Tunicamycin (TM), an inhibitor of dolichylphosphate-mediated protein glycosylation, was injected intravitreally into the eyes of diurnal rodents with cone-dominated retinas. Injection of 1 μg of the B2 isomer led to a progressive degeneration of the photoreceptor outer segments and disruption of the RPE-photoreceptor interface that took place over a 10-day period. Cone outer segments were shortened by postinjection day 6 and virtually absent by day 9. The microvilli that normally protrude from the apical surface of the retinal pigment epithelium were replaced by a fringe of shortened processes. The other retinal layers showed no morphological evidence of disruption. Retinal sensitivity, as measured by electroretinographic b-wave threshold, showed a significant and progressive decline over the 10-day course of the experiment that paralleled the disruption of retinal morphology. These results suggest that TM leads to similar morphological and electrophysiological effects on rod and cone photoreceptors.

The sensitivity of rod- and cone-driven responses was studied in the isolated frog retina during the period of rapid dark adaptation following a conditioning flash which bleached a negligible amount of visual pigment. Following a conditioning flash, cone-driven b-wave responses were first enhanced and then depressed. The time courses of the enhancement and subsequent depression of cone-drive responses varies greatly with the intensity and wavelength of the conditioning flash, but were identical when the conditioning' flashes were matched for equal excitation of 502 nm rods. These changes in cone-driven response sensitivity were correlated with the desensitization and recovery of rods following the conditioning flash. When signal transmission from rods to second-order cells was interrupted by the addition of L-glutamate, the conditioning flash did not produce the above-described enhancement and subsequent depression of long-wavelength receptor potential responses. The suppression of cone-driven response therefore appears to be due to a synaptically mediated influence from 502 nm rods which is maximal when the rods are in the dark-adapted state, with little or no contribution from 433 nm rods, and no involvement of the pigment epithelium.

The action spectrum of the ERG b-wave was measured under dark-adapted conditions in intact goldfish (Carassius auratus). It is substantially broader than the absorption spectrum of goldfish rod porphyropsin. Neither prolonged dark adaptation nor removal of possible efferent neural activity affected its shape. Moreover, a 682-nm background did not produce a selective loss of sensitivity to long wavelengths. The results imply that the spectral sensitivity of the b-wave in dark-adapted goldfish reflects the influence of at least two photoreceptor types which act as a single univariant mechanism near absolute threshold.

A method for constructing nonlinear models for light adaptation in the retina is introduced. The components of the models are linear filters and static (instantaneous) nonlinear elements configured in a feedback arrangement. The signals in the models are combined through algebraic addition or multiplication. We apply the method to model light adaptation measured in turtle horizontal cells. Given a particular wiring diagram for the components, the functional forms of the static nonlinearities and frequency responses of the linear filters are determined by constraining the model to give temporal frequency responses (linear regime behavior) consistent with a family of linear feedback models which has been shown to provide a good description of adaptation in these cells. Two particular models, quite different in structure, are presented. Each model responds to perturbations around a mean light level as a feedback circuit in which the gain (strength) of feedback is adjusted to be proportional to the mean light level, but neither model has a separate pathway for measuring the mean light level. Thus, each of these nonlinear models embeds an entire family of linear models parametric in mean light level. Harmonic distortion in the responses of these models to sinusoidal input is found to be qualitatively consistent with physiological data. An alternative class of nonlinear models in which feedback gain is set by a separate slow pathway which tracks the mean light level is ruled out on the basis of its incorrect steady-state input-output behavior. The methods presented can be used to develop specific physical models for light adaptation based on the chemical kinetics of phototransduction or on nonlinear neural feedback. The relevance of the nonlinear models and construction techniques to modeling phototransduction is discussed.

The spectral sensitivities of rods and cones in macaque and human retinas were determined by recording the membrane current from single outer segments. In the macaque retina, the wavelengths of maximum sensitivity were at about 430, 530, and 561 nm for the blue, green, and red cones, respectively, and at 491 nm for the rods. The shapes of the spectra of the three cones were similar when plotted on a log wavenumber scale; the rod spectrum was slightly broader. Spectral sensitivities of the red and green cones from a human retina were virtually identical to those of macaque cones. For comparison with human psychophysical measurements, the rod and cone spectra were adjusted to give the sensitivities expected for light incident on the cornea of the human eye. These functions satisfactorily predicted the scotopic and photopic luminosity functions as well as results from human color-matching experiments. The adjusted spectra of the red and green cones also agreed well with the π-mechanisms of Stiles (1953, 1959).

