Two hierarchically organized neural systems for object information in human visual cortex

The primate visual system is broadly organized into two segregated processing pathways, a ventral stream for object vision and a dorsal stream for space vision. Here, evidence from functional brain imaging in humans demonstrates that object representations are not confined to the ventral pathway, but can also be found in several areas along the dorsal pathway. In both streams, areas at intermediate processing stages in extrastriate cortex (V4, V3A, MT and V7) showed object-selective but viewpoint- and size-specific responses. In contrast, higher-order areas in lateral occipital and posterior parietal cortex (LOC, IPS1 and IPS2) responded selectively to objects independent of image transformations. Contrary to the two-pathways hypothesis, our findings indicate that basic object information related to shape, size and viewpoint may be represented similarly in two parallel and hierarchically organized neural systems in the ventral and dorsal visual pathways.     

Neural correlates of disparity-defined shape discrimination in the human brain

Binocular disparity, the slight differences between the images registered by our two eyes, provides an important cue when estimating the three-dimensional (3D) structure of the complex environment we inhabit. Sensitivity to binocular disparity is evident at multiple levels of the visual hierarchy in the primate brain, from early visual cortex to parietal and temporal areas. However, the relationship between activity in these areas and key perceptual functions that exploit disparity information for 3D shape perception remains an important open question. Here we investigate the link between human cortical activity and the perception of disparity-defined shape, measuring fMRI responses concurrently with psychophysical shape judgments. We parametrically degraded the coherence of shapes by shuffling the spatial position of dots whose disparity defined the 3D structure and investigated the effect of this stimulus manipulation on both cortical activity and shape discrimination. We report significant relationships between shape coherence and fMRI response in both dorsal (V3, hMT+/V5) and ventral (LOC) visual areas that correspond to the observers' discrimination performance. In contrast to previous suggestions of a dichotomy of disparity-related processes in the ventral and dorsal streams, these findings are consistent with proposed interactions between these pathways that may mediate a continuum of processes important in perceiving 3D shape from coarse contour segmentation to fine curvature estimation.     

Neural representation of three-dimensional features of manipulation objects with stereopsis

In the first part of this article, we review our neurophysiological studies of the hand-manipulation-related neurons in the anterior part of the lateral bank of the intraparietal sulcus (area AIP) . We describe the properties of visually responsive neurons in area AIP. Object-type visual-dominant neurons responded to the sight of objects and showed selectivity not only for simple geometrical shapes, but also for complex objects such as a knob-in-groove and a plate-in-groove. Some of the object-type visual-dominant neurons showed selectivity for the orientation of the longitudinal axis or the plane (surface) of a plate or a ring. In the second part of this article, we review our study of binocular visual neurons in the caudal part of the lateral bank of the intraparietal sulcus (c-IPS area), in particular, of axis-orientation-selective (AOS) neurons and surface-orientation-selective (SOS) neurons. AOS neurons preferred long and thin stimuli, were sensitive to binocular disparity, and tuned to the axis orientation in three-dimensional (3D) space. SOS neurons preferred broad and flat stimuli and were tuned to the surface orientation in depth. Some SOS neurons responded to a square in a random dot stereogram (RDS) with orientation tuning, suggesting that they encode surface orientation from a disparity gradient. Others responded to solid figure stereograms with orientation disparity and/or width disparity. It was concluded that the c-IPS area is a higher center for stereopsis, which integrates various binocular disparity signals received from the V3 complex and other prestriate areas to represent the neural code for 3D features. It may send projections to the AIP area and contribute to visual adjustment of the shape of the handgrip and/or hand orientation for manipulation and grasping. Neurons of the AIP area may also receive monocular cues of depth from the ventral visual pathway to discriminate the 3D shape of the object of manipulation. Received: 21 September 1998 / Accepted: 23 March 1999     

Responses to contour features in macaque area V4

The ventral pathway in visual cortex is responsible for the perception of shape. Area V4 is an important intermediate stage in this pathway, and provides the major input to the final stages in inferotemporal cortex. The role of V4 in processing shape information is not yet clear. We studied V4 responses to contour features (angles and curves), which many theorists have proposed as intermediate shape primitives. We used a large parametric set of contour features to test the responses of 152 V4 cells in two awake macaque monkeys. Most cells responded better to contour features than to edges or bars, and about one-third exhibited systematic tuning for contour features. In particular, many cells were selective for contour feature orientation, responding to angles and curves pointing in a particular direction. There was a strong bias toward convex (as opposed to concave) features, implying a neural basis for the well-known perceptual dominance of convexity. Our results suggest that V4 processes information about contour features as a step toward complex shape recognition.     

