specificity among species differing in their phylogenetic status and/or the visual ecological niche that they occupy. Based on a review of the available evidence, we have formulated an hypothesis to account for the varying degrees of developmental specificity that characterize different visual systems. It is suggested that extremely precise systems required for high-acuity binocular vision exhibit fewer presumed developmental errors than do visual systems characterized by poorer acuity and relatively crude depth perception. The developmental implications of the hypothesis are considered, and specific experiments are proposed to further test its validity.

Prosencephalic connections of striate and extrastriate areas of rat visual cortex

SummaryAfferent connections of rat primary visual cortex (area 17 or V1 area) and the rostral and caudal parts of areas 18a and 18b were studied, by placing in each of the areas, small electrophoretic injections of enzyme horseradish peroxidase (HRP) or wheat germ agglutinated-HRP. The results indicate that: 1) each of the areas has a distinct pattern of distribution of afferent neurons in the ipsilateral visual thalamus — area 17 receives its principal thalamic input from the dorsal lateral geniculate nucleus, the caudal parts of areas 18a and 18b receive a major thalamic input from the lateral posterior nucleus and a minor input from the posterior nucleus, while the rostral parts of areas 18a and 18b receive a major input from the posterior nucleus, and a minor projection from the lateral posterior nucleus; 2) the rostral and caudal parts of areas 18a and 18b each receive an associational input from area 17; 3) the rostral parts of areas 18a and 18b each receive associational input from three different extrastriate regions, the caudal part of the same extrastriate area, and the rostral and caudal parts of the other extrastriate area, whereas the caudal parts of areas 18a and 18b receive associational inputs only from one or two extrastriate regions; 4) area 17, area 18b and rostral area 18a each receive a substantial associational input from lamina V of the caudal part of the frontal eye field (FEF) in the motor cortex; however the input from the FEF to caudal area 18a (if present) is very small; 5) The extrastriate areas studied receive associational input from the restrosplenial cingulate area 29d; however, the input from area 29d to area 17 appears to be very small. The distinct patterns of distribution of prosencephalic afferents suggest to us that multiple retinotopically organized areas described previously in the rat cortex (cf Montero 1981; Espinoza and Thomas 1983) represent functionally distinct areas.