Pattern Propagation Speed in Synfire Chains with Excitatory-Inhibitory Couplings

Abstract

The speed of firing pattern propagation in a synfire chain, composed of non-leaky integrate-and-fire neurons, and assuming homogenous connection delays, is studied. An explicit relation, relating the propagation speed to the connecting weights distribution and other network parameters, is derived. The analytic results are then checked with a computer simulation. When the network is fed with a fully synchronized input pattern, the pattern propagation speed is independent of the weight parameters. When the fed input is asynchronous, depending on the weight parameters, the propagation speed is more than or less than the synchronous case. In this case the propagation speed increases by increasing the mean or standard deviation of connecting weights. The biological relevance of these findings and their relevance to the notion of synfire chains are discussed.