The N-methyl-D-aspartate receptor (NMDAR) is an ionotropic glutamate receptor that is important in neurotransmission as well as in processes of synaptic plasticity in the mammalian superior colliculus (SC). Despite the importance of this receptor in synaptic transmission, there is as yet no evidence that demonstrates directly the synaptic localization of the NMDAR receptor in SC. We have used electron-microscope (EM) immunocytochemistry to localize the NMDAR1 subunit of this receptor protein and its association with sensory afferents in the cat SC. Retinal synaptic terminals were identified by normal morphology and cortical synaptic terminals by degeneration after lesions of areas 17–18 of the visual cortex. At the light-microscope level, label was densest within the superficial gray and upper optic layers, but also present in all other layers. Label was contained within cell bodies, dendrites, and a few putative axons. At the EM level, antibody labeling was found along postsynaptic densifications and internalized within the cytoplasm of a variety of dendrites and some cell bodies. Postsynaptic profiles labeled by NMDAR1 included conventional dendrites and presynaptic dendrites which contained pleomorphic synaptic vesicles and are known to be GABAergic. Many of the labeled postsynaptic densifications of both of these profile types received synaptic inputs from retinal or cortical terminals. Virtually no NMDAR1 immunoreactivity was found on thin dendritic thorns or putative spines, even when these were postsynaptic to retinal or cortical terminals. In summary, these results show that the NMDAR1 subunit is postsynaptic to both retinal and cortical afferents, which are known to be glutamatergic, and are consistent with physiological evidence showing that stimulation of either pathway can activate the NMDA receptor.

In normal adult cats, a monoclonal antibody directed toward the NR-1 subunit of the N-methyl-d-aspartate (NMDA) receptor (Pharmingen, clone 54.1) produced dense cellular and neuropil labeling throughout all layers of the lateral geniculate nucleus (LGN) and adjacent thalamic nuclei, including the thalamic reticular, perigeniculate, medial intralaminar, and ventral lateral geniculate nuclei. Cellular staining revealed well-defined somata, and in some cases proximal dendrites. NMDAR-1 cell labeling was also evident in the LGN of early postnatal kittens, suggesting that developing LGN cells possess this receptor subunit at or before eye opening. Within the A-layers of the adult LGN, staining encompassed a wide range of soma sizes. Soma size comparisons of NMDAR-1 stained cells with those stained with an antibody directed toward a nonphosphorylated neurofilament protein (SMI-32), which selectively stains Y-relay cells (Bickford et al., 1998), or an antibody to glutamic acid decarboxylase (GAD), which stains for GABAergic interneurons, suggested that NMDA receptors are utilized by relay cells and interneurons. NMDAR-1 staining was also observed in the LGN of cats with early monocular lid suture. Although labeling was apparent in both deprived and nondeprived A-layers of LGN, the distribution of soma sizes was significantly different. In the deprived A-layers of LGN, staining was limited to small- and medium-sized cells. Cells with relatively large soma were lacking. However, cell density measurements as well as soma size comparisons with cells stained for Nissl substance suggested these differences were due to deprivation-induced cell shrinkage and not to a loss of NMDAR-1 staining in Y-cells. Taken together, these results suggest that NMDA receptors are utilized by both relay cells and interneurons in LGN and that alterations in early visual experience do not necessarily affect the expression of NMDA receptors in the LGN.

Patterns of terminals labeled after WGA-HRP injections in the superior colliculus (SC) in squirrel monkeys and macaque monkeys, and after DiI application in marmosets, were related to the architecture of the pulvinar and dorsal lateral geniculate nucleus (LGN). In all studied species, the SC projects densely to two architectonic subdivisions of the inferior pulvinar, the posterior inferior pulvinar nucleus (PIp) and central medial inferior pulvinar nucleus (PIcm). These projection zones expressed substance P. Thus, sections processed for substance P reveal SC termination zones in the inferior pulvinar. The medial subdivision of the inferior pulvinar, PIm, which is known to project to visual area MT, does not receive a significant collicular input. Injections in MT of a squirrel monkey revealed no overlap between SC terminals and neurons projecting to area MT. Thus, PIm is not the significant relay station of visual input from the SC to MT. The SC also sends an input to the LGN, however, this projection is sparser than the input directed to pulvinar.

