The purpose of this study was to evaluate the use of slow multifocal m-sequence stimulation in analyzing the topographic distribution and underlying mechanisms (including nonlinearities) of the retinal oscillatory potentials (OPs). In giving us access to the response topography and the nonlinear characteristics of the OP,, the m-sequence technique provides us with two important means for the identification and characterization of the signal sources. In this study, we analyzed the OPs into the first- and second-order components and investigated their topographies and luminance dependence. The distribution of both the first- and second-order OP components differed significantly from that of the nicker ERG investigated by Sutter and Tran (1992). At eccentricities and luminance levels favoring activity by both rods and cones, the second-order OPs were particularly prominent showing the most clear-defined and complex waveform The topographic distribution of the second-order OPs showed combined features of both rod and cone distributions On a strong rod-bleaching background the second-order OPs were eliminated and the first-order OPs showed a reduced amplitude and a shifted latency These results are consistent with the notion that the second-order component of the OPs is dominated by contributions from rod-cone interactions.

A neural network model is developed to explain how visual thalamocortical interactions give rise to boundary percepts such as illusory contours and surface percepts such as filled-in brightnesses. Top-down feedback interactions are needed in addition to bottom-up feed-forward interactions to simulate these data. One feedback loop is modeled between lateral geniculate nucleus (LGN) and cortical area V1, and another within cortical areas V1 and V2. The first feedback loop realizes a matching process which enhances LGN cell activities that are consistent with those of active cortical cells, and suppresses LGN activities that are not. This corticogeniculate feedback, being endstopped and oriented, also enhances LGN ON cell activations at the ends of thin dark lines, thereby leading to enhanced cortical brightness percepts when the lines group into closed illusory contours. The second feedback loop generates boundary representations, including illusory contours, that coherently bind distributed cortical features together. Brightness percepts form within the surface representations through a diffusive filling-in process that is contained by resistive gating signals from the boundary representations. The model is used to simulate illusory contours and surface brightnesses induced by Ehrenstein disks, Kanizsa squares, Glass patterns, and cafe wall patterns in single contrast, reverse contrast, and mixed contrast configurations. These examples illustrate how boundary and surface mechanisms can generate percepts that are highly context-sensitive, including how illusory contours can be amodally recognized without being seen, how model simple cells in V1 respond preferentially to luminance discontinuities using inputs from both LGN ON and OFF cells, how model bipole cells in V2 with two colinear receptive fields can help to complete curved illusory contours, how short-range simple cell groupings and long-range bipole cell groupings can sometimes generate different outcomes, and how model double-opponent, filling-in and boundary segmentation mechanisms in V4 interact to generate surface brightness percepts in which filling-in of enhanced brightness and darkness can occur before the net brightness distribution is computed by double-opponent interactions.

The cholinergic identity of retinal starburst amacrine neurons is well established, but recent evidence suggests that these cells are GABAergic as well. Confirmation of this dual transmitter function requires the demonstration of glutamate decarboxylase (GAD), the biosynthetic enzyme for GABA, within starburst cells. The current work was undertaken to determine whether rabbit retinal starburst amacrine neurons contain either of the two known isoforms of GAD. To do this, we have examined the localization of the following: (1) the 65-kDa isoform of GAD; (2) the 67-kDa isoform of GAD; (3) choline acetyltransferase; and (4) the fluorescent dye DAPI, a marker for cholinergic amacrine cells. In addition, we labeled displaced starburst neurons directly, by injecting them with Lucifer Yellow in vitro. Four strata within the inner plexiform layer contained immunoreactive GAD65. A non-GAD65-immunoreactive zone separated the two innermost strata (G3 and G4); this zone contained (1) the dendrites of individual Lucifer Yellow-injected, displaced starburst amacrine cells; (2) dendrites immunoreactive for choline acetyltransferase; and (3) processes of DAPI-labeled amacrine cells. Immunoreactive GAD67 appeared in the same strata that contained GAD65, and in at least two additional strata, one of which lay at precisely the same depth as the proximal cholinergic stratum. In addition, the somas of displaced starburst cells were strongly immunoreactive for GAD67, but not for GAD65. These results demonstrate (1) that displaced starburst amacrine cells contain the 67-kDa isoform of GAD, but not the 65-kDa isoform; and (2) that the dendrites of starburst (67-kDa GAD) amacrines, and the dendrites of 65-kDa-GAD-containing amacrines, occupy different strata within the inner plexiform layer. Thus, displaced starburst cells do contain GAD, and can, presumably, manufacture GABA. The reasons for their preferential use of the 67-kDa GAD isoform remain to be elucidated.

