摘要

The regularity of heart rates has a loss in cases of illness and aging. Assessing the dynamics of heart rate fluctuations can provide valuable information about cardiac system. In this paper, heart rate fluctuations and its wavelet entropy (WE) are analyzed to demonstrate its potentials for risk stratification of cardiac diseases. The regularity of heart rate fluctuations is estimated by exploiting the time-frequency localization ability of wavelet analysis and the ability of entropy. The results show that WE for patients with congestive heart failure show a very low value and can be completely separated from health subjects. In addition, the values of WE decrease with aging. The lower the values of WE, the higher the risk of heart disease is. The values of WE also reflect the distribution of the energy of heart rhythm. Significant correlations are demonstrated between WE and the power in the three frequency bands. The results have shown that WE can be used to analyze short, non-stationary data time series both in time domain and in frequency domain simultaneously and can be feasible for the discrimination of the differences of heart rate fluctuations between healthy groups and CHF groups as a diagnostic tool.