Steady discharges were collected from ganglion cells of the magnocellular (MC) and parvocellular (PC) pathways of the macaque while their receptive fields were uniformly illuminated with a 4.7-deg steady yellow light of photopic illuminance. The mean rates, coefficients of variation, interval distributions, serial correlation coefficients, and power spectra of these discharges were determined. The results presented permit one to estimate the noise power in the discharges of macaque ganglion cells and hence determine how visual signals of different amplitudes will be affected by the noise resident in their discharges.

Although there was some small serial correlation in the discharges of both MC- and PC-pathway cells, their discharges can be considered to result from renewal processes with reasonable accuracy. As with the discharges of cat ganglion cells, macaque ganglion cell discharges can be considered to have approximately gamma-distributed intervals. Steady discharges of MC- and PC-pathway cells show considerable overlap in their statistics, although small but significant differences are present.

Two types of field potentials were identified in cat visual cortex using contrast reversal of oriented bar gratings: a short-latency fast-local component with a retinotopic organization similar to that seen with single-unit discharges at the same cortical site, and a slow, nonretinotopic component with a longer peak latency. The slow-distributed component had an extensive receptive field mapped by measuring the amplitude of binary kernels and showed strong inhibitory interactions within the receptive field. The peak latency of the slow-local component increased with distance from the retinotopic center, suggesting a possible conduction delay. Both components showed some orientation bias depending on the laminar location, but the bias could be independent of the orientation preferred by single units in the immediate vicinity. The present findings indicate that locally generated field potentials reflect cortical mechanisms for nonlinear integration over wide areas of the visual field.

We have studied light-evoked changes in extracellular Ca2+ concentration in the intact cat eye using ion-sensitive double-barreled microelectrodes. Two prominent changes in Ca2+ concentration were observed that differed in retinal location. There was a light-evoked increase in accompanied by brief ON and OFF transients, which was maximal in the inner plexiform layer and was not further studied. There was an unexpected sustained light-evoked decrease in of relatively rapid onset and offset, which was maximal in the distalmost region of the subretinal space (SRS). in the SRS was 1.0 mM higher than in the vitreous humor during dark adaptation and this transretinal gradient disappeared during rod-saturating illumination. After correcting for the light-evoked increase in the volume of the SRS, an increase in the total Ca2+ content of the SRS during illumination was revealed, which presumably represents the Ca2+ released by rods. To explain the light-evoked changes, we used the diffusion model described in the accompanying paper (Li et al., 1994b), with the addition of light-dependent sources of Ca2+ at the retina/retinal pigment epithelium (RPE) border and rod outer segments. We conclude that a drop in around photoreceptors, which persists during illumination and reduces a transretinal Ca2+ gradient, is the combined effect of the light-evoked SRS volume increase, Ca2+ release from photoreceptors, and an unidentified mechanism(s), which is presumably Ca2+ transport by the RPE. The relatively rapid onset and offset of the decrease remains unexplained. These steady-state shifts in should have significant effects on photoreceptor function, especially adaptation.

Rod on-bipolar cell light responses are mediated by a class of metabotropic glutamate receptor which is coupled via a G-protein to the control of a cGMP cascade, with cGMP acting to open cation channels, whilst off-bipolar cells possess ionotropic glutamate receptors. Whole-cell voltage-clamp recordings were obtained from on- and off-bipolar cells of dark-adapted dogfish retinal slices, identified by their light responses. Isolated cells were exposed to concentration-jumps of glutamate. At negative voltage-clamp potentials, on-bipolar cells responded to glutamate with outward currents with a mean delay of 10.8 ms, whilst off-bipolar cells responded with inward currents without any delay. Neither cell type showed desensitization to applied steps of glutamate. The dose-response relation for on-bipolar cells showed no gradual saturation, but increased linearly with a sharp cutoff above 200 μM glutamate. This dose-response relation could be fitted with a theoretical expression assuming Michaelis-Menten kinetics for the action of glutamate on receptors and a linear relation between the concentration of receptors bound to glutamate and the fall in cGMP this induces. The dose-response relation of off-bipolar cells showed saturation with a limiting slope of 2 at low glutamate concentrations, suggesting that two molecules of glutamate are required to open each channel by a cooperative mechanism. The glutamate receptor coupled cGMP cascade of rod on-bipolar cells can account for high synaptic voltage gain.

