Automatic detection of sleep apnea events based on inter‐band energy ratio obtained from multi‐band EEG signal

Abstract

Sleep apnea is a potentially serious sleep disorder characterised by abnormal pauses in breathing. Electroencephalogram (EEG) signal analysis plays an important role for detecting sleep apnea events. In this research work, a method is proposed on the basis of inter‐band energy ratio features obtained from multi‐band EEG signals for subject‐specific classification of sleep apnea and non‐apnea events. The K ‐nearest neighbourhood classifier is used for classification purpose. Unlike conventional methods, instead of classifying apnea patient and healthy person, the objective here is to differentiate apnea and non‐apnea events of an apnea patient, which makes the task very challenging. Extensive experimentation is carried out on EEG data of several subjects obtained from a publicly available database. Comprehensive experimental results reveal that the proposed method offers very satisfactory classification performance in terms of sensitivity, specificity and accuracy.