Multiscale permutation entropy analysis of electrocardiogram

• Published in: Physica A Vol. 471; pp. 492 - 498
• Main Authors: Liu, Tiebing, Yao, Wenpo, Wu, Min, Shi, Zhaorong, Wang, Jun, Ning, Xinbao
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2017
• Subjects: Congestive heart failure; Electrocardiogram; Multiscale entropy; Permutation entropy; Scale factor; Multiscale entropy; Permutation entropy; Congestive heart failure; Scale factor; Electrocardiogram
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2016.11.102

To make a comprehensive nonlinear analysis to ECG, multiscale permutation entropy (MPE) was applied to ECG characteristics extraction to make a comprehensive nonlinear analysis of ECG. Three kinds of ECG from PhysioNet database, congestive heart failure (CHF) patients, healthy young and elderly subjects, are applied in this paper. We set embedding dimension to 4 and adjust scale factor from 2 to 100 with a step size of 2, and compare MPE with multiscale entropy (MSE). As increase of scale factor, MPE complexity of the three ECG signals are showing first-decrease and last-increase trends. When scale factor is between 10 and 32, complexities of the three ECG had biggest difference, entropy of the elderly is 0.146 less than the CHF patients and 0.025 larger than the healthy young in average, in line with normal physiological characteristics. Test results showed that MPE can effectively apply in ECG nonlinear analysis, and can effectively distinguish different ECG signals. •We make MPE analysis of the ECG of CHF, healthy young and elderly people and compare to MSE.•As scale factor increase from 2 to 100, we analyze relationships of the three ECG signals’ complexity.•MPE can distinguish the three kinds of ECG effectively.•MPE shows some advantages to MSE in my contribution.