Neuronal responses in cat visual area 21a were analyzed when the primary visual cortex (areas 17 and 18) was deactivated by cooling. Ipsilateral and contralateral cortices were deactivated separately. Results established that (1) cooling the ipsilateral primary cortex diminished the activity of all area 21a cells and, in 30%, blocked responsiveness altogether, and (2) cooling the contralateral primary cortex initially increased activity in area 21a cells but, with further cooling, reduced it to below the original level although only 9% of cells ceased responding. These findings were then compared to earlier results in which bilateral deactivation of the primary cortex greatly reduced and, in most cases, blocked the activity of area 21a cells (Michalski et al., 1993). Despite the response attenuation following cooling of the primary visual cortex (either ipsilateral or contralateral), neurons of area 21a retained their original orientation specificity and sharpness of tuning (measured as the half-width at half-height of the orientation tuning curve). Direction selectivity also tended to remain unchanged. We concluded that for area 21a cells (1) the ipsilateral primary cortex provides the main excitatory input; (2) the contralateral primary cortex supplies a large inhibitory input; and (3) the nature of orientation specificity, sharpness of orientation tuning, and direction selectivity are largely unaffected by removal of the ipsilateral hemisphere excitatory input or the contralateral hemisphere inhibitory input.

It has been proposed that flying foxes (genus Pteropus) have a primate-like pattern of representation in the superficial layers of the superior colliculus (SC), whereby the visual representation in this structure is limited by the same decussation line that limits the retino-geniculo-cortical projection (Pettigrew, 1986). To test this hypothesis, visual receptive fields were plotted based on single- and multi-unit recordings in the SC of ten flying foxes. A complete representation of the contralateral hemifield was observed in the SC. Although the binocular hemifield of vision in Pteropus is 54 deg wide, receptive-field centers invaded the ipsilateral hemifield by only 8 deg, and the receptive-field borders by 13 deg. This invasion is similar to that observed at the border between visual areas VI and V2 in the occipital cortex. The extent of the ipsilateral invasion was not affected by a lesion that completely ablated the occipital visual areas, thus suggesting that this invasion may be consequence of a zone of nasotemporal overlap in the retinal projections to the two colliculi. Neurones located in the superficial layers typically responded briskly to stimulation of both eyes, with a bias towards the contralateral eye. After cortical lesions the neuronal responses to the ipsilateral eye were depressed, and the ocular-dominance histograms shifted towards an even stronger dominance by the contralateral eye. However, cells located in the rostral pole of the SC remained responsive to the ipsilateral eye after cortical lesions. Responses in the stratum opticum and stratum griseum intermediate were more severely affected by cortical lesions than those in the stratum griseum superficiale. Our results demonstrate that the SC in flying foxes retain some generalized mammalian characteristics, such as the stronger direct projections of the contralateral eye and the location of the upper, lower, central, and peripheral representations in the SC. Nonetheless, the extent of visual representation in the SC demonstrates a specialized, primate-like pattern. These observations are consistent with the hypothesis that megachiropterans are members of a group that branched off early during the differentiation of primates from basal mammals.

Smooth pursuit eye movements allow primates to keep gaze pointed at small objects moving across stationary surroundings. In monkeys trained to track a small moving target, we have injected brief perturbations of target motion under different initial conditions as probes to read out the state of the visuo-motor pathways that guide pursuit. A large eye movement response was evoked if the perturbation was applied to a moving target the monkey was tracking. A small response was evoked if the same perturbation was applied to a stationary target the monkey was fixating. The gain of the response to the perturbation increased as a function of the initial speed of target motion and as a function of the interval from the onset of target motion to the time of the perturbation. The response to the perturbation also was direction selective. Gain was largest if the perturbation was along the axis of ongoing target motion and smallest if the perturbation was orthogonal to the axis of target motion. We suggest that two parallel sets of visual motion pathways through the extrastriate visual cortex may mediate, respectively, the visuo-motor processing for pursuit and the modulation of the gain of transmission through those pathways.

