Quantifying the Dynamics of Interpersonal Interaction: A Primer on Cross-Recurrence Quantification Analysis using R

Humans live in a very interactive context, requiring very frequent exchange of information with con-specifics, which itself requires subtle temporal calibration of linguistic and nonlinguistic activities. Over the last decade, the dynamics of interpersonal interaction has become a growing topic in cognitive science, precisely because of the important implications that interpersonal dynamics carry on shaping our ‘social cognitive system.’ Research on this topic has helped us understand, for example, how overt behavior such as body sway and eye movements of interacting individuals converge or diverge in various ways (Shockley, Santana, & Fowler, 2003; Richardson & Dale, 2005), whether temporal ‘calibration’ occurs over multiple behavioral scales (Louwerse, Dale, Bard, & Jeuniaux, 2012); as well as, developmental childcaregiver dynamics (Yu & Smith, 2013). Many important advances on this research topic have been possible through the application of concepts and statistical methods, which provide quantification for the dynamic structure of cognitive responses observed when individuals interact. Recurrence Quantification Analysis (RQA) is one of such framework, and has received growing attention for research on interpersonal dynamics. Conceptually, RQA makes it possible to quantify how, and the extent to which, a signal is revisiting a similar state in time (Marwan, Carmen Romano, Thiel, & Kurths, 2007). When RQA is applied on two different streams of the same information, such as the eye-movement trajectories of two interlocutors, it takes the name of Cross-Recurrence Quantification Analysis (C/RQA). C/RQA can be used, for example, to examine the temporal organization of eye-movement trajectories of dyads of interlocutors as they complete a communicative task, and establish their attentional correspondence, their feedback dynamics (e.g., leader-follower lag), as well as examine how experimental variables might foster or disrupt such synchronism. In a sense, this makes C/RQA a very comprehensive time series technique for obtaining new descriptive statistics, and some have referred to it as a sort of generalized nonlinear crosscorrelation function (Marwan et al., 2007). In this tutorial, cognitive scientists from different fields, and at different stages of their career (from graduates students to senior scientists) will learn a dynamical systems framework to interpret and understand interpersonal interaction, and acquire the analytical principles of C/RQA, which helps framing this approach.