Because the visual system is a collection of information processing networks, a particularly useful way to express our knowledge of its function is in terms of a well defined dynamical network model whose interacting elements reflect the quantitative physiological properties of visual neurons. As such a description evolves, it will not only broaden our view and deepen our insight concerning the nature of visual activity; it will also more sharply relate the symptoms of visual abnormalities to specific departures from normal dynamics at the level of neurons, in a way that has clear potential for clinical usefulness. The goal of the present proposal is to further develop such a model of interacting visual neurons. Major emphasis in the next five years will be upon detailed dynamical modeling of retinal activity in cat and in primates. This will be carried forward in close conjunction with experimental work, primarily visual neuron recordings under stimulation from computer-generated displays of a form devised to expedite modeling. From these data dynamical equations will be developed, which will summarize the behavior and indicate the physiological structure of the piece of nervous system under investigation. The dynamical equations will be translated into computer simulations which will make further predictions to be compared against further experiments, and in this way the process of modeling will be extended.
Showing the most recent 10 out of 38 publications