Efferent fibers from a central circadian clock innervate photoreceptors along the ventral nerve of Limulus and release octopamine when active. We have recorded ERG-like responses from the ventral eye in vivo over several day periods. We have also used intracellular microelectrodes to study changes in ventral photoreceptor function during exogenous applications of octopamine (the putative efferent neurotransmitter), IBMX (a phosphodiesterase inhibitor), and forskolin (an adenylate cyclase activator): (1) Responses to light measured at night from ventral photoreceptors in vivo are greater in amplitude than those recorded during the day; (2) Octopamine and agents that increase intracellular levels of cAMP in ventral photoreceptors decrease the rate of spontaneous (dark) bumps, increase photoreceptor response to light without changing threshold, and often increase the bump duration; and (3) These changes in function of ventral photoreceptors are similar to those that have been observed in the photoreceptor of the lateral eye during circadian clock activity at night, and in vitro in the presence of those same pharmacological agents.

The synaptic organization of dopaminergic interplexiform cells (DA-IPC) in the goldfish retina was studied by a combined double-label electron-microscopical (EM) immunocytochemical/autoradiographical study. DA-IPCs were labeled with antisera against tyrosine hydroxylase. The possibility of synaptic contact with GABAergic amacrine cells in the proximal inner plexiform layer (IPL) was studied by using 3H-GABA uptake. Most synaptic input and output from DA-IPC processes involved amacrine cell processes. In addition, synaptic interactions were observed between DA-IPC processes and bipolar cell terminals, other DA-IPC processes, very small dendrites in the IPL, ganglion cell and optic fiber layers (OFL), and cell bodies in the ganglion cell layer (GCL). Input and output synapses with GABAergic amacrine processes also were observed. Two-thirds of the DA-IPC boutons in the proximal IPL were involved in “junctional appositions,” that is, the junctions appeared to be specialized but they were different than classical chemical synapses. The synaptic organization of DA-IPCs in the goldfish IPL appears to be far more complex than previously thought. Although earlier studies have attempted to explain the action of dopamine in terms of interaction only with amacrine cells, the present study shows that effects involving bipolar cells, other DA-IPCs, unidentified processes and cell bodies in the GCL and OFL must be considered as well.

Using fractions of the protein spectrum of the cat retina as immunogens, we have generated antibodies with substantial specificity for the Müller cells of the retina of cat, rabbit, guinea pig, and rat. The antibodies appear to bind to the filamentous components of the Müller cells and allow demonstration of the pattern of Müller cell endfeet at the inner surface of the retina, best seen in wholemount preparations. In sections and at the edge of wholemount preparations the somas and processes of the cells can be observed. Müller cells are more evenly distributed over the retina than ganglion cells, indicating that their proliferation continues during the differential growth of retina which continues into postnatal life. The morphology and distribution of the endfeet varies with the structures present at the inner surface of the retina. Where the axon bundles are thick, the endfeet are relatively small and are confined to narrow rows between bundles. Müller cell endfeet are also separated widely by large blood vessels. In both situations, it seems likely that Müller cells and astrocytes both contribute, perhaps competitively, to form the glia limitans of the inner surface of the retina. Where the somas of neurones are densely packed in the ganglion cell layer, the endfeet are small and numerous, forming rings around the somas. Where axon bundles, vessels, and somas are sparse, the endfeet appear largest and form a regular array.

Monoclonal antibodies directed against nicotinic acetylcholine receptors (nAChRs) were used to identify and characterize cholinoceptive neurons in the chick retina. Two monoclonal antibodies (mAbs), mAb 210 and mAb 270, stained many neurons in both the inner nuclear layer (INL) and ganglion cell layer (GCL). A class of large labeled cells in the inner INL were positioned at the INL/IPL (inner plexiform layer) border and resembled displaced ganglion cells (DGCs). Their identity was confirmed with injections of rhodamine-labeled microspheres into the ventral tectum and nucleus of the basal optic root (nBOR). Four days after the injection, large nAChR-positive neurons in the inner INL were labeled with beads. The distribution of these cells matched that reported for DGCs in the chicken and pigeon (Reiner et al., 1979; Fite et al., 1981). Many smaller cells in the INL also exhibited nAChR immunoreactivity. These cells were not retrogradely labeled after bead injections into retinal recipient areas. Their processes entered the IPL where they arborized in a band comprised of the inner leaflet of lamina 1 and all of lamina 2. In some instances, a process continued inward to lamina 4. These neurons were tentatively identified as amacrine cells because of their position and branching pattern.

