Connexin26 (Cx26) is a member of the family of integral membrane proteins that normally form intercellular gap junctional channels. We have used Western blotting, immunofluorescence, immunoelectron microscopy, and single-cell reverse-transcriptase polymerase chain reaction amplification (RT-PCR) to analyze the expression and cellular localization of Cx26 in the carp retina. In the outer plexiform layer, strong clustered Cx26 immunolabeling was concentrated at and restricted to the terminal dendrites of horizontal cells. Single-cell RT-PCR confirmed the expression of Cx26 in carp retinal horizontal cells. 248-bp fragments amplified from cDNAs of four different horizontal cells were cloned and each nucleotide sequence encodes a protein fragment (AA 104-185) with highly significant homology to rat and mouse Cx26. Immunoelectron microscopy revealed that only the invaginating dendrites of horizontal cells in intimate lateral association with the presynaptic ribbon complex were labeled. No labeling was found at the photoreceptor membrane and there was no septalaminar structure, indicative of gap junctions, between photoreceptors and horizontal cells. The focal location of Cx26 at the membrane of the dendritic tips of horizontal cells and the lack of gap junctional morphology suggests that Cx26 might form hemichannels.

The superfused retinal slice preparation was used to examine the morphology and glutamate-activated whole-cell currents of rabbit bipolar cells. There were six morphologically distinct types of cone bipolar cells and a rod bipolar cell that had axon terminals stratifying in stratum 3 to 5 of sublamina-b. All of these bipolar cell types exhibited an outward current in response to the application of the metabotropic glutamate receptor, mGluR6, agonist AP-4 (APB), and had I/V curves indicative of membrane channel closure. Conversely, there were no currents activated during the application of kainate, the AMPA/kainate receptor agonist. These data demonstrate they were on-bipolar cells. In addition, there were six morphologically distinct cone bipolar cells that stratified in sublamina-a. Every cell with axonal arborizations in stratum 1 and 2 exhibited an inward current when the ionotropic glutamate receptor agonist kainate was applied. This current was blocked by application of the AMPA/kainate receptor antagonist CNQX. These cells also decreased their membrane resistance in response to kainate, a characteristic of the opening of channels within the plasma membrane. Without exception, no cells stratifying in sublamina-a responded to the mGluR6 agonist AP-4, further identifying them as off-bipolar cells.

The distribution of local preferred directions and motion sensitivities within the receptive fields of so-called tangential neurons in the fly visual system was previously found to match optic flow fields as induced by certain self-motions. The complex receptive-field organization of the tangential neurons and the recent evidence showing that the orderly development of the fly's peripheral visual system depends on visual experience led us to investigate whether or not early visual input is required to establish the functional receptive-field properties of such tangential neurons. In electrophysiological investigations of two identified tangential neurons, it turned out that dark-hatched flies which were kept in complete darkness for 2 days develop basically the same receptive-field organization as flies which were raised under seasonal light/dark conditions and were free to move in their cages. We did not find any evidence that the development of the sophisticated receptive-field organization of tangential neurons depends on sensory experience. Instead, the input to the tangential neurons seems to be “hardwired” and the specificity of these cells to optic flow induced during self-motions of the animal may have evolved on a phylogenetical time scale.

We recorded electroretinograms (ERGs) under stimulus conditions that only modulated one cone type: either the L- or the M-cones. In these conditions the rods were also modulated. We measured the ERG responses at different temporal frequencies. A simple model that assumes that the first harmonic components of the responses are the result of a vector addition of rod- and cone-driven ERG responses can explain the data satisfactorily for temporal frequencies at and above 6 Hz. From fits of the model to the data, estimates of the gains and phases of the rod- and cone-driven responses can be obtained. At 6 Hz, the fundamental responses originate exclusively in the rods. The gains of the cone-driven responses are substantial at the other temporal frequencies, are maximal at 12 or 18 Hz, and then decrease with increasing temporal frequencies. The gains of the rod-driven responses decrease more steeply with increasing temporal frequencies than the cone gains. Furthermore, the rod and cone phases decrease approximately linearly with temporal frequency, suggesting that they are mainly determined by a response delay. The response delay in the rod-driven ERGs is larger than in the cone-driven ERGs.

