GATSBY COMPUTATIONAL NEUROSCIENCE UNIT
UCL Logo

Synchrony in large groups encodes contours

A B Bonds, Z Zhou, M R Bernard and J M Samonds

Depts. of Electrical Engineering and Computer Science and Biomedical Engineering, Vanderbilt University, USA

The presence of synchrony across larger (>2) assemblies of cells is strongly dependent on contour structure. We have studied synchrony in cell populations in primary visual cortex by recording with a 100-electrode array in anesthetized and paralyzed cats. We quantified synchrony using information-based Type Analysis, the Joint PST Histogram as well as Coherence (for two-cell assemblies) and a novel heuristic for larger populations. Most of the cooperative information in a population takes the form of coordinated firing within about a ten-millisecond window. This synchrony is stimulus-dependent and information grows nearly exponentially with the size of assemblies. Membership in a synchronous assembly is dynamic, reorganizing with different stimulus configurations. The stimulus features most conducive to generation of synchrony are well-defined, contiguous contours that follow the principles of the association field and that are matched to the collective receptive fields of the assembly. We have also explored the mechanisms supporting synchrony by comparing firing synchrony and coherent gamma oscillation. Gamma oscillation develops much more slowly than synchrony, but when it is present appears to serve as an organizational framework that enhances synchrony. These results are consistent with a model in which coherent stimulus structure triggers synchrony that is sustained by network interactions if the stimulus is a good fit to the aggregate receptive field of the assembly. We have no idea whether this is relevant to the process of vision.

NIH:R01EY014680-03

BACK