Approximately 12–18% of the cells in the GCL exhibited nAChR immunoreactivity. Many of these cells could be classified as ganglion cells as their axons were also labeled following exposure to nAChR antibodies. Their distribution mirrored that of all ganglion cells with a higher density of cells in the central retina than in the periphery (Ehrlich, 1981). A “double label” technique was used to compare the distribution of nAChR-positive neurons with that of the choline acetyltransferase-positive (ChAT), cholinergic neurons in the chick retina. The two antigens were visualized with two different fluorophores: FITC and RITC. We were unable to find any cells in either the INL or GCL that exhibited both ChAT- and nAChR-like immunoreactivity. The nAChR-positive cells and the ChAT-positive cells both arborized in two bands within the IPL. The patterns were in perfect register in the inner IPL (lamina 4). But, in the outer IPL, the nAChR-positive dendrites were observed in the inner leaflet of lamina 1 and in all of lamina 2 while the ChAT-positive dendrites did not extend into the innermost portion of lamina 2.

The response to spatially focal flicker is enhanced by dim, spatially diffuse, rod-stimulating backgrounds. This effect is called suppressive rod-cone interaction (SRCI) as it reflects a tonic, suppressive influence of dark-adapted rods upon cone pathways which is removed by selective rod-light adaptation. SRCI is observed in amphibian retina with intracellular recordings from most cone-driven cells including the cones themselves, and is most obvious using stimuli flickering at frequencies too rapid for rods to follow. SRCI is blocked by glutamate analogs which selectively block the photic response of horizontal cells (HCs). In the presence of these agents, flicker responses from bipolar cells and cones are enhanced to levels normally seen only with selective rod-light adaptation. In the HCs themselves, SRCI is similarly blocked by lead chloride which blocks rod-, but not cone-related activity.

In amphibian and cat HCs and in human observers, SRCI is limited by a space constant of very similar value (between 100 and 150 μm). We suggest that SRCI in all three species is mediated by HCs: in amphibians, SRCI must at least partially reflect rod-modulation of HC feedback onto cones.

Selective agonists and antagonists were employed to determine the role of indoleaminergic amacrine cells in the generation of the light-evoked responses and spontaneous activity of direction and orientation selective cells. Perfusion with 5-HT2 antagonists reduced the spontaneous activity and both the leading and trailing edge responses of ON/OFF direction selective cells. 5-HT1a agonists had a similar effect on this class of cell, namely, a reduction of light-evoked and spontaneous activity. Results from ON-center and OFF-center orientation selective cells were consistent with those obtained from direction selective cells in that no disruption of direction or orientation selectivity was observed during perfusion of these drugs. These data suggest that the indoleaminergic cells are not directly involved in the generation of the trigger features of complex ganglion cells, but may be facilitating synaptic transmission in the inner retina. This function is discussed relative to the connectivity of the rod bipolar cells and the putative indoleaminergic amacrine cells. The similarity of the effects of 5-HT1a agonists and 5-HT2 antagonists supports the hypothesis, developed during our prior studies of brisk ganglion cells, that these two receptor classes mediate antagonistic processes in the target neurons.

Action potentials were recorded from isolated goldfish retinae maintained in a superfusate of Ringer's solution. Responses to flashes of light and maintained discharges were obtained from 84 cells. The properties of these cells were compared to those in two other goldfish preparations: the isolated retina maintained in a flow of moist oxygen and the self-respiring fish. Maintained discharges of cells in the superfused retinae tended to have lower mean firing rates, higher variability, and weaker high-pass properties than had been observed in the previous preparations. These properties seemed insensitive to the particular formulae used to superfuse the retinae.