We analyzed the immunohistochemical distribution of the two calcium-binding proteins, parvalbumin (PV) and calbindin D-28k (CB), in the primary visual cortex and lateral dorsal geniculate nucleus (dLGN) of monocularly enucleated macaque monkeys (Macaca fascicularis and Macaca nemestrind) in order to determine how the expression of PV and CB is affected by functional inactivity. The monkeys survived 1–17 weeks after monocular enucleation. The distribution pattern of each of the proteins was examined immunocytochemically using monoclonal antibodies and compared with that of the metabolic marker cytochrome oxidase (CO). We recorded manually the number of immunostained neurons and estimated the concentration of immunoreactive staining product using a computerized image-acquisition system. Our results indicate a decrease of approximately 30% in the labeling of PV-immunoreactive (ir) neuropil particularly in those layers of denervated ocular-dominance columns receiving the geniculocortical input. There was no change in the number of PV-ir neurons in any compartment irrespective of the enucleation interval. For CB-ir, we found a 20% decrease in the neuropil labeling in layer 2/3 of the denervated ocular-dominance columns. In addition, a subset of pyramidal CB-ir neurons in layers 2 and 4B, which are weakly stained in control animals, showed decreased labeling. In the dLGN of enucleated animals, PV-ir and CB-ir were decreased only in the neuropil of the denervated layers.

From these results, we conclude that cortical interneurons and geniculate projection neurons still express PV and CB in their cell bodies after disruption of the direct functional input from one eye. The only distinct decrease of PV and CB expression is seen in axon terminals from retinal ganglion cells in the dLGN, and in the axons and terminals of both geniculocortical projection cells and cortical interneurons in the cerebral cortex.

The purpose of this study was to obtain quantitative measures of the influence of callosal input to cells at the area 17/18 border region where transcallosal axons terminate most densely. Single-cell recordings were performed at the area 17/18 border region of the right hemisphere, while gamma-aminobutyric acid (GABA) or its antagonist, bicuculline, were applied to the transcallosal projecting regions of the left hemisphere to either block or overactivate the cells which projected to the neurons at the recording site. The results showed that visually evoked responses of the cells at the area 17/18 border were affected by administration of GABA or bicuculline to the contralateral hemisphere. Blockade of transcallosal input by application of GABA in the left hemisphere diminished the visually evoked responses of 51% of the neurons in the right hemisphere, and led to an increase in response magnitude for 17% of the neurons. Disinhibition of transcallosal input by application of bicuculline increased the evoked activity of 40% of the neurons and diminished the response magnitude of 20% of the neurons in the right hemisphere. GABA and bicuculline failed to show antagonistic effects on some cells. Thirty-two percent of the cells were affected by only one type of drug administration, and 13% of the cells showed either an increase or a decrease in responses after both GABA and then bicuculline administration. This study demonstrated complex interactions between neurons connected by the transcallosal pathway. A model of the transcallosal circuitry was proposed to explain the results.

We have used monolayer cultures prepared from early postnatal rabbit retinae (days 2–5) by the sandwich technique to study the capacity of immature neurons to express specific neuronal phenotypes in a homogeneous in vitro environment. Applying morphological, immunocytochemical, and autoradiographic criteria, we demonstrate that a variety of phenotypes could be distinguished after 7–14 days in vitro, and correlated with known retinal cell types. Bipolar cell-like neurons (approximately 4% of total cell number) were identified by cell type-specific monoclonal antibodies (115A10) and their characteristic bipolar morphology. Small subpopulations (about 1%) of GABA-immunoreactive neurons acquired elaborate morphologies strikingly similar to those of A- and B-type horizontal cells. Amongst putative amacrine cells several different subpopulations could be classified. GABA-immunoreactive amacrine-like neurons (6.5%), which also showed high affinity [3H]-GABA uptake, comprised cells of varying size and shape and could be subdivided into subpopulations with respect to their response to different glutamate receptor agonists (NMDA, kainic acid, quisqualic acid). In addition, a small percentage of [3H]-GABA accumulating cells with large dendritic fields showed tyrosine-hydroxylase immunoreactivity. Presumptive glycinergic amacrine cells (18.5%) were rather uniform in shape and had small dendritic fields. Release of [3H]-glycine from these neurons was evoked by kainic and quisqualic acid but not by NMDA. Small [3H]-glutamate accumulating neurons with few short processes were the most frequent cell type (73%). This cell type also exhibited opsin immunoreactivity and probably represented incompletely differentiated photoreceptor cells. Summing the numbers of characterized cells indicated that we were able to attribute a defined retinal phenotype to most, if not all of the cultured neurons. Thus, we have demonstrated that immature neuronal cells growing in monolayer cultures, in the absence of a structured environment, are capable of maintaining or producing specific morphological and functional properties corresponding to those expressed in vivo. These results stress the importance of intrinsic factors for the regulation of neuronal differentiation. On the other hand, morphological differentiation was far from perfect indicating the requirement for regulatory factors.

