The spatial receptive fields of simple cells in mammalian striate cortex have been reasonably well described physiologically and can be characterized as being localized, oriented, and ban@ass, comparable with the basis functions of wavelet transforms. Previously, we have shown that these receptive field properties may be accounted for in terms of a strategy for producing a sparse distribution of output activity in response to natural images. Here, in addition to describing this work in a more expansive fashion, we examine the neurobiological implications of sparse coding. Of particular interest is the case when the code is overcomplete--i.e., when the number of code elements is greater than the effective dimensionality of the input space. Because the basis functions are non-orthogonal and not linearly independent of each other, sparsifying the code will recruit only those basis functions necessary for representing a given input, and so the input-output function will deviate from being purely linear. These deviations from linearity provide a potential explanation for the weak forms of non-linearity observed in the response properties of cortical simple cells, and they further make predictions about the expected interactions among units in

Natural time-varying images possess substantial spatiotemporal correlations. We measure these correlations --- or equivalently the power spectrum --- for an ensemble of more than a thousand segments of motion pictures, and we find significant regularities. More precisely, our measurements show that the dependence of the power spectrum on the spatial frequency, f , and temporal frequency, w, is in general nonseparable and is given by f 0m01 F (w=f ), where F (w=f ) is a nontrivial function of the ratio w=f . We give a theoretical derivation of this scaling behaviour and show that it emerges from objects with a static power spectrum ¸ f 0m , appearing at a wide range of depths and moving with a distribution of velocities relative to the observer. We show that in the regime of relatively high temporal and low spatial frequencies, the power spectrum becomes independent of the details of the velocity distribution and it is separable into the product of spatial and temporal power spectra...

Natural time-varying images possess significant temporal correlations when sampled frame by frame by the photoreceptors. These correlations persist even after retinal processing and hence, under natural activation conditions, the signal sent to the lateral geniculate nucleus is temporally redundant or inefficient. We explore the hypothesis that the LGN is concerned, among other things, with improving efficiency of visual representation through active temporal decorrelation of the retinal signal much in the same way that the retina improves efficiency by spatially decorrelating incoming images. Using some recently measured statistical properties of time-varying images, we predict the spatio-temporal receptive fields that achieve this decorrelation. It is shown that, because of neuronal nonlinearities, temporal decorrelation requires two response types, the lagged and nonlagged, just as spatial decorrelation requires on and off response types. The tuning and response properties of the p...

plore several parallels to the function and architecture of the visual system that help interpret the neural representation of odors. Key words: olfaction; receptor; glomerulus; map; tuning; imaging Our olfactory system senses chemicals in the ambient air through specialized receptor cells that are located in the neural epithelium lining the upper reaches of the nasal cavity. Each of these olfactory neurons is thought to express a single type of receptor protein that spans the plasma membrane and whose binding properties determine the interaction with extracellular ligands (Buck and Axel, 1991; Buck, 1996; Mombaerts, 1999). In this respect, olfactory receptors are similar to photoreceptors: each photoreceptor cell makes a single type of opsin protein, whose absorption properties determine the interaction of the cell with photons of different wavelength. However, photoreceptors come in only a handful of varieties, whereas mammals seem to have up to 1000 types of olfact

Abstract not found

This study compares the efficiency and plausibility of published computational models of familiarity discrimination in the perirhinal cortex. Substantial evidence indicates that the perirhinal cortex is involved in both the familiarity discrimination aspect of recognition memory and in perceptual functions involved with representations of complete stimuli (i.e., object identification). Published models of how the perirhinal cortex may perform familiarity discrimination can be divided into two groups. The first group assumes that a proportion of perirhinal neurons form a network specialised just for familiarity discrimination (these models may be based on Hebbian or anti-Hebbian synaptic plasticity). In contrast, the second group assumes that both familiarity discrimination and learning representations of complete stimuli are performed within a single combined network. This study establishes that when the responses of neurons that provide input to the familiarity discrimination network are correlated (as indicated by experimental data), specialised networks based on anti-Hebbian learning may recognise the previous occurrence of many more stimuli (i.e., have a capacity up to thousands of times larger) than specialised networks based on Hebbian learning. The currently published combined models do not learn an optimal stimulus representation (they do not fully extract statistically independent features), and hence their capacities are even lower than those of the specialised models based on Hebbian learning. Hence, the combined models published thus far are critically less efficient than the specialised models based on anti-Hebbian learning. This study also compares the consistency of the models with experimental observations concerning what is kno...

