Representations In Visual Cortex Research Articles

Dissociable Neural Effects of Long-term Stimulus–Reward Pairing in Macaque Visual Cortex

It has been proposed that perceptual learning may occur through a reinforcement process, in which consistently pairing stimuli with reward is sufficient for learning. We tested whether stimulus-reward pairing is sufficient to increase the sensorial representation of a stimulus by recording local field potentials (LFPs) in macaque extrastriate area V4 with chronically implanted electrodes. Two oriented gratings were repeatedly presented; one was paired with a fluid reward, whereas no reward was given at any other time. During the course of conditioning the LFP increased for the rewarded compared to the unrewarded orientation. The time course of the effect of stimulus-reward pairing and its reversal differed between an early and late interval of the LFP response: a fast change in the later part of the neural response that was dissociated from a slower change in the early part of the response. The fast change of the late interval LFP suggests that this late LFP change is related to enhanced attention during the presentation of the rewarded stimulus. The slower time course of the early interval response suggests an effect of sensorial learning. Thus, simple stimulus-reward pairing is sufficient to strengthen stimulus representations in visual cortex and does this by means of two dissociable mechanisms.

Shape-specific preparatory activity mediates attention to targets in human visual cortex

The mechanisms of attention prioritize sensory input for efficient perceptual processing. Influential theories suggest that attentional biases are mediated via preparatory activation of task-relevant perceptual representations in visual cortex, but the neural evidence for a preparatory coding model of attention remains incomplete. In this experiment, we tested core assumptions underlying a preparatory coding model for attentional bias. Exploiting multivoxel pattern analysis of functional neuroimaging data obtained during a non-spatial attention task, we examined the locus, time-course, and functional significance of shape-specific preparatory attention in the human brain. Following an attentional cue, yet before the onset of a visual target, we observed selective activation of target-specific neural subpopulations within shape-processing visual cortex (lateral occipital complex). Target-specific modulation of baseline activity was sustained throughout the duration of the attention trial and the degree of target specificity that characterized preparatory activation patterns correlated with perceptual performance. We conclude that top-down attention selectively activates target-specific neural codes, providing a competitive bias favoring task-relevant representations over competing representations distributed within the same subregion of visual cortex.

The topography of training-induced visual field recovery: Perimetric maps and subjective representations

The cognitive representation of blind regions varies considerably between patients with vision loss and may influence compensatory behaviour and treatment motivation. We therefore measured “objective” visual field topography (perimetry) in 19 patients with postgeniculate visual system lesions and related this to the subjective scotoma representation as expressed by patients’ drawings of the defect and monitored changes of these measures during training-induced recovery of function. Blind regions were mostly adequately represented; however, central regions were overestimated and peripheral areas underestimated in size. Perimetric and subjective defect size decreased significantly during training. Again, training-induced visual field border shifts in central regions were larger in subjective than in perimetric maps but vice versa in the peripheral field. Thus, vision restoration therapy improves “objective” visual field size along with its cognitive representation. The subjective topography is shaped by the functional importance of visual field regions and is a function of cortical magnification, thus resembling the neural representation in visual cortex.

Unsupervised natural experience rapidly alters invariant object representation in visual cortex.

Object recognition is challenging because each object produces myriad retinal images. Responses of neurons from the inferior temporal cortex (IT) are selective to different objects, yet tolerant ("invariant") to changes in object position, scale, and pose. How does the brain construct this neuronal tolerance? We report a form of neuronal learning that suggests the underlying solution. Targeted alteration of the natural temporal contiguity of visual experience caused specific changes in IT position tolerance. This unsupervised temporal slowness learning (UTL) was substantial, increased with experience, and was significant in single IT neurons after just 1 hour. Together with previous theoretical work and human object perception experiments, we speculate that UTL may reflect the mechanism by which the visual stream builds and maintains tolerant object representations.

Sparse representation of global features of visual images in human primary visual cortex: Evidence from fMRI

In fMRI experiments on object representation in visual cortex, we designed two types of stimuli: one is the gray face image and its line drawing, and the other is the illusion and its corresponding completed illusion. Both of them have the same global features with different minute details so that the results of fMRI experiments can be compared with each other. The first kind of visual stimuli was used in a block design fMRI experiment, and the second was used in an event-related fMRI experiment. Comparing and analyzing interesting visual cortex activity patterns and blood oxygenation level dependent (BOLD)-fMRI signal, we obtained results to show some invariance of global features of visual images. A plausible explanation about the invariant mechanism is related with the cooperation of synchronized response to the global features of the visual image with a feedback of shape perception from higher cortex to cortex V1, namely the integration of global features and embodiment of sparse representation and distributed population code.