Chickens were raised with either translucent occluders or lenses, both under normal light cycles (12–h light/12–h dark) and in constant light (CL). Under normal light cycles, eyes with occluders became very myopic, and eyes with lenses became either relatively hyperopic (positive lenses) or myopic (negative lenses). After the treatment, retinal dopamine (DA), DOPAC, and serotonin levels were measured by high-pressure liquid chromatography (HPLC-EC). A significant drop in daytime retinal DOPAC (-20%) was observed after 1 week of deprivation, and in both DOPAC (-40%) and DA (-30%) after 2 weeks of deprivation. No changes in retinal serotonin levels were found. Retinal DA or DOPAC content remained unchanged after 2 or 4 days of lens wearing even though the lenses had already exerted their maximal effect on axial eye growth. When the chickens were raised in CL, development of deprivation myopia was reduced (8 days CL) or entirely blocked (13 days CL). Lens-induced changes in eye growth were not different after either 6 or 11 days in CL, compared to animals raised in a normal light cycle. Thirteen days of CL resulted in a dramatic reduction of DA and DOPAC levels, but serotonin levels were also lowered. The results suggest that lens-induced changes in refraction may not be dependent on dopaminergic pathways whereas deprivation myopia requires normal diurnal DA rhythms to develop.

Axon diameters of retinal ganglion cells were measured from electron micrographs of the nerve fiber layer of the cat. Three adult retinae were examined which had mean axonal diameters of 1.18 ± 0.86 (n = 5553), 1.12 ± 0.79 (n = 7265), and 1.47 ±1.11 μm (n = 10,867). Cumulative histograms from several locations adjacent to the optic disc were unimodal (modal peaks: 0.6–0.8 μm). This unimodal distribution, however, did not reflect the regional differences in axonal diameters found throughout the retina. In many locations, especially those related to axons of the temporal retina, axon diameter distributions were clearly bimodal or even trimodal (modal peaks: 0.6–0.8, 1.4–2.1, and 3.3 μm). Measurements from one retina indicated that the mean diameters of axons arising from the area centralis and visual streak (0.94 ± 0.63 and 0.98 ± 0.68, respectively) were not significantly different from each other; however, when compared to other areas around the optic disc, the percentage of fibers with diameters between 1.5–2.0 μm was highest in the sample adjacent to the area centralis. Axons temporal to the optic disc were found to be on average larger than those nasal to the optic disc; similarly superior axons were larger than inferior axons. Axonal distributions at the retinal periphery were found to be significantly different from those at the optic disc (p ≤ 0.05) and contained a higher percentage of medium-sized axons and fewer small axons. In each of the three retinae the proportions small, medium, and large axons were respectively γ: 46; 47; 48, β: 50; 49; 48, and α: 4; 4; 4; regional differences in the proportions of each axonal class are compared to previously published ganglion cell density maps. Differences between axonal bundles within each sample location were not significantly different; however, in one retina axons in the scleral half of the fiber layer were significantly larger (P ≤ 0.01) than axons in the vitreal half of the nerve fiber layer adjacent to the optic disc. When compared to the axonal diameter distributions found within the optic nerve (Cottee et al., 1991) and optic tract (Reese et al., 1991), our data indicates that the diameter of retinal axons may increase by up to 30% along the length of the visual pathway.

Tritiated proline and leucine were injected into areas 17 and 18 of intact cats and into the medial bank of the lateral suprasylvian (LS) cortex of intact cats and cats from which areas 17 and 18 had been removed on postnatal day 1 (P1), P28, or in adulthood (A). The density of label transported to the superior colliculus was quantified using image-analysis equipment. The results from the intact cats confirmed previous reports that areas 17 and 18 project most heavily to stratum zonale (SZ) and stratum griseum superficiale (SGS) and LS cortex projects most heavily to stratum opticum (SO) of the superior colliculus. However, in cats with lesions of areas 17 and 18, the projections from LS cortex showed an age-dependent reorganization. LS projections to SGS and SZ were enhanced following ablation of areas 17 and 18 on P1, and projections to SGS were enhanced following an ablation on P28. The pattern of LS-collicular projection following ablations incurred in adulthood was indistinguishable from the pattern presented by intact cats. This study demonstrates that the LS corticocollicular projection expands in SGS and possibly substitutes for inputs eliminated by the removal of areas 17 and 18 from the immature brain. This enhanced pathway may contribute to compensatory neuronal changes and to spared behaviors that accompany damage of immature cortex.

