Figure 5.

Relative affective impacts of reward and performance predict vmPFC activity. A, top, BOLD activity in vmPFC was parametrically modulated by reward magnitude (peak: −3, 38, −1). Bottom, Bold activity in an overlapping region of vmPFC was modulated by trial-by-trial successful skilled performance (peak: −3, 50, −1). B, An independent vmPFC ROI shows modulation by both reward magnitude and skilled performance (both p < 0.01). C, In the independent vmPFC ROI, subjects with higher performance than reward weights in the computational analysis of affective dynamics displayed stronger neural responses in the vmPFC for performance than subjects with higher reward than performance weights (p = 0.003). D, The difference between performance and reward weights in the happiness computational model predicted the difference in vmPFC neural responses for successful skilled performance relative to reward magnitude (ρ(31) = 0.50, p = 0.003); *p < 0.05, **p < 0.01.