Macroscale dynamics of EEG microstates determine the periodic and aperiodic features of the neural power spectrum

Electroencephalography (EEG) recordings from the scalp reveal two main components in the neural power spectrum: periodic oscillatory rhythms and aperiodic broadband fluctuations. Researchers have long used spectral decomposition to analyze these features, but current interpretations lack a comprehensive mechanism and often mix signals from different neural sources.

The study suggests that the periodic rhythms and most of the broadband spectral power in the brain's primary frequencies are generated by the network architecture responsible for creating EEG microstates. Microstates are a limited set of quasi-stable topographic voltage patterns that represent the cortex's current functional state, and their dynamics produce both periodic and aperiodic spectral features.

To examine this mechanism, the researchers isolated and eliminated the spatial projections of microstates from high-density EEG recordings using orthogonal subspace projection on both surface scalp data and modeled cortical generators. Removing seven specific microstates significantly reduced both alpha and theta rhythms, as well as elements of the aperiodic 1/f background.

Further analysis indicated that each microstate has unique oscillatory generators and distinct 1/f aperiodic structures. Overall, the findings indicate that the dominant periodic and aperiodic spectral features are best understood as the frequency-domain manifestations of the collective brain networks that produce EEG microstates.