Prestriate Cortex Research Articles

Impaired reading in patients with right hemianopia

A left occipital stroke may result in alexia for two reasons, which may coexist depending on the distribution of the lesion. A lesion of the left lateroventral prestriate cortex or its afferents impairs word recognition ("pure" alexia). If the left primary visual cortex or its afferents are destroyed, resulting in a complete right homonymous hemianopia, rightward saccades during text reading are disrupted ("hemianopic" alexia). By using functional imaging, we showed two separate but interdependent systems involved in reading. The first, subserving word recognition, involved the representation of foveal vision in the left and right primary visual cortex and the ventral prestriate cortex. The second system, responsible for the planning and execution of reading saccades, consisted of the representation of right parafoveal vision in the left visual cortex, the bilateral posterior parietal cortex (left > right), and the frontal eye fields (right > left). Disruption of this distributed neural system was demonstrated in patients with severe right homonymous hemianopia, commensurate with their inability to perform normal reading eye movements. Text reading, before processes involved in comprehension, requires the integration of perceptual and motor processes. We have demonstrated these distributed neural systems in normal readers and have shown how a right homonymous hemianopia disrupts the motor preparation of reading saccades during text reading.

Orientation Constancy in Neurons of Monkey Visual Cortex

We studied the effect of body tilt on the orientation selectivity of single neurons in the visual cortex of an alert monkey. The monkey performed a visual fixation task either in the upright position or with its whole body tilted about the naso-occipital (roll) axis by ±25° or ±30°. We determined the preferred stimulus orientation for 51 of 117 neurons in two or, if possible, three body positions (i.e. with the whole body upright, and tilted either left ear or right ear down). In striate cortex, most of the neurons were of a non-compensatory type, showing a change in the preferred orientation according to the body tilt and the estimated counterrolling of the eye. By contrast, about 40% of the neurons in prestriate cortex were of a compensatory type, preferring similar orientations in all body positions. This suggests that mechanisms which produce orientation constancy with respect to the direction of gravity are implemented at an early stage of cortical processing.

Colour constancy impairments in patients with lesions of the prestriate cortex.

Colour matching and colour constancy were studied in seven patients and 46 control subjects. Subjects were required to match Munsell Colour Chips presented under either identical or different illumination. Three of the patients had deficits in colour constancy, i.e. failure to compensate for the change in the wavelength composition of the illumination. Two of the patients with defective constancy had suffered bilateral cortical damage to the posterior lingual and fusiform gyri, and one patient had a lesion restricted to the same regions of the right hemisphere. Our observations indicate that these cortical areas, which include part of putative human area V4, play an important role in colour constancy.

Space-based and object-based visual attention: shared and specific neural domains.

Visual attention can be primarily allocated to either where an object is in space (with little emphasis on the structure of the object itself) or to the structure of the object (with little emphasis on where in space the object is located). Using PET measures of regional cerebral blood flow (rCBF) to index neural activity, we investigated the shared and specific functional anatomy underlying both of these types of visual attention in a controlled non-cueing non-blocked paradigm that involved identical stimuli across the conditions of interest. The interaction of eye movements with these attentional systems was studied by introducing fixation or free vision as an additional factor. Relative to the control condition, object-based and space-based attention showed significant activations of the left and right medial superior parietal cortex and the left lateral inferior parietal cortex, the left prefrontal cortex and the cerebellar vermis. Significant differential activations were observed during object-based attention in the left striate and prestriate cortex. Space-based attention activated the right prefrontal cortex and the right inferior temporal-occipital cortex. Differential neural activity due to free vision or fixation was observed in occipital areas only. Significant interactions of free vision/fixation on activations due to object-based and space-based attention were observed in the right medial superior parietal cortex and left lateral inferior parietal cortex, respectively. The study provides direct evidence for the importance of the parietal cortex in the control of object-based and space-based visual attention. The results show that object-based and space-based attention share common neural mechanisms in the parietal lobes, in addition to task specific mechanisms in early visual processing areas of temporal and occipital cortices.

Neural mechanisms involved in the processing of global and local aspects of hierarchically organized visual stimuli.