Neural coding of global form in the human visual cortex

Extensive psychophysical and computational work proposes that the perception of coherent and meaningful structures in natural images relies on neural processes that convert information about local edges in primary visual cortex to complex object features represented in the temporal cortex. However, the neural basis of these mid-level vision mechanisms in the human brain remains largely unknown. Here, we examine functional MRI (fMRI) selectivity for global forms in the human visual pathways using sensitive multivariate analysis methods that take advantage of information across brain activation patterns. We use Glass patterns, parametrically varying the perceived global form (concentric, radial, translational) while ensuring that the local statistics remain similar. Our findings show a continuum of integration processes that convert selectivity for local signals (orientation, position) in early visual areas to selectivity for global form structure in higher occipitotemporal areas. Interestingly, higher occipitotemporal areas discern differences in global form structure rather than low-level stimulus properties with higher accuracy than early visual areas while relying on information from smaller but more selective neural populations (smaller voxel pattern size), consistent with global pooling mechanisms of local orientation signals. These findings suggest that the human visual system uses a code of increasing efficiency across stages of analysis that is critical for the successful detection and recognition of objects in complex environments.     

Visuo-haptic object-related activation in the ventral visual pathway

The ventral pathway is involved in primate visual object recognition. In humans, a central stage in this pathway is an occipito-temporal region termed the lateral occipital complex (LOC), which is preferentially activated by visual objects compared to scrambled images or textures. However, objects have characteristic attributes (such as three-dimensional shape) that can be perceived both visually and haptically. Therefore, object-related brain areas may hold a representation of objects in both modalities. Using fMRI to map object-related brain regions, we found robust and consistent somatosensory activation in the occipito-temporal cortex. This region showed clear preference for objects compared to textures in both modalities. Most somatosensory object-selective voxels overlapped a part of the visual object-related region LOC. Thus, we suggest that neuronal populations in the occipito-temporal cortex may constitute a multimodal object-related network.     

Distributed population mechanism for the 3-D oculomotor reference frame transformation

Human saccades require a nonlinear, eye orientation-dependent reference frame transformation to transform visual codes to the motor commands for eye muscles. Primate neurophysiology suggests that this transformation is performed between the superior colliculus and brain stem burst neurons, but provides little clues as to how this is done. To understand how the brain might accomplish this, we trained a 3-layer neural net to generate accurate commands for kinematically correct 3-D saccades. The inputs to the network were a 2-D, eye-centered, topographic map of Gaussian visual receptive fields and an efference copy of eye position in 6-dimensional, push-pull "neural integrator" coordinates. The output was an eye orientation displacement command in similar coordinates appropriate to drive brain stem burst neurons. The network learned to generate accurate, kinematically correct saccades, including the eye orientation-dependent tilts in saccade motor error commands required to match saccade trajectories to their visual input. Our analysis showed that the hidden units developed complex, eye-centered visual receptive fields, widely distributed fixed-vector motor commands, and "gain field"-like eye position sensitivities. The latter evoked subtle adjustments in the relative motor contributions of each hidden unit, thereby rotating the population motor vector into the correct correspondence with the visual target input for each eye orientation: a distributed population mechanism for the visuomotor reference frame transformation. These findings were robust; there was little variation across networks with between 9 and 49 hidden units. Because essentially the same observations have been reported in the visuomotor transformations of the real oculomotor system, as well as other visuomotor systems (although interpreted elsewhere in terms of other models) we suggest that the mechanism for visuomotor reference frame transformations identified here is the same solution used in the real brain.     

Direction selectivity in V1 of alert monkeys: evidence for parallel pathways for motion processing

In primary visual cortex (V1) of macaque monkeys, motion selective cells form three parallel pathways. Two sets of direction selective cells, one in layer 4B, and the other in layer 6, send parallel direct outputs to area MT in the dorsal cortical stream. We show that these two outputs carry different types of spatial information. Direction selective cells in layer 4B have smaller receptive fields than those in layer 6, and layer 4B cells are more selective for orientation. We present evidence for a third direction selective pathway that flows through V1 layers 4Cm (the middle tier of layer 4C) to layer 3. Cells in layer 3 are very selective for orientation, have the smallest receptive fields in V1, and send direct outputs to area V2. Layer 3 neurons are well suited to contribute to detection and recognition of small objects by the ventral cortical stream, as well as to sense subtle motions within objects, such as changes in facial expressions.     