We examined the laminar distribution of corticogeniculate neurons in the macaque striate cortex labeled by axonal transport following injections of retrograde tracers into the lateral geniculate nucleus (LGN). Large injections of retrograde tracers involving all layers of the LGN resulted in a distinctive bilaminar distribution of labeled cells in cortical layer 6. One tier of labeled neurons was located along the layer 5–6 border and a second was located near the bottom of the layer, leaving the middle of layer 6 largely free of labeled neurons. Following injections of tracers that were restricted to the magnocellular layers of the LGN, almost all of the labeled neurons were located in the lower tier. In contrast, following injections of retrograde tracers confined to the parvocellular layers of the LGN, labeled cells were found in both tiers, with the greatest number in the upper tier. Thus, layer 6 of macaque striate cortex consists of three distinct sublayers only two of which are the source of descending projections to the LGN: an upper tier that projects exclusively to the parvocellular layers and a lower tier that projects to both magnocellular and parvocellular layers.

Both A- and B-type horizontal cells in the rabbit retina were labeled by brief in vitro incubations of the isolated retina in the blue fluorescent dye 4,6–diamino-2–phenylindole. Intracellular injection of Lucifer Yellow into the somata revealed the morphology of the individual cells. Dye-coupling with Lucifer Yellow was seen only between A-type horizontal cells. By contrast, injection of the tracer Neurobiotin showed dye-coupling between both A- and B-type horizontal cells. There also appeared to be coupling between the axon terminals of B-type horizontal cells.

The extensive dye-coupling seen following injection of Neurobiotin into a single horizontal cell soma can be used to obtain population counts of each cell type. Staining of large numbers of each cell type across the retina showed that each type increased in number and declined in dendritic diameter as the visual streak was approached, such that relatively constant coverage across the retina was maintained. In the visual streak, A-type horizontal cells numbered 555 cells/mm2 and averaged 120 μm in diameter, compared to 1375 cells/mm2 and 100 μm for B-type horizontal cells. In the periphery, the A- and B-types numbered 250 cells/mm2 and 400 cells/mm2, respectively. The average diameters of the dendritic trees at these locations were 225 μm for the A-type and 175 μm for the B-type. Coverage across the retina averaged almost six for A-type horizontal cells and 8–10 for the B-type. A-type horizontal cells in the visual streak whose elliptical dendritic fields were shown by Bloomfield (1992) to correlate physiologically with orientation bias were shown to be dye-coupled to cells with symmetrical dendritic fields.

