A neuronal network model of associative memory and a brain theory.

Abstract

There are facts suggesting that the changes of excitability of neurons and conductance of synapses are responsible for learning and memory. In accordance with the idea, the author proposed a neuronal network model of associative memory, which can learn conditioned reflexes repeatedly together with its adaptive properties, without much interference with the previously existing memories. Further, a modified model can memorize dynamic images and establish motor patterns. An analysis of the structure of the model is presented, and some conclusions about this sort of model are drawn. A theory of memory and learning suggesting that the brain works by means of a mechanism similar to that of the model is proposed, as well as the explanations for some behavioral and clinical findings from the viewpoint of the model.