A study on the bursting firing phenomena in neural network with the hardware neuron model

Abstract

For information communication in a biological system, bursting firing phenomena are very often confirmed in thalamic and neocortex neurons of the brain, and contain various information on processing. We have been studying bursting firing phenomena of neurons with particular reference to the delay feedback of signals in the nervous system. We discuss bursting firing phenomena in the recurrent neural network. First, we show the negative resistance circuit and pulse type hardware neuron model with the circuit, which is an element of the neural network. Next, we show the dynamics in the self recurrent circuit and relay recurrent circuit, which are basic construction units of the recurrent neural network with the neuron model. As a result, it is verified that bursting firing phenomena can be simulated by the neural network models, where the dynamics is controlled by the external input and circuit parameters.