EEG-based brain connectivity analysis of states of unawareness

摘要

This work investigates phase synchrony as a neuro-marker for the identification of two brain states: coma and quasi-brain-death. Scalp electroencephalography (EEG) data of 34 patients were recorded in an intensive care unit (ICU), with 17 recordings for patients in a coma state, and 17 recordings for patients in a quasi-brain-death state. Phase synchrony was used for feature extraction from EEG recording by comparing the phase value between pairs of electrodes using an entropy based measure. In particular, we performed phase synchrony analysis in five standard frequency bands and provide visualization of the phase synchronies in matrices. The effectiveness of the phase synchrony features in each of the frequency bands are evaluated with statistical analysis. Results suggest phase synchrony for coma patients has a significant increase in the theta / alpha band compared to quasi-brain-death patients. Hence, we propose phase synchrony as a candidate for the identification of consciousness states between coma and quasi-brain-death.