The immunocytochemical labeling of arrestin (S-antigen) in photoreceptors of the ovine retina was examined following incubation of the retina with hydroxylamine (NH2OH), an agent known to inhibit the phosphorylation of photoactivated rhodopsin. Intact, isolated retinas bathed in medium containing 20 mM NH2OH, or in control medium lacking NH2OH, were maintained in darkness or exposed to bright light for 3 min (dark-adapted and light-adapted conditions, respectively); further incubated in darkness for 10 min; and then fixed and prepared for cryosectioning. Cryosections were incubated with anti-S-antigen monoclonal antibody MAb A2G5; with secondary antibodies that were conjugated with horseradish peroxidase; and with either 3–amino-9–ethyl carbazole or diaminobenzidine as chromogen. Anti-arrestin labeling in cryosections was then analyzed densitometrically using a light-microscopic image processing system. In dark-adapted control retinas, labeling density of the photoreceptor outer segment (OS) layer (0.061 ± 0.004; average ± S.e.m.) was less than that of the inner segment (IS) layer (0.138 ± 0.011). In light-adapted control retinas, OS labeling density (0.139 ± 0.007) exceeded IS labeling density (0.095 ± 0.005). Incubation with NH2OH eliminated this light-dependent increase in labeling of the OS relative to that of the IS, i.e. eliminated the increase in relative OS/IS labeling. Densities of labeling were 0.110 ± 0.006 (OS) and 0.183 ± 0.006 (IS) in NH2OH-treated dark-adapted retinas vs. 0.078 ± 0.004 (OS) and 0.182 ± 0.008 (IS) in NH2OH-treated light-adapted retinas. Anti-arrestin labeling was also examined in retinas that were exposed to 3 min or 13 min of bright light and then immediately fixed. Among retinas incubated in the absence of NH2OH, an increase in OS/IS labeling density was evident after 3 min of illumination, and retinas illuminated for 13 min exhibited an even larger increase in OS/IS labeling. An increase in OS/IS labeling was also exhibited by NH2OH-treated retinas that had been illuminated for 3 min; by comparison with dark-adapted NH2OH-treated controls (average value of OS/IS labeling: 0.60), OS/IS labeling in these illuminated retinas was 0.97. However, OS/IS labeling in NH2OH-treated retinas that had been illuminated for 13 min (average value: 0.35) was lower than that of the dark-adapted controls. The results indicate that, within intact rods, NH2OH inhibits the light-dependent increase in OS/IS anti-arrestin labeling that is ordinarily expressed at long times (~10 min) after major bleaching of the visual pigment. Among the possible bases for the effect of NH2OH are a reduction in the driving force for the movement of arrestin from the inner to the outer segment and/or a facilitation of the degradation of arrestin in the outer segment.

The aim of the present study was to determine the number of cone classes in the Xenopus retina. We examined the dimensions and staining properties of cones, utilizing two monoclonal antibodies, COS-1 and OS-2, developed by Szel and Rohlich (1985). Living cones also were reacted with the plant lectins peanut agglutinin (PNA) and wheat germ agglutinin (WGA) and with a fluorescent stilbene dye, DIDS, which binds selectively to red-sensitive cones (Kleinschmidt, 1991; Kleinschmidt & Harosi, 1992a, b). Three cone populations were distinguished based on differences in size and staining properties. Eighty-eight percent of all cones were stained strongly by COS-1, PNA, and DIDS, but weakly by OS-2. The group of cones stained by COS-1 had the largest mean dimensions of outer segment length, width, and oil droplet diameter. COS-1 negative cones were divisible into two groups: a subclass of miniature cones (approximately 4% total cones) was stained strongly by OS-2, PNA, and DIDS. The balance, constituting approximately 9% total cones, were of intermediate size, were not stained by PNA and reacted weakly to OS-2 and DIDS. WGA stained all cones. Large, COS-1 + cones appear to be red-sensitive and belong to the class of anion-tunable cone pigments. We suggest that the intermediate size, COS-1 negative cones are blue-sensitive based on the finding that blue-sensitive chromatic horizontal cells connect to them preferentially (Witkovsky et al., work in progress). The remaining class of miniature cones may be UV-sensitive, since another amphibian, the salamander, has been shown to possess red-, blue-, and UV-sensitive cones (Perry & McNaughton, 1991).

Inferior temporal (IT) cortex is a “high-order” region of extrastriate visual cortex important for visual form perception and recognition in adult primates. The pattern of cortical afferents from both ipsilateral and contralateral hemispheres to anterior IT cortex was determined in infant macaque monkeys 7–18 weeks of age following injections of wheat-germ agglutinin-HRP. Within the ipsilateral hemisphere, the locations and laminar distribution of labeled cells were similar to those observed after comparable injections in adult monkeys. Specifically, ipsilateral afferents derived from visual areas V4, TEO, anterior and posterior IT, and STP, from parahippocampal, perirhinal, and parietal zones, and from several anterior zones including lateral and ventral frontal cortex, the insula, and cingulate cortex. Within the contralateral hemisphere, we observed labeled cells in homotopic regions of IT and in parahippocampal and perirhinal areas, as has been reported for adult monkeys. However, we also identified additional contralateral regions not previously known to provide input to anterior IT, including lateral and ventral frontal cortex, cingulate cortex, and STP. Overall, the strongest and most widespread projections from outside the temporal lobe were found in the youngest monkey, suggesting that some of these projections may represent transient circuitry necessary for the development of complex visual response properties in anterior IT.