The purpose of the present study was to examine the contributions made by cat posterior parietal cortex to the analyses of the relative position of objects in visual space. Two cats were trained on a landmark task in which they learned to report the position of a landmark object relative to a right or left food-reward chamber. Subsequently, three pairs of cooling loops were implanted bilaterally in contact with visuoparietal cortices forming the crown of the middle suprasylvian gyrus (MSg; architectonic area 7) and the banks of the posterior-middle suprasylvian sulcus (pMS sulcal cortex) and in contact with the ventral-posterior suprasylvian (vPS) region of visuotemporal cortex. Bilateral deactivation of pMS sulcal cortex resulted in a profound impairment for all six tested positions of the landmark, yet bilateral deactivation of neither area 7 nor vPS cortex yielded any deficits. In control tasks (visual orienting and object discrimination), there was no evidence for any degree of attentional blindness or impairment of form discrimination during bilateral deactivation of pMS cortex. Therefore, we conclude that bilateral cooling of pMS cortex, but neither area 7 nor vPS cortex, induces a specific deficit in spatial localization as examined with the landmark task. These observations have significant bearing on our understanding of visuospatial processing in cat, monkey, and human cortices.

The present study uses cell-attached patch-recording techniques to study the single-channel properties of Ca2+ channels in isolated salamander photoreceptors and investigate their sensitivity to reductions in intracellular Cl−. The results show that photoreceptor Ca2+ channels possess properties similar to L-type Ca2+ channels in other preparations, including (1) enhancement of openings by the dihydropyridine agonist, (−)BayK8644; (2) suppression by a dihydropyridine antagonist, nisoldipine; (3) single-channel conductance of 22 pS with 82 mM Ba2+ as the charge carrier; (4) mean open probability of 0.1; (5) open-time distribution fit with a single exponential (τ0 = 1.1 ms) consistent with a single open state; and (6) closed time distribution fit with two exponentials (τc1 = 0.7 ms, τc2 = 25.4 ms) consistent with at least two closed states. Using a Cl−-sensitive dye to measure intracellular [Cl−], it was found that perfusion with gluconate-containing, low Cl− medium depleted intracellular [Cl−]. It was therefore possible to reduce intracellular [Cl−] by perfusion with a low Cl− solution while maintaining the extracellular channel surface in high Cl− pipette solution. Under these conditions, the single-channel conductance was unchanged, but the mean open probability fell to 0.03. This reduction can account for the 66% reduction in whole-cell Ca2+ currents produced by perfusion with low Cl− solutions. Examination of the open and closed time distributions suggests that the reduction in open probability arises from increases in closed-state dwell times. Changes in intracellular [Cl−] may thus modulate photoreceptor Ca2+ channels.

Thresholds for the detection of differences in the duration of visual stimuli were determined for a variety of programs of stimulus onset and offset. Performance suffers when a time interval begins with an ON step and ends with another ON stimulus, compared to the standard ON–OFF stimulation, but the decrement is reversed when the light is ramped down to background during the interval. Neither the magnocellular nor the parvocellular streams can be excluded because there is relatively little impairment of duration discrimination when the stimulus has low contrast or is heterochromatic at isoluminance. Performance at a variety of intensity levels suggests that sustained neural firing in an early stage of visual processing provides a background activity, which prevents good temporal precision of signals.

The important visual stimulus parameters for a given cell are defined by the classical receptive field (CRF). However, cells are also influenced by visual stimuli presented in areas surrounding the CRF. The experiments described here were conducted to determine the incidence and nature of CRF surround influences in the primary visual cortex. From extracellular recordings in the cat's striate cortex, we find that for over half of the cells investigated (56%, 153/271), the effect of stimulation in the surround of the CRF is to suppress the neuron's activity by at least 10% compared to the response to a grating presented within the CRF alone. For the remainder of the cells, the interactions were minimal and a few were of a facilitatory nature. In this paper, we focus on the suppressive interactions. Simple and complex cell types exhibit equal incidences of surround suppression. Suppression is observed for cells in all layers, and its degree is strongly correlated between the two eyes for binocular neurons. These results show that surround suppression is a prevalent form of inhibition and may play an important role in visual processing.

