Spike–Gamma Phase Relationship in the Visual Cortex

Abstract

Gamma oscillations (30–70 Hz) have been hypothesized to play a role in cortical function. Most of the proposed mechanisms involve rhythmic modulation of neuronal excitability at gamma frequencies, leading to modulation of spike timing relative to the rhythm. I first show that the gamma band could be more privileged than other frequencies in observing spike–field interactions even in the absence of genuine gamma rhythmicity and discuss several biases in spike–gamma phase estimation. I then discuss the expected spike–gamma phase according to several hypotheses. Inconsistent with the phase-coding hypothesis (but not with others), the spike–gamma phase does not change with changes in stimulus intensity or attentional state, with spikes preferentially occurring 2–4 ms before the trough, but with substantial variability. However, this phase relationship is expected even when gamma is a byproduct of excitatory–inhibitory interactions. Given that gamma occurs in short bursts, I argue that the debate over the role of gamma is a matter of semantics.