Extended Data Fig. 9: Hierarchical Bayesian models.
From: Individual risk attitudes arise from noise in neurocognitive magnitude representations
Graphical representations of the hierarchical Bayesian (a) noisy logarithmic coding model and (b) mediation analysis. Clear circles represent latent variables while filled circles are observed variables, such as trialwise choice (rcsi) data, subject-wise behavioural and neural measurements (yj), and numerosity/payoff inputs (X,C). See Supplementary Note for more details.