An intracranial event-related potential study on transformational apparent motion. Does its neural processing differ from real motion?

How the brain processes visual stimuli has been extensively studied using scalp surface electrodes and magnetic resonance imaging. Using these and other methods, complex gratings have been shown to activate the ventral visual stream, whereas moving stimuli preferentially activate the dorsal stream. In the current study, a first experiment assessed brain activations evoked by complex gratings using intracranial electroencephalography in 10 epileptic patients implanted with subdural electrodes. These stimuli of intermediate levels of complexity were presented in such a way that transformational apparent motion (TAM) was perceived. Responses from both the ventral and the dorsal pathways were obtained. The response characteristics of visual area 4 and the fusiform cortex were of similar amplitudes, suggesting that both ventral areas are recruited for the processing of complex gratings. On the other hand, TAM-induced responses of dorsal pathway areas were relatively noisier and of lower amplitudes, suggesting that TAM does not activate motion-specific structures to the same extent as does real motion. To test this hypothesis, we examined the activity evoked by TAM in comparison to the one produced by real motion in a patient implanted with the same subdural electrodes. Findings demonstrated that neural response to real motion was much stronger than that evoked by TAM, in both the primary visual cortex (V1) and other motion-sensitive areas within the dorsal pathway. These results support the conclusion that apparent motion, even if perceptually similar to real motion, is not processed in a similar manner.     

Processing of kinetically defined boundaries in areas V1 and V2 of the macaque monkey

We recorded responses in 107 cells in the primary visual area V1 and 113 cells in the extrastriate visual area V2 while presenting a kinetically defined edge or a luminance contrast edge. Cells meeting statistical criteria for responsiveness and orientation selectivity were classified as selective for the orientation of the kinetic edge if the preferred orientation for a kinetic boundary stimulus remained essentially the same even when the directions of the two motion components defining that boundary were changed by 90 degrees. In area V2, 13 of the 113 cells met all three requirements, whereas in V1, only 4 cells met the criteria of 107 that were tested, and even these demonstrated relatively weak selectivity. Correlation analysis showed that V1 and V2 populations differed greatly (P < 1.0 x 10(-6), Student's t-test) in their selectively for specific orientations of kinetic edge stimuli. Neurons in V2 that were selective for the orientation of a kinetic boundary were further distinguished from their counterparts in V1 in displaying a strong, sharply tuned response to a luminance edge of the same orientation. We concluded that selectivity for the orientation of kinetically defined boundaries first emerges in area V2 rather than in primary visual cortex. An analysis of response onset latencies in V2 revealed that cells selective for the orientation of the motion-defined boundary responded about 40 ms more slowly, on average, to the kinetic edge stimulus than to a luminance edge. In nonselective cells, that is, those presumably responding only to the local motion in the stimulus, this difference was only about 20 ms. Response latencies for the luminance edge were indistinguishable in KE-selective and -nonselective neurons. We infer that while responses to luminance edges or local motion are indigenous to V2, KE-selective responses may involve feedback entering the ventral stream at a point downstream with respect to V2.     