We investigated the functional anatomy involved in sustaining or switching visual attention between different perceptual levels, using functional imaging measures of neural activity. Two experiments were carried out using hierarchically organized letters (i.e. large letters made out of small letters). In a divided-attention task, subjects were required to switch attention between local and global levels. The number of successive stimuli for which subjects had to sustain attention to either the global or local level co-varied significantly with temporal-parietal activations bilaterally. Other activations were also observed in the right orbitofrontal cortex, the right dorsolateral prefrontal cortex, and the right middle temporal gyrus. The number of switches between levels co-varied significantly with activations in the left supplementary motor area and the left medial parietal cortex. In the directed-attention task, subjects were required to attend to either the global or local level of the stimuli throughout all trials; attention to the global aspect resulted in significant activation of the right lingual gyrus while attention to the local aspect significantly activated the left inferior occipital cortex. We suggest that left hemisphere activations with increasing numbers of switches between perceptual levels reflect increased demands on an executive attentional system, while sustained attention to either level activates a predominantly right hemispheric network involving temporal-parietal and dorsolateral prefrontal regions. Overall, the results provide evidence for relative hemispheric specialization for global and local processing in accordance with previous neuropsychological studies. In addition, the findings demonstrate that early visual processing mechanisms in the prestriate cortex are influenced by an attentional system in temporal-parietal areas.

Figure-ground segregation at contours: a neural mechanism in the visual cortex of the alert monkey.

An important task of vision is the segregation of figure and ground in situations of spatial occlusion. Psychophysical evidence suggests that the depth order at contours is defined early in visual processing. We have analysed this process in the visual cortex of the alert monkey. The animals were trained on a visual fixation task which reinforced foveal viewing. During periods of active visual fixation, we recorded the responses of single neurons in striate and prestriate cortex (areas V1, V2, and V3/V3A). The stimuli mimicked situations of spatial occlusion, usually a uniform light (or dark) rectangle overlaying a grating texture of opposite contrast. The direction of figure and ground at the borders of these rectangles was defined by the direction of the terminating grating lines (occlusion cues). Neuronal responses were analysed with respect to figure-ground direction and contrast polarity at such contours. Striate neurons often failed to respond to such stimuli, or were selective for contrast polarity; others were non-selective. Some neurons preferred a certain combination of figure-ground direction and contrast polarity. These neurons were rare both in striate and prestriate cortex. The majority of neurons signalled figure-ground direction independent of contrast polarity. These neurons were only found in prestriate cortex. We explain these responses in terms of a model which also explains neuronal signals of illusory contours. These results suggest that occlusion cues are used at an early level of processing to segregate figure and ground at contours.

Visual field representation in striate and prestriate cortices of a prosimian primate (Galago garnetti).

Microelectrode mapping techniques were used to study the visuotopic organization of the first and second visual areas (V1 and V2, respectively) in anesthetized Galago garnetti, alorisiform prosimian primate. 1) V1 occupies approximately 200 mm2 of cortex, and is pear shaped, rather than elliptical as in simian primates. Neurons in V1 form a continuous (1st-order) representation of the visual field, with the vertical meridian forming most of its perimeter. The representation of the horizontal meridian divides V1 into nearly equal sectors representing the upper quadrant ventrally, and the lower quadrant dorsally. 2) The emphasis on representation of central vision is less marked in Galago than in simian primates, both diurnal and nocturnal. The decay of cortical magnification factor with increasing eccentricity is almost exactly counterbalanced by an increase in average receptive field size, such that a point anywhere in the visual field is represented by a compartment of similar diameter in V1. 3) Although most of the cortex surrounding V1 corresponds to V2, one-quarter of the perimeter of V1 is formed by agranular cortex within the rostral calcarine sulcus, including area prostriata. Although under our recording conditions virtually every recording site in V2 yielded visually responsive cells, only a minority of those in area prostriata revealed such responses. 4) V2 forms a cortical belt of variable width, being narrowest (approximately 1 mm) in the representation of the area centralis and widest (2.5-3 mm) in the representation of the midperiphery (>20 degrees eccentricity) of the visual field. V2 forms a second-order representation of the visual field, with the area centralis being represented laterally and the visual field periphery medially, near the calcarine sulcus. Unlike in simians, the line of field discontinuity in Galago V2 does not exactly coincide with the horizontal meridian: a portion of the lower quadrant immediately adjacent to the horizontal meridian is represented at the rostral border of ventral V2, instead of in dorsal V2. Despite the absence of cytochrome oxidase stripes, the visual field map in Galago V2 resembles the ones described in simians in that the magnification factor is anisotropic. 5) Receptive field progressions in cortex rostral to dorsal V2 suggest the presence of a homologue of the dorsomedial area, including representations of both quadrants of the visual field. These results indicate that many aspects of organization of V1 and V2 in simian primates are shared with lorisiform prosimians, and are therefore likely to have been present in the last common ancestor of living primates. However, some aspects of organization of the caudal visual areas in Galago are intermediate between nonprimates and simian primates, reflecting either an intermediate stage of differentiation or adaptations to a nocturnal niche. These include the shape and the small size of V1 and V2, the modest degree of emphasis on central visual field representation, and the relatively large area prostriata.

