An asynchronous pulse neural network model with finite pulse width for VLSI implementation

摘要

It has been pointed out that a spatio-temporal coding is very important in information processing a brain (1),(2). In this respect, an asynchronous chaotic neural network model (3), which uses spatio-temporal patterns of neural activities for computation, has been proposed. In this paper, we propose a VLSI circuit compatible asynchronous pulse neural network model by modifying the asynchronous chaotic neural network model. In the proposed model, a pulse with finite width is used instead of the impulse. Response characteristics of a single neuron and a network are investigated with computer simulations.