When an image feature moves with sufficient speed it should become smeared across space, due to temporal integration in the visual system, effectively creating a spatial motion pattern that is oriented in the direction of the motion. Recent psychophysical evidence shows that such “motion streak signals” exist in the human visual system. In this study, we report neurophysiological evidence that these motion streak signals also exist in the primary visual cortex of cat and monkey. Single neuron responses were recorded for two kinds of moving stimuli: single spots presented at different velocities and drifting plaid patterns presented at different spatial and temporal frequencies. Measurements were made for motion perpendicular to the spatial orientation of the receptive field (“perpendicular motion”) and for motion parallel to the spatial orientation of the receptive field (“parallel motion”). For moving spot stimuli, as the speed increases, the ratio of the responses to parallel versus perpendicular motion increases, and above some critical speed, the response to parallel motion exceeds the response to perpendicular motion. For moving plaid patterns, the average temporal tuning function is approximately the same for both parallel motion and perpendicular motion; in contrast, the spatial tuning function is quite different for parallel motion and perpendicular motion (band pass for the former and low pass for the latter). In general, the responses to spots and plaids are consistent with the conventional model of cortical neurons with one rather surprising exception: Many cortical neurons appear to be direction selective for parallel motion. We propose a simple explanation for “parallel motion direction selectivity” and discuss its implications for the motion streak hypothesis. Taken as a whole, we find that the measured response properties of cortical neurons to moving spot and plaid patterns agree with the recent psychophysics and support the hypothesis that motion streak signals are present in V1.

Nitric oxide (NO) activates soluble guanylyl cyclase (sGC) and the resulting increase in cyclic guanosine monophosphate (cGMP) is an important intracellular signalling pathway in the vertebrate retina. Immunocytochemical detection of cGMP following exposure to NO donors has proven an effective method of identifying cells that express sGC. While such an approach has proven useful for the study of several vertebrate retinas, it has not been applied to the well-characterized teleost retina. Therefore, in the present study, we have applied this approach to the retina of the goldfish (Carassius auratus). In the presence of the phosphodiesterase (PDE) inhibitor 3-isobutyl-1-methylxanthine (IBMX), incubation of goldfish eyecups in Ringer's solution containing (±)-S-nitroso-N-acetylpenicillamine (SNAP) increased cGMP-like immunoreactivity (cG-ir) in bipolar, horizontal, amacrine, and ganglion cells and in ganglion cell axons and optic nerve. Weak labeling was observed in horizontal cells but no change in cG-ir was noted within photoreceptors. The NO donor-stimulated increases of cG-ir in horizontal, bipolar, amacrine, and ganglion cells are consistent with known physiological effects of NO on these neurons. The physiological significance of NO action at the level of optic nerve is not known. The lack of an effect of SNAP on cG-ir in photoreceptors was unexpected, as there are known physiological actions of NO, mediated by cGMP, on these neurons. Although this may be due to insufficient sensitivity of immunolabeling, this result may indicate a difference between isoforms of sGC or cGMP PDE in these neurons, compared to neurons where exogenous NO increased cG-ir.

