Predicting Response to McGurk Illusion Based on Periodic and Aperiodic Prestimulus EEG Activity

Abstract

Studies have reported that prestimulus brain oscillations guide perceptual experiences during AV speech perception. However, what features of these prestimulus activity drives perception remains unknown. In this electroencephalography (EEG) study, we investigated the relationship between prestimulus periodic oscillations and aperiodic components with subsequent perception of the McGurk illusion at single-trial levels. Using logistic mixed-effect models, we determined which spectral features from different sensor regions predicted response to the illusory perception. We found that lower alpha (8-12 Hz) and beta (15-30 Hz) band oscillations over parieto-occipital sensors predicted illusion perception. We also found lower aperiodic offset values over parieto-temporal sensors, and lower global effect of exponent over the scalp, that predicted the response to McGurk illusion. We conclude that the predominant source of the prestimulus oscillatory state is aperiodic background activity, and that variations in these arrhythmic components account for inter-trial and inter-individual variability in perception of the McGurk illusion.