Deterministic SR phenomena in autoassociative chaotic neural networks

In the previous works, we have set up a signal-driven scheme of the neural network with the coupling constants corresponding to certain information, and investigated the SR (stochastic resonance)-like effects under the aperiodic stored/non-stored pattern signals, comparing the deterministic case with the conventional case of the Hopfield network with stochastic noise. In this paper, we focus on deterministic SR-type responses in a simple system which consists of three neurons. This case is considered as a effective double-well model with chaotic dynamics. We demonstrate that SR-like effect can be enhanced by modulating the control parameter (bias strength) for a periodic input signal, in the phases of intermittency of chaos-chaos type.