EEG And ECG Based Sleep Stage Analysis

Various kinds of cardiovascular & cerebrocascular diseases and social aging make the evaluating of the sleep quality and diagnoses of sleep diseases become the most challenging research in the 21st century. It is crucial of decreasing the incidence of cardiovascular & cerebrocascular diseases to improve the efficiency of diagnosis in early stage.The dissertation mainly focuses on the feature extraction of EEG &ECG and sleep stage processing. First of all, the paper disserts the standard of sleep staging and analysis methods inside and outside, in general. The chapter 2 depicts the theory of EEG & ECG and their physiological characteristic, which are the bases for this paper. The first Lyapunov exponent and ApEn are discussed, and used in EEG basic feature extraction. Comparation between the two methods is done to find the most proper parameter. SampEn and DFA are used to find the correlation between HRV and sleep stages .According to the chapter 1, four parameters found above are calculated for sleep segmentation processing.The emulation results with Matlab show that the fist Lyapunov exponent, particularly the ApEn can partly reveal the sleep features used in EEG. Search on finding the relation between HRV and sleep features couldn't get what we wanted, that is to say, it is not a good method to extracting the sleep feature only with the HRV. According to the SVM, we find that the stage of Wake and the sleep is distinct, however changes between the NREM and the REM is very small. The dissertation predicts the future trend of EEG processing and sleep stages in the end.