Hemispheric specialization for global and local processing: the effect of stimulus category.

Neuropsychological evidence indicates that the global aspect of complex visual scenes is preferentially processed by the right hemisphere, and local aspects are preferentially processed by the left hemisphere. Using letter-based hierarchical stimuli (Navon figures), we recently demonstrated, in a directed-attention task, lateralized neural activity (assessed by positron emission tomography) in the left prestriate cortex during local processing, and in the right prestriate cortex during global processing. Furthermore, temporal-parietal cortex was critically activated bilaterally in a divided-attention task that involved varying the number of target switches between local and global levels of letter-based hierarchical stimuli. Little is known about whether such stimulus categories influence such hemispheric lateralization. We now present data on brain activity, derived from positron emission tomography, in normal subjects scanned during either local or global processing of object-based hierarchical stimuli. We again observe attentional modulation of neural activity in prestriate cortex. There is now greater right-sided activation for local processing and greater left-sided activation for global processing, which is the opposite of that seen with letter-based stimuli. The results suggest that the relative differential hemispheric activations in the prestriate areas during global and local processing are modified by stimulus category.

Spatial cortical patterns of metabolic activity in monkeys performing a visually guided reaching task with one forelimb

The 2-[14C]deoxyglucose method was used to map the metabolic activity in the neocortex of monkeys (Macaca nemestrina) performing a visually guided reaching task with one forelimb. Monkeys received liquid reward for correct, single directional reaching movements, which were required at a rate of about 10 per minute. We estimated the weighted average of local glucose consumption within several neocortical areas, and we reconstructed quantitative, high-resolution, two-dimensional maps of the detailed spatiointensive patterns of activity. Our findings demonstrate the involvement of the striate and prestriate cortices, the inferior intraparietal and superior temporal visual association areas, the frontal eye field and the caudal periprincipal cortex, the primary somatosensory and the related superior intraparietal area, the primary and association auditory cortices, the superior temporal multimodal region, and the premotor, primary, supplementary, and cingulate motor areas. The visual cortex in the superior temporal and the intraparietal sulci, which is concerned with “where”, was activated during visually guided reaching. In contrast, the inferior temporal visual association cortex, which is concerned with “what”, was not involved in our study.We suggest that the activated direction-selective layer four of V1 and the thick stripes of V2 convey visuomotor information to the activated cortex in the posterior bank and the floor of the superior temporal sulcus, which may encode the constantly updated position of the moving forelimb. In parallel, the activated cortex in the ventral part and the lateral bank of the intraparietal sulcus may encode visuospatial information related to the localization of the visual target in the extrapersonal space. Furthermore, the dorsal part of the medial bank of the intraparietal sulcus may be involved in proprioceptive guidance of movement, based on the parallel metabolic effects shown only contralateral to the moving forelimb within this region and the forelimb representations in the primary somatosensory and motor cortices. Finally, the bilaterally activated network including the inferior postarcuate skeletomotor and prearcuate oculomotor cortical fields and the caudal periprincipal region 46 may participate in sensory and oculomotor to motor transformations, in parallel with the medial and lateral intraparietal cortices with which this network is reciprocally interconnected.