The role of excitatory amino acid (EAA) receptors in primate retinal ganglion cell function was analyzed in a superfused retina-eyecup preparation using single-unit, extracellular recording techniques. The effects of bath applied L-2–amino-4–phosphonobutyrate (APB), N-methyl-D-aspartate (NMDA), and non-NMDA EAA receptor agonists and antagonists were examined on the light-evoked responses and resting firing rates of ganglion cells. APB (30–100 μM) reduced or blocked the light-evoked responses and resting firing rates of all ON-center ganglion cells; higher doses of APB (100 μM) were required to block the light-evoked responses of ON-transient cells. In contrast, an increase in resting firing rates was observed when L-APB was applied to some OFF-center ganglion cells. The EAA agonists kainate (KA) (10–20 μM) and NMDA (200–350 μM) increased the firing rate of virtually all ganglion cells examined. Quisqualate (10–20 μM) increased firing in most cells, but occasionally (4/13 cases) produced inhibition. The NMDA antagonist D-amino-phosphono-heptanoic acid (D-AP7) (200–250 μM) reduced the light-evoked responses of ganglion cells by an average of 12% from control levels, while resting firing rates declined 37%. In the presence of D-AP7, the basic receptive-field characteristics of cells were not significantly altered. In contrast, two non-NMDA receptor antagonists, NBQX (2,3–Dihydroxy-6–nitro-7–sulfamoyl-benzo-(F)-quinoxalinedione) and DNQX (6,7–dinitro-quinoxaline-2,3–dione), produced substantial reductions in the light-evoked responses (82%) and resting firing rates (87%) of all ganglion cell classes. A striking observation in some neurons was the recovery of a persistent transient light-evoked response in the presence of NBQX. This NBQX-insensitive, light-evoked response was always blocked by adding D-AP7. Thus, neurotransmission from bipolar to ganglion cells in primates is mediated predominantly by non-NMDA EAA receptors, with NMDA receptors forming a minor component of the light-evoked response.

The hypothesis that a large, possibly toxic, increase in cellular calcium accompanies photoreceptor cell degeneration in several different Drosophila mutants was tested. The calcium content of wild type and mutant photoreceptors of Drosophila was measured using rapid freezing of the eyes and energy-dispersive x-ray analysis (e.d.x.) of cryosections and semithin sections of cryosubstituted material. Light- and dark-raised mutants of the following strains were studied: retinal degeneration B (rdgB); retinal degeneration C (rdgC); neither inactivation nor afterpotential C (ninaC), and no receptor potential A (norpA). These are light-dependent retinal degeneration mutants in which the affected gene products had been previously shown as myosin-kinase (ninaC), calcium-dependent phosphoprotein phosphatase (rdgC), phosphoinositide transfer protein (rdgB), and phospholipase C (norpA). In light-raised mutants, ommatidia of variable degrees of degeneration were observed. Mass-dense globular bodies of 200–500 nm diameter in relatively large quantities were found in the degenerating photoreceptor of all the mutants tested. These subcellular globules were found to have a very high calcium content, which was not found in wild type or in nondegenerating photoreceptors of the mutants. Nondegenerating photoreceptors were found not only in dark-raised mutants, but in smaller quantities also in light-raised mutants. Usually these globular structures contained high levels of phosphorus, indicating that at least part of the calcium in the mutant photoreceptors is precipitated as calcium phosphate. The results indicate that a large increase in cellular calcium accompanies light-induced photoreceptor degeneration in degenerating Drosophila mutants even when induced by very different mutations, suggesting that the calcium accumulation is a secondary rather than a primary effect in the degeneration process.

