, 2008; Frith and Frith, 1999, 2006; Hampton et al., 2008; Saxe, 2006). While vmPFC responses to valuation and goal-directed choice are the subject of several computational theories (Boorman et al., 2009; Hare et al., 2011; Hunt et al., 2012; Levy and Glimcher, 2011; Lim et al., 2011),

only scant attention has been given to computational mechanisms underlying dmPFC social responses (Behrens et al., 2008; Hampton et al., 2008; Yoshida et al., 2010). One possibility is that the ability to impute the intentions (Frith and Frith, 2006) and predict the actions (Behrens et al., 2008) of others derives from the same mechanisms that allow us to reflect on our own goals and actions (Amodio and Frith, 2006; Buckner and Carroll, 2007; Mitchell, 2009). Such a theory is appealing, as it would allow a computational understanding of goal-directed choice to be extended to social inferences. Ruxolitinib However, this idea is difficult to reconcile with the existence of brain regions that appear specialized for processing self and other. Instead, it raises the intriguing possibility that a functional specialization in rostromedial prefrontal cortex (and in related temporoparietal cortex; Mitchell, 2008; Saxe and Wexler, 2005) is driven more by differences

Dasatinib molecular weight between choices that are executed versus those that are imagined or modeled. Teasing these two possible functional architectures apart is problematic as they are almost always aligned in cognitive tasks, where self-choices tend to be executed and others’ actions and intentions modeled. Here, we describe neural signals that compute the choice preferences of another individual, whether or not they are relevant to the current choice. These signals precisely mirror well-studied signals reflecting

personal choice preferences. Furthermore, by designing situations in which values of self and other may be either modeled or executed, we show that the functional gradient in the medial prefrontal cortex does not align with the individual, but is dependent on whether choices are executed by the subject or instead are modeled without overt execution. not To examine neural computations for self and others in medial prefrontal cortex, we designed a delegated intertemporal decision-making task (Figure 1A). Subjects chose between a large monetary reward delivered later, and a small reward delivered sooner. Critically, we asked subjects to choose for themselves in some trials, but in other trials to choose on the basis of what a partner participant would have chosen in the same context (Figure 1A). It is known that different individuals display significant variability in their preferences, with “low-discounters” preferring to wait for a later higher-value option and “high-discounters” preferring the more immediate smaller reward (Kable and Glimcher, 2007).