Injury detection for central nervous system via EEG with higher order crossing-based methods.

摘要

Higher order crossing (HOC) is a powerful tool for time series analysis. Two HOC-based EEG analysis methods are developed for brain injury detection and quantification. The first method explores EEG spectrum characteristics via an estimate of the dominant frequency of a pre-processed EEG signal. The second method is based on the norm of the AHOC, an HOC obtained from the alpha-filter prefiltered EEG signal. Both methods are shown to be effective in detecting hypoxic/asphyxic injuries as well as assessing the severity of the injury.