multiSyncPy: A Python Package for Assessing Multivariate Coordination Dynamics

Abstract

In order to support the burgeoning field of research into interpersonal synchrony, we present an open-source software package: multiSyncPy. Multivariate synchrony goes beyond the bivariate case and can be useful for quantifying how groups, teams, and families coordinate their behaviors, or estimating the degree to which multiple modalities from an individual become synchronized. Our package includes state-of-the-art multivariate methods including symbolic entropy, multidimensional recurrence quantification, coherence, the cluster-phase ‘Rho’ metric, and a statistical test based on the Kuramoto order parameter. We also include functions for two surrogation techniques to compare the observed coordination dynamics with chance levels. Taken together, our collation and presentation of these make the study of interpersonal synchronization and coordination dynamics applicable to larger, more complex and often more ecologically valid study designs. In this work, we summarize the relevant theoretical background and present illustrative practical examples as well as guidance for the usage of our package - using synthetic as well as empirical data. Furthermore, we provide a discussion of our work and software and outline interesting further directions and perspectives. multiSyncPy is freely available under the LGPL license at: https://github.com/cslab-hub/multiSyncPy, and also available at the Python package index.

Prospects for Augmenting Team Interactions with Real‐Time Coordination‐Based Measures in Human‐Autonomy Teams

“…While a more in depth and systematic investigation and comparison of these methods is warranted, this nonexhaustive collation and overview of these methods is meant to serve as inspiration, and to provide a useful guide for interested researchers to consider candidate order parameters to model team coordination dynamics that could be implemented in HAT systems. Note that some of these methods require preprocessing to get multivariate signals into a format that is suitable for a given analysis; however, sufficient details are contained in the referenced source articles and some of these multivariate methods were collated and correlated in a recently released Python package: multiSyncPy (Hudson, Wiltshire, & Atzmueller, 2021). In terms of our envisioned HAT system, given input from key sensors of the team members, such as their neurophysiological state, this step would involve an estimation of the coordination of the team members at a sufficient resolution for utilization in the subsequent steps.…”

Never waste a good crisis : Examining team coordination breakdowns during crisis situations

“…Thus, for our analyses, for each window of the physiological signals, a recurrence plot was generated using the multiSyncPy package (Hudson et al, 2021), from which multiple properties can be derived. An overview of the properties is provided by Wallot (2019).…”