Temporal sequence storage capacity of time-summating neural networks
Paul C. Bressloff, John G. Taylor · Journal of Physics A Mathematical and General · 1992
The maximum capacity for storage of temporal sequences in a time-summating neural network is analysed within a statistical-mechanical framework. Each neuron of such a network maintains an activity trace consisting of a decaying sum of all previous inputs to that neuron. The maximum storage capacity is expressed as a function of the temporal correlations of the local fields of the neurons and evaluated using a perturbation expansion in powers of the decay rate.