Chattering cells: superficial pyramidal neurons contributing to the generation of synchronous oscillations in the visual cortex.

In response to visual stimulation, a subset of neurons in the striate and prestriate cortex displays synchronous rhythmic firing in the gamma frequency band (20 to 70 hertz). This finding has raised two fundamental questions: What is the functional significance of synchronous gamma-band activity and how is it generated? This report addresses the second of these two questions. By means of intracellular recording and staining of single cells in the cat striate cortex in vivo, a biophysically distinct class of pyramidal neuron termed "chattering cells" is described. These neurons are located in the superficial layers of the cortex, intrinsically generate 20- to 70-hertz repetitive burst firing in response to suprathreshold depolarizing current injection, and exhibit pronounced oscillations in membrane potential during visual stimulation that are largely absent during periods of spontaneous activity. These properties suggest that chattering cells may make a substantial intracortical contribution to the generation of synchronous cortical oscillations and thus participate in the recruitment of large populations of cells into synchronously firing assemblies.

Where in the brain does visual attention select the forest and the trees?

The perceptual world is organized hierarchically: the forest consists of trees, which in turn have leaves. Visual attention can emphasize the overall picture (global form) or the focal details of a scene (local components). Neuropsychological studies have indicated that the left hemisphere is biased towards local and the right towards global processing. The underlying attentional and perceptual mechanisms are maximally impaired by unilateral lesions to the temporal and parietal cortex. We measured brain activity of normal subjects during two experiments using 'hierarchically' organized figures. In a directed attention task, early visual processing (prestriate) areas were activated: attention to the global aspect of the figures activated the right lingual gyrus whereas locally directed attention activated the left inferior occipital cortex. In a subsequent divided attention task, the number of target switches from local to global (and vice versa) covaried with temporal-parietal activation. The findings provide direct evidence for hemispheric specialization in global and local perception; furthermore, they indicate that temporal-parietal areas exert attentional control over the neural transformations occurring in prestriate cortex.

Memory for object-features versus memory for object-location: A positron emission tomography study of encoding and retrieval processes

Abstract Regional cerebral blood flow was measured with positron-emission tomography during two encoding and two retrieval tasks that were designed to compare memory for object features with memory for object locations. Bilateral increases in regional cerebral blood flow were observed in both anterior and posterior regions of inferior temporal cortex and in ventral regions of prestriate cortex, when the condition that required retrieval of object locations was subtracted from the condition that required retrieval of object features. During encoding, these changes were less pronounced and were restricted to the left inferior temporal cortex and right ventral prestriate cortex. In contrast, both encoding and retrieval of object location were associated with bilateral changes in dorsal prestriate and posterior parietal cortex. Finally, the two encoding conditions activated left frontal lobe regions preferentially, whereas the two retrieval conditions activated right frontal lobe regions. These findings confirm that, in human subjects, memory for object features is mediated by a distributed system that includes ventral prestriate cortex and both anterior and posterior regions of the inferior temporal gyrus. In contrast, memory for the locations of objects appears to be mediated by an anatomically distinct system that includes more dorsal regions of prestriate cortex and posterior regions of the parietal lobe.

Anatomical substrates for early stages in cortical processing of visual information in the macaque monkey

