The goal of the proposed work is to understand retinal processing through a careful analysis of the dynamic (space-time) patterns of activity that are generated at each retinal level. The spatiotemporal patterns will be generated in two complementary ways: (1) patterns will be measured and constructed through a novel method of physiological recording of excitation and inhibition over a large region for every cell type at each retinal level and extending the result to a population of that cell type, and (2) the patterns will be generated as the emergent properties of retinal models constructed by incorporating the space, time pharmacological and morphological properties derived from single cell studies. The measured and modeled patterns will be compared to evaluate the quantitative hypothesis of retinal embodied in the model. The modeling infrastructure, the Cellular Neural Network, (CNN) is now well-established massively parallel analog array processor designed with an architecture quite similar to the vertebrate retina. These studies will show how both simple and complex stimuli are represented in terms of physical, electrical activity (excitatory and inhibitory membrane currents and voltages) across arrays of thousands of elements at each retinal level. Analyses of these data will yield insights into the underlying neural mechanisms that mediate visual function and provide some clues about the strategies undertaken via neuronal interactions at each retinal level.
Showing the most recent 10 out of 33 publications