We investigated chromatic processing in the outer retina of the cichlid fish Aequidens pulcher. Intracellular recordings from cone-specific horizontal cells (CHCs) revealed that the two morphologically identified types (H1 and H2) also have different spectral responses. H1-L cells hyperpolarize to all wavelengths (“luminosity”). H2-Cb cells depolarize to long wavelengths and hyperpolarize to short wavelengths (“chromaticity”, biphasic). Furthermore, we verified by immunocytochemistry that H2-Cb cells of A. pulcher predominantly contact the middle-wavelength-sensitive (MWS) members of double cones. Developmental plasticity in the cone–CHC networks was induced by rearing fish under conditions of spectral deprivation and different levels of white light. H1-L spectral responses were unaffected by the rearing conditions. Different intensity levels of white light and deprivation of long wavelengths during rearing both induced changes in the spectral responses of H2-Cb. Deprivation of short and middle wavelengths had no effect. Our results indicate that spectral processing in the outer retina of fishes can be modulated in response to different visual experiences and suggest that developmental fine tuning of the color-vision system occurs at early levels of visual processing.

The visual cortex of freshwater turtles contains pyramidal cells, which have a regular spiking (RS) firing pattern, and several categories of aspiny, inhibitory interneurons. The interneurons show diverse firing patterns, including the fast spiking (FS) pattern. Postsynaptic potentials (PSPs) evoked in FS cells by visual stimulation of the retina reach their peak amplitudes as much as 200 ms before PSPs in RS cells (Mancilla et al., 1998). FS cells could, consequently, control the amplitudes of light-evoked PSPs in RS cells by producing disynaptic, feedforward inhibitory postsynaptic potentials (IPSPs) that overlap in time with geniculocortical excitatory postsynaptic potentials (EPSPs). Since FS cells receive recurrent, excitatory inputs from RS cells, they could also control the amplitudes of light-evoked PSPs in RS cells via polysynaptic, feedback inhibition. The in vitro geniculocortical preparation of Pseudemys scripta was used to characterize the temporal relationships of EPSPs and IPSPs produced in RS cells by electrical activation of geniculate afferents and by diffuse light flashes presented to the retina. GABAA receptor-mediated inhibition was blocked using extracellular application of bicuculline (3.5 mM) or intracellular perfusion of picrotoxin (1 μM) in individual RS cells. Electrical stimulation of thalamic afferents produced compound PSPs. Blockade of GABAA receptor-mediated IPSPs with either bicuculline or picrotoxin provided evidence for both early and late IPSPs in RS cells. Analysis of the apparent reversal potentials of light-evoked PSPs indicated the existence of early IPSPs during the first 140–300 ms following light onset. Light responses of cells perfused with picrotoxin diverged from control light responses at about 300 ms after light onset and had maximum amplitudes that were significantly different from control light responses. These experiments indicate that the responses of RS cells to both electrical and natural stimulation of geniculate afferents are controlled by both early and late IPSPs, consistent with activation of both feedforward and feedback pathways.

The phototransduction cascade in invertebrate photoreceptors has not been fully elucidated. It has been proposed that in Limulus ventral photoreceptor cGMP is the intracellular second messenger that directly controls the gating of the light-dependent channels (Johnson et al., 1986; Bacigalupo et al., 1991). Recently, a putative cGMP-gated channel cDNA, Lcng1, has been cloned from Limulus and shown to be expressed in the brain and the ventral eye (Chen et al., 1999). In this study, we sought to more specifically localize the LCNG1 transcript and protein. In situ hybridization was used to determine whether the gene is expressed in glia or photoreceptor cells in the ventral eye. The results clearly demonstrated that Lcng1 mRNA is transcribed in the ventral photoreceptors. On Western blots probed with a polyclonal antibody raised against the C-terminus of LCNG1, a 100-kDa band and an 80-kDa band was labeled in the membrane protein preparations from brain and ventral eye, respectively. The labeling of these bands was blocked by preabsorption of the antibody with the antigen, indicating the labeling specificity. Immunocytochemistry and confocal microscopy were applied to investigate the subcellular localization of this antigen. Immunolabeling was highly localized in the transducing lobes of ventral eye photoreceptors and lateral eye photoreceptors. In both cases, the labeling was associated with membrane regions specialized for phototransduction, but the exact pattern appeared to be somewhat different in the two eyes. Preabsorption of the antiserum with antigen abolished the labeling, confirming specificity. The results lend support to the hypothesis that a cGMP-gated channel is directly involved in the phototransduction process.