Primate; Parallel pathway; Neuroanatomy; Striate cortex; Prestriate cortex; Visual system 1. Introduction It is known that in many features of its anatomy and physiology, the macaque monkey provides an excellent animal model for the human visual system. Visual psy- chophysical studies of m~Ln and macaque show them to have very similar functional capacities (e.g., [17]). In recent years, researchers' attention has shifted away from primary visual cortex, area V1, to visual pathways extending forward into the temporal and parietal cortices and to their connections with the prefrontal association cortex. Imaging studies suggest that like the monkey, the human cerebral cortex has particular areas special- ized for analysis of different qualities of the visual stimulus (see [107]). A broader knowledge of cortical visual processing suggests that an important factor in determining each area's ,distinctive response properties and overall functional relevance is the channelling of particular types of visual information to that area (for example, direction-selective V1 inputs to area V5; [83]), together with uniquely structured associations between different types of information within each area. This places less onus on each .area being anatomically 'tailor made' for specific tasks, although the degree to which the intrinsic neuropil of different visual cortical areas is uniquely structured still remains to be examined in detail. Channelling of different types of visual information via different routes through the cortical neuropil begins with the distribution of functionally distinct types of thalamic inputs to different laminae in area V1. From the neurons that receive these different inputs, intrinsic V1 pathways lead to different sets of efferent neurons, * Corresponding author. Tel: (+44 1771) 608 6914; Fax: (+44 1771) 608 6850; E-mail: j.levitt@ucl.ac.uk. 0166-4328/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved

Functional streams in occipito-frontal connections in the monkey

It is known that the prestriate cortical regions that project to area LIP in parietal cortex and to areas TEO and TE in temporal cortex are mostly separated. Two separate streams of information transfer from occipital cortex can thus be distinguished. We wished to determine whether the parietal and temporal streams remain segregated in their projections to frontal cortex. Paired injections of retrograde fluorescent tracers were placed in parietal and temporal cortex, or in the lateral and medial parts of the frontal eye field (FEF). The cortical regions containing retrogradely labeled cells were reconstructed in two-dimensional maps. The results show that temporal cortex mainly projects to lateral FEF (area 45). Parietal cortex sends projections to medial FEF (area 8a) and to lateral FEF, as well as to area 46. Thus, the parietal and temporal streams converge in lateral FEF. Most of the occipital regions projecting to medial FEF are the same as those projecting to parietal cortex, whereas lateral FEF receives afferents from the same occipital regions as those sending projections to temporal cortex. Thus, one can distinguish two interconnected networks. One is associated with the inferotemporal cortex and includes areas of the ventral bank and fundus of the superior temporal sulcus (STS), lateral FEF and ventral prestriate cortex. This network emphasizes central vision, small saccades and form recognition. The other network is linked to cortex of the intraparietal sulcus. It consists of areas of the upper bank and fundus of STS, medial FEF and dorsal prestriate cortex. These areas encode peripheral visual field and are active during large saccades.

Cerebral blood oxygenation changes induced by visual stimulation in humans

We examined local changes of cerebral oxygenation in response to visual stimuli by means of near infrared spectroscopy. A sharply outlined colored moving stimulus which is expected to evoke a broad activation of the striate and prestriate cortex was presented to sixteen healthy subjects. Six of these subjects were also exposed to a colored stationary and a gray stationary stimulus. In two subjects the colored moving stimulus was tested against the colored stationary with an optode position presumably over area V5/MT. As a control condition, subjects performed a simple finger opposition task. Since the calcarine fissure varies greatly with respect to bony landmarks, optodes were positioned individually according to 3-D reconstructed magnetic resonance imaging (MRI). Concentration changes in oxyhemoglobin ([oxy-Hb]) and deoxyhemoglobin ([deoxy-Hb]) were continuously monitored with a temporal resolution of 1 s, using an NIRO 500 (Hamamatsu Photonics, KK, Japan). In response to the visual stimulus, the grand average across all sixteen subjects resulted in a significant increase in [oxy-Hb] of 0.3360.09 arbitrary units (mean6S.E.M.) mirrored by a significant decrease in [deoxy-Hb] of −0.1860.02 arbitrary units, while the motor control condition elicited no significant changes in any parameters. When the near infrared spectroscopy probes were positioned over area V5/MT, the drop of [deoxy-Hb] associated with the moving stimulus was significantly more pronounced than with the stationary stimulus in both subjects examined. No significant differences between the visual stimuli were observed at the optode position close to the calcarine fissure. The oxygenation changes observed in this study are consistent with the pattern we have reported for motor activation. They are in line with physiological considerations and functional MRI studies relying on blood oxygenation level-dependent contrast.

Age-related changes in cortical blood flow activation during perception and memory.