Cobalt, which disables synapses, dramatically reduced maintained firing and eliminated photic responses. Cells that did fire in the presence of cobalt generally had low variabilities before cobalt was added; their firing in cobalt was considerably more variable than the baseline. Nevertheless, cobalt did not seem to change the temporal dependency (high-pass properties) of the maintained discharges. The cholinergic agonist carbachol had an excitatory effect upon 71% of the cells tested. Bursty or oscillatory firing in cobalt was rendered more regular by the addition of carbachol. With the exception of the mean firing rate, none of the statistical properties of the maintained discharge differed in cobalt plus carbachol from those in normal Ringer's solution.

There was a tendency for the statistical properties of the maintained discharge after the treatment to approach those of the previously reported preparations; the treatment was at least partially responsible for the drift in properties. The results are discussed in terms of the possible sources of variability in the ganglion cell's discharge, with particular reference to the high-pass filter that appears to act upon it.

In order to elucidate the role of putative indoleaminergic amacrine cells in visual processing, we have employed pharmacological agents specific for the two classes of serotonin receptor, 5-HT2 and 5-HT1, which have been identified in both the retina and brain. Perfusion of the rabbit retina with 5-HT2 selective antagonists decreases the ON-excitation of all classes of ganglion cell as well as the spontaneous activity of these cells. The effect on OFF-responses depends on the cell type: OFF-excitation of center-surround brisk and sluggish cells is increased or not affected by these drugs, but OFF excitation of ON/OFF direction selective cells is reduced. No disruption of the trigger features of direction selective or orientation selective cells was discovered, suggesting that indoleaminergic amacrine cells do not play a role in the generation of the complex properties of these cells.

5-HT1 receptors are heterogeneous and classified as a, b, or c subtypes. Since no selective antagonists are available for these sites, we have employed specific agonists. The most specific of these are for the 5-HT1A receptor. Perfusion with these agonists had physiological effects similar to those with perfusion of 5-HT2 antagonists. Thus, we have suggested that these two classes of serotonin receptors mediate opponent processes in the neural pathway. Since indoleaminergic cells make reciprocal synaptic connections with rod bipolar cell terminals, which are depolarizing in the rabbit retina, we hypothesize that 5-HT2 receptors facilitate the synaptic transmission from the depolarizing rod bipolar cell thus facilitating ON-excitation in the retinal network while 5-HT1A receptors mediate an inhibitory process. Similar self-opponent processing appears to take place in the hypothalamic and hippocampal serotonergic systems as well as in the dopaminergic retinal system. Thus, it is likely that this organization is a general feature of monoamine signal transmission in the central nervous system.

We have examined the soma size, number, and distribution of cholinergic amacrine cells in the retinas of albino and pigmented rats and of Siamese and common cats, using an antibody against choline acetyl transferase (ChAT). In the pigmented strains of rat and cat, ChAT-immunoreactive (ChAT-IR) somata were located in both the inner part of the inner nuclear layer (INL) and ganglion cell layer (GCL), and their processes spread in distinct strata of the inner plexiform layer (IPL). The diameters of the somata in the INL and GCL did not differ significantly at any retinal location. Furthermore, soma diameter did not vary with eccentricity, except at the area centralis of the common cat, where ChAT-IR somata in both layers were relatively smaller. In both species, ChAT-IR somata in the GCL outnumbered those in the INL at all retinal locations. Both populations of cells tended to concentrate at the area of peak ganglion cell density and along the visual streak. Additionally, areas of relatively high density extended superiorly from the area of peak density. The same features of morphology and distribution were identifiable in the hypopigmented strains of rat and cat, but the numbers of ChAT-IR cells were consistently higher.

