Table 2.

Integrated Bayesian information criteria (iBIC) for tested models

Model no	Model parameters	No. of parameters	Likelihood	Pseudo-R2	iBIC
1	ε, ρ	2	− 23,463	0.498	46,970
2	ε, ρ, ξ	3	− 23,314	0.501	46,695
3	ε, ρ, ξ, b	4	− 21,798	0.534	43,685
4	ε, ρwin, ρlose, ξ, b	5	− 21,334	0.544	42,779
5	ε, ρwin, ρlose, ξ, b, πvariable	6	− 21,137	0.548	42,406
6	ε, ρwin, ρlose, ξ, b, πconstant	6	− 21,106	0.549	42,346

Boldface type: winning model statistics, ε: learning rate, ρ_win: weighting of reward on win trials, ρ_lose: weighting of punishments on lose trials. ξ: irreducible noise, b: go bias, π: Pavlovian bias, iBIC: integrated Bayesian information criterion (smaller iBIC values indicate a better model fit). Descriptives for the parameters in the winning model (M ± SD): ε = 0.26 ± 0.15, ρ_win = 15.32 ± 13.30, ρ_lose = 7.51 ± 4.03, ξ = 0.96 ± 0.06, b = 1.10 ± 0.74, π_constant = 0.65 ± 0.57