EEG Based Sleep Staging Analysis of BECTS Subjects

Spike-wave index (SWI) in electroencephalograms (EEGs) of different sleep stages is an important criterion for the diagnosis and prognosis evaluation of benign childhood epilepsy with centrotemporal spikes (BECTS). Accurate sleep staging thus becomes a prerequisite for statistical SWI analysis in different sleep stages of BECTS subjects. In this study, a novel BECTS sleep staging method based on multi-domain EEG features and statistical significance analysis has been developed. Particularly, 29 features from single-channel signal are taken for EEG characterization and statistical methods are adopted to describe the mechanisms underlying the variation in sleep in different types of subjects. Analysis on 11 BECTS and 7 normal subjects reveals that the proposed method can achieve 86 % and 86.7% F1 score, 87.2% and 92.7% kappa for BECTS and normal subjects, respectively.