Izumi Ohzawa

Receptive-field dynamics in the central visual pathways

Neurons in the central visual pathways process visual images within a localized region of space, and a restricted epoch of time. Although the receptive field (RF) of a visually responsive neuron is inherently a spatiotemporal entity, most studies have focused exclusively on spatial aspects of RF structure. Recently, however, the application of sophisticated RF-mapping techniques has enabled neurophysiologists to characterize RFs in the joint domain of space and time. Studies that use these techniques have revealed that neurons in the geniculostriate pathway exhibit striking RF dynamics. For a majority of cells, the spatial structure of the RF changes as a function of time; thus, these RFs can be characterized adequately only in the space-time domain. In this review, the spatiotemporal RF structure of neurons in the lateral geniculate nucleus and primary visual cortex is discussed.

On the neurophysiological organization of binocular vision

The considerable mixing in the visual cortex, of signals from left and right eyes, provides an abundant population of binocularly activated neurons. Based on this and on the fact that cortical cells respond best to different ranges of retinal disparities, it has been proposed that these neurons form the physiological substrate of stereoscopic depth discrimination. We outline reasons here for addressing first the more fundamental issue of the rules of convergence in the visual cortex, for input from the two eyes. We show that most of this convergence may be described by a linear summation process. However, there is a nonlinear mechanism that maintains binocular interaction regardless of large differences in stimulus strength between the eyes. This finding suggests that a cell which appears to be dominated by one eye, when monocular tests are conducted, may respond equally under binocular conditions. In this case, binocular processing for all cortical cells could be uniform and independent of the ocular dominance values determined monocularly. With respect to a neural mechanism for the processing of information concerning different depths in space, we propose an alternative to the conventional notion. First, we identify fundamental problems with the current view. Second, we describe a procedure which allows us to distinguish between the conventional view and our alternative proposal. Standard receptive field mapping techniques are not adequate for determining phase-disparity relationships of the type we require. Therefore, we have employed a reverse correlation procedure which enables efficient and detailed mapping of receptive field structure. Third, we describe preliminary data concerning the physiological mechanism of stereoscopic depth discrimination.

Suppression outside the classical cortical receptive field.

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 cats 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 neurons 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.

Linear and nonlinear contributions to orientation tuning of simple cells in the cat's striate cortex

Orientation selectivity is one of the most conspicuous receptive-field (RF) properties that distinguishes neurons in the striate cortex from those in the lateral geniculate nucleus (LGN). It has been suggested that orientation selectivity arises from an elongated array of feedforward LGN inputs (Hubel & Wiesel, 1962). Others have argued that cortical mechanisms underlie orientation selectivity (e.g. Sillito, 1975; Somers et al., 1995). However, isolation of each mechanism is experimentally difficult and no single study has analyzed both processes simultaneously to address their relative roles. An alternative approach, which we have employed in this study, is to examine the relative contributions of linear and nonlinear mechanisms in sharpening orientation tuning. Since the input stage of simple cells is remarkably linear, the nonlinear contribution can be attributed solely to cortical factors. Therefore, if the nonlinear component is substantial compared to the linear contribution, it can be concluded that cortical factors play a prominent role in sharpening orientation tuning. To obtain the linear contribution, we first measure RF profiles of simple cells in the cats striate cortex using a binary m-sequence noise stimulus. Then, based on linear spatial summation of the RF profile, we obtain a predicted orientation-tuning curve, which represents the linear contribution. The nonlinear contribution is estimated as the difference between the predicted tuning curve and that measured with drifting sinusoidal gratings. We find that measured tuning curves are generally more sharply tuned for orientation than predicted curves, which indicates that the linear mechanism is not enough to account for the sharpness of orientation-tuning. Therefore, cortical factors must play an important role in sharpening orientation tuning of simple cells. We also examine the relationship of RF shape (subregion aspect ratio) and size (subregion length and width) to orientation-tuning halfwidth. As expected, predicted tuning halfwidths are found to depend strongly on both subregion length and subregion aspect ratio. However, we find that measured tuning halfwidths show only a weak correlation with subregion aspect ratio, and no significant correlation with RF length and width. These results suggest that cortical mechanisms not only serve to sharpen orientation tuning, but also serve to make orientation tuning less dependent on the size and shape of the RF. This ensures that orientation is represented equally well regardless of RF size and shape.

A comparison of contrast detection and discrimination.

In order to complement previous studies of contrast detection, we have examined the effects of three stimulus variables (spatial frequency, retinal illuminance, retinal locus) and one visual disorder (amblyopia) on contrast discrimination. Although each factor has a profound effect on the detection of gratings on otherwise unpatterned displays, we find a similar dipper-shaped contrast discrimination function and similar supra-threshold Weber fractions for contrast under all these conditions.

Joint-encoding of motion and depth by visual cortical neurons: neural basis of the Pulfrich effect.