We examined whether lateral spread of adaptation can be observed in the electroretinogram in humans. Specifically, we tested whether the luminance level of a surrounding, nonmodulated annulus affects the multifocal electroretinogram (ERG) response of a modulated central area. Multifocal electroretinograms were recorded in response to an array of 37 unscaled hexagons subtending a retinal area of 38 deg × 35 deg. Responses were recorded in six control subjects. In the first series of experiments, only the center hexagon was modulated, while the surrounding 36 hexagons were held constant at either 0.45, 172, or 340 cd/m2. In a subsequent series of control experiments, modulation depth of the center hexagon was varied and the proximity of the surrounding hexagon systematically altered. For the center-modulated condition, response amplitude and implicit time for the first-order kernel response significantly decreased as a function of increasing surround luminance. Control experiments demonstrated that the effect of the surround illumination was not due to scattered light but was influenced by the proximity of the surrounding annulus. These results demonstrate that lateral adaptation influences can be measured using the multifocal ERG.

Here we examine the patterns of connections between the zona incerta (ZI) of the thalamus and the major visual centers of the rat brain, namely the retina, dorsal lateral geniculate nucleus (LGd), superficial layers of the superior colliculus (SCs), and occipital cortex (Oc1). Injections of the tracers biotinylated dextran or cholera toxin subunit b were made into each of these centers, as well as ZI itself, by using stereotaxic coordinates. Rat brains were then aldehyde-fixed and processed using standard methods. We show that the retina, LGd, SCs, and Oc1 all have connections with ZI; moreover, that each of these connections make a very distinct territory or subsector within the most lateral ZI regions. This subsector of connectivity with the visual centers does not respect the well-defined cytoarchitectonic sectors of ZI, being made up of small zones in the dorsal, ventral, and caudal sectors. Often, a distinctive “horse-shoe” pattern is evident, particularly after retinal and Oc1 injections. Tracer injections into topographically distinct regions of the LGd, SCs, or Oc1 results in no shift in the spatial location of labelling within ZI; after each injection, labelling is always seen within the lateral edge of the nucleus. Labelled terminals and cells are seen after LGd and SCs injections, while only labelled terminals are seen after retinal and Oc1 injections. Although the precise function of this novel visual subsector is not known, these early findings suggest that ZI may be in a position to integrate visual information together with the other somatosensory, motor, and visceral information that it receives.

It has often been assumed that the recovery of the a-wave from its trough is caused by the intrusion of the b-wave. This study examined the recovery following the a-wave trough using intraretinal recordings in dark-adapted intact cat retina. Adult cats were anesthetized and paralyzed. The vitreal ERG was recorded between the vitreous humor and a reference electrode near the eye. Intraretinal recordings were made by referencing a microelectrode to the vitreal electrode. Bright flashes of diffuse white light were used to elicit a- and b-waves. Intravitreal injections of 2-amino-4-phosphonobutyrate (APB), cis 2,3-piperidine dicarboxylic acid (PDA), and kynurenic acid (KYN) were used to block the responses of bipolar and horizontal cells. Intravitreal injections of UL-FS 49 or DK-AH 269 were used to block Ih, a hyperpolarization-activated potassium current. Since the microelectrode was referenced to the vitreal electrode, recordings from the inner retina showed only the oscillatory potentials and b-waves. In the inner retina, the potential was flat until the b-wave became measurable, ∼17 ms from the onset of the flash. The a-wave started to appear as the microelectrode reached the photoreceptors and its amplitude increased with depth until the microelectrode reached the choroid. The a-wave peaked at ∼8 ms in response to flashes that saturated its amplitude and then began to recover well before any inner retinal responses were apparent. After injections of APB, PDA, and KYN, vitreal and intraretinal recordings showed only the a-wave, which consisted of an increase to peak at ∼10 ms followed by a recovery to a plateau which was reached at ∼25 ms. Blockers of Ih reduced the recovery, but did not eliminate it. The a-wave peaks and partially recovers before the b-wave intrudes. Both phases survive blockers of second-order neurons which implies that the photoreceptors generate both the rising and recovery phase of the a-wave. The recovery phase may be due to a current generated by the inner segment of photoreceptors.

