Interpersonal Physiological Synchrony (IPS) refers to the temporal coordination of autonomic states during social interactions. Despite the potential importance of IPS in modulating social functions, a clear understanding of its functional role as well as the contextual and individual factors that may promote or hinder its emergency, is missing. In the present study, we recorded the cardiac activity of dyads engaged in a cooperative joint action task and measured the degree of IPS occurring during the task. To investigate the role played by contextual factors, we engaged participants in different versions of the task (peer-to-peer vs leader-follower, complementary vs imitative movements) and explored whether switching compared to repeating the same task version would affect IPS. Moreover, we assessed the influence of dyad members’ Social Anxiety and Perspective Taking abilities on IPS. Our results revealed a significant increase in IPS when participants switched to a new version of the task (i.e., during switch blocks). In addition, dyad-level Social Anxiety negatively predicted IPS, indicating that more socially anxious dyads were less likely to become physiologically aligned. We also observed that individual heart rate variability (HRV), an autonomic marker of attentional performance in cognitive tasks, increased during switch blocks. We submit that switching to a new interpersonal task requires additional attentional resources to monitor and predict the action of the partner, and that Social Anxiety is characterised by a focus of attention aberrantly shifted toward the self. Thus, we propose that IPS during joint action may represent an index of reciprocal performance monitoring.