Spatial summation and degree of center-surround antagonism were examined in the receptive field of nonlagged cells in the dorsal lateral geniculate nucleus (dLGN). We recorded responses to stationary light or dark circular spots that were stepwise varied in width. The spots were centered on the receptive field. For a sample of nonlagged X-cells, we made simultaneous recordings of action potentials and S-potentials, and could thereby compare spatial summation in the dLGN cell and in the retinal input to the cell. Plots of response versus spot diameter showed that the response for a dLGN cell was consistently below the response in the retinal input at all spot sizes. There was a marked increase of antagonism at the retinogeniculate relay. The difference between the retinal input and dLGN cell response suggested that the direct retinal input to a relay cell is counteracted in dLGN by an inhibitory field that has an antagonistic center-surround organization. The inhibitory field seems to have the same center sign (ON- or OFF-center), but a wider receptive-field center than the direct retinal input to the relay cell. The broader center of the inhibitory field can explain the increased center-surround antagonism at the retinogeniculate relay. The ratio between the response of a dLGN cell and its retinal input (transfer ratio) varied with spot width. This variation did not necessarily reflect a nonlinearity at the retinogeniculate relay. Plots of dLGN cell response against retinal input were piecewise linear, suggesting that both excitatory and inhibitory transmission in dLGN are close to linear. The variation in transfer ratio could be explained by sustained suppression evoked by the background stimulation, because such suppression has relatively stronger effect on the response to a spot evoking weak response than to a spot evoking a strong response. A simple model for the spatial receptive-field organization of nonlagged X-cells, that is consistent with our findings, is presented.

In many vertebrate retinas the outer segments of rod photoreceptors have multiple incisures, that is, there are numerous indentations in the highly curved membrane forming the edge of their disks and in the plasma membrane enclosing the entire stack of disks. Immunofluorescent localization of tubulin in amphibian photoreceptors yielded a novel series of thin, parallel, fluorescent lines in rod outer segments that extended their full length and coincided with their multiple incisures. Electron-microscopic examination of amphibian retinas revealed the structures responsible for this fluorescence: longitudinally oriented microtubules were associated with incisures at heights throughout rod outer segments. These microtubules were located between the disk rims and the overlying plasma membrane, in the small cytoplasmic compartment at the mouth of incisures; the microtubules and membranes were separated from each other by distances that were uniform, as though interconnected by filaments described in other studies. Thus, in amphibian rod outer segments the incisures mark the site of a cytoskeletal system containing longitudinal microtubules distinct from those of the ciliary axoneme, linked by filaments to the adjacent membranes. This cytoskeleton is expected to be important for the normal structure, function, and renewal of rod outer segments. In amphibian cone outer segments, which do not have incisures, the only anti-tubulin immunofluorescence and the only microtubules were at the axoneme. These findings may help elucidate the diverse properties of rods and cones in many vertebrate retinas and could prove relevant for human retinal degenerations.

Thalamic relay cells fire in two distinct modes, burst or tonic, and the operative mode is dictated by the inactivation state of low-threshold, voltage-gated, transient (or T-type) Ca2+ channels. Tonic firing is seen when the T channels are inactivated via membrane depolarization, and burst firing is seen when the T channels are activated from a hyperpolarized state. These response modes have very different effects on the relay of information to the cortex. It had been thought that only tonic firing is seen in the awake, alert animal, but recent evidence from several species suggests that bursting may also occur. We have begun to explore this issue in macaque monkeys by recording from thalamic relay cells of unanesthetized, behaving animals. In the lateral geniculate nucleus, the thalamic relay for retinal information, we found that tonic mode dominated responses both during alert behavior as well as during sleep. We nonetheless found burst firing present during the vigilant, waking state. There was, however, considerably more burst mode firing during sleep than wakefulness. Surprisingly, we did not find the bursting during sleep to be rhythmic. We also recorded from relay cells of the somatosensory thalamus. Interestingly, not only did these somatosensory neurons exhibit much more burst mode activity than did geniculate cells, but bursting during sleep was highly rhythmic. It thus appears that the level and nature of relay cell bursting may not be constant across all thalamic nuclei.

