Quantifying Heart Rate Variability Using Multiscale Fuzzy Dispersion Entropy

Heart rate variability (HRV), which is the variation of inter-beat intervals, exhibits complex characteristics on multiple temporal scales due to the balancing function of the autonomic nervous system. Although there are various nonlinear analysis methods for assessing the complexity of HRV, quantifying HRV over multiple scales is lacking. Here, we present a novel multiscale fuzzy dispersion entropy (MFDE) measure that incorporates quantifying fuzzy dispersion entropy over multiple temporal scales. The proposed MFDE comprises two steps: First, a coarse-graining procedure is carried out for the multiscale decomposition of an inter-beat interval. Second, it conducts FDE computation for each coarse-grained time series. It results in the quantification of complexity, reflecting the long-range correlations inherent in HRV. Using synthetic signals and actual electrocardiogram (ECG), we evaluate the performance of MFDE and compare it to the traditional multiscale entropy methods. The results using synthetic signals show better robustness of MFDE for quantifying complexity with various lengths and predefined parameters. The results using ECGs demonstrate that the proposed MFDE leads to more significant discrimination of HRVs of different cardiovascular states regarding
$p$
-values from the Mann-Whitney U test. The capability of MFDE can provide a prospective tool for real-time and practical computer-aided diagnosis using HRV analysis.