We have developed a new model of phototransduction that accounts for the dynamics of primate and human cone flash responses in both their linear and saturating range. The model incorporates many of the known elements of the phototransduction cascade in vertebrate photoreceptors. The input stage is a new analytic expression for the activation and inactivation of cGMP-phosphodiesterase (PDE). Although the Lamb and Pugh (1992) model (of a delayed ramp for the rising phase of the PDE* response in amphibian rods) provided a good fit for the first 2 log units of stimulus intensity without parameter adjustments, the remaining 4 log units of the data required nonlinear modifications of both delay and gain (slope). We show that this nonlinear behavior is a consequence of the delay approximation and develop a completely linear model to account for the rising phase of amphibian rod photocurrent responses over the full intensity range (~6 log units). We use the same dynamic model to account for primate cone responses by decreasing the time constants of PDE activation and introducing an enhanced inactivation process. This PDE* response activates a nonlinear calcium feedback stage that modulates guanylate cyclase synthesis of cyclic GMP. By adjustment of the throughput and feedback parameters, the full model successfully captures most of the features of the primate and human cone flash responses throughout their dynamic range. Our analysis suggests that initial processes in the transduction cascade may be qualitatively different from comparable processes in rods.

Centrifugal fibers innervating the retina have been shown in all classes of vertebrate, except for mammals where conventional tract-tracing methods have not been able to unmistakably demonstrate their existence. In a previous study, a unilateral, intravitreal injection of 5, 7-dihydroxytryptamine was used to reveal indoleamine-accumulating centrifugal fibers which were visualized by an immunoreaction against serotonin. In the present study, I employed a modification of this method to stain retinopetal neurons in the rat. Terminals were located preferentially in the outer retina; labeled fibers could be traced back along an ipsilateral pathway to somata in the dorso-caudal portions of the chiasm or the medio-lateral preoptic area, and thence towards the suprachiasmatic nuclei. The unique beaded appearance of the fibers distinguishes them from retinal ganglion cell axons. The labeling of central cell bodies strongly suggests that they possess terminals in the retina. Thus, at least some mammalian retinas receive centrifugal innervation. This indoleamine-accumulating retinopetal pathway may be involved in retinal melatonin synthesis, coordination of circadian rhythms, and interocular phenomena.

The cholinergic circuit within the tectum and the cholinergic input from the nucleus isthmi mediate a presynaptic augmentation of retinotectal transmitter release via nicotinic receptors. In this study, the cholinergic systems were either eliminated using the cholinergic neurotoxin AF64A or blocked using nicotinic antagonists to test for effects on the activity-driven sharpening of the regenerating retinotectal projection. The effectiveness of the AF64A was verified by recording field potentials elicited by optic tract stimulation and by immunohistochemical staining for choline acetyltransferase (ChAT). At 1 week after intracranial (IC) injection of AF64A (12 to 144 nmoles) into the fluid above the tectum, field potentials showed a selective dose-dependent decrement of the cholinergic polysynaptic component with no effect on the amplitude of the glutamatergic monosynaptic component. The decrement was only partially recovered in recordings at 2 and 6 weeks. In normal fish, the ChAT antibody stains a population of periventricular neurons, their apical dendrites, and a dense plexus within the optic terminal lamina that consists of their local axons and fine dendrites and of input fibers from the nucleus isthmi. One week after IC AF64A injection (48–72 nmoles), most immunostaining in superficial tectum was lost but most neuronal somas in the deep tectum could still be seen, and staining in the tegmentum below the tectum was completely intact. At 2 weeks and later, the staining of neuronal somata largely recovered, but staining of the superficial plexus did not. AF64A treatment at 18 days after nerve crush, when regenerating retinal fibers are beginning to form synapses, prevented retinotopic sharpening of the projection. Recordings showed a rough retinotopic map on the tectum but the multiunit receptive fields (MURFs) at each tectal point averaged 34 deg vs. 11 deg in vehicle-injected control regenerates. AF64A treatment before nerve crush also blocked sharpening, ruling out a direct effect on retinal growth cones or retinal fibers, as AF64A rapidly decomposes, whereas its effect on the cholinergic fibers is long-lasting. IC injection or minipump infusion of the nicotinic antagonists α-bungarotoxin (αBTX), neuronal bungarotoxin (nBTX), and pancuronium during regeneration also prevented sharpening (MURFs averaging 29.4 deg, 33.0 deg, and 31.4 deg, respectively). Control Ringer≈s solution infusions or injections over the same period (19–37 days postcrush) had no effect on regenerated MURF size (11.7 deg). The results show that the cholinergic innervation, which modulates transmitter release, is required for activity-driven retinotopic sharpening, thought to be triggered by NMDA receptor activation.