To provide a quantitative description of the postnatal development of dendritic trees in alpha ganglion cells of the rabbit retina, these cells were stained either by intracellular injection of Lucifer yellow or by application of the lipophilic dye Oil. This was done at three developmental stages: postnatal day (P) 8/9, P 16/17, and in adults. For different retinal locations we quantified the alpha cell dendritic field area, the number of dendritic branch points, and the average dendritic length between branch points. According to the alpha cell location, the data were collected in three groups representing the retinal center, midperiphery, and far periphery, respectively. The data were then correlated with the postnatal retinal expansion which is known to differ among the above topographic regions of the retinae (Reichenbach et al., 1993). Our results show that the growth of alpha ganglion cell dendrites is not proportional to, but significantly exceeds, that of the local retinal tissue. Between P 8/9 and adulthood, the area of central alpha cells increases almost six-fold from 26,000 to 144,000 μm2 (retinal expansion: 2.2-fold), and that of peripheral cells more than 15-fold from 35,000 to 556,000 μm2 (retinal expansion: four-fold). During this period, the coverage factor of alpha cell dendritic fields increases about three-fold, and reaches adult levels of about 3 (retinal center) and 2.2 (periphery), respectively. The number of dendritic branch points remains nearly constant, and the distance between them increases by a factor close to the square root of the factor by which the dendritic field area grows. Thus, it appears that, from the second postnatal week on, dendritic trees of rabbit alpha ganglion cells increase by intense “interstitial growth,” rather than by outgrowth of (new) dendritic branches. This growth pattern is different from that of some other rabbit retinal ganglion cell types, and of alpha ganglion cells of the cat retina, whose dendritic trees expand at a rate equal to or less than that of the surrounding retinal tissue. The consequences for synaptic contacts with bipolar and amacrine cells are discussed; they suggest a high degree of synaptic plasticity during normal postnatal retinal growth.

The optical alignment of individual cuticular cones in the dioptric array of the lateral eye of Limulus polyphemus was determined with a precision two-circle goniometer constructed and mounted to the stage of a compound microscope and using a new formaldehyde-induced fluorescence procedure. All measurements were made from the corneal surface of the excised eye mounted in seawater through an air/water interface perpendicular to the optic axis of the microscope. Our results revealed two variants of visual field and eye curvature which can actually be discriminated in casual examination of adult animals. We call animals possessing these two variants “morlocks” and “eloi.” Adult male and female morlocks about 25 cm across the carapace have eyes which are relatively elongated, often darker in pigmentation, smaller, and relatively flatter in curvature. Morlocks have a monocular field of view of about 3.13 steradians or 50% of a hemisphere. The coverage averages 115 deg along the vertical axis and 168 deg along the horizontal axis of the eye, with maximum resolution in the anteroventral quadrant. Adult male and female eloi of comparable size have eyes which are relatively more round, often lighter in pigmentation, larger with more ommatidia, and relatively more bulged. Eloi have a monocular field of view of approximately 3.83 steradians or 61% of a hemisphere that covers 145 deg vertically and 185 deg horizontally. Eloi have more uniform resolution than morlocks with best resolution in the posteroventral quadrant. All horseshoe crabs examined, whether morlocks or eloi, have an identical orientation of the margin of the eye relative to the animals’ coordinates.

The distribution of acetylcholinesterase (AChE) in the lateral eye and brain of the horseshoe crab was investigated with histochemical means using standard controls to eliminate butyrylcholinesterase and nonspecific staining. Intense staining was observed in the neural plexus of the lateral compound eye, in the lateral optic nerve, and in various neuropils of the brain. Nerve fibers with moderate to weak staining were widespread in the brain. No sornata were stained in either the lateral eye or the brain. The distribution of acetylcholinesterase in the supraesophageal ganglia and nerves of the giant barnacle was also investigated for comparison. Although both the median optic nerve of the barnacle and the lateral optic nerve of the horseshoe crab appear to contain the fibers of histaminergic neurons, only the lateral optic nerve of the horseshoe crab shows AChE staining. Other parts of the barnacle nervous system, however, showed intense AChE staining. These results along with the histochemical controls eliminate the possibility that some molecule found in histaminergic neurons accounted for the AChE staining but support the possibility that acetylcholine might be involved as a neurotransmitter in lateral inhibition in the horseshoe crab retina. Two reasonable neurotransmitter candidates for lateral inhibition, histamine and acetylcholine, must now be investigated.

