Information Processing in a Network of Model Neutrons: A Computer Simulation Study

摘要

An interactive computer simulation package intended for the simulation of neuronal network activity is described. The networks simulated are comprised of excitatory and inhibitory model neurons connected with each other via conventional chemical synapses. The model neurons were designed to be as biologically realistic as possible within the limits imposed by the computational capacity of an ordinary mainframe computer. The package allows interactive manipulation of network configuration and system parameters and it includes routines for display of results on a graphic terminal or plotter. The biological relevance of the information processing principles suggested by the so called cell assembly theory is discussed. Some simulation results using a simplified model of the mammalian cerebral cortex are reported and discussed. The formation of subsets of neurons (cell assemblies), that serve as internal representations of different stimulus patterns could be simulated.