Study of analogue neural networks that obey Dale's law using mean-field theory

Abstract

The mean field formalism of attractor neural networks described in terms of spike rates and currents is extended to the study of a network of analogue excitatory neurons in which the effect of the inhibitory neurons is modelled as a function of the excitation. It is shown that such a network of integrate-and-fire neurons has attractors with uniform low firing rates that correspond to the retrieval of single patterns. The analysis is carried out for extensively many patterns using the replica symmetric approximation.