Date Published: April 22, 2010
Publisher: Public Library of Science
Author(s): Jenny C. A. Read, Laurence T. Maloney
Abstract: Primary visual cortex is often viewed as a “cyclopean retina”, performing the initial encoding of binocular disparities between left and right images. Because the eyes are set apart horizontally in the head, binocular disparities are predominantly horizontal. Yet, especially in the visual periphery, a range of non-zero vertical disparities do occur and can influence perception. It has therefore been assumed that primary visual cortex must contain neurons tuned to a range of vertical disparities. Here, I show that this is not necessarily the case. Many disparity-selective neurons are most sensitive to changes in disparity orthogonal to their preferred orientation. That is, the disparity tuning surfaces, mapping their response to different two-dimensional (2D) disparities, are elongated along the cell’s preferred orientation. Because of this, even if a neuron’s optimal 2D disparity has zero vertical component, the neuron will still respond best to a non-zero vertical disparity when probed with a sub-optimal horizontal disparity. This property can be used to decode 2D disparity, even allowing for realistic levels of neuronal noise. Even if all V1 neurons at a particular retinotopic location are tuned to the expected vertical disparity there (for example, zero at the fovea), the brain could still decode the magnitude and sign of departures from that expected value. This provides an intriguing counter-example to the common wisdom that, in order for a neuronal population to encode a quantity, its members must be tuned to a range of values of that quantity. It demonstrates that populations of disparity-selective neurons encode much richer information than previously appreciated. It suggests a possible strategy for the brain to extract rarely-occurring stimulus values, while concentrating neuronal resources on the most commonly-occurring situations.
Partial Text: It is commonly accepted that in order for a neuronal population to encode the value of a quantity x, it must contain cells tuned to a range of values of x. Thus for example the retina can encode information about the wavelength of light because it contains three different types of cones with different tuning to wavelength, and the primary visual cortex can encode feature orientation because it contains neurons tuned to a range of orientations. This is unproblematic because natural images contain a wide range of light wavelengths and object orientations. However, the same argument applied to stereo vision produces some more challenging conclusions.
This paper has implemented a simple physiologically-inspired two-dimensional stereo correspondence algorithm. It consists of two model “brain areas”: one which performs the initial encoding of binocular disparity between left and right images, and one which decodes this activity so as to arrive at an estimate of the two-dimensional disparity in the images. The unusual feature of this model is that the encoding neurons are all tuned to the same vertical disparity (zero). Despite this, the decoding neurons are able to successfully recover 2D stimulus disparity. This is possible because vertical disparity causes distinctive patterns of activity across the encoding population. The model uses its stored knowledge about these patterns, in the form of templates of expected activity, to deduce the stimulus disparity.