This study examined the consequences of visual system lesions on visual aftereffects produced by achromatic stimuli of various luminance contrasts and chromatic stimuli of various wavelength compositions. The effects of repeated exposure to such adapting stimuli were assessed using probes whose luminance contrast and wavelength composition were systematically varied using both detection and discrimination paradigms. Interocular tests revealed that both peripheral and central mechanisms contribute to the visual aftereffects produced by the adapting stimulus arrays used in this study. Contrary to the hypothesis according to which the midget system of the retina is the conveyor of visual afterimages, we found that blocking this system with lesions of parvocellular lateral geniculate nucleus, through which the midget cells make their way to the striate cortex in primates, did not eliminate the visual aftereffects. It appears therefore that the parasol system of the retina, which courses through the magnocellular layers of the lateral geniculate nucleus to cortex, can convey the necessary signals for the generation of visual aftereffects. Lesions of areas V4 and MT did not have significant effects on the visual aftereffects studied suggesting that the central factors that contribute to the visual aftereffects occur either already in area VI or are conveyed to higher centers through regions other than areas V4 and MT.

The superior colliculus (SC) projects to all layers of the cat's lateral geniculate nucleus (LGN) and thus is in a position to influence information transmission through the LGN. We investigated the function of the tecto-geniculate pathway by studying the responses of cat LGN neurons before, during, and after inactivating the SC with microinjections of lidocaine. The LGN cells were stimulated with drifting sine-wave gratings that varied in spatial frequency and contrast. Among 71 LGN neurons that were studied, 53 showed a statistically significant change in response during SC inactivation. Control experiments with mock injections indicated that some changes could be attributed to slow waxing and waning of responsiveness over time. However, this could not account for all of the effects of SC inactivation that were observed. Forty cells showed changes that were attributed to the removal of tecto-geniculate influences. About equal numbers of cells showed increases (22 cells) and decreases (18 cells) in some aspect of their response to visual stimuli during SC inactivation. The proportion of cells that showed tecto-geniculate influences was somewhat higher in the C layers (68% of the cells) than in the A layers (44% of the cells). In addition, among cells that showed a significant change in maximal response to visual stimulation, the change was larger for cells in the C layers (64% average change) than in the A layers (26% average change) and it was larger for W cells (61% average change) than for X and Y cells (29% average change). Nearly all of the X cells that showed changes had an increase in response, and nearly all of the Y cells had a decrease in response. In addition, across all cell classes, 80% of the cells with receptive fields < 15 deg from the area centralis had an increase in response, and 80% of the cells with receptive fields > 15 deg from the area centralis had a decrease in response. None of the LGN cells had significant changes in spatial resolution, and only three cells had changes in optimal spatial frequency. Ten cells had a change in contrast threshold, 25 cells had a change in contrast gain, and 29 cells had a change in the maximal response to a high-contrast stimulus. Thus, our results suggest that the tecto-geniculate pathway has little or no effect on spatial processing by LGN neurons. Rather, the major influence is on maximal response levels and the relationship between response and stimulus contrast. Several hypotheses about the role of the tecto-geniculate pathway in visual behavior are considered.

Calcium (Ca2+) plays an integral role in the light response of the photoreceptors in both vertebrate and invertebrate organisms. In the ventral eye of the horseshoe crab, Limulus polyphemus, a flash of light delivered to a dark-adapted photoreceptor stimulates a rapid rise in intracellular free calcium concentration ([Ca2+]i), which in turn mediates light adaptation. It has previously been demonstrated that in Limulus photoreceptors light, via Ca2+, activate s a calcium/calmodulin (Ca2+/CaM)-dependent protein kinase which increases the phosphorylation of arrestin. We now have identifie d biochemically, a calcium/calmodulin-dependent protein phosphatase (Ca2+/CaM PP ) in homogenates of the Limulus lateral and ventral eye, brain, and lateral optic nerve using as a substrate, a 32P-labeled peptide fragment of the regulatory subunit of cAMP-dependent protein kinase (RII). This protein phosphatase shares biochemical properties with calcineurin, a Ca2+/CaM-dependent protein phosphatase (type-2B). Its activity is enhanced by Ca2+, calmodulin and Mn2+; and is inhibited by mastoparan, a calmodulin antagonist, and a synthetic peptide corresponding to the autoinhibitory domain of mammalian calcineurin. Most importantly, light regulates the Ca2+/CaM PP activity in the lateral eye. While there is no difference in basal activity in long-term dark- or light-adapted preparations, Ca2+ enhances Ca2+/CaM PP activity only in long-term light-adapted eyes.