Although the retinotectal projection of goldfish has long been known to have a high degree of retinotopic order, the structural basis for this in terms of the precise positioning of axonal arbors from neighboring retinal ganglion cells has not been determined. In studying this, a small number of neighboring retinal ganglion cells was selectively labeled by a microinjection of DiI into the retina. Following axoplasmic transport for several days, the tectum was removed and flat-mounted for fluorescence microscopy. The injection labeled a small number of axons and their terminal arbors which ranged in size from 108 × 134 μm to 394 × 331 μm with a mean of 233 × 219 μm. This mean size corresponds to about 1/15 of the length of one tectal axis. Although individual arbors labeled from one small retinal injection were always observed near the same retinotopic position, they were almost never coextensive. Overlap between pairs of arbors along the lines of projection perpendicular to the tectal surface averaged 57% of the area of a single arbor. These results indicate that neighboring retinal ganglion cells do not converge onto the same locus but instead project as a continuous retinotopic array of partially overlapping terminal fields.

In the retina, nitric oxide (NO) functions in network coupling, light adaptation, neurotransmitter receptor function, and synaptic release. Neuronal nitric oxide synthase (nNOS) is present in the retina of every vertebrate species investigated. However, although nNOS can be found in every retinal cell type, little is known about the production of NO in specific cells or about the diffusion of NO within the retina. We used diaminofluorescein-2 (DAF-2) to image real-time NO production in turtle retina in response to stimulation with N-methyl-D-aspartate (NMDA). In response to NMDA, NO was produced in somata in the ganglion cell and inner nuclear layers, in synaptic boutons and processes in the inner plexiform layer, in processes in the outer plexiform layer, and in photoreceptor inner segments. This NO-dependent fluorescence production quickly reached transient peaks and declined more slowly toward baseline levels at different rates in different cells. In some cases, the NO signal was primarily confined to within 10 μm of the source, which suggests that NO may not diffuse freely through the retina. Such limited spread was not predicted and suggests that NO signal transduction may be more selective than suggested, and that NO may play significant intracellular roles in cells that produce it. Because NO-dependent fluorescence within amacrine cells can be confined to the soma, specific dendritic sites, or both with distinct kinetics, NO may function at specific synapses, modulate gene expression, or coordinate events throughout the cell.

Spatial receptive fields of relay cells in dorsal lateral geniculate nucleus (dLGN) have commonly been modeled as a difference of two Gaussian functions. We present alternative models for dLGN cells which take known physiological couplings between retina and dLGN and within dLGN into account. The models include excitatory input from a single retinal ganglion cell and feedforward inhibition via intrageniculate interneurons. Mathematical formulas describing the receptive field and response to circular spot stimuli are found both for models with a finite and an infinite number of ganglion-cell inputs to dLGN neurons. The advantage of these models compared to the common difference-of-Gaussians model is that they, in addition to providing mathematical descriptions of the receptive fields of dLGN neurons, also make explicit contributions from the geniculate circuit. Moreover, the model parameters have direct physiological relevance and can be manipulated and measured experimentally. The discrete model is applied to recently published data (Ruksenas et al., 2000) on response versus spot-diameter curves for dLGN cells and for the retinal input to the cell (S-potentials). The models are found to account well for the results for the X-cells in these experiments. Moreover, predictions from the discrete model regarding receptive-field sizes of interneurons, the amount of center-surround antagonism for interneurons compared to relay cells, and distance between neighboring retinal ganglion cells providing input to interneurons, are all compatible with data available in the literature.

Distribution of the mt1 melatonin receptor in the guinea pig retina was immunocytochemically investigated using peptide-specific anti-mt1 receptor antibody. Western blots of the guinea pig retina showed a single band at approximately 37 kilodalton (kD) immunoreactive to the anti-mt1 antibody. The most intense immunoreactivity for the mt1 receptor was detected in the cell bodies of ganglion cells. Their dendrites and axons were also immunolabeled. Subpopulations of amacrine cells, the inner plexiform layer, and the outer plexiform layer also exhibited moderate to weak immunolabeling. The mt1-positive amacrine cells were located either at the vitreal border of the inner nuclear layer or displaced in the ganglion cell layer. Double immunolabeling using antibodies to the mt1 receptor and tyrosine hydroxylase revealed that the majority of dopaminergic amacrine cells showed mt1 immunoreactivity. Almost all the 1CA type dopaminergic cells were mt1 positive while the 2CA type cells less frequently exhibited mt1 immunoreaction. By double immunolabeling for the mt1 receptor and GABA, more than 50% of the mt1-immunoreactive amacrine cells were shown to be GABAergic neurons. Approximately one-third of the GABAergic amacrine cells were immunolabeled for the mt1 receptor. The present results demonstrate expression of the mt1 receptor in diverse neuronal cell types in the guinea pig retina and provide the first evidence for the direct effect of melatonin on dopaminergic and GABAergic amacrine cells via the mt1 receptor.

