Automatic sleep staging by simultaneous analysis of ECG and respiratory signals in long epochs. (April 2015)

• Record Type: Journal Article
• Authors: Ebrahimi, Farideh; Setarehdan, Seyed-Kamaledin; Nazeran, Homer
• Is Part Of: Biomedical signal processing and control. Volume 18(2015)
• Page Start: 69
• Page End: 79
• Publication Date: 2015-04
• Subjects: Heart Rate Variability (HRV) -- ECG-Derived Respiration (EDR) -- Empirical Mode Decomposition (EMD) -- Discrete Wavelet Transform (DWT) -- Detrended Fluctuation Analysis (DFA) -- Support Vector Machine-Recursive Feature Elimination (SVM-RFE); Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2014.12.003
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: We evaluated the potential utility of combining ECG and respiratory signals in sleep staging. These signals are easily recorded and can be useful in home sleep monitoring systems. The SVM-RFE and SVM were used for the ranking and classification of extracted features, respectively. The extracted features from thoracic respiratory signals have higher ranking than ECG-derived respiratory features. The best result was obtained by using HRV and thoracic respiratory signals. Abstract: EEG, EMG, and EOG are very informative signals recorded in polysomnography (PSG) and used for sleep staging. Their reliable acquisition at home, however, is difficult. In comparison, ECG and thoracic respiratory (R) signals are easier to record and can be useful in home sleep monitoring systems. The simultaneous utilization of Heart Rate Variability (HRV) and respiratory (R) signals seems a plausible scenario as both heart rate (HR) and respiration rate (RR) vary during different sleep states. Therefore, we explored the combined discriminative capacity (accuracy, sensitivity, and specificity) of ECG/R signals in automatic sleep staging. As baseline, we classified the wakefulness, Stage 2, SWS (slow wave sleep) and REM sleep by using a Support Vector Machine (SVM) fed with a set of features extracted from: (a) HRV (34-features), (b) HRV/ECG-Derived Respiration (45-features), and (c) combined HRV/R (45-features) signals. Approach (a) produced reasonable discriminative capacity, while approach (b) significantly improved the classification; however, the best outcomes were achieved by using approach (c). Then, we enhanced the SVM classifier with the Recursive Feature Elimination (RFE) method. The classification results were improved with 35 out of the 45 HRV/RS-EDR features. In comparison, best results were obtained by combining 27 out of the 45 features derived from HRV/R signals, in which the optimal feature set selected by the SVM-RFE method, included a combination of time domain, time-frequency, and fractal features, as well as entropies. Overall, these improvements revealed that it is possible to simplify home monitoring of sleep disorders and achieve high discriminative capacity (accuracy = 89.32%, specificity = 92.88%, and sensitivity = 78.64%) in automatic sleep staging by the exclusive recording of cardiorespiratory signals.