Nonlinear dynamical and entropic complexity measures as indicators of nonstationarities in short-term ECG signals

Heart Rate Variability (HRV) is among important characteristics of general cardiac health. While 24-hour Holter monitoring is well recognized as a comprehensive analytical technique, short-term (up to 30 min.) electrocardiogram (ECG), recorded in presence of controlled environmental stimuli, remains rather unexplored in diagnostic practice. The presented method of change point detection in such class of signals addresses urgent quest from cardiologists for a simple, real time analysis of short-term ECG. Based on the concepts of Sample Entropy and Lempel-Ziv compression, the methodology combines sensitivity to subtle dynamical changes of the heart rate with robustness to noise. It was evaluated on 20 ECG signals from two groups of subjects: healthy and with cardiac problems. The method helps improve accuracy in discerning between the two groups and the results support hypothesis that healthy individuals are more responsive to environmental stimuli.