GABAergic responses of rabbit rod bipolar cells were reexamined by using whole-cell recordings in the superfused slice preparation to determine if there is GABAC receptor input to their axon terminal and to characterize the contribution that GABAA and GABAC receptors make to the total GABA current on the axon terminals of these cells. Pharmacological agents specifically blocking GABAA and GABAC receptor currents demonstrated that 37% of the GABA-activated current was blocked by either the GABAA antagonists bicuculline or SR-95531, whereas the remaining 63% of the GABA current was blocked by a mixture of bicuculline and the GABAC antagonist TPMPA. This indicated that GABAC receptors were present on the axon terminal of the rabbit rod bipolar cell and that they were responsible for mediating the bicuculline insensitive GABA current.

This study is concerned with how information about the direction of visual motion is encoded by motion-sensitive neurons. Motion-sensitive neurons are usually studied using stimuli unchanging in speed and direction over several seconds. Recently, it has been suggested that neuronal responses to more naturalistic stimuli cannot be understood on the basis of experiments with constant-motion stimuli (de Ruyter van Steveninck et al., 1997). We measured the variability and information content of spike trains recorded from directional neurons in the nucleus of the optic tract (NOT) of the wallaby, Macropus eugenii, in response to constant and time-varying motion. While the NOT forms part of the mammalian optokinetic system, we have shown previously that the responses of its directional neurons resemble those of insect H1 in many respects (Ibbotson et al., 1994). We find that directional neurons in the wallaby NOT respond with lower variability and higher rates of information transmission to time-varying stimuli than to constant motion. The difference in response variability is predicted by an inhomogeneous Poisson model of neuronal spiking incorporating an absolute refractory period of 2 ms during which no subsequent spike can be fired. Refractoriness imposes structure on the spike train, reducing variability (de Ruyter van Steveninck & Bialek, 1988; Berry & Meister, 1998). A given refractory period has a greater impact when firing rates are high, as for the responses of NOT neurons to time-varying stimuli. It is in just these cases that variability in experimentally observed spike trains is lowest. Thus, differences in response variability do not necessarily imply that different models are required to predict neuronal responses to constant- and time-varying motion stimuli.

How neurons in the primary visual cortex (V1) of primates process parallel inputs from the magnocellular (M) and parvocellular (P) layers of the lateral geniculate nucleus (LGN) is not completely understood. To investigate whether signals from the two pathways are integrated in the cortex, we recorded contrast-response functions (CRFs) from 20 bush baby V1 neurons before, during, and after pharmacologically inactivating neural activity in either the contralateral LGN M or P layers. Inactivating the M layer reduced the responses of V1 neurons (n = 10) to all stimulus contrasts and significantly elevated (t = 8.15, P < 0.01) their average contrast threshold from 8.04 (± 4.1)% contrast to 22.46 (± 6.28)% contrast. M layer inactivation also significantly reduced (t = 4.06, P < 0.01) the average peak response amplitude. Inactivating the P layer did not elevate the average contrast threshold of V1 neurons (n = 10), but significantly reduced (t = 4.34, P < 0.01) their average peak response amplitude. These data demonstrate that input from the M pathway can account for the responses of V1 neurons to low stimulus contrasts and also contributes to responses to high stimulus contrasts. The P pathway appears to influence mainly the responses of V1 neurons to high stimulus contrasts. None of the cells in our sample, which included cells in all output layers of V1, appeared to receive input from only one pathway. These findings support the view that many V1 neurons integrate information about stimulus contrast carried by the LGN M and P pathways.