Visually guided adjustments of body posture in the roll plane

Body position relative to gravity is continuously updated to prevent falls. Therefore, the brain integrates input from the otoliths, truncal graviceptors, proprioception and vision. Without visual cues estimated direction of gravity mainly depends on otolith input and becomes more variable with increasing roll-tilt. Contrary, the discrimination threshold for object orientation shows little modulation with varying roll orientation of the visual stimulus. Providing earth-stationary visual cues, this retinal input may be sufficient to perform self-adjustment tasks successfully, with resulting variability being independent of whole-body roll orientation. We compared conditions with informative (earth-fixed) and non-informative (body-fixed) visual cues. If the brain uses exclusively retinal input (if earth-stationary) to solve the task, trial-to-trial variability will be independent from the subject’s roll orientation. Alternatively, central integration of both retinal (earth-fixed) and extra-retinal inputs will lead to increasing variability when roll-tilted. Subjects, seated on a motorized chair, were instructed to (1) align themselves parallel to an earth-fixed line oriented earth-vertical or roll-tilted 75° clockwise; (2) move a body-fixed line (aligned with the body-longitudinal axis or roll-tilted 75° counter-clockwise to it) by adjusting their body position until the line was perceived earth-vertical. At 75° right-ear-down position, variability increased significantly (p < 0.05) compared to upright in both paradigms, suggesting that, despite the earth-stationary retinal cues, extra-retinal input is integrated. Self-adjustments in the roll-tilted position were significantly (p < 0.01) more precise for earth-fixed cues than for body-fixed cues, underlining the importance of earth-stable visual cues when estimates of gravity become more variable with increasing whole-body roll.     

Population coding of shape in area V4

Shape is represented in the visual system by patterns of activity across populations of neurons. We studied the population code for shape in area V4 of macaque monkeys, which is part of the ventral (object-related) pathway in primate visual cortex. We have previously found that many macaque V4 neurons are tuned for the curvature and object-centered position of boundary fragments (such as 'concavity on the right'). Here we tested the hypothesis that populations of such cells represent complete shapes as aggregates of boundary fragments. To estimate the population representation of a given shape, we scaled each cell's tuning peak by its response to that shape, summed across cells and smoothed. The resulting population response surface contained 3-8 peaks that represented major boundary features and could be used to reconstruct (approximately) the original shape. This exemplifies how a multi-peaked neural population response can represent a complex stimulus in terms of its constituent elements.     

Neuronal responsiveness in areas 19 and 21a,and the posteromedial lateral suprasylvian cortex of the cat

Responsiveness to slits and pattern stimuli was quantified in a total of 68 cells sampled in the posterior extreme of the lateral suprasylvian (PS) cortex as response indices. The cells were studied in relationship to their locations in several subareas of the PS cortex, including areas 19 (n=15) and 21a (n=32) and the posteromedial lateral suprasylvian cortex (PMLS; n=21). These subareas were identified based on retrograde labelling from area 17 and also supplemented with photic responsiveness. This analysis revealed that each cortical area contains cells expressing different combinations of stimulus features. Area 19 contained two major groups of cells: (1) those with strong end-stop selectivity combined with moderate orientation or direction selectivity, and (2) those with weak end-stop selectivity combined with strong orientation selectivity. The groups of cells with strong or moderate orientation selectivity showed a strong preference for stripe over visual noise patterns and relatively large modulatory responses to motion of individual stripes. The PMLS contained one major group of cells with strong end-stop and direction selectivities and with poor orientation selectivity. They also showed stronger preference for visual noise than cells in the other cortical areas and rather weak modulatory responses. Area 21a contained only one group of cells with strong orientation selectivity and length summation property rather than end-stop selectivity, and they also lacked direction selectivity. These cells exhibited a strong preference for stripe patterns and moderate or weak modulatory responses. Altogether, these findings indicate that each cortical area is specialized in expressing different stimulus features. The two groups of cells in area 19 may encode the position and motion of discontinuous visual elements such as corners and line ends and continuous elements such as lines and edges. PMLS cells may encode the motion of single elements or associated motion of multiple discontinuous elements such as textures and backgrounds. Area 21a cells may specifically encode the orientation of long, continuous elements such as lines and edges. In support of this view, two types of statistical analyses demonstrated that the combinations of the response properties expressed in individual PS cells are highly correlated with their locations in cortical areas and that the anatomical locations of individual PS cells are reliably predicted from the sets of response indices expressed in these cells.     

The role of temporo-parietal junction (TPJ) in global Gestalt perception

Grouping processes enable the coherent perception of our environment. A number of brain areas has been suggested to be involved in the integration of elements into objects including early and higher visual areas along the ventral visual pathway as well as motion-processing areas of the dorsal visual pathway. However, integration not only is required for the cortical representation of individual objects, but is also essential for the perception of more complex visual scenes consisting of several different objects and/or shapes. The present fMRI experiments aimed to address such integration processes. We investigated the neural correlates underlying the global Gestalt perception of hierarchically organized stimuli that allowed parametrical degrading of the object at the global level. The comparison of intact versus disturbed perception of the global Gestalt revealed a network of cortical areas including the temporo-parietal junction (TPJ), anterior cingulate cortex and the precuneus. The TPJ location corresponds well with the areas known to be typically lesioned in stroke patients with simultanagnosia following bilateral brain damage. These patients typically show a deficit in identifying the global Gestalt of a visual scene. Further, we found the closest relation between behavioral performance and fMRI activation for the TPJ. Our data thus argue for a significant role of the TPJ in human global Gestalt perception.     

