The neural correlates of theory of mind within interpersonal interactions

Abstract

Tasks that engage a theory of mind seem to activate a consistent set of brain areas. In this study, we sought to determine whether two different interactive tasks, both of which involve receiving consequential feedback from social partners that can be used to infer intent, similarly engaged the putative theory of mind neural network. Participants were scanned using fMRI as they played the Ultimatum Game (UG) and the Prisoner's Dilemma Game (PDG) with both alleged human and computer partners who were outside the scanner. We observed a remarkable degree of overlap in brain areas that activated to partner decisions in the two games, including commonly observed theory of mind areas, as well as several brain areas that have not been reported previously and may relate to immersion of participants in real social interactions that have personally meaningful consequences. Although computer partners elicited activation in some of the same areas activated by human partners, most of these activations were stronger for human partners.

Introduction

One of the distinctive attributes of human social cognition is our propensity to build models of other minds: to make inferences about the mental states of others. This ability has become widely known as theory of mind (Premack and Woodruff, 1978). Several neuroimaging studies have attempted to elucidate the neural substrates that support this distinctively human ability. At least four separate studies have asked participants to make inferences about the mental states of characters in stories or cartoons Brunet et al., 2000, Fletcher et al., 1995, Gallagher et al., 2000, Vogeley et al., 2001. Others have asked participants to infer mental states from expressions in photographs (Baron-Cohen et al., 1999), or to attribute mental states to animations of geometric shapes (Castelli et al., 2000). Collectively, these studies have implicated a consistent network of brain areas in theory of mind, including anterior paracingulate cortex, the posterior superior temporal sulcus at the temporo-parietal junction, and the temporal pole (Gallagher and Frith, 2003). Two more recent studies have probed the neural correlates of theory of mind in participants who are actually immersed in a social interaction with a partner who is outside the scanner. Both reported activation in anterior paracingulate cortex Gallagher et al., 2002, McCabe et al., 2001, but not posterior superior temporal sulcus or temporal pole.

As part of an effort to explore the “social brain” with interactive games, we scanned a group of participants using fMRI as they played two different social games, the Ultimatum (UG) and Prisoner's Dilemma (PDG) games, each of which assesses cooperative intent in a different way. In the Ultimatum Game, two players are asked to split a sum of money. One player proposes how the sum should be divided and the other player either accepts or rejects the offer. If the offer is accepted, it is divided as proposed. However, if it is rejected, both players receive nothing. Although the game theoretic prediction is for the proposer to offer the smallest possible amount based on the assumption that any monetary amount will be accepted by a rational responder, cooperative proposers will offer an even split. In the present study, scanned participants were always in the role of responder in the UG game. In contrast to the Ultimatum Game, the Prisoner's Dilemma Game confronts each of two players with the same decision: cooperate or defect. Each player is awarded a sum of money that depends upon the interaction of their two choices, such that individual earnings are maximized by defection but collective earnings are maximized by cooperation. In both tasks, participants witness a decision by their partner, an offer in the UG game or a choice in the PDG game, which reveals something about the partner's intentions. Are they generous or greedy in the UG game? Are they cooperative or selfish in the PDG game? To the extent that this feedback provokes inferences about the partners' intentions, it is expected to engage theory of mind neural systems.

Our goals in this study were threefold. First, to determine whether learning about the intentions of others activates the putative theory of mind neural network outlined above. Second, to assess the generality and reproducibility of these findings by comparing the neural correlates of two different tasks conducted within the same session that both involve learning about the mental states of others. Third, to utilize a task with high face validity in which participants are engaged in actual social interactions and compare our results with those of other studies in which participants were not similarly engaged. As noted above, this is not the first theory of mind imaging study to immerse participants in genuine social interactions. In both the Gallagher et al. (2002) and McCabe et al. (2001) studies, participants received feedback from human partners in interactive games. However, McCabe et al. did not focus their analysis on epochs in which partner decisions are revealed, and the Gallagher et al. (2002) PET study lacked the temporal resolution to distinguish epochs involving feedback from others. Here, we use fMRI to specifically focus on the blood oxygen level dependent (BOLD) response to receiving feedback from a partner that reveals something about the partner's intent.