Just a phase? Causal probing reveals spurious phasic dependence of sensory perception

Abstract

For over a decade, electrophysiological studies have reported correlations between sensory perception and the phase of spontaneous pre-stimulus brain oscillations. To date, these findings have been interpreted as evidence that the brain uses neural oscillations to sample and predict upcoming stimuli. Yet, evidence from simulations have shown that analysis artefacts could also lead to spurious pre-stimulus oscillations that appear to predict future brain responses. To address this discrepancy, we conducted an experiment in which visual stimuli were presented timed to specific phases of spontaneous alpha and theta oscillations. This allowed us to causally probe the role of ongoing neural activity in visual processing independent of the stimulus-evoked dynamics. Our findings did not support a causal link between spontaneous alpha / theta rhythms and behaviour. However, spurious correlations between theta phase and behaviour emerged offline using gold-standard acausal time-frequency analyses. These findings demonstrate that care should be taken when inferring causal relationships between neural activity and behaviour using acausal analyses. HIGHLIGHTSO_LIWe causally probed the role of spontaneous EEG rhythms in visual attention via phase-locking. C_LIO_LITheta phase predicted behaviour offline, but cues presented in real-time with theta had no effect on behaviour. C_LIO_LISpurious theta rhythmic sampling is an artefact of the evoked potential and acausal filtering. C_LIO_LIecHT accurately computes the phase in real-time and mitigates erroneous phase-behaviour correlations. C_LI