Reconstructing computational principles in a vertebrate adaptive reflex system

This research is based on the premise that computational principles evolved by biological systems are applicable to engineering devices, that an interaction of ongoing neurobiological experiments with computer simulation will be productive in extracting computational principles, and that tools for the analysis of neurobiological systems and for their modeling and simulation have evolved to the point where success is probable. The approach is based on neuroanatomical experiments to establish the connectional circuit. However, it is found that knowledge of network architecture (connectivity) is not enough. The biophysics of neurons leads to important diversity of intrinsic properties among constituents of a network. Consequently, neurons have dynamical response properties that can sensitively affect, or even completely alter, the functioning of a distributed network. Experimentally determined facts of both connection and biophysics must inform computational models in order to capture network dynamics/performance. Simulation results in turn generate hypotheses which drive further testing in biological experiments