, 2009 and Rushworth Bortezomib et al., 2011). Theoretical models predict that precommitment arises as a function of learning about one’s own self-control abilities (Kurth-Nelson and Redish, 2010, Kurth-Nelson and Redish, 2012 and Ali, 2011). In the current study,

we were able to show that between-subject differences in self-control abilities moderated precommitment-related neural activity. Future work might examine the within-subject dynamics of learning about one’s own self-control abilities and how such learning relates to precommitment. For example, one might dynamically manipulate the difficulty of resisting temptations (thus making precommitment more valuable at some times than others) and examine how activation in LFPC and its connectivity with willpower regions tracks with the expected value of precommitment on a trial-to-trial basis. The LFPC may be involved in such learning processes, given its role in self-awareness and metacognition (Fleming et al., 2010 and De Martino et al., 2013). Although the anterior prefrontal cortex (BA 10) is cytoarchitechtonically homogeneous, it may be functionally heterogeneous (Gilbert et al., 2006 and Liu et al., 2013); for instance, studies of metacognition (Fleming et al., 2010 and De Martino et al., 2013) find more have reported activations in anterior prefrontal cortex

that are situated dorsal and medial to those reported in studies of counterfactual value processing (Boorman et al., Mephenoxalone 2009 and Boorman et al., 2011). A recent study of connectivity patterns within FPC found

that the lateral FPC (FPCl) showed strongest connectivity to DLPFC, while the orbital FPC (FPCo) showed strongest connectivity to the OFC and subgenual ACC (Liu et al., 2013). Notably, the region we found to be associated with precommitment is located precisely in the transition zone between FPCl and FPCo. This region is therefore ideally situated to arbitrate between regions involved in calculating expected value (OFC, subgenual ACC) and regions involved in implementing self-control (DLPFC). Fitting with this notion, we observed that LFPC was functionally connected to DLPFC during precommitment and that the strength of this connectivity was moderated by activation in the vmPFC. Precommitment decisions in the real world often involve longer delays (in the order of weeks to months), in contrast with the shorter delays used in the current study. Future studies might examine whether the precommitment to large rewards with much longer delays engage similar neural processes as those described in the current study. Given the role of the LFPC in forward planning (Daw et al., 2006, Burgess et al., 2007, Koechlin and Hyafil, 2007, Boorman et al., 2009, Boorman et al., 2011, Rushworth et al., 2011 and Tsujimoto et al., 2011), we might expect to see even stronger effects in LFPC with longer delays than in our current design, in which the shorter delays placed relatively low demands on prospective cognition.