It has been assumed that the ‘ON’ and ‘OFF’ responses of the human electroretinogram (ERG) interact to produce a single waveform when brief flashes are used. To test this assumption, we examined the separate effects of stimulus intensity, the level of retinal illuminance and stimulus duration on the ERG. Both ‘ON’ and ‘OFF’ response amplitude decreased as stimulus intensity was decreased and as the level of retinal illumination was reduced. When stimulus duration was reduced, the amplitude of the ‘OFF’ response decreased; however, the amplitude of the ‘ON’ response increased. Summing of the ‘ON’ and ‘OFF’ response waveforms could not account for the increased amplitude in response to brief stimuli or for the changes in ERG wave shape. These results indicate that there is not a significant corneally recordable ‘OFF’ response elicited by brief stimuli, such as those commonly used to record the flash ERG.

Dark-adapted thresholds of albino and pigmented rats were estimated using behavioral methods. Albino and pigmented rats who had been water deprived learned to bar press for water reinforcement when a light stimulus was presented. Absolute threshold was defined to be the light intensity at which bar pressing behavior was significantly modified by the presence of the light stimulus. Albino rats had an average threshold of −5.23 log cd/m2 and the pigmented rats had a threshold of −5.0 log cd/m2. These values are close to −5.3 log cd/m2, the psychophysical threshold of human observers in the same apparatus. Consistent with our earlier electrophysiology, these behavioral experiments provide no evidence for an albino/pigmented sensitivity difference. Comparisons are made between behavioral and electrophysiological determinations of absolute threshold in albino and pigmented rats. Thresholds determined behaviorally agree remarkably well with those derived from visual evoked potentials.

A fundamental anatomical feature of retinal neurons is that they form planar mosaics. Each mosaic can be described by its density, pattern, and regularity (non-randomness). As part of ongoing studies to quantitatively describe the anatomy of regenerated retina in the goldfish, we determined the planimetric density and regularity of the mosaic of dopaminergic interplexiform cells in patches of regenerated retina and compared this to the mosaic generated de novo. In addition, we selectively ablated dopaminergic neurons with the neurotoxin 6–hydroxydopamine (6–OHDA) before inducing local regeneration and determined whether or not the absence of the extant dopaminergic neurons modulated the planimetric density or number of regenerated ones. The results showed that dopaminergic neurons are regenerated at higher planimetric densities and in less orderly arrays than normal. Furthermore, there was no statistical difference in the density or number of regenerated cells in normal retinas and retinas treated with 6–OHDA.

Anterograde transport of cholera toxin subunit B (CTb) was used to study the retinal projections in birds, with an emphasis on retinohypothalamic connections. Pigeons (Columba livia) were deeply anesthetized and received unilateral intraocular injections of CTb. In addition to known contralateral retinorecipient regions, CTb-immunoreactive fibers and presumptive terminals were found in several ipsilateral regions, such as the nucleus of the basal optic root, ventral lateral geniculate nucleus, intergeniculate leaflet, nucleus lateralis anterior, area pretectalis, and nucleus pretectalis diffusus. In the hypothalamus, CTb-immunoreactive fibers were observed in at least two contralateral cell groups, a medial hypothalamic retinorecipient nucleus, and a lateral hypothalamic retinorecipient nucleus. To compare retinorecipient hypothalamic nuclei in pigeons with the mammalian suprachiasmatic nucleus, double-label experiments were conducted to study the existence of neurophysin-like immunoreactivity in the retinorecipient avian hypothalamus. The results showed that only cell bodies in the medial hypothalamic nucleus contained neurophysin-like immunoreactivity. The results demonstrate CTb to be a sensitive anterograde tracer and provide further anatomical information on the avian equivalent of the mammalian suprachiasmatic nucleus.