Retinotopic representations of human visual cortex revealed by functional magnetic resonance imaging

To establish a set of methodological standard of retinotopic mapping using functional magnetic resonance imaging (MRI), to discuss anatomical features of functionally delineated V1/V2 boundary based on retinotopic criteria. Three adults aged 25-30 (1 male and 2 females), right-handed, without contraindications of MRI were studied. Visual acuity was 20/20 or corrected to 20/20. Visual stimuli subtended a field of view of approximately 12 degrees, consisted of high contrast, achromatic and drifting checkerboards. Two types of stimuli were used: the rotating wedge, either clockwise or counterclockwise, to map the polar angle and the annular ring, either contracting or expanding, to map eccentricity. MRI parameters: GE signa VH/i 3.0T scanner. Functional data: GRE-EPI sequence, 20 slices lay perpendicular to the calcarine sulcus, TR: 2 s, TE: 40 ms, FA: 90 degrees, Phase: 128, FOV: 19 cm, matrix: 64x64, slice: 3 mm, gap: 0, resolution: 2.9 mmx2.9 mmx3.0 mm. Anatomical data was obtained using 3D-SPGR sequence to acquire high resolution. Data was analyzed using AFNI. In order to avoid the disadvantages of restricted view due to highly convoluted occipital cortex, cortical surface was unfolded and then cut and flattened. Functional data was presented to this flattened surface and subsequently analyzed. By using the phase-encoded visual stimuli, retinotopic organizations were exactly revealed both in the eccentricity and polar angle dimensions. Eccentricity maps showed that foveal representation lay in the occipital poles and the representation appeared further anterior as eccentricity increased. Polar angle maps showed that early retinotopically organized areas had either a mirror or non-mirror image representation of visual field. Boundaries of different visual areas were manually delineated based on retinotopic criteria called visual field sign, and the functionally delineated V1/V2 boundary located on the crests of gyrus. Phase-encoded stimuli is validated to be an efficient way to reveal retinotopic representation of visual cortex, promising a solid foundation for the future research under pathological conditions. Anatomical features of functionally delineated V1/V2 boundary support the wiring optimization principle.

Stimulus context modulates competition in human extrastriate cortex

When multiple stimuli appear simultaneously in the visual field, they are not processed independently, but rather interact in a mutually suppressive way, suggesting that they compete for neural representation in visual cortex. The biased competition model of selective attention predicts that the competition can be influenced by both top-down and bottom-up mechanisms. Directed attention has been shown to bias competition in favor of the attended stimulus in extrastriate cortex. Here, we show that suppressive interactions among multiple stimuli are eliminated in extrastriate cortex when they are presented in the context of pop-out displays, in which a single item differs from the others, but not in heterogeneous displays, in which all items differ from each other. The pop-out effects seemed to originate in early visual cortex and were independent of attentional top-down control, suggesting that stimulus context may provide a powerful influence on neural competition in human visual cortex.

Does Neuronal Synchrony Underlie Visual Feature Grouping?

Previous research suggests that synchronous neural activity underlies perceptual grouping of visual image features. The generality of this mechanism is unclear, however, as previous studies have focused on pairs of neurons with overlapping or collinear receptive fields. By sampling more broadly and employing stimuli that contain partially occluded objects, we have conducted a more incisive test of the binding by synchrony hypothesis in area MT. We find that synchrony in spiking activity shows little dependence on feature grouping, whereas gamma band synchrony in field potentials can be significantly stronger when features are grouped. However, these changes in gamma band synchrony are small relative to the variability of synchrony across recording sites and do not provide a robust population signal for feature grouping. Moreover, these effects are reduced when stimulus differences nearby the receptive fields are eliminated using partial occlusion. Our findings suggest that synchrony does not constitute a general mechanism of visual feature binding.

Local correlations, information redundancy, and sufficient pixel depth in natural images.

A mathematical framework that permits the factorization of a joint probability distribution into its localized components for a two-dimensional array of pixels is presented. The factorization was used to estimate the contribution to mutual information of two- (I2) and three-pixel (I3) luminance correlations for a large ensemble of natural images analyzed at various spatial scales and pixel depths b. It is shown that both I2 and I3 saturate at b approximately 6 bits per pixel. Three-pixel correlations are shown to produce only a marginal increase of information redundancy (4%) over two-pixel correlations (50%). Implications for neural representation in visual cortex are discussed.

Specialized representations in visual cortex: a role for binding?

Specialized representations in visual cortex: a role for binding?

The binding problem.

