Further Explorations of a Minimal Polychronous Memory

The study of temporal spiking dynamics in biologically inspired neural networks (polychronous groups discovered by Izhikevich(2)) exhibits complex dynamics that makes it difficult to study. A minimal model of polychronous groups in neural networks was proposed by Maier and Miller (7) who discovered that a very minimal neural network model, without the synaptic weights was sufficient to produce PCGs. In this paper we expand on their study and propose ways to condition the network to produce sets of PCGs that are more unique; hence theoretically more descriptive of an input signal.