Figure 5
From: Uncertainty of treatment efficacy moderates placebo effects on reinforcement learning
Learning rates per treatment condition across four compared models. Learning rates of each model, either for gain or general updating, are depicted on the y-axis. Purported probabilities of treatment efficacy are denoted on the x-axis. Bold dots represent group-level means, and error bars indicate the standard error. Models are sorted from lowest to highest LOOIC, from top left to bottom rights. LOOIC and its standard errors are shown below the respective figure cells. We compared an extended standard model (ES; top left) with a forgetting-RL model including a choice kernel component (RLf-CK; top right), the translated standard model with small priors using the hBayesDM R package (SSh; bottom left) as well as its variant with wide priors (SWh; bottom right).
