GM volume is uncorrelated with preferences for altruism in the domain of disadvantageous inequality α (p = 0.551, small volume [SV] FWE corrected) or with preferences for positive reciprocity θ (p = 0.581, SV FWE corrected) or negative reciprocity δ (p = 0.629, SV FWE corrected). Finally, note that all our results are robust to the exclusion of the participant with extreme values of β and α (top left data point in Figure 2). When we repeat the analyses without the data from this participant, our main findings remain the same: using the independent ROI specified

above, β correlates significantly (r = 0.57, p = 0.0013) with TPJ GM volume, while all other parameters do not (p > 0.10). These findings suggest that GM volume in TPJ may be a crucial neuroanatomical basis for subjects’ baseline willingness to behave altruistically because BTK inhibitor the preference parameter β determines a subject’s generosity in the domain of advantageous inequality. This parameter determines, in particular, the maximal cost (denoted by w¯) a subject is willing to bear to increase the partner’s payoff by a given amount (say by one unit). The higher β, the higher the subject’s maximum willingness to pay w¯ to increase the partner’s payoff by one unit (see Figure S2). Therefore, subjects with a high β are generally willing to consider behaving altruistically for a much larger

range Thiazovivin of costly altruistic actions than those with a low value of β. In other words, if the costs of an altruistic act are relatively high, a subject with a relatively high value of β is still willing to consider behaving altruistically, while a subject with a low Thymidine kinase value of β will behave selfishly in this situation. This means that w¯ represents a subject-specific cutoff value such that if the actual cost of the altruistic act is below w¯, the subject will consider making an altruistic choice, while the subject behaves selfishly if the actual cost is above w¯. This insight about the role of β (and the

implied role of w¯), together with the known functional role of the TPJ in perspective-taking tasks (Decety and Lamm, 2007, Frith and Frith, 2007, Saxe and Kanwisher, 2003 and Young et al., 2010), can help us establish a link between GM volume in the TPJ and functional activations in TPJ during decision making in our task (in which subjects faced many different cost levels across trials). A high value of β implies a high maximum willingness to pay w¯, meaning that the correlation between GM volume in right TPJ and β should translate into a correlation between GM volume and w¯ (see Figure 4A). In addition, taking the other individual’s perspective seems particularly necessary in those cost situations in which a subject is in principle willing to behave altruistically (i.e., when the actual cost is below w¯) but in which self-interest provides a strong obstacle for altruistic acts because the cost is close to w¯.