Abstract

Electroencephalography (EEG) is an electrophysiological monitoring method to record electrical activity of the brain. Since the patterns of EEG activity changed with sleep stages, the feature extraction of EEG signals plays an important role in numerous automatic sleep scoring methods. EEG signals are quasi-stationary, i.e., they are considered stationary only within short-time intervals over longer periods of time, so numerous feature extraction methods rely on the short-time windowing techniques, including the short-time Fourier transform (STFT) and the wavelet transform. In order to reduce the blurring effect caused by the use of windows, we will introduce the background of STFT-based synchrosqueezed transform (SST) and apply this modern signal processing tool on the publicly accessible EEG signal database PhysioNet to demonstrate its usefulness on the subsequent classification of sleep stages. This is a joint work with Hautieng Wu (Duke University) and Yuan-Chung Sheu (National Chiao Tung University).