Modeling Brain States and Functions with a Chaotic Neural Network

An analog neural network model with time delays was studied which demonstrated chaos similar to human and animal EEGs. Measures of chaos and correlation were used to investigate chaotic neural network behavior under the different external conditions and internal structural changes. External sinusoidal and pulsed periodic forces were used for chaos control. They produced transitions, "chaos-order" and "chaos-chaos", which were similar to those observed experimentally in the brain. A chaotic neural network state was introduced which is similar to that proposed for the brain. It is characterized by two types of memory and correlation structure of neuronal activities. Pathological non-chaotic neural network activity and activity with decreased chaos is induced by weakening interneuronal connections. Possible application to Parkinson's disease and dementia is discussed.