Fig. 5: Mood regulates the memory-enhancing effect of RPEs.
From: Positive affect modulates memory by regulating the influence of reward prediction errors
a Top: Model comparison of the three mixed-effects regression models showing that the model assessing the influence of factor 1 on memory fits the empirical data better than the model assessing the influence of factors 2 or 3. The gray triangle indicates the difference in WAIC scores. Bottom: Posterior distributions for fixed effects in mixed-effects model examining how transdiagnostic factor score f1 impacts memory. The shaded portion represents the 95% high-density (HDI) interval. The vertical line indicates a coefficient of 0. Posterior distributions that include 0 are shaded gray, while those that do not are shaded red, indicating a meaningful effect. b) Top: the relationship between βRPE (subject-level random effect from the mixed-effects model) and memory performance, organized by a tercile split (for visualization only) of positive affect (measured by factor 1 score). The asterisk indicates a significant interaction between factor score and βRPE in predicting memory performance (linear-regression coefficient = 1.4, p = 0.011). Bottom: the relationship between βPM (subject-level random effect from the mixed-effects model) and memory performance, arranged by a tercile split of positive affect (measured by factor 1 score). Dots denote values for individual participants. The solid line indicates linear model fit to participant data.