We have studied the steady-state PERG in human subjects in response to red-green plaid patterns modulated either in luminance or in chromaticity or both. By varying the relative luminance of the red and green components, a value could be obtained at which the PERG amplitude was either minimum or locally maximum. This always occurred at equiluminance, as measured by standard psychophysical techniques. PERG amplitude and phase were measured as a function of spatial and temporal frequency of sinusoidal contrast reversal. In both space and time, the response to chromatic patterns was low-pass, while that to luminance was band-pass, and extended to higher spatial and temporal frequencies. The phase of the PERG to chromatic stimuli was systematically lagged compared with that to luminance stimuli, by an amount corresponding to about 20 ms under our experimental conditions. The variation of phase with temporal frequency suggested an apparent latency of about 67 ms for color contrast compared with 47 ms for luminance. These estimates were confirmed with separate measurements of transient PERGs to abrupt contrast reversal. For both luminance and chromatic stimuli, the amplitude of PERGs increases with increasing stimulus contrast. By summing vectorially the responses to appropriate luminance and chromatic contrasts, we were able to predict with accuracy the response as a function of color ratio (ratio of red to total luminance). The above findings all agree with those reported in the accompanying paper for the monkey PERG (Morrone et al., 1994), and indicate that the differences in response latency and integration time of luminance and chromatic stimuli observed by psychophysical and VEP techniques may arise at least in part from the properties of retinal mechanisms.

Four experiments were performed to assess the effects of ON channel blockade with the glutamate analog 2-amino-4-phosphonobutyrate (APB) on brightness and contrast perception in monkeys. In Experiment 1, we demonstrate that stimuli brighter than background (incremental stimuli) appear less bright following ON channel blockade. This decrease in brightness is not enough to account for the previously observed threshold increase for detection of incremental stimuli following APB administration (Schiller et al., 1986; Dolan & Schiller, 1989). Experiment 2 examines the role of the ON and OFF channels in the interaction between local contrast and apparent brightness. The phenomenon of simultaneous contrast was examined under normal conditions and following APB administration. We find that even following ON channel blockade, the brightness of a stimulus is determined primarily by its contrast with its immediate background. This indicates that the lateral processes involved in simultaneous contrast can operate even when one channel has been compromised. In Experiment 3, we examined the role of the ON channel in detection of stimuli that appear by virtue of changes in background vs. foreground luminance. We find that the ON channel selectively conveys information pertaining not only to the temporal nature that defines the stimulus as incremental but also to the spatial features that define it as incremental. In Experiment 4, we test the hypothesis that incremental and decremental temporal luminance ramps are differentially processed by the ON and OFF channels to a higher degree than are step-luminance changes. We find that the detection of incremental ramps is no more affected than is the detection of incremental steps following APB administration.

Naturally occurring cell death has been hypothesized to sculpt various features of the organization of the mature visual pathways, including the recent proposal that the selective elimination of ganglion cells in the temporal retina shapes the formation of decussation patterns. Through a class-specific interocular competition, ganglion cells in the two temporal hemiretinae are selectively lost to produce the decussation patterns characteristic of each individual cell class (Leventhal et al., 1988). The present study has tested this hypothesis by asking whether the removal of one retina in newborn ferrets, which should disrupt binocular interactions at the level of the terminals, alters the decussation pattern of the alpha cells, a cell class that is entirely decussating in the normal adult ferret. Enucleation on the day of birth was found to increase the uncrossed projection by ≈50%, but not a single uncrossed alpha cell was found in the temporal retina. Either alpha cells never project ipsilaterally during development, or if they do, they cannot be rescued by early enucleation. While naturally occurring cell death plays many roles during development, creating the decussation pattern of the ferreth's alpha cell class via a binocular competition at the level of the targets is unlikely to be one of them.

The molecules which control the patterns of cell division, growth, and precise interconnections characteristic of the central nervous system still remain largely unidentified. The protein platelet-derived growth factor (PDGF) has been shown to mediate interactions among glial cells in vitro. More recent evidence has indicated that PDGF may also be involved in controlling communication between neurons and glial cells and among neurons. The presence of receptors for PDGF on neurons of the developing nervous system is an essential piece of evidence in this chain of events. Ganglion cells are labeled with antibodies to PDGF receptor only during the period of active process outgrowth. These findings suggest that PDGF is used as a mediator of intercellular signaling during neuronal development.