Dissociation of BOLD responses to reward prediction errors and reward receipt by a model comparison

Data S1. Materials and methods

Table S1. Scanning parameters in the three study parts

Table S2. Effects of the different reward contrasts in the ROIs, separately computed for the three samples

Table S3. Residual effects of RR and RPE in the ROIs

Fig. S1. Mean (± 1 SEM) parameter estimates of the RPE and the RR regressors in MB, VS and MOFC in each of the three samples. Voxel-wise parameter estimates are averaged across all voxels of an anatomically predefined ROI and across Paradigm A and B. Parameter estimates of the RPE and RR regressors represent common and regressor-specific effects. Significant differences (P < 0.05 in a paired t test) between RPE and RR within a ROI are marked with an asterisk. The interaction effects of the factors ROI and RPE/RR model in a repeated measures ANOVA were significant in each sample: (A) Sample 1: F_3,57 = 6.910, P < 0.001, partial η² = 0.267 (B) Sample 2: F_1.46,20.37 = 5.587, P = 0.018, partial η² = 0.285, Greenhouse-Geisser corrected. (C) Sample 3: F_1.86,42.85 = 20.893, P < 0.001, partial η² = 0.476, Greenhouse-Geisser corrected.

Fig. S2. (A) Mean (± 1 SEM) parameter estimates of the orthogonalized RPE and RR regressors representing residual regressor-specific effects in MB, VS and MOFC. Voxel-wise parameter estimates are averaged across all voxels of an anatomically predefined ROI and across Paradigm A and B. The interaction effect of the factors ROI and RPE/RR model in a repeated measures ANOVA was significant, F_2.39,138.41 = 27.545, P < 0.001, partial η² = 0.322, Greenhouse-Geisser corrected. Significance of the parameter estimates in one-sample t tests against zero (P < 0.05) is indicated by an asterisk. (B) Mean logarithmic residual variance (± 1 SEM) of the GLMs containing either the RPE or the RR regressor, averaged across all voxels of an anatomically predefined ROI and across Paradigm A and B. Additionally, the main effect of each ROI has been subtracted to increase the presentability of the subtle differences between both models. Smaller logarithmic residual variances represent superior model fit. The interaction effect of the factors ROI and RPE/RR model in a repeated measures ANOVA was significant, F_2.51,145.47 = 5.202, P = 0.003, partial η² = 0.082, Greenhouse-Geisser corrected. Significant differences (P < 0.05 in a paired t test) between logarithmic residual variances of RPE and RR within a ROI are marked with an asterisk.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset by Wiley-Blackwell. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Filename	Format	Size	Description
EJN_8125_sm_FigS1-S2andTableS1-S3andMethods_1.doc		409K	Supporting info item

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.