A comparison of different time series techniques to analyze phasic coupling: A case study of cardiac and electrodermal activity

Abstract

This study examined the efficiency of different time series analysis techniques to extract information on the coupling of spontaneous phasic physiological responses. We compared four bivariate approaches, cross-spectral, cross-covariance, cross-covariance with prewhitening, and dynamic factor analysis, in their ability to yield unbiased estimates of (a) shared variance, (b) covariance, (c) strength of relationship, and (d) interchannel time-lag in empirical and simulated interbeat interval–electrodermal activity (IBI–EDA) time series. All methods produced similar estimates of the grand-averaged IBI–EDA dynamics, but only the measures of covariance produced reliable and unbiased estimates of the interindividual distribution of IBI–EDA coupling. We conclude that the extraction of phasic response patterns during continuous and unrestricted experimental situations may considerably facilitate psychophysiological research.