Although many cognitive functions are affected by age, some are relatively maintained. There also are numerous age-related changes in brain structure, but not much is known about how these changes impact upon the alterations seen in cognition. In order to understand the basis for the reductions and sparings of cognitive function in the aged, experiments were carried out to compare regional cerebral blood flow (rCBF) and performance in young and old subjects on visual perceptual and memory tasks. In the first experiment, which examined the perception of faces and spatial locations, old subjects were as accurate as young subjects, and both groups had rCBF activation in fusiform gyrus during face matching and in superior parietal cortex during location matching. However, old subjects had less activation of prestriate cortex and more activation of frontal and lateral temporal cortex than did young subjects. The second experiment tested recognition memory for faces, a task on which old subjects were impaired. Young subjects had rCBF activation in left prefrontal and inferior temporal cortex and in right hippocampus during encoding of the faces, and in right prefrontal and parietal cortex during recognition. Old subjects showed no significant activation of the areas involved in encoding, but did have increased rCBF in right prefrontal cortex during recognition. These results suggest that the aging brain can demonstrate alterations of the functional systems involved in some cognitive processes, such as perception, that may serve as a compensatory mechanism to maintain performance. Failure to show such compensation, coupled with dysfunction of the areas primarily involved in processing may lead to more marked deficits in performance.

Motion specific responses from a blind hemifield.

In a previous study we showed that fast moving stimuli activate V5, an area specialized for motion, at very short latencies through a pathway that reaches it without passing through V1. Using the same technique of visual evoked responses, we have tested our conclusions by studying patient GY, whose V1 is damaged but whose V5 is intact. In spite of the contralateral hemi-blindness due to his V1 lesion, GY has a residual visual capacity that allows him to perceive, consciously, fast but not slow moving stimuli presented in his affected hemifield. By stimulating GY's 'blind' hemifield and comparing the responses with those obtained from normal subjects, we were able to study the relative contribution of V1 and V5 to the visual evoked response to motion in normal subjects. We found that GY's early response to fast motion is preserved and correlates with activity elicited in control subjects over area V5, while slow motion, pattern offset, and pattern reversal stimuli failed to elicit responses in GY. The results confirm our previous conclusions: namely, that the early part of the motion evoked response is generated in area V5 and that signals reach this area through a dynamically parallel pathway that bypasses area V1. They go on to demonstrate that neurophysiological activity in the prestriate cortex correlates with the conscious visual perception of motion.

Projections from visual areas of the cerebral cortex to pretectal nuclear complex, terminal accessory optic nuclei, and superior colliculus in macaque monkey.

The purpose of this study was to analyze the projections from visually related areas of the cerebral cortex of rhesus monkey to subcortical nuclei involved in eye-movement control; i.e., the pretectal nuclear complex, the terminal nuclei of the accessory optic system (AOS), and the superior colliculus (SC). The anterograde tracer 3H-leucine was pressure injected bilaterally into the cortex of six monkeys (for a total of 12 cases) involving the primary visual cortex (area 17); the medial prestriate cortex (medial 18/19); dorsomedial area 19; the caudal portion of the cortex of the superior temporal sulcus, upper bank (cytoarchitectural area OAa) and lower bank (area PGa); the lower bank of the caudal lateral intraparietal sulcus (area POa); and the inferior parietal lobule (area 7). The results revealed that the pretectal nucleus of the optic tract received inputs from medial prestriate cortex, dorsomedial part of area 19, OAa, and PGa. The posterior pretectal nucleus received sparse projections from area 7 and the cortex lining the intraparietal sulcus (dorsomedial part of area 19 and POa). The pretectal olivary nucleus was targeted by neurons in cortex of dorsomedial area 19, and the anterior pretectal nucleus was targeted by neurons in both dorsomedial 19 and area 7. The nuclei of the AOS (dorsal terminal; lateral terminal; and interstitial nuclei of the superior fasciculus, posterior and medial fibers) received projections exclusively from areas OAa and PGa. Furthermore, in one case with PGa injection, the medial terminal nucleus, dorsal portion, was also labeled. The visual cortical areas studied projected differentially upon the SC laminae. The primary visual area 17 projected only to the superficial laminae, i.e., stratum zonale (SZ), stratum griseum superficiale (SGS), and stratum opticum (SO). On the other hand, the medial portion of the prestriate cortex and caudal OAa and PGa targeted the superficial and intermediate laminae, i.e., SZ, SGS, SO, and stratum griseum intermediale (SGI), whereas caudal area POa projected primarily to the intermediate layer SGI. Rostral area 7 (mainly 7b) neurons terminated in the stratum album intermediale (SAI); no SC terminals were found in a case in which caudal area 7 (mainly 7a) was injected.

