Date created: | Last Updated:

Category: Project

Description: Socially relevant tasks reliably elicit activity in the theory of mind network (ToMN), but the computations implemented within these regions have remained unclear. Following recent suggestions that ToMN activity tracks prediction error in social contexts, we examined the by-stimuli relationship between ToMN activity and dispositional, prescriptive, and descriptive prediction error, all of which have been identified as critical sources of information in classic social psychological research. Participants read moral narratives in which agents made moral (or immoral) decisions. Additional information reframed these narratives, reversing participants’ initial moral judgments. Study 1 normed these stimuli in an online sample (N = 554), identifying underlying dimensions of dispositional, prescriptive, and descriptive prediction error using principal components analysis. Study 2 compared by-stimuli prediction error with ToMN activity as participants (N=20) read the same scenarios inside the scanner. Multilevel models provided estimates of by-stimuli estimates of behavioral ratings in Study 1, and ToMN activity in Study 2, allowing the samples to be compared. ToMN activity was related to dispositional and prescriptive prediction error, but not to descriptive prediction error. Critically, significant relationships were observed only when morally relevant information was presented (as opposed to morally irrelevant, control information). Thus, the present work provides evidence that the ToMN encodes prediction error in social contexts.