A 6.4–kb sequence upstream of the mouse blue visual pigment gene has been assayed in transgenic mice for the ability to direct cell-type-specific expression of a linked β-galactosidase (lacZ) reporter. The construct is expressed specifically in cone photoreceptors in three independent lines. Transgene expression is found in the developing retina on the first postnatal day, increases rapidly in subsequent days, and persists through adulthood. A gradient of transgene expression is observed across the retina, with the transgene-expressing cones found almost exclusively in the lower retina and rarely in the upper retina, a pattern that parallels the distribution of blue cones in the mouse retina. Double-labeling with anti-cone pigment antibodies shows that transgene expression is confined to blue cones. These results imply that all of the sequence elements necessary for the control of blue cone-specific expression are encoded within the 6.4–kb DNA fragment tested.

The neurotransmitter used by horizontal cells in mammals has not been identified. GABA has been the leading candidate, but doubt has remained because of failure to clearly demonstrate the GABA synthetic enzyme, glutamic acid decarboxylase (GAD) in these cells. Because GAD was recently shown to exist as two isoforms, 65 kDa and 67 kDa, we considered whether there might be a mismatch between the forms of GAD expressed in horizontal cells and the probes used to detect it. Accordingly, we stained sections of mammalian retina with antibodies specific for each isoform. Cat horizontal cells of both types (A and B) were immunoreactive for GAD67 but negative for GAD65; monkey horizontal cells of both types (H1 and H11) were positive for GAD65 and negative for GAD67. The findings reconcile previous, apparently conflicting, observations and strengthen considerably the hypothesis that mammalian horizontal cells are GABAergic.

Immunohistochemical and in situ hybridization techniques were used to localize the β4 subunit of the neuronal nicotinic acetylcholine receptors (nAChRs) in the ground squirrel retina. The β4 nAChR subunit was detected in both transverse and horizontal sections of the retina using a subunit-specific antiserum and the avidin-biotin complex technique. Two bands of labeled processes were seen in the inner plexiform layer, corresponding approximately to the laminae where the cholinergic cells arborize. Labeled cells were found in the ganglion cell layer and the inner third of the inner nuclear layer. The cells in the ganglion cell layer were medium- to large-sized and were frequently observed to give rise to axon-like processes. Most of the labeled neurons in the inner nuclear layer were small presumptive amacrine cells, but a few medium-to-large cells were also labeled. These could constitute a different class of amacrine cells or displaced ganglion cells. The latter possibility is supported by the existence of nAChR-containing displaced ganglion cells in the avian retina. In situ hybridization with a 35S-labeled cRNA probe revealed the expression of mRNA coding for the nAChR β4 subunit in the ganglion cell layer and the inner third of the inner nuclear layer. This finding confirmed the immunohistochemical data of the cellular localization of β4 nAChR subunit.

These results indicate that the β4 nAChR subunit is expressed by specific subtypes of neurons on the ground squirrel retina. As the expression of that particular nAChR subunit appears to be very limited in the brain, the present data suggest that the retina might represent a useful model to study the function of nAChRs containing the β4 subunit.

The presence of inhibitory bipolar cells in salamander retina was investigated by a comparative analysis of the distribution of glutamate- and GABA-immunoreactivities (GLU-IR; GABA-IR) using a postembedding immunocytochemical method. GLU-IR was found in virtually all photoreceptors, bipolar cells and ganglion cells, neuronal elements that transfer information vertically through the retina. GLU-IR also was found in numerous amacrine cells in the mid and proximal inner nuclear layer as well as in the cytoplasm of horizontal cells, while the nucleus of horizontal cells was either lightly labeled or not labeled at all. GLU-IR was found in the outer plexiform layer and intensely in the inner plexiform layer, in which there was no apparent sublamination. Forty-seven percent of Type IB bipolar cells in the distal inner nuclear layer and 13% of the displaced bipolar cells were GABA-IR. All bipolar cells were also GLU-IR, indicating that GABA-IR bipolar cells were a subset of GLU-IR bipolar cells rather than a separate population. About 12% of the Type IB bipolar cells were moderately GABA-IR and likely comprised a GABAergic subtype. GLU-IR levels in the presumed GABAergic bipolar cells were higher than in other purely GLU-IR bipolar cells suggesting that these GABA-IR bipolar cells are glutamatergic as well. All of the displaced bipolar cells were only lightly GABA-IR, indicating that displaced bipolar cells comprise a more homogeneous class of glutamatergic cell than orthotopic bipolar cells. GAD-IR co-localized with GABA-IR in orthotopic but not displaced bipolar cells, further supporting the idea that some orthotopic bipolar cells are GABAergic. A small proportion of bipolar cells in salamander retina contain relatively high levels of both GABA and glutamate. Co-release of these substances by bipolar cells could contribute to the “push-pull” modulation of ganglion cell responses.

