Mu or alpha? Source decomposition for social perception analysis

The sensorimotor mu rhythm is widely used as an electroencephalography (EEG) marker of mirror neuron system activity during action and emotion perception. However, its frequency overlap with occipital-parietal alpha raises concerns that sensor-level measurements over central electrodes may be contaminated by visual alpha activity. Here, we directly compared conventional sensor-level analyses with independent component analysis (ICA)-based source separation to examine how methodological choices affect interpretations of mu dynamics during emotional face perception. Thirty-three participants viewed static and dynamic facial expressions (smile, smirk) and nonsocial control stimuli while EEG was recorded. Time-frequency dynamics were analyzed using both conventional sensor-level measures at central and parietal electrodes and ICA mu and alpha sources. Cluster-based permutation statistics were applied to assess condition effects. Sensor-level analyses showed similar dynamics in central and parietal electrodes, with robust desynchronization patterns consistent with visual alpha modulation. In contrast, ICA revealed 2 functionally distinct components: an alpha source showing desynchronization and a sensorimotor mu component showing synchronization across most conditions. The mu component demonstrated minimal sensitivity to stimulus category. These results demonstrate that conventional sensor-level analyses do not provide a selective measure of mu activity and that prior reports of mu suppression to emotional faces may reflect alpha contamination. HighlightsConventional central-electrode electroencephalography analyses confound sensorimotor mu with visual alpha activity.Independent component analysis reveals distinct mu and alpha sources with opposite spectral dynamics during face perception.Mu rhythm shows synchronization and limited sensitivity to facial emotion stimuli during passive viewing.