Expectations about future motions can influence both perceptual judgements and pursuit eye movements. However, it is not known whether these two effects are due to shared processing, or to separate mechanisms with similar properties. We have addressed this question by providing subjects with prior information about the likely direction of motion in an upcoming random-dot motion display and measuring both the perceptual judgements and pursuit eye movements elicited by the stimulus. We quantified the subjects' responses by computing oculometric curves from their pursuit eye movements and psychometric curves from their perceptual decisions. Our results show that directional cues caused similar shifts in both the oculometric and psychometric curves toward the expected motion direction, with little change in the shapes of the curves. Prior information therefore biased the outcome of both eye movement and perceptual decisions without systematically changing their thresholds. We also found that eye movement and perceptual decisions tended to be the same on a trial-by-trial basis, at a higher frequency than would be expected by chance. Furthermore, the effects of prior information were evident during pursuit initiation, as well as during pursuit maintenance, indicating that prior information likely influenced the early processing of visual motion. We conclude that, in our experiments, expectations caused similar effects on both pursuit and perception by altering the activity of visual motion detectors that are read out by both the oculomotor and perceptual systems. Applying cognitive factors such as expectations at relatively early stages of visual processing could act to coordinate the metrics of eye movements with perceptual judgements.

In the retina of the African clawed frog (Xenopus laevis), melatonin is synthesized by the photoreceptors at night, and binds to receptors that likely mediate paracrine responses. Melatonin appears to alter the sensitivity of the retinal cells to light, and may play a key role in regulating important circadian events that occur in the eye. A polyclonal antibody was raised against a 13 amino acid peptide corresponding to a region of the third cytoplasmic loop of the Xenopus laevis Mel1c melatonin receptor. Western blot analysis revealed a major immunoreactive band of approximately 60 kD in neural retina and retinal pigment epithelium (RPE) membranes. Immunocytochemical labeling of sections of Xenopus eyes demonstrated intense melatonin receptor-like immunoreactivity in the inner plexiform layer (IPL). Immunolabeling with antibodies to glutamate decarboxylase (GAD) or tyrosine hydroxylase (TOH) appeared to co-localize with the melatonin receptor immunoreactivity in different sublaminas of the IPL. This suggests that both GABAergic and dopaminergic amacrine cells express melatonin receptor protein. There were also some melatonin receptor immunoreactive varicose fibers in the IPL that did not co-localize with either TOH or GAD, and may represent efferent fibers, since they could be followed into the optic nerve. Melatonin receptor immunoreactivity was also present on cell soma in the ganglion cell layer. Furthermore, a moderate level of melatonin receptor immunoreactivity was observed in the RPE and rod and cone photoreceptor cells. The presence of melatonin receptor immunoreactivity in these cells supports previous observations of melatonin receptor RNA expression in multiple cell types in the Xenopus retina. Expression of melatonin receptor protein in the photoreceptors suggests that melatonin may have a direct action on these cells.