Endogenous dopamine release in the retina of the African clawed frog (Xenopus laevis) increases in light and decreases in darkness. The roles of the inhibitory amino acid transmitters gamma-aminobutyric acid (GABA) and glycine in regulating this light/dark difference in dopamine release were explored in the present study. Exogenous GABA, the GABA-A receptor agonist muscimol, the GABA-B receptor agonist baclofen, and the GABA-C receptor agonist cis-aminocrotonic acid (CACA) suppressed light-evoked dopamine overflow from eyecups. The effects of GABA-A and -B receptor agonists were selectively reversed by their respective receptor-specific antagonists, whereas the effect of CACA was reversed by the competitive GABA-A receptor antagonist bicuculline. The benzodiazepine diazepam enhanced the effect of muscimol on light-evoked dopamine release. Both GABA-A and -B receptor antagonists stimulated dopamine release in light or darkness. Bicuculline was more potent in light than in darkness. These data suggest that retinal dopaminergic neurons are inhibited by GABA-A and -B receptor activation in both light and darkness but that GABA-mediated inhibitory tone may be greater in darkness than in light.

Exogenous glycine inhibited light-stimulated dopamine release in a concentration-dependent and strychnine-sensitive manner. However, strychnine alone did not increase dopamine release in light or darkness, nor did it augment bicuculline-stimulated release in darkness. Additionally, both strychnine and 7-chlorokynurenate, an antagonist of the strychnine-insensitive glycine-binding site of the N-methyl-D-aspartate subtype of glutamate receptor, suppressed light-evoked dopamine release. Thus, the role of endogenous glycine in the regulation of dopamine release remains unclear.

To determine whether central communication of retinal signals is necessary for the development of an experimentally induced myopia, tree shrews were exposed to monocular deprivation (MD) while the action potentials of retinal cells in the deprived eye were blocked with intravitreally injected tetrodotoxin (TTX-MD animals). TTX injections (0.6 μ 3 μL) and MD began about 15 days after eye opening, at the start of the susceptible period for the development of lid-suture myopia. Six injections were given, one every second day to produce 12 days of MD and TTX-blockade. Control TTX animals (TTX-open) received TTX in one eye, but not MD, on the same injection schedule and were always found to be behaviorally unresponsive to visual stimuli through the injected eye indicating that TTX blocked central communication of action potentials. Other control animals received intravitreally injected saline in either an open eye (saline-open), or an MD eye (saline-MD). A sham-injected group (sham-inj-MD) received MD and all anesthetic and surgical manipulations except for penetration of the sclera. In all groups, one eye in each animal was an untreated control.

Two effects were found. All MD groups, including the TTX-MD animals, developed a significant vitreous chamber elongation in the deprived eye, indicating that an experimental myopia developed despite ganglion cell blockade. Thus, retinal mechanisms in tree shrew can detect the presence of a degraded visual image and produce an experimental myopia that does not depend on the receipt of visual messages by central neural structures. In addition, eyes in which the sclera was punctured had smaller vitreous chamber depths than comparable uninjected eyes, indicating that puncturing the sclera reduced the normal elongation. These data suggest that forces within the eye normally contribute to its expansion and may be resisted by the choroid and/or the sclera.