Multi-dimensional networks as a tool to quantify role imbalance in EEG-hyperscanning data

Social interactions are fundamental to human life, shaping our behaviour, our social relationships, and even our cognitive and emotional development. They can take several forms, including cooperation, competition and collective decision making. Given their nuanced and complex nature, social interactions are seldom symmetrical, and leader-follower dynamics can spontaneously emerge, even if the task and the instructions are balanced and no specific role is pre-assigned. EEG hyperscanning was developed as a technique able to simultaneously record brain activity from multiple subjects and to develop a multi-subject model including inter-brain functional connectivity. A critical challenge in hyperscanning studies is identifying roles when they emerge spontaneously rather than being pre-assigned. In this study, we use a framework based on the combination of multiple-brain connectivity and multi-dimensional network modelling to characterize roles within dyads of subjects. In this context, we defined and implemented a new measure, which we called the inter-imbalance index, to retrospectively realign multiple-brain connectivity networks across the dyads based on the roles established during the interaction. This index was tested on multi-subject connectivity matrices extracted from 16 dyads of neurotypical volunteers performing a joint action task. Results showed a higher similarity between the multi-subject networks after the realignment with respect to the original configuration. This enhanced similarity suggests that our approach effectively identifies and accounts for spontaneously emerging role dynamics. Although future work will be needed to fully elucidate its relationship with the leader-follower dynamics, our results suggest that the proposed index can be an effective tool to capture the subjects’ roles as they emerge spontaneously during the social interaction.