Capturing Spatially Organized Oscillatory Cliques as Signatures of Neuronal Assemblies

Abstract

State-of-the-art neural recording technologies now enable dense sampling of neuronal activity, demonstrating meaningful spiking variability from moment to moment. Yet spiking activity is sparse by nature and allows only partial access to collective neuronal dynamics. Local field potentials (LFPs) provide complementary information by capturing coordinated activity from large neuronal populations, but methods to characterize the fine-scale spatio-temporal organization of LFPs are lacking. Starting from the notion that neural oscillations are brief and burst-like, we introduce a framework to detect and analyze SPatially Organized Oscillatory Cliques (SPOOCs), oscillatory events that are cohesive in space, time, and frequency. SPOOCs displayed diverse dynamics in space and frequency and were differentially modulated during stimulus processing. Linking SPOOCs to local spiking, we demonstrate that these events index rapid reconfigurations of neuronal assemblies. With SPOOChunter, we provide an open-source toolbox that enables systematic detection and analysis of transient, spatially organized population dynamics in high-density electrophysiological recordings.