The goal of this chapter is to present computational theories of scene coding by image segmentation and to suggest their relevance for understanding visual cortical function and mechanisms. We will first introduce computational theories of image and scene segmentation and show their relationship to efficient

Retinal ganglion cells of a given type overlap their dendritic fields such that every point in space is covered by three to four cells. We investigated what function is served by such extensive overlap. Recording from pairs of ON or OFF brisk-transient ganglion cells at photopic intensities, we confirmed that this overlap causes the Gaussian receptive field centers to be spaced at �2 SDs (�). This, together with response nonlinearities and variability, was just sufficient to provide an ideal observer with uniform contrast sensitivity across the retina for both threshold and suprathreshold stimuli. We hypothesized that overlap might maximize the information represented from natural images, thereby optimizing retinal performance for many tasks. Indeed, tested with natural images (which contain statistical correlations), a model ganglion cell array maximized information represented in its population responses with �2 � spacing, i.e., the overlap observed in the retina. Yet, tested with white noise (which lacks statistical correlations), an array maximized its information by minimizing overlap. In both cases, optimal overlap balanced greater signal-to-noise ratio (from larger receptive fields) against greater redundancy (because of larger receptive field overlap). Thus, dendritic overlap improves vision by taking optimal advantage of the statistical correlations of natural scenes. Key words: retina; ganglion cell; paired recording; contrast sensitivity; ideal observer; natural scenes; optimal coding

Decoding the activity of a population of neurons is a fundamental problem in neuroscience. A key aspect of this problem is determining whether correlations in the activity, i.e., noise correlations, are important. If they are important, then the decoding problem is high dimensional: decoding algorithms must take the correlational structure in the activity into account. If they are not important, or if they play a minor role, then the decoding problem can be reduced to lower dimension and thus made more tractable. The issue of whether correlations are important has been a subject of heated debate. The debate centers around the validity of the measures used to address it. Here, we evaluate three of the most commonly used ones: synergy, �Ishuffled, and �I. We show that synergy and �Ishuffled are confounded measures: they can be zero when correlations are clearly important for decoding and positive when they are not. In contrast, �I is not confounded. It is zero only when correlations are not important for decoding and positive only when they are; that is, it is zero only when one can decode exactly as well using a decoder that ignores correlations as one can using a decoder that does not, and it is positive only when one cannot decode as well. Finally, we show that �I has an information theoretic interpretation; it is an upper bound on the information lost when correlations are ignored. Key words: retina; encoding; decoding; neural code; information theory; population coding; signal correlations; noise correlations

Cross-adaptation is widely used to probe whether different stimuli share common neural mechanisms. For example, that adaptation to second-order stimuli usually produces little aftereffect on first-order stimuli has been interpreted as reflecting their separate processing. However, such results appear to contradict the cue-invariant responses of many visual cells. We tested the novel hypothesis that the null aftereffect arises from the large difference in the backgrounds of first- and second-order stimuli. We created second-order faces with happy and sad facial expressions specified solely by local directions of moving random dots on a static-dot background, without any luminance-defined form cues. As expected, adaptation to such a second-order face did not produce a facial-expression aftereffect on the first-order faces. However, consistent with our hypothesis, simply adding static random dots to the first-order faces to render their backgrounds more similar to that of the adapting motion face led to a significant aftereffect. This background similarity effect also occurred between different types of first-order stimuli: real-face adaptation transferred to cartoon faces only when noise with correlation statistics of real faces or natural images was added to the cartoon faces. These findings suggest the following: (1) statistical similarities between the featureless backgrounds of the adapting and test stimuli can influence aftereffects, as in contingent adaptation; (2) weak or null cross-adaptation aftereffects should be interpreted with caution; and (3) luminance- and motion-direction-defined forms, and local features and global statistics, converge in the representation of faces.