Stratified Multivariate Multiscale Dispersion Entropy for Physiological Signal Analysis

Abstract

Multivariate Entropy quantification algorithms are becoming a prominent tool for the extraction of information from multi-channel physiological time-series. However, during the analysis of physiological signals from heterogeneous organ systems, certain channels may overshadow the patterns of others, resulting in information loss. Here, we introduce the framework of Stratified Entropy to control the prioritization of each channels' dynamics based on their allocation to respective strata, leading to a richer description of the multi-channel signal. As an implementation of the framework, three algorithmic variations of the Stratified Multivariate Multiscale Dispersion Entropy are introduced. These variations and the original algorithm are applied to synthetic and physiological time-series, formulated from electroencephalogram, arterial blood pressure, electrocardiogram, and nasal respiratory signals. The results of experiments conducted on synthetic time-series indicate that the variations successfully prioritize channels based on their strata allocation while maintaining the low computation time of the original algorithm. Based on the physiological time-series results, the distributions of features extracted from healthy sleep versus sleep with obstructive sleep apnea display increased statistical difference for certain strata allocations in the variations.This suggests improved physiological state monitoring by the variations. Furthermore, stratified algorithms can be modified to utilize a priori knowledge for the stratification of channels. Thus, our research provides a novel approach of multivariate analysis for the extraction of previously inaccessible information from heterogeneous systems.