Abstract
This methodological paper examines the assessment of interpersonal synchrony during a joint dancing task between mothers and their children (aged 4 to 5 years) using OpenPose. This pose estimation tool captures movement in naturalistic settings. The study analyzes 45 mother-child dyads, comparing two analytical methods for assessing synchrony, and examines their correlation with the Coding Interactive Behavior (CIB) measure of interaction quality. The first method employs cross-wavelet transform (CWT) coherence to assess synchrony based on vertical head movement. This straightforward and computationally efficient approach reveals a significant correlation between interpersonal synchrony and CIB scores, thus implying its potential as a reliable indicator of interaction quality and suggesting its potential as a measure of interaction quality. The second method, the generalized cross-wavelet transform (GCWT), analyzes synchrony across multiple body parts, offering a more complex and detailed analysis of interpersonal dynamics. However, it did not significantly correlate with the CIB scores. Our findings suggest that focusing on head movement using CWT can effectively capture critical elements of interpersonal synchrony linked to interaction quality. In contrast, despite its richness, the more complex GCWT approach may not align as closely with observed interactive behaviors as the CIB scores indicate. This study underscores the need to balance methodological complexity and ecological validity in research, offering insights into selecting analytical techniques based on research objectives and the nuances of interpersonal dynamics. Our results contribute to the field of interpersonal synchrony research, emphasizing the benefits of efficient methods in understanding mother-child interactions and interaction relationships in general.