Unconsciousness State Identification Using Phase Information Extracted By Wavelet And Hilbert Transform

This work aims to determine features for the distinction of coma and quasi brain death (QBD) consciousness states by implementing an algorithm for extracting phase information from EEG data using Wavelet/Hilbert Transform. The relationship between the EEG data recorded from pairs of different electrodes is then quantified by calculating phase synchrony using Shannon entropy, as a phase synchrony index (PSI). Statistical analysis was used to evaluate the significant pairs of electrodes for the features extracted in the different frequency bands. The findings suggest confirm that both Wavelet and Hilbert Transform based phase synchrony analysis provide similar results. In particular, Hilbert Transform might be a more suitable method for phase synchrony analysis to characterize coma or QBD brain states for lower frequency bands. Using non-parametric statistical tools is reliable and does not require strong assumption on the dataset distribution. The algorithm is not designed to be a diagnosis tool; it rather serves as a secondary test to confirm diagnosis. It also has a potential contribution for the systematic distinction of brain states in other areas of EEG-based research.