Mechanisms of purely subjective contour tilt aftereffects

Neurones tuned for second-order stimuli—those which have edges defined by properties other than luminance and colour—have been frequently observed in prestriate cortex and in area V2 there are neurones which explicitly and unambiguously signal the orientation of purely subjective contours, i.e. contours with no Fourier components at the orientation of the perceived edge [von der Heydt, R. & Peterhans, E. (1989) Journal of Neuroscience, 9, 1731–1748]. No neurones in area V1 showed similar tuning characteristics. In addition, it has been demonstrated that like real contours, purely subjective test contours are subject to tilt aftereffects following prolonged viewing of an adapting stimulus, whether that stimulus is real or subjective [Paradiso, M. A., Shimojo, S. & Nakayama, K. (1989) Vision Research, 29, 1205–1213]. This result supports the assertion that the cortical processes responsible for real contour perception are also those giving rise to subjective contour perception. The data reported here further examined this hypothesis. Four experiments show that purely subjective contours exhibit both direct and indirect tilt aftereffects and tilt illusions like those observed with real contours. Further, they provide evidence that direct and indirect subjective contour effects, like direct and indirect real contour effects, arise via the operation of two mechanisms: a low level process, possibly lateral inhibition between orientation channels, and a second “higher-order” process. The data suggest that processing of orientation information beyond the striate cortex is similar to that which occurs in area V1 and the data are consistent with models of contour processing which assume that all perceived contours, both real and subjective, arise from a common mechanism.

Mechanisms of stereopsis in monkey visual cortex.

A substantial proportion of neurons in the striate and prestriate cortex of monkeys have stereoscopic properties; that is, they respond differentially to binocular stimuli that are known in humans to provide cues for stereoscopic depth perception. Stereoscopic neurons, as these cells may be called, are selective for horizontal positional disparity (i.e., display disparity selectivity) and for the textural correlation between images over their receptive fields (i.e., they show correlation selectivity). Many neurons have tuned disparity response profiles that collectively cover the entire range of physiological disparities. Neurons with peak responses at or about the zero disparity ("tuned zero neurons," excitatory or inhibitory) have narrow and symmetrical profiles. Neurons that are tuned to larger disparities, either crossed ("tuned near neurons") or uncrossed ("tuned far neurons"), have broader excitatory profiles that are asymmetrically wider toward the smaller disparities, and commonly include an inhibitory component about the zero disparity. Other stereoscopic neurons have reciprocal profiles ("near" or "far" neurons, respectively) in the sense that they respond with excitation to crossed or uncrossed disparities, and with suppression to disparities of opposite sign. Stereoscopic neurons can also signal the textural correlation between paired retinal images by giving different responses to random-dot patterns that have, and to those that do not have, the same dot distribution over the neuron's left and right receptive fields. Tuned-zero excitatory neurons characteristically respond to uncorrelation with suppression; tuned-zero inhibitory neurons, with excitation; and both types give the opposite responses to correlated stereopatterns. Neurons selective for nonzero disparities, both tuned and reciprocal, also give excitatory responses to uncorrelated stimuli, but these responses are smaller and more variable than those evoked by correlated patterns at the effective disparities. These findings suggest that stereoscopic neurons in the visual cortex of the macaque comprise three operational systems: (1) a zero-disparity system that is involved in fine depth discrimination with the obligatory singleness of vision, and the maintenance of vergence; and (2) a near-, and (3) a far-disparity system that together signal qualitative estimates of depth with double vision, and vergence responses to large disparities.