For all ROIs identified in the group level PPI analyses, an indirect effect of connectivity was tested with a bias-corrected bootstrapped mediation analysis using an SPSS macro (Preacher and Hayes, 2004). The dependent variable was DV, the independent variable was sensation-seeking score, and the mediator was connectivity with NAcc. As post-hoc analysis revealed no main effect of valence on connectivity (with the exception of right NAcc-right lingual gyrus; see Results), connectivity for reward and loss trials was linearly combined for this analysis. A point estimate of the indirect effect was derived from the mean of n=5000 estimates and 95% confidence intervals computed using the 2.5% highest and lowest scores of the empirical distribution. Indirect effects were considered significant when the bias-corrected and accelerated confidence interval did not include zero (Preacher and Hayes, 2004).
