Table 1. Design table.

Question	Hypothesis	Sampling plan (e.g., power analysis)	Analysis plan	Interpretation given to different outcomes	Observed outcome
(1) Does DA signaling during the experience of positive treatment effects modulate PA?	H1. Pharmacologic manipulation of DA signaling during the experience of pain relief associated with the placebo treatment in the conditioning session modulates PA.	Sample size estimation was calculated to meet requirements for sufficient power in frequentist analysis. At an expected effect size of pharmacological manipulation (Cohen’s d of 0.3), the total sample size recommended for rmANOVA in G*Power with an alpha error probability of 0.05 and a power of 1-ß = 0.9 is n = 144. Considering a dropout rate of approximately 10%, we aimed to include a total of N = 165 subjects (55 per group). Bayesian Inference was conducted using the fixed-n-approach of N = 165 as determined by frequentist sample size estimation. This number is deemed sufficiently large to generate a meaningful Bayes Factor to inform about the strength of evidence for and against our hypothesis.	Frequentist analysis:  Two-way rmANOVA with medication group (sulpiride vs. L-dopa vs. inactive pill) and experimental condition (placebo vs. control) to compare pain ratings on day 2. Bayesian Analysis: Bayesian repeated-measures ANOVA comparing models with medication group (sulpiride vs. L-dopa vs. inactive pill) and experimental condition (placebo vs. control) to define the evidence for or against the inclusion of model terms to model pain ratings on day 2 (BFincl). JASP’s default priors were used (Cauchy distribution with r scale for fixed effects = 0.5 and r scale for random effects = 1).	Frequentist analysis: An F-test determined significance of variance explained by the specified factors in the ANOVA. For the interaction effect:   - p > 0.05: reject H1   - p < 0.05: reject H0.  Post hoc Bonferroni–Holm corrections were applied for multiplicity-adjusted pairwise comparisons if F-statistic yields significant results. Effect sizes were quantified using partial eta squared ηp2:   - ηp2 = 0.01: small effect   - ηp2 = 0.06: medium effect   - ηp2 = 0.14: large effect  Bayesian Analysis: BFincl   - >100: Extreme evidence for inclusion/H1.   - 30–100: Very strong evidence for inclusion/H1   - 10–30: Strong evidence for inclusion/H1   - 3–10: Moderate evidence for inclusion/H1   - 1–3: Anecdotal evidence for inclusion/H1   - 1/3–1: Anecdotal evidence against inclusion/for H0   - 1/10–1/3: Moderate evidence against inclusion/for H0   - 1/30–1/10: Strong evidence against inclusion/for H0   - 1/100–1/30: Very strong evidence against inclusion/for H0   - <1/100: Extreme evidence against inclusion/for H0	Frequentist analysis:  Significant main effect of experimental condition (F(1,151) = 8.29, p = 0.004, ηp2 = 0.05) No main effect of medication group (F(2,151) = 0.64, p = 0.53, ηp2 = 0.01) No interaction between medication and experimental condition (F(2,151) = 0.35, p = 0.71, ηp2 < 0.01).   - p > 0.05: H1 is disconfirmed  Bayesian analysis: The BFincl of experimental condition is 4.32. The BFincl of medication group is 0.17. The BFincl of the interaction between experimental condition and medication group is 0.06.   - BF = 1/30–1/10: Strong evidence against inclusion/for H0
(2) Does DA signaling during the experience of pain relief influence the persistence of PA?	H2. Pharmacologic manipulation of DA signaling during conditioning affects the persistence of PA up to 6 days after test session 1 (day 8).	The sample size for this analysis was determined by the analysis of H1.	Frequentist analysis:  Two-way rmANOVA with medication group (sulpiride vs. L-dopa vs. inactive pill) and experimental condition (placebo vs. control) to compare pain ratings on day 8. Bayesian analysis: Bayesian repeated-measures ANOVA comparing models with medication group (sulpiride vs. L-dopa vs. inactive pill) and experimental condition (placebo vs. control) to define the evidence for or against the inclusion of model terms to model pain ratings on day 8 (BFincl). JASP’s default priors were used (Cauchy distribution with r scale for fixed effects = 0.5 and r scale for random effects = 1).	The interpretation of different outcomes relies on the same specifications as shown for H1.	Trend for effect of experimental condition (F(1,150) = 3.19, p = 0.08, ηp2 = 0.02) No main effect of medication group (F(2,150) = 0.33, p = 0.71, ηp2< 0.01). No interaction between medication and experimental condition (F(2,151) = 1.05, p = 0.35, ηp2 = 0.01).   - p > 0.05: H2 is disconfirmed  Bayesian analysis: The BFincl of experimental condition is 0.38. The BFincl of medication group is 0.11. The BFincl of the interaction between experimental condition and medication group is 0.04   - BF = 1/30–1/10: Strong evidence against inclusion/for H0
(3) Does DA signaling during the experience of pain relief alter treatment expectation?	H3. DA-manipulation during conditioning differentially alters the establishment of treatment expectation.	The sample size for this analysis was determined by the analysis of H1.	Frequentist analysis:  A two-way rmANOVA with medication group (sulpiride vs. L-dopa vs. inactive pill) and rating time point (pre conditioning vs. post conditioning) to compare the EXPECT scores between groups. Bayesian analysis: Bayesian repeated-measures ANOVA comparing models with medication group (sulpiride vs. L-dopa vs. inactive pill) and rating time point (pre conditioning vs. post conditioning) to define the evidence for or against the inclusion of model terms to model EXPECT scores (BFincl). JASP’s default priors were used (Cauchy distribution with r scale for fixed effects = 0.5 and r scale for random effects = 1).	The interpretation of different outcomes relies on the same specifications as shown for H1.	Significant main effect of experimental condition (F(1,149) = 20.11, p < 0.001, ηp2 = 0.12). No main effect of medication group (F(2,149) = 1.44, p = 0.24, ηp2 = 0.02). No interaction between medication and experimental condition (F(2,149) = 0.38, p = 0.68, ηp2 < 0.01).   - p > 0.05: H3 is disconfirmed  Bayesian analysis: The BFincl of rating time point is 512.31. The BFincl of medication group is 0.19. The BFincl of the interaction between rating time point and medication group is 0.07   - BF = 1/30–1/10: Strong evidence against inclusion/for H0