Both Ghose and Maunsell and Shadlen and Movshon point out what is perhaps the fundamental problem with the binding hypothesis: even supposing that temporal coding is the vehicle for signaling which neural populations should be bound together, the theory does not adequately address how those combinations are computed. In a sense, the binding problem is pushed back one level, for it must be solved at least partially for the necessary temporal correlations to be established. Are neurons with appropriate receptive field properties and anatomical connectivity a sufficient basis for the generation of correlated signals? What is the role of top-down connections in establishing proper synchrony? These questions are difficult to answer and difficult to model with biological realism. A similar problem rests with the output side of the temporal binding hypothesis—if correlations are the signal for binding, how are those signals read out? The readout problem is one of the most puzzling and fundamental problems for systems neuroscience in general: how is the firing of populations of neurons interpreted and transformed by other neurons to result in decision, action, perception, etc? This problem plagues most, if not all, models of brain function, for in modeling it is the modeler that attributes semantics to nodes in a network. One way to potentially avoid that problem would be to close the loop between the model and the world, so that the world impinges upon the network, which then acts upon and affects the world, thus generating semantics through action.This brings me to mention (by virtue of my editorial license) what is perhaps the most mystifying binding problem of all: the problem of consciousness. How does something as simple and mechanistic as neural firing add up to subjectivity, raw feelings, a self? Are the mechanisms that allow us to attribute the correct color and shape to an object the same ones that lead to the unity of phenomenal experience? Will the solution of the binding problem be the solution to the mystery of consciousness? I will not belabor the point, since answers will be long in coming, but although none of the scientists who authored the reviews that follow discuss binding with respect to consciousness, I will wager that the a good part of the interest, excitement, and contentiousness that surrounds the binding debate is attributable to the magnitude of the issues with which it is connected. The following pieces provide a comprehensive review of the status of the binding problem at the dawn of the new millenium. It will be extremely interesting to repeat this exercise in a decade or two, to chart the progress that we as scientists make on one of the most puzzling and fascinating issues that the brain and cognitive sciences have ever faced.*E-mail: aroskies@cell.com.

Solutions to the binding problem: progress through controversy and convergence.

Solutions to the binding problem: progress through controversy and convergence.

The atoms of vision: Cartesian or polar?

The inherent structure of the encoding in early stages of the visual system is investigated from a combined information-theoretical, psychophysical, and neurophysiological perspective. We argue that the classical modeling in terms of linear spatial filters is equivalent to the assumption of a Cartesian organization of the feature space of early vision. We show that such a linear Cartesian feature space would be suboptimal for the exploitation of the statistical redundancies of natural images since these have a radially separable probability-density function. Therefore a more efficient representation can be obtained by a nonlinear encoding that yields a feature space with polar organization. This prediction of the information-theoretical approach regarding the organization of the feature space of early vision is confirmed by our psychophysical measurements of basic discrimination capabilities for elementary Gabor patches, and the necessary nonlinear operations are shown to be closely related to cortical gain control and to the phase invariance of complex cells. Finally, we point out some striking similarities between the polar representation in visual cortex and basic image-coding strategies pursued in shape-gain vector quantization schemes.

Suppression of visual perception by magnetic coil stimulation of human occipital cortex

Magnetic coil (MC) stimulation percutaneously of human occipital cortex was tested on perception of 3 briefly presented, randomly generated alphabetical characters. When the visual stimulus-MC pulse interval was less than 40–60 msec, or more than 120–140 msec, letters were correctly reported; at test intervals of 80–100 msec, a blur or nothing was seen. Shifting the MC location in the transverse and rostro-caudal axes had effects consistent with the topographical representation in visual cortex, but incompatible with an effect on attention or suppression from an eyeblink. The MC pulse probably acts by eliciting IPSPs in visual stimulus.

A hexagonal orthogonal-oriented pyramid as a model of image representation in visual cortex.

Retinal ganglion cells represent the visual image with a spatial code, in which each cell conveys information about a small region in the image. In contrast, cells of primary visual cortex employ a hybrid space-frequency code in which each cell conveys information about a region that is local in space, spatial frequency, and orientation. Despite the presumable importance of this transformation, we lack any comprehensive notion of how it occurs. Here we describe a mathematical model for this transformation. The hexagonal orthogonal-oriented quadrature pyramid (HOP) transform, which operates on a hexagonal input lattice, employs basis functions that are orthogonal, self-similar, and localized in space, spatial frequency, orientation, and phase. The basis functions, which are generated from seven basic types through a recursive process, form an image code of the pyramid type. The seven basis functions, six bandpass and one low-pass, occupy a point and a hexagon of six nearest neighbors on a hexagonal sample lattice. The six bandpass basis functions consist of three with even symmetry, and three with odd symmetry. The three even kernels are rotations of 0, 60, and 120 degrees of a common kernel; likewise for the three odd kernels. At the lowest level, the inputs are image samples. At each higher level, the input lattice is provided by the low-pass coefficients computed at the previous level. At each level, the output is subsampled in such a way as to yield a new hexagonal lattice with a spacing square root 7 larger than the previous level, so that the number of coefficients is reduced by a factor of seven at each level. In the biological model, the input lattice is the retinal ganglion cell array. The resulting scheme provides a compact, efficient code of the image and generates receptive fields that resemble those of the primary visual cortex.