Quantification of the spontaneous emergence of leader-follower dynamics in EEG hyperscanning data

Leader-follower dynamics play a fundamental role in coordinated interactions between two or more subjects. Due to the complexity of social interactions, they can emerge spontaneously in social settings even without an explicit role assignment. Investigating these dynamics at the neural level requires techniques capable of capturing inter-brain coordination, such as hyperscanning, which enables the simultaneous recording of brain activity from multiple subjects. While some studies have used the direction of the causal relations in the inter-brain model as an indication of the roles assumed, no standardized method has yet been developed to identify leader and follower roles as they emerge naturally during an interaction. In this study, we link the results of an alignment method based on multi-dimensional network framework with behavioral data collected from 32 neurotypical volunteers during a Joint Action task in an EEG-based hyperscanning setting. By analyzing multiple-subject networks with multi-dimensional graph theory, we quantify differences in inter-brain connectivity and derive a measure, the inter imbalance index, that reflects role asymmetries within each dyad. These asymmetries were characterized as indicative of a leader-follower relation by means of a statistical comparison with a behavioral measure of leadership tendency. Although future work will be needed to further characterize the subjects’ roles from behavioral measure, our results shows how the proposed index effectively realigns subjects according to their roles.