Rhythmic Features of Movement Synchrony for Bonding Individuals in Dyadic Interaction

Abstract

This study examined ways in which rhythmic features of movement contribute to bonding between individuals. Though previous studies have described synchrony as a form of social glue, this research extends those findings to consider the impact of fast versus slow tempo on movement synchrony. This two-part experiment examined dyadic interactions as they occurred between same-sex strangers (Study 1) and friends (Study 2). Participants were video-recorded as they engaged in 5-or 6-min chats, and synchrony was evaluated using wavelet transform via calculations of cross-wavelet coherence. Study 1 employed regression commonality analysis and hierarchical linear modeling and found that among various frequency bands, rapport between individuals was positively associated with synchrony under 0.025 Hz (i.e., slower than once every 40 s) and 0.5-1.5 Hz (i.e., once every 0.67-2 s). On the contrary, Study 2 determined that synchrony of 0.5-1.5 Hz was not impactful among friend dyads and only predictive of the motivation to cultivate a friendly relationship during interactions with strangers. These results indicate the existence of a distinctive rhythm for bonding individuals, and the role of pre-existing friendship as a moderator of the bonding effect of synchrony. However, the role of relative phase (i.e., timing of movement; same versus opposite timing) remains unclear, as the ratio of in-and anti-phase patterning had no significant influence on perceived rapport and motivation to develop relationships. On the basis of the research results, a theoretical contribution is proposed to the study of interpersonal coordination.

Comparing Manual and Automated Coding Methods of Nonverbal Synchrony

“…This technique does not capture the coordinate point of each joint part but calculates the change in gray-scale pixels between adjacent video frames in a region of interest (ROI), which can be encoded as a movement if the camcorder was fixed and neither background nor the lightning condition changed through the clip. Previous studies suggest that the reliability of this technique for capturing movement is high (e.g., Fujiwara et al, 2019Fujiwara et al, , 2020Schoenherr et al, 2019).…”

“…Spectrum analysis is an analysis technique for time-series signals in the frequency domain, which deconstructs complex time-series into rhythmic components. Thus, cross-spectrum analysis enables the researcher to assess synchrony from the perspective of rhythm (Fujiwara et al, 2020). It should be noted that these kinds of time-series analyses offer information regarding the convergence of two time-series, not the similarity of "behavior" that is categorized and counted/rated via manual coding.…”

“…This is similar to how previous studies use the absolute value of cross-correlation analysis (Schoenherr et al, 2019;Tschacher et al, 2014), where the negative correlation is considered as one pattern of synchrony. Overall, this measurement can be interpreted as Pearson's correlation, but it indicates the convergence of rhythm or the degree of synchrony (Fujiwara & Daibo, 2016;Fujiwara et al, 2020;Schmidt et al, 2012Schmidt et al, , 2014.…”

“…Previous studies on structured/unstructured conversations have employed this approach to assess synchrony (Fujiwara & Daibo, 2016;Fujiwara et al, 2020;Schmidt et al, 2012Schmidt et al, , 2014, which added new insight to the synchrony research that has been conducted by manual rating/coding. Fujiwara et al (2020) revealed that synchrony, the cross-wavelet coherence score (the WTC score) calculated via the cross-wavelet transform, in unstructured conversation was significantly associated with rapport between speakers, which supported the previous studies (Bernieri, 1988;Tickle-Degnen & Rosenthal, 1990). It also illustrated that the WTC score in a specific frequency band (i.e., .5-1.5 Hz, the rhythm of once every 0.67-2 s) had a distinctive association with rapport.…”

“…C-BAS was employed for the manual coding to count the occurrence of synchrony such as gestures, postural shifts, and nodding. For automated coding, this study features Motion Energy Analysis (Ramseyer & Tschacher, 2011) because it has been widely used in previous studies (e.g., Fujiwara et al, 2019Fujiwara et al, , 2020Schoenherr et al, 2019) and is widely available to many researchers. The MEA software has excellent usability but does not require a written code for analysis, and the current version (4.11; June 30, 2020) is available for Mac and Windows users.…”