Beyond one Bayesian brain: Modeling intra- and inter-personal processes during social interaction: Commentary on “Mentalizing homeostasis: The social origins of interoceptive inference” by Fotopoulou & Tsakiris

In their thought-provoking and integrative target article, Fotopoulou and Tsakiris cut across different bodies of literature to argue for a second-person account of interoception and social cognition. More specifically, they argue for the constitutive role of embodied social interactions in the development of interoceptive abilities and the ability for self-other differentiation. Furthermore, they review evidence to suggest that social interactions have a specific role in binding together subjective feeling states with the perception of the body and the (social) world. Along this line of reasoning, they refer to predictive coding and active inference frameworks of the “Bayesian brain” function to suggest that basic inferential processes of embodied perception and action may be turned into more advanced forms of social understanding. We applaud Fotopoulou and Tsakiris for advocating a second-person account of social cognition and for connecting the predictive coding account of interoception to the dynamics of social interaction conceptually. We discuss these contributions in light of the existing literature and encourage the authors to be more precise about the computational processes, which they suggest may connect social interaction and interoception at the individual level. Furthermore, we describe a multilevel Bayesian framework that could be used to formally test a proposal, such as the one discussed by the authors, but also allows for going beyond one Bayesian brain, by modeling interpersonal processes during social interaction.