Preserved temporal hierarchy but frequency-specific alterations in dynamical regimes of EEG microstate multimers during reversible unconsciousness.

Summary

EEG microstate sequences, which reflect brief patterns of brain activity, show precise temporal rhythms across theta, alpha, beta, and gamma frequency bands. These rhythms persist during both wakefulness and reversible unconsciousness induced by anesthesia or deep sleep. The periodic patterns arise from repeating multimer structures and the duration of each microstate, as shown by surrogate data analysis and a generative model. During deep sedation and N3 sleep, beta-band microstates specifically increase in peak power and shift to a lower center frequency. A new algorithm extracted these multimer sequences, revealing distinct frequency-dependent changes during unconsciousness that point to a shift toward specific dynamical brain regimes. The findings offer potential neurophysiological biomarkers for assessing consciousness.

Study at a glance

Abstract

Employing a spectral analysis framework based on Chaos Game Representation (CGR), we investigated the multimer-based dynamics of EEG microstates across broadband and canonical frequency bands during reversible unconsciousness (anesthesia and sleep). Robust periodic components consistently emerged within microstate sequences across theta, alpha, beta, and gamma bands, persisting across distinct states of consciousness. Converging evidence from both deconstruction via surrogate data and reconstruction via a hierarchical generative model demonstrates that the multimer structure, along with the conditional duration distribution, constitutes the underlying mechanism of microstate periodicity. Furthermore, we show that temporal smoothing abolishes these intrinsic periodic components. Most notably, during both deep sedation and N3 sleep, the beta band microstate sequence exhibited a consistent increase in peak power and a decrease in center frequency, resulting in highly characteristic patterns in the CGR spectra. To dissect the structural basis of these periodicities, we developed a data-driven algorithm to extract multimers and calculate their metrics. We identified distinct, frequency-dependent alterations in multimer dynamics during reversible unconsciousness, suggesting that the transition to unconsciousness marks a shift towards specific dynamical regimes. Collectively, our findings confirm that microstate sequences exhibit precise temporal orchestration. By elucidating the generative mechanisms of microstate periodicity and establishing a multimer-based analytical framework, this study provides a solid methodological foundation for investigating higher-order temporal structures, while offering promising neurophysiological biomarkers for consciousness assessment and novel insights into the temporal organization of large-scale neural dynamics.