The classical model of the visual cortex is that it is made up a set quasi-linear, independent spatial of a set of filters. The feature discrimination of individual cells is however quite broad and does not match the fine performance of visual behavior. There currently exists little hard evidence linking the synthesis of what we perceive with the fairly coarse selectivity of visual cortical cells. Because of the extremely high connectivity within A17, we believe that the neural """"""""code"""""""" cannot be adequately expressed by the analysis of isolated responses from single cells. One linking hypothesis is that visual structures is related by linking cells together dynamically in functional groups, and one historic theory is that the groups are defined by synchronous activity. This proposal is a direct exploration of these ideas. We will record simultaneously from large numbers of individual cells in primary visual cortex (A17) of the cat using 25 and 100 electrode arrays. We will stimulate the cells with patterns designed to recruit grouped activity and to show how groups change as patterns change. We will conduct these experiments for cells in both supra-granular and sub-granular layers, since lateral connectivity patterns differ between these regions. We will analyze the cell responses for synchrony using gravitational clustering and type analysis, which shows how cooperativity between cells may change over time and space. We will display dependency between neurons as a function of time and explore the relationship between trial-by-trial spike trains. Thus far we have found that cells as far separated as 1.5 mm can synchronize under certain stimulus conditions, and that the membership within synchronous groupings can change as stimuli change. We will extend these results from responses to simple stimuli to more complex multiple stimuli to explore the physiological basis for visual phenomena such as transparency and figure-ground discrimination. Our global goal is to provide a basis for how visual information is represented in the context of ensemble, as opposed to individual cell, activity. This is an important step towards understanding how the brain processes sensory information, and will contribute to our knowledge of brain pathologies as well as provide a strong base for development of sensory prosthetics.
