Spiking neural network for recognizing spatiotemporal sequences of spikes

A decoding scheme using a spiking neural network for recognizing spatiotemporal sequences of spikes was discussed. The network consists of excitory neurons and two globally inhibitory interneurons of different types that provide delayed feedforward and fast feedback inhibition. It was suggested that the neural network recognition scheme is invariant to variations in the intervals between input spikes within some range. It was also suggested that the proposed network provides a simple way to decode spatiotemporal spikes with diverse types of neurons.