Motion and stereoscopic depth are fundamental parameters of the structural analysis of visual scenes. Because they are defined by a difference in object position, either over time or across the eyes, a common neural machinery may be used for encoding these attributes. To examine this idea, we analyzed responses of binocular complex cells in the cat striate cortex to stimuli of various intra- and interocular spatial and temporal shifts. We found that most neurons exhibit space–time-oriented response profiles in both monocular and binocular domains. This indicates that these neurons encode motion and depth jointly, and it explains phenomena such as the Pulfrich effect. We also found that the relationship between neuronal tuning of motion and depth conforms to that predicted by the use of motion parallax as a depth cue. These results demonstrate a joint-encoding of motion and depth at an early cortical stage.

Mechanisms of stereoscopic vision: the disparity energy model

The past year has seen significant advances in our understanding of the role played by the primary visual cortex (V1) in stereoscopic vision. Recently, the mechanism by which complex cells in V1 respond to random-dot stereograms has been characterized; it appears that their response properties greatly reduce the complexity of one of the critical links for stereopsis, the correspondence problem.

Neuronal Mechanisms Underlying Stereopsis: How Do Simple Cells in the Visual Cortex Encode Binocular Disparity?

Binocular neurons in the visual cortex are thought to form the neural substrate for stereoscopic depth perception. How are the receptive fields of these binocular neurons organized to encode the retinal position disparities that arise from binocular parallax? The conventional notion is that the two receptive fields of a binocular neuron have identical shapes, but are spatially offset from the point of retinal correspondence (zero disparity). We consider an alternative disparity-encoding scheme, in which the two receptive fields may differ in shape (or phase), but are centered at corresponding retinal locations. Using a reverse-correlation technique to obtain detailed spatiotemporal receptive-field maps, we provide support for the latter scheme. Specifically, we show that receptive-field profiles for the left and right eyes are matched for cells that are tuned to horizontal orientations of image contours. However, for neurons tuned to vertical orientations, the left and right receptive fields are predominantly dissimilar in shape. These results show that the striate cortex possesses a specialized mechanism for processing vertical contours, which carry the horizontal–disparity information needed for stereopsis. Thus, in a major modification to the traditional notion of the neural basis of stereopsis, we propose that binocular simple cells encode horizontal disparities in terms of phase at multiple spatial scales. Implications of this scheme are discussed with respect to the size–disparity correlation observed in psychophysical studies.

Receptive field properties of cells in area 19 of the cat

SummaryWe have recorded extracellularly from single cells in area 19 of the cat for the purpose of providing a quantitative description of response characteristics. A prominent feature of this area is a high incidence of cells that are end-stopped. Drifting sinusoidal gratings were used to determine spatial and temporal characteristics of the discharge region. In addition, we have conducted independent tests to characterize end zones of receptive fields. When a grating patch was used to stimulate the discharge region alone, all of the cells showed a band-pass spatial frequency tuning characteristic. The optimal spatial frequency ranged from 0.1 to 1.13 cycles/deg, and the distribution had a peak at 0.4 cycles/deg. The bandwidth at half peak amplitude ranged widely from 0.7 to 3.3 octaves (mean 2.0 octaves). When gratings were also presented to the end zones, responses to stimulation of the central region were suppressed. The surround was phase-insensitive in that the relative phase between the grating in the two regions generally did not affect the strength of the suppression. To determine spatial characteristics of the end-zone inhibition, the spatial frequency of the end-zone grating was changed while that for the central pattern was fixed. All cells showed a bandpass characteristic for end-zone inhibition, but in each case, the tuning width was broader than that for excitation. The mean spatial frequency bandwidth of end-zone inhibition was 2.7 octaves. The peak of the inhibition generally coincided with the peak of the excitatory spatial frequency tuning of the discharge center. Considered together, these results show that neurons in area 19 share common properties with those in areas 17 and 18, but they exhibit phase-insensitve end-zone inhibition more frequently.

Neural Basis for Stereopsis from Second-Order Contrast Cues

Humans and animals use visual cues such as brightness and color boundaries to identify objects and navigate through environments. However, even when these cues are not available, we can effortlessly perform these tasks by using second-order cues such as contrast variation (envelope) of patterns on surfaces. Previously, numerous psychophysical studies examined properties of binocular depth processing based on the contrast-envelope cues and suggested the existence of a stereo system that uses these cues. However, its physiological substrate has not been identified yet. Here, we show that a subset of cortical neurons in cat area 18 show binocular interactions for the contrast-envelope stimuli. These neurons are capable of representing a variety of depths in the three-dimensional space based on the information available from contrast cues alone. Furthermore, these neurons show similar disparity-tuning curves for borders defined by both luminance and contrast cues. This cue-invariant tuning is consistent with a linear binocular convergence model for monocular luminance and contrast-envelope processing pathways.