INTEGRATION OF CHAOTIC BEHAVIOR MODULES INTO A FLOW-TYPED 2-BODY NEURAL-NETWORK MODEL

Features of information processing in the brain include distributed fu nctional structure, as in the areas and columns, as well as integrated processing among those parts. This paper presents a model, where one- way coupling is provided between two-body modules, as a basic structur e for the neural network model composed of multiple modules. The effec tive role of chaos is demonstrated as the integration mechanism of the neural network model. The relation between the coupling density betwe en modules and the pattern recall behavior in the two-body one-way cou pling model is analyzed. It is verified that there exists all area, ca lled the A area, of the intermodule coupling parameters, where the pat terns can be retained in both modules. (C) 1998 Scripta Technica.