The effects of kainate receptor-preferring glutamate ligands were tested on the electroretinogram (ERG) of the Xenopus retina. Kainate, domoic acid, and 5-iodowillardiine (20–100 μM) acted similarly in every respect. They increased peak amplitudes of the ERG a-, b-, and d-waves significantly over controls. The AMPA-specific antagonist, GYKI 53655, prevented a kainate-induced increase in ERG a- and d-waves, but was without effect on an increase in the b-wave. Once the effect of agonist on the b-wave had peaked, the ERG began to subside, leading to its nearly complete disappearance within 20 min. Prior exposure to GYKI followed by a combination of GYKI + agonist did not significantly slow the rate of b-wave disappearance. Our results indicate that (1) AMPA receptors contribute to ERG a- and d-waves. (2) The kainate-evoked increase in ERG a-, b-, and d-waves probably results, in part, from an excitotoxic swelling of inner retinal processes. (3) The inner retina has a population of GYKI-resistant, kainate-sensitive receptors which may contribute to b-wave generation.

To characterize the temporal and spatial parameters of gamma activity evoked by visual stimuli in areas V1 and V4 of the monkey cortex, we recorded the electrocorticogram (ECoG) with an implanted array of 28 and 31 subdural electrodes placed over the surface of the operculum in two anesthetized monkeys. This intermediate level of recordings should help to bridge the gap between multiunit and scalp recordings. Both averaged and single-trial responses to small flashed stimuli, for which we varied the retinotopic position, the luminance and the color, were analyzed in the time-frequency domain using a wavelet-based decomposition of the signal. Large gamma oscillations (40–55 Hz), not phase locked to stimulus onset, were observed during the whole stimulus presentation, whereas visual evoked potentials (VEPs) were present mainly at stimulus onset and offset. Cortical mapping showed that both activities were restricted in spatial extent and followed the retinotopic organization of area V1 on the operculum, thus strongly suggesting they were generated in the underlying cortex. Oscillatory burst detection in single trials showed that one to two bursts lasting from 100 ms to 500 ms occurred in the first 500 ms following stimulus onset, and that bursts occurring during the subsequent phases of the response had a smaller amplitude and duration. Finally, we showed that gamma activity was stronger with higher luminances and for red than for green, yellow, or white stimuli of same luminance. In one animal we recorded gamma activity over area V4. This was of lower magnitude than the activity recorded over V1 and was delayed by 40 ms with respect to the beginning of gamma activity in V1, in contrast with the VEPs that were delayed by 20 ms only. Both gamma oscillations and early VEP followed the retinotopic organization of V4 over the prelunate gyrus. The results show that gamma oscillations are dependent upon the same parameters as the VEPs (retinotopic position, luminance, and color). However, the differences in the time course of VEPs and gamma activity (transient vs. sustained) suggests that these two responses may reflect different cell populations, different networks, or different firing modes.

We model feedback from primary visual cortex to the dorsal lateral geniculate nucleus (dLGN). This feedback makes dLGN neurons sensitive to orientation discontinuity (Sillito et al., 1993; Cudeiro & Sillito, 1996). In the model, each dLGN neuron receives retinotopic input driven by layer 6 cortical neurons in a full set of orientation columns. Excitation is monosynaptic, while inhibition is through perigeniculate neurons and dLGN interneurons. The stimulus consists of drifting gratings, one within and the other outside a circular region centered over the receptive field of the model dLGN relay neuron we study. They appear as a single grating when they are aligned with equal contrast. The model reproduces experimental results showing an increasing inhibitory effect of feedback on the firing rate of dLGN neurons as the two gratings move towards the aligned position. Moreover, enhancement of dLGN cell center-surround antagonism by feedback is revealed by measuring the responses to drifting gratings inside a circular window, as a function of window radius. This effect is related to the observed length tuning of dLGN cells. Sensitivity to orientation discontinuity could be mediated in the model by feedback from either simple or complex cells. The model puts constraints on the feedback synaptic footprint and shows that its elongated shape does not play a crucial role in sensitivity to orientation discontinuity. The inhibitory component of feedback must predominate overall, but the feedback signal from a cortical neuron to a dLGN neuron with the same or nearby receptive-field center can be dominated by excitation. Predictions of the model include (1) robust stimuli for layer 6 cortical neurons give pronounced nonlinearities in the responses of dLGN neurons; (2) the sensitivity to orientation discontinuity at low contrast is twice that at high contrast.

