Table 1. Model parameters.

Subject’s behavior on the first free choice