Delayed accumulation of inhibitory input explains gamma frequency variation with changing contrast in an Inhibition Stabilized Network

Abstract

AbstractGamma rhythm (30-70 Hz), thought to represent the push-pull activity of excitatory and inhibitory population, can be induced by presenting achromatic gratings in the primary visual cortex (V1) and is sensitive to stimulus properties such as size and contrast. In addition, gamma occurs in short bursts, and shows a “frequency-falloff” effect where its peak frequency is high after stimulus onset and slowly decreases to a steady state. Recently, these size-contrast properties and temporal characteristics were replicated in a self-oscillating Wilson-Cowan (WC) model operating as an Inhibition stabilized network (ISN), stimulated by Ornstein-Uhlenbeck (OU)-type inputs. In particular, frequency-falloff was explained by delayed and slowly accumulated inputs arriving at local inhibitory populations. We hypothesized that if the stimulus is preceded by another higher contrast stimulus, frequency-falloff could be abolished or reversed, since the excessive inhibition will now take more time to dissipate. We presented gratings at different contrasts consecutively to two female monkeys while recording gamma using microelectrode arrays in V1 and confirmed this prediction. Further, this model also replicated a characteristic pattern of gamma frequency modulation to counter-phasing stimuli as reported previously. Thus, the ISN model with delayed surround input replicates gamma frequency responses to time-varying contrasts.Significance statementGamma rhythms represent sustained push-pull dynamics between excitatory and inhibitory populations during visual stimulation. Gamma power and centre frequency varies depending on stimulus features, and onset of stimulus produces a “frequency-fall” trend where onset frequency is higher and subsequently plateaus to a lower value. In an earlier work, we argued, using a noisy rate-model of V1, that a delayed onset of inhibition-drive from the surround populations produced the gamma ‘frequency-falloff’. We tested a key prediction of this hypothesis that the frequency-falloff can be abolished or reversed if the stimulus is preceded by a higher contrast stimulus, and confirmed the same by recording from primate primary visual cortex while presenting multiple stimuli consecutively at varying contrasts.