The two-brain approach: how can mutually interacting brains teach us something about social interaction?

TLDR

Simultaneous two‑person brain imaging is appealing but methodologically challenging, with only a few fMRI, EEG, or NIRS studies to date. The review examines two primary two‑brain methodological approaches and proposes new analytic tools for studying interpersonal neural effects. It contrasts fMRI studies of turn‑based or pseudo‑interactive paradigms with dual EEG/NIRS studies of face‑to‑face or continuous interactions that assess functional connectivity and neural coupling. These approaches may enhance understanding of social cognition as both representation and participation.

Abstract

Measuring brain activity simultaneously from two people interacting is intuitively appealing if one is interested in putative neural markers of social interaction. However, given the complex nature of interactions, it has proven difficult to carry out two-person brain imaging experiments in a methodologically feasible and conceptually relevant way. Only a small number of recent studies have put this into practice, using fMRI, EEG, or NIRS. Here, we review two main two-brain methodological approaches, each with two conceptual strategies. The first group has employed two-brain fMRI recordings, studying (1) turn-based interactions on the order of seconds, or (2) pseudo-interactive scenarios, where only one person is scanned at a time, investigating the flow of information between brains. The second group of studies has recorded dual EEG/NIRS from two people interacting, in (1) face-to-face turn-based interactions, investigating functional connectivity between theory-of-mind regions of interacting partners, or in (2) continuous mutual interactions on millisecond timescales, to measure coupling between the activity in one person's brain and the activity in the other's brain. We discuss the questions these approaches have addressed, and consider scenarios when simultaneous two-brain recordings are needed. Furthermore, we suggest that (1) quantification of inter-personal neural effects via measures of emergence, and (2) multivariate decoding models that generalize source-specific features of interaction, may provide novel tools to study brains in interaction. This may allow for a better understanding of social cognition as both representation and participation.