Neural correlates of consciousness: a definition of the dorsal and ventral streams and their relation to phenomenology

The paper presents a hypothesis for a neural correlate of consciousness. A proposal is made that both the dorsal and ventral streams must be concurrently active to generate conscious awareness and that V1 (striate cortex) provides a serial link between them. An argument is presented against a true extrastriate communication between the dorsal and ventral streams. Secondly, a detailed theory is developed for the structure of the visual hierarchy. Premotor theory states that each organism-object interaction can be described by the two quantitative measures of torque and change in joint position served by the basal ganglia and cerebellum, respectively. This leads to a component theory of motor efference copy providing a fundamental tool for categorizing dorsal and ventral stream networks. The rationale for this is that the dorsal stream specifies spatial coordinates of the external world, which can be coded by the reafference of changes in joint position. The ventral stream is concerned with object recognition and is coded for by forces exerted on the world during a developmental exploratory phase of the organism. The proposed pathways for a component motor efference copy from both the cerebellum and basal ganglia converge on the thalamus and modulate thalamocortical projections via the thalamic reticular nucleus. The origin of the corticopontine projections, which are a massive pathway for cortical information to reach the cerebellum, coincides with the area typically considered as part of the dorsal stream, whereas the entire cortex projects to the striatum. This adds empirical support for a new conceptualization of the visual streams. The model also presents a solution to the binding problem of a neural correlate of consciousness, that is, how a distributed neural network synchronizes its activity during a cognitive event. It represents a reinterpretation of the current status of the visual hierarchy.     

THE PERCEPTION OF SPATIAL RELATIONS IN A PATIENT WITH VISUAL FORM AGNOSIA

It is generally believed that neuropsychological patients presenting with visual agnosia, a deficit on object perception/recognition, have suffered damage to the ventral visual cortical pathway (Milner & Goodale, 1995). Rarely has the ability of such patients to perceive the spatial location of objects been investigated-perhaps because “spatial vision” is thought by some researchers to be mediated exclusively by the dorsal visual cortical pathway. Here we present data on spatial perception in a patient DF, who has a profound visual form agnosia. DF and two control subjects were required to make a copy of the spatial arrangement of a target display of five differently coloured circular tokens using a duplicate set of the same tokens. Spatial performance was analysed in two ways: (1) relative location measured the ability to reconstruct the relative spatial relations between the tokens such as left versus right, above versus below, and nearer versus farther; (2) absolute location measured the exact displacement in millimetres of each token's copied position relative to its true location. DF was able to copy some of relative location relations between the tokens although her abilityto do so was not nearly as accurate as that of the control subjects. Nevertheless, DF's limited appreciation of relative location was enough to enable her to discriminate rather well between spatial patterns of tokens. She could not, however, reconstruct the absolute distance relations between the tokens and showed large displacements of token position compared to the control subjects. Interestingly, although Df was not “normal” in her ability to appreciate the allocentric spatial relations between the locations of the tokens relative to one another, she could accurately process token location egocentrically (i.e. relative to her own body and hand position). Thus, like controls, she was perfectly able to point to and touch all the tokens in an array. These results demonstrate deficits in the ability to perceive spatial relations between objects in a patient with visual form agnosia and suggest that the ventral steam also plays a functional role in spatial vision, particularly allocentric spatial vision.     