The distribution and ultrastructure of neurons and neuropil labeled by an antiserum to gamma-aminobutyric acid (GABA) were examined in the lateral geniculate nucleus (LGN) of the tree shrew (Tupaia belangeri). The LGN of this species segregates center type and cell class into three distinct pairs of laminae: a medial pair (laminae 1 and 2) containing ON-center cells, a more lateral pair (4, 5) containing OFF-center cells, and 2 laminae (3, 6) containing W-like cells. The relationship between this laminar segregation and the distribution of GABA immunoreactivity was investigated in the present study. GABA-immunoreactive neurons and neuropil were present in all six of the laminae. However, both the density of labeled cells (adjusted for neuronal density across laminae) and the density of labeled neuropil showed a medial-to-lateral gradient. The adjusted density of labeled cells was higher laterally than medially, and the density of labeled neuropil was significantly greater in the more lateral OFF-center laminae and W-like laminae than in the medial two ON-center laminae. Thus, inhibitory, GABAergic influences may modulate to different degrees the visual signals in the ON, OFF, and W pathways. Labeled cells had a mean cross-sectional area (107 μm2) approximately one-half that of unlabeled cells (216 μm2). They constitute 16–34% of the neurons in the LGN. At the electron microscope level, three different kinds of labeled profile were observed. Vesicle containing profiles like the F2 profiles of cat were postsynaptic to retinal terminals and presynaptic to conventional dendrites. Fl axon terminals with dense clusters of vesicles were also labeled as were some myelinated axons. Another labeled profile, which we suggest should be called an F3 process, was a large dendrite of irregular caliber with punctate groups of vesicles near the synapse. Our results suggest that GABAergic circuitry is an important part of the functional organization in the LGN of the tree shrew.

The organization of the retina of the turtle species Mauremys caspica, found in fresh water ponds of Israel, has been examined by light microscopical techniques including examination of fresh wholemount retina, one micron blue-stained vertical sections and Golgi-stained material. The anatomical findings on Mauremys retina have been compared with those of the Pseudemys retina (Kolb, 1982) which is more commonly used for electrophysiological and neurochemical studies in the USA. The photoreceptors of Mauremys are similar in type and oil droplet content to Pseudemys photoreceptors except for the double cone in Mauremys. This cone type appears more abundant than in Pseudemys and the principal member contains a yellow oil droplet instead of an orange oil droplet. Golgi staining reveals that all the cell types that have been seen in Pseudemys are found in Mauremys with identical morphology. In addition, two amacrine cell types that were not before described for Pseudemys have been added to the classification. One of these is the tristratified dopaminergic amacrine cell described in immunocytochemical studies (Witkovsky et al., 1984; Nguyen-Legros et al., 1985; Kolb et al., 1987). We have used these anatomical studies on Pseudemys and Mauremys retina to form a catalogue of neural types for the turtle retina in general. We conclude with an attempt to combine findings from anatomy, electrophysiology, and neurochemistry to form an overview of the organization of this reptilian retina.

A new form of receptive field map, the response-component map, was developed to identify points within a receptive field that produce similar response patterns. The fields were probed with discretely flashed small spots of light. The magnitudes of the responses to stimulus onset and to stimulus offset elicited at each point were represented on the map by a vector radiating from the position representing the location of that point. Thus, response-component maps preserve the spatial distributions of responsivity and temporal nonlinearities. Points with similar response patterns were identified from a scatterplot in which the response at each spatial position was located in a plane representing the angles of the response-component vectors. Points with similar response characteristics that were also spatially contiguous were considered as a distinct response subregion within the receptive field.

Barely 10% of the receptive fields of goldfish ganglion cells mapped with this technique proved as simple as the traditional concentric field described for retinal cells. In at least 17% of the cases, the field showed three concentric rings, with a very small “inner center” within the center of the field. In at least 50% of the cases, response subregions of different type lay side by side, rather than in a concentric configuration. Some subregions could be differentiated by the relative strengths of the responses to onset and offset of the stimulus spot, supporting the hypothesis that a push-pull system generates ganglion cell responses. Subregions were evident in successive mappings of the same cell, demonstrating they are not due to the vagaries of individual responses. They probably represent the spatial domains (or their intersections) of individual interneurons distal to the retinal ganglion cells. It is possible that position within the receptive field may be coded by the temporal pattern of the responses.

In 1827, plant biologist Robert Brown discovered what is known as Brownian motion, a class of chaos. Formal derivative of Brownian motion is Gaussian white-noise. In 1938, Norbert Wiener proposed to use the Gaussian white-noise as an input probe to identify a system by a series of orthogonal functionals known as the Wiener G-functionals.

White-noise analysis is uniquely suited for studying the response dynamics of retinal neurons because (1) white-noise light stimulus is a modulation around a mean luminance, as are the natural photic inputs, and it is a highly efficient input; and (2) the analysis defines the response dynamics and can be extended to spike trains, the final output of the retina. Demonstrated here are typical examples and results from applications of white-noise analysis to a visual system.