Empirically Constrained Network Models for Contrast-dependent Modulation of Gamma Rhythm in V1

Abstract

Here we present experimentally constrained computational models of gamma rhythm and use these to investigate gamma oscillation instability. To this end, we extracted empirical constraints for PING (Pyramidal Interneuron Network Gamma) models from monkey single-unit and LFP responses recorded during contrast variation. These constraints implied weak rather than strong PING, connectivity between excitatory (E) and inhibitory (I) cells within specific bounds, and input strength variations that modulated E but not I cells. Constrained models showed valid behaviours, including gamma frequency increases with contrast and power saturation or decay at high contrasts. The route to gamma instability involved increased heterogeneity of E cells with increasing input triggering a breakdown of I cell pacemaker function. We illustrate the models capacity to resolve disputes in the literature. Our work is relevant for the range of cognitive operations to which gamma oscillations contribute and could serve as a basis for future, more complex models.