To study the detailed morphology of human retinal ganglion cells, we used intracellular injection of horseradish peroxidase and Neurobiotin to label over 1000 cells in an in vitro, wholemount preparation of the human retina. This study reports on the morphology of 119 wide-field bistratified and 42 diffuse ganglion cells. Cells were analyzed quantitatively on the basis of dendritic-field size, soma size, and the extent of dendritic branching. Bistratified cells were similar in dendritic-field diameter (mean ± s.d. = 682 ± 130 μm) and soma diameter (mean ± s.d. = 18 ± 3.3 μm) but showed a broad distribution in the extent of dendritic branching (mean ± s.d. branch point number = 67 ± 32; range = 15–167). Differences in the extent of branching and in dendritic morphology and the pattern of branching suggest that the human retina may contain at least three types of wide-field bistratified cells. Diffuse ganglion cells comprised a largely homogeneous group whose dendrites ramified throughout the inner plexiform layer. The diffuse cells had similar dendritic-field diameters (mean ± s.d. = 486 ± 113 μm), soma diameters (mean ± s.d. = 16 ± 2.3 μm), and branch points numbers (mean ± s.d. = 92 ± 32). The majority had densely branched dendritic trees and thin, very spiny dendrites with many short, fine, twig-like thorny processes. Five of the diffuse cells had much more sparsely branched dendritic trees (<50 branch points) and less spiny dendrites, suggesting that there are possibly two types of diffuse ganglion cells in human retina. Although the presence of a diversity of large bistratified and diffuse ganglion cells has been observed in a variety of mammalian retinas, little is known about the number of cell types, their physiological properties, or their central projections. Some of the human wide-field bistratified cells in the present study, however, show morphological similarities to monkey large bistratified cells that are known to project to the superior colliculus.

Most animals experience daily changes in light and darkness. The retinas of many of these animals show concomitant rhythmic changes in the levels of mRNAs that encode proteins involved in the photoresponse. These changes may be circadian and independent of light, independent of circadian clocks and regulated by light, or regulated by a circadian clock and light. We have taken advantage of the organization of the Limulus visual system to examine the separate and combined effects of signals from a circadian clock and light on arrestin mRNA levels in photoreceptors. The clock that regulates photoreceptors in the lateral eye of Limulus is in the brain, and signals from the clock reach the lateral eye via activation of a well-characterized efferent projection in the lateral optic nerve. In the experiments described, clock-driven efferent input to the lateral eye was eliminated by cutting the lateral optic nerve, and light input to the lateral eye was eliminated by placing an opaque patch over the eye. Arrestin mRNA levels were quantified relative to 18s rRNA with a ribonuclease protection assay. We observed the following. In lateral eyes exposed to natural diurnal light and endogenous efferent nerve activity, the level of arrestin mRNA was higher during the day in the light than during the night in the dark. Circadian efferent nerve activity was necessary and sufficient to produce normal daily fluctuations in the level of arrestin mRNA. Light influenced arrestin mRNA levels only in eyes with intact and active efferent projections. We conclude that arrestin mRNA levels in lateral eye photoreceptors are controlled entirely by efferent nerve activity, and that light exerts its effects by modulating this output from the circadian clock. Light-stimulated changes in arrestin mRNA in the vertebrate retina may likewise require interactions between light-driven biochemical cascades and clock output.

Previous studies have shown that the expression of glutamate receptor subunits is developmentally regulated and have been implicated in processes of cell differentiation during postnatal life. The tissue localization and developmental pattern of the glutamate receptor 2 subunit of the α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionate (AMPA) receptor were investigated by means of immunohistochemistry and immunoblotting. Labeling of amacrine and ganglion cells and the inner plexiform layer appeared early during development, while glutamate receptor 2 subunit expression in the outer plexiform layer started after the first postnatal week. The distribution of labeling within the inner plexiform layer changed from nonorganized to laminated appearance prior to eye-opening. There was an increasing number of positive amacrine and ganglion cell somata during the first 2 weeks, but their number decreased considerably as the retina matured and were seen at least up to 35 days of postnatal development. Little labeling was found in the ganglion cell layer and in the inner plexiform layer of late postnatal and adult retina. Labeling in the outer plexiform layer and of bipolar cell somata appeared to increase in the developing retina. Glur2 labeling of these cells and the outer plexiform layer became discernible during the second postnatal week, and this labeling was present in the adult as well. Immunoblotting showed that GluR2 protein levels were similar at postnatal days 7 and 10, but slightly decreased between the second and fourth postnatal weeks. Our data imply that the immunological expression of glutamate receptor 2 subunit in the inner plexiform layer decreases as a function of age, and is correlated with developmental event(s) in the postnatal retina.