Based on cytoarchitectonic criteria, the primate pulvinar nucleus has been subdivided into medial (PM), lateral (PL), and inferior (PI) regions. However, these subdivisions show no correlation with those established by electrophysiological, immunocytochemical, or neuroanatomical tracer studies. In this work, we studied the connections of the pulvinar nucleus of Cebus monkey with visual areas V1, V2, V4, MT, and PO by means of retrograde fluorescent tracers injected into these areas. Based on the projection zones to cortical visual areas, the visual portion of the pulvinar of Cebus monkey was subdivided into three subregions: P1, P2, and P3, similar to those described in the macaque (Ungerleider et al., 1984). In Cebus, P1 includes the centrolateral portion of traditionally defined PI and adjacent portion of PL. P2 is located in the dorsal portion of PL and P3 includes the medial portion of PI and extends dorsally into adjacent PL and PM. In addition, we studied the histology of the pulvinar using multiple criteria, such as cytoarchitecture and myeloarchitecture; histochemistry for cytochrome oxidase, NADPH-diaphorase, and acetylcholinesterase; and immunocytochemistry for two calcium-binding proteins, calbindin and parvalbumin, and for a neurofilament recognized by the SMI-32 antibody. Some of these stains, mainly calbindin, showed additional subdivisions of the Cebus pulvinar, beyond the traditional PI, PL, and PM. Based on this immunohistochemical staining, the border of PI is moved dorsally above the brachium of the superior colliculus and PI can be subdivided in five regions (PIP, PIM, PIC, PIL, and PILS). Regions P1, P2, and P3 defined based on efferent connections with cortical visual areas are not architectonically/neurochemically homogeneous. Rather they appear to consist of further chemoarchitectonic subdivisions. These distinct histochemical regions might be related to different functional modules of visual processing within one connectional area.

A psychophysical procedure, classical conditioning of respiration, was used to measure contrast sensitivity to positive- and negative-contrast discs (8-deg diameter) in goldfish after crushing one optic nerve intraorbitally. In five out of six fish, sensitivity recovered to normal. Recovery times for negative contrasts were significantly shorter than for positive contrasts. The average times postcrush of initial responding to negative and positive contrasts were 23.8 and 30.6 days, and for threshold to come within 0.5 log unit of control values was 29.8 and 39.8 days, respectively. Thereafter, recovery to normal sensitivity was significantly faster for positive contrasts. These results parallel prior observations of neural activity in tectum after optic nerve crush: an early phase of OFF responding followed by a more sudden recovery of ON responding.

The contribution of interocular orientation differences to depth perception, at either the neuronal or the psychophysical level, is unclear. To understand the responses of binocular neurons to orientation disparity, we extended the energy model of Ohzawa et al. (1990) to incorporate binocular differences in receptive-field orientation. The responses of the model to grating stimuli with interocular orientation differences were examined, along with the responses to random dot stereograms (RDS) depicting slanted surfaces. The responses to combinations of stimulus orientations in the two eyes were left–right separable, which means there was no consistent response to the binocular orientation difference. All existing neuronal data concerning orientation disparity can be well described by this type of model (even a version with no disparity selectivity). The disparity sensitive model is nonetheless sensitive to changes in RDS slant, although it requires narrow orientation bandwidth to produce substantial modulation. The disparity-insensitive model shows no selectivity to slant in this stimulus. Several modifications to the model were attempted to improve its selectivity for orientation disparity and/or slant. A model built by summing several disparity-sensitive models showed left–right inseparable responses, responding maximally to a consistent orientation difference. Despite this property, the selectivity for slant in RDS stimuli was no better than the simple disparity-selective model. The range of models evaluated here demonstrate that interocular orientation differences are neither necessary nor sufficient for signaling slant. In contrast, within the framework of the energy model, positional disparity sensitivity appears to be both necessary and sufficient.