Time-dependent multivariate multiscale entropy based analysis on brain consciousness diagnosis

Abstract

The recently introduced multivariate multiscale sample entropy (MMSE) well evaluates the long correlations in multiple channels, so that it can reveal the complexity of multivariate biological signals. The existing MMSE algorithm deals with short time series statically whereas long time series are common for real-time computation in practical use. As a solution, we novelly proposed our time-dependent MMSE as an extension of MMSE. This helps us gain greater insight into the complexity of each section of time series, respectively, producing multifaceted and more robust estimates than the standard MMSE. The simulation results illustrated the effectiveness and well performance in the brain death diagnosis. © 2013 Springer-Verlag.