An Automatic Sleep-Wake Classifier Using ECG Signals

Abstract

Sleep stage influence autonomic nervous system, this influence can be investigated by analysis of ECG signal. This paper presents system aimed to score sleep-wake stages using only the electrocardiogram (ECG) records. The feature extraction stage described in this paper was performed using methods of Heart Rate Variability analysis (HRV) and Detrended fluctuation analysis (DFA). These features are based on QRS detection times. Therefore, this detection was generated automatically for all recordings using a new algorithm based on the detection of singularities through the local maxima in order to construct the RR series. We illustrate the performance of this method using a neural network algorithm called Extreme Learning Machine (ELM). We make a comparative study of our algorithm using two classifiers back propagation neural network (BPNN) and support vector machine (SVM). The proposed method shows significantly better performance than back propagation neural network (BPNN), and almost same result than support vector machine (SVM), it achieves the classification accuracy of 78.33%.