The perception of spatial relations in a patient with visual form agnosia

It is generally believed that neuropsychological patients presenting with visual agnosia, a deficit on object perception/recognition, have suffered damage to the ventral visual cortical pathway (Milner & Goodale, 1995). Rarely has the ability of such patients to perceive the spatial location of objects been investigated-perhaps because "spatial vision" is thought by some researchers to be mediated exclusively by the dorsal visual cortical pathway. Here we present data on spatial perception in a patient DF, who has a profound visual form agnosia. DF and two control subjects were required to make a copy of the spatial arrangement of a target display of five differently coloured circular tokens using a duplicate set of the same tokens. Spatial performance was analysed in two ways: (1) relative location measured the ability to reconstruct the relative spatial relations between the tokens such as left versus right, above versus below, and nearer versus farther; (2) absolute location measured the exact displacement in millimetres of each token's copied position relative to its true location. DF was able to copy some of relative location relations between the tokens although her abilityto do so was not nearly as accurate as that of the control subjects. Nevertheless, DF's limited appreciation of relative location was enough to enable her to discriminate rather well between spatial patterns of tokens. She could not, however, reconstruct the absolute distance relations between the tokens and showed large displacements of token position compared to the control subjects. Interestingly, although Df was not "normal" in her ability to appreciate the allocentric spatial relations between the locations of the tokens relative to one another, she could accurately process token location egocentrically (i.e. relative to her own body and hand position). Thus, like controls, she was perfectly able to point to and touch all the tokens in an array. These results demonstrate deficits in the ability to perceive spatial relations between objects in a patient with visual form agnosia and suggest that the ventral steam also plays a functional role in spatial vision, particularly allocentric spatial vision.     

Evidence for a widespread dopaminergic innervation of the human cerebral neocortex

The recent finding that D1 dopamine receptors are present in all neocortical areas of the human brain, does not fit in with the generally held view that the mesocortical dopaminergic pathway is restricted to prefrontal areas. We investigated the brains of 3 patients who died with a unilateral infarction in the ventral midbrain, including the substantia nigra and ventral tegmental area. Compared to the intact side, the D1 receptors in frontal, temporal, parietal and occipital cortices and caudate nucleus at the lesioned side were increased by 27-37%, which is consistent with an up-regulation in response to a depletion of dopamine. These data provide evidence for a more widespread dopaminergic innervation of the human neocortex.     

Laminar processing of stimulus orientation in cat visual cortex

One of the most salient features to emerge in visual cortex is sensitivity to stimulus orientation. Here we asked if orientation selectivity, once established, is altered by successive stages of cortical processing. We measured patterns of orientation selectivity at all depths of the cat's visual cortex by making whole-cell recordings with dye-filled electrodes. Our results show that the synaptic representation of orientation indeed changes with position in the microcircuit, as information passes from layer 4 to layer 2+3 to layer 5. At the earliest cortical stage, for simple cells in layer 4, orientation tuning curves for excitation (depolarization) and inhibition (hyperpolarization) had similar peaks (within 0–7 deg,   n = 11  ) and bandwidths. Further, the sharpness of orientation selectivity covaried with receptive field geometry (   r = 0.74  ) - the more elongated the strongest subregion, the shaper the tuning. Tuning curves for complex cells in layer 2+3 also had similar peaks (within 0–4 deg,   n = 7  ) and bandwidths. By contrast, at a later station, layer 5, the preferred orientation for excitation and inhibition diverged such that the peaks of the tuning curves could be as far as 90 deg apart (average separation, 54 deg;   n = 6  ). Our results support the growing consensus that orientation selectivity is generated at the earliest cortical level and structured similarly for excitation and inhibition. Moreover, our novel finding that the relative tuning of excitation and inhibition changes with laminar position helps resolve prior controversy about orientation selectivity at later phases of processing and gives a mechanistic view of how the cortical circuitry recodes orientation.     

Disparity-selective neurons in area V4 of macaque monkeys

Area V4 is an intermediate stage of the ventral visual pathway providing major input to the final stages in the inferior temporal cortex (IT). This pathway is involved in the processing of shape, color, and texture. IT neurons are also sensitive to horizontal binocular disparity, suggesting that binocular disparity is processed along the ventral visual pathway. In the present study, we examined the processing of binocular disparity information by V4 neurons. We recorded responses of V4 neurons to binocularly disparate stimuli. A population of V4 neurons modified their responses according to changes of stimulus disparity; neither monocular responses nor eye movements could account for this modulation. Disparity-tuning curves were similar for different locations within a neuron's receptive field. Neighboring neurons recorded using a single electrode displayed similar disparity-tuning properties. These findings indicate that a population of V4 neurons is selective for binocular disparity, invariant for the position of the stimulus within the receptive field. The finding that V4 neurons with similar disparity selectivity are clustered suggests the existence of functional modules for disparity processing in V4.