Surround enhancement (sensitization) is a poorly understood form of network adaptation in which the kinetics of the responses of retinal neurons to test stimuli become faster, and absolute sensitivity of the responses increases with increasing level of steady, surrounding light. Surround enhancement has been observed in all classes of retinal neurons in lower vertebrates except cones, in some primate retinal ganglion cells, and in human psychophysical studies. In theory, surround enhancement could be mediated by two broad classes of mechanisms, which are not mutually exclusive: one in which the kinetics of the transduction linking cone voltage to postsynaptic current in second-order neurons is modulated, and another in which the transformation of postsynaptic current to membrane voltage is modulated. We report here that both classes of mechanism play a role in surround enhancement measured in turtle horizontal cells (HCs). We stimulated the retina by modulating sinusoidally the illuminance of a bar placed at various positions in the HC receptive field. The bar was surrounded by either equally luminant or dim, steady light. Interpretation of responses in the context of a model for the cone-HC network led to the conclusion that the speeding up of response kinetics —due to selective increase in response gain at high temporal frequencies — by surround illuminance is almost completely accounted for by the change in the kinetics of the transduction linking cone membrane potential to HC postsynaptic current. However, surround illuminance also had an additional, surprising effect on the transformation between postsynaptic current and voltage: the space constant for signal spread in the HC network for the dim-surround condition was roughly twice as large as that for the bright-surround condition. Thus, increasing surround illuminance had analogous effects in the spatial and temporal domains: it restricted the time course and the spatial spread of signal. Both effects were dependent on the contrast between the mean bar illuminance and that of the surround, rather than on overall light level. When the stimulus with the bright surround was dimmed uniformly by a neutral density filter, the space constant did not increase, and response gain at high temporal frequencies did not decrease. Pharmacological experiments performed with dopamine and various agonists and antagonists indicated that, although exogenous dopamine can influence surround enhancement, endogenous dopamine does not play an important role in surround enhancement. We conclude that contrast in background light modulates the spatiotemporal properties of signal processing in the outer retina, and does so by a non-dopaminergic mechanism.

When intense adapting lights are turned off, the pupil of the turtle, Pseudemys scripta elegans, enlarges. The recovery functions for pupillary dilation have different time constants that are defined by red- and green-sensitive cones and rods as they are affected by prior light adaptation and time in the dark. Pupillary area related to dilation responds over at least a three- to four-fold range. Following white-light adaptation, the course of pupil dilation in the dark shows a three-legged curve of differing time constants. With spectral-light adaptations, the contributions of separate classes of photoreceptors can be isolated. Red- and green-sensitive cones contribute shorter time constants of 3.31 and 3.65 min to prior white-light adaptation—4.81 and 4.18 min to prior spectral-light adaptations. Rods contribute a much longer time constant of 6.69 min to prior white-light adaptation—7.60 min to prior spectral-light adaptation. The ratios are in keeping with the flash sensitivities of photoreceptors in this same animal, as well as with psychophysical visual threshold mechanisms of color sensitivity.

The physiological actions of biogenic amine and amino-acid neurotransmitters are terminated by their removal from the synaptic cleft by specific high-affinity transport proteins. The members of the Na+- and Cl−-dependent neurotransmitter transporter family expressed in bovine retina and responsible for the uptake of biogenic amine and amino-acid neurotransmitters were identified using a reverse transcriptase-polymerase chain reaction-based approach. cDNA clones encoding bovine homologues of glycine (GLYT-1), γ-aminobutyric acid (GAT-1) creatine (CreaT), and orphan (NTT4) transporters were identified using this strategy. The expression pattern of mRNAs encoding these proteins in the retina was determined by in situ hybridization histochemistry GLYT-1 CreaT NTT4 and GAT-1 mRNAs were expressed in the retina by cells in the inner nuclear inner plex, iform and ganglion cell layers They were not expressed at detectable levels in the photoreceptor cells whose cell bodies are in the outer nuclear layer and are the most abundant cell type in the retina GLYT-1 mRNA was present exclusively in the proximal inner nuclear layer GAT-1 mRNA was localized to both the inner nuclear and ganglion cell layers CreaT mRNA was expressed in all cell types in the retina except photoreceptors and NTT4 mRNA was expressed by a sub subpoulation of cells in the ganglion cell laver. Elucidation of the expression pattern of these neurotransmitter transporter mRNAs in the retina provides a basis for studies of the role of glycine γ-aminobutyric acid and creatine transporters in retinal function.

We have examined, by light-microscopic immunocytochemistry, the distribution of GABA in the optic nerves of adult rabbits, rats, and cats. Within the optic nerves, immunoreactivity for GABA was restricted to a small subset of axons; some axons were strongly labelled, others weakly labelled, whilst most axons were unlabelled. Glia and other non-neuronal elements were always unlabelled. Our ability to detect GABA in optic nerve axons of adult mammals contrasts with previous reports that indicate a lack of GABA immunoreactivity in such axons. We suggest that this discrepancy may be due to the sensitivity of our immunocytochemical techniques which enable us to detect low concentrations of GABA.

Evidence from several sources indicates that the photoreceptors of rabbit retina include rods, green cones and blue cones, and that blue-green color opponency occurs in select retinal ganglion cells. One of us (Famiglietti) has identified wide-field cone bipolar cells as probable blue-cone-selective bipolars, and type C horizontal cells as possibly connected to blue cones. We wished to extend the analysis of blue cone pathways in rabbit retina and to characterize the topographic distribution of blue and green cones. Two monoclonal antibodies raised against chicken visual pigments are reported to label medium- and long-wavelength cones (COS-1) and short-wavelength cones (OS-2) in all mammalian retinas studied thus far (Szél and colleagues). Using selective labeling with these two antibodies and a nonselective method in nasal and temporal halves of the same retinas, we have found that densities of photoreceptors vary systematically, depending upon the size of the eye and age of the animal. In ‘standard’ New Zealand rabbits of 2–3 kg (2–3 months old), rods reached a peak density of about 300,000/mm2 just dorsal to the visual streak, while cones exhibit peak density at mid-visual streak of about 18,000/mm2. Published measurements of visual acuity in rabbit are less than predicted by this calculation. The ratio of cones to rods is significantly higher in ventral retina, where the density of cones declines to a plateau of 10,000–12,000/mm2, when compared to dorsal retina, where cones are uniformly distributed at a density of about 7000/mm2. The density of OS-2 labeled (presumably “blue”) cones is uniformly low, 1000–1500/mm2, in a wide expanse that includes dorsal retina, the visual streak, and much of ventral retina, except for a region of higher density along the vertical midline. We confirm that there is a far ventral horizontal region near the perimeter that is populated exclusively by a high density (about 13,000/mm2) of OS-2-positive cones (Juliusson and colleagues). This region does not extend to the ventral retinal margin, however, where cone density drops precipitously. Transitional zones between COS-1 and OS-2 labeling, in a region of relatively high and uniform cone density, where sums of COS-1 and OS-2 labeling are higher than expected and in which weakly and strongly labeled cones are intermixed, raise questions about the identities of the visual pigment epitopes, the possibility of double labeling, and therefore the possibility of dual expression of pigments in single cones. The “inverted- T -shaped” topography of higher density OS-2 labeling raises doubts about the significance of a ventral concentration of blue cones for visual function in rabbit retina.

Voltage-gated potassium channels are differentially expressed in the brain, and recent studies have shown that K+ channels show subcellular localization. We characterized the distribution of five different K+ channels in the mouse retina. Each channel was distributed in a unique pattern in the retina and was localized to specific subcellular domains within a given retinal neuron. Kvl.4 and Kv4.2 were consistently found in axonal and somatodendritic portions, respectively, consistent with previous studies in brain. In contrast, Kvl.2, Kvl.3, and Kv2.1 showed variable subcellular distribution depending upon cellular context. These results suggest that no one K+ channel is distributed over the entire length of the neuron to provide a “housekeeping” level of membrane potential stabilization. Instead, we propose that each K+ channel is associated with a specific subcellular functional module, and each local K+ conductance responds uniquely to local voltage and second messenger signals.

The responses of simple cells (recorded from within the striate visual cortex) were measured as a function of the contrast and the frequency of sine-wave grating patterns in order to explore the effect of contrast on the spatial and temporal phase transfer functions and on the spatiotemporal receptive field. In general, as the contrast increased, the phase of the response advanced by approximately 45 ms (approximately one-quarter of a cycle for frequencies near 5 Hz), although the exact value varied from cell to cell. The dynamics of this phase-advance were similar to the dynamics of the amplitude: the amplitude and the phase increased in an accelerating fashion at lower contrasts and then saturated at higher contrasts. Further, the gain for both the amplitude and the phase appeared to be governed by the magnitude of the contrast rather than the magnitude of the response. For the spatial phase transfer function, variations in contrast had little or no systematic effect; all of the phase responses clustered around a single straight line, with a common slope and intercept. This implies that the phase-advance was not due to a change in the spatial properties of the neuron; it also implies that the phase-advance was not systematically related to the magnitude of the response amplitude. On the other hand, for the temporal phase transfer function, the phase responses fell on five straight lines, related to the five steps in contrast. As the contrast increased, the phase responses advanced such that both the slope and the intercept were affected. This implies that the phase-advance was a result of contrast-induced changes in both the response latency and the shape/symmetry of the temporal receptive field.