Dendritic microprocesses and neural activity: implications for neural network modeling

The author shows how considerations at the basic neuronal level, involving neural-dendritic coupling, have a substantial influence on habituation and other neurophysiological processes. Traditional neural network models assume that neural interactions can be modeled in terms of axonal connections with different strengths. In contrast, there is considerable biological evidence to support the realization that brain processes such as perception are represented not so much by the activity of individual neurons as they are by patterns of polarization across ensembles of synaptic and post-synaptic interactions. At synapses, cooperativity of processes, even among spatially separated events, occurs. By incorporating models of dendritic field polarizations into neural network systems, a richer set of possibilities is created