Inter-Brain Neural Correlates of Self–Other Integration in Joint Statistical Learning

While statistical learning has often been investigated in an individual context, it remains unclear whether humans are able to integrate information from both the self and from another to build a collective representation of regularities. Here, we investigated the dynamic self–other integration process and its multi-brain mechanism by recording EEG activity simultaneously from dyads. Participants (N = 112) each responded repeatedly to one half of a fixed stimulus sequence either with an active partner (i.e., joint context) or with a passive observer (i.e., baseline context). At individual level, we found that a significant statistical learning effect in the joint context characterized by decreased trends in reaction time (RT) and intra-brain neural responses (e.g., ERPs and functional connectivities) as well as a subsequent modulation by an insertion of an interference sequence. At dyad level, Brain-to-Brain Coupling (BtBC) in the theta band first showed an increasing trend followed by a subsequent modulation, providing direct neural evidence for the occurrence of a dynamic self–other integration process. Critically, the strength of BtBC was negatively correlated with RT and positively correlated with intra-brain functional connectivities. These findings suggest that BtBC serves as a crucial neural correlate of self–other integration underpinning the joint statistical learning effect, and that statistical regularity can both implicitly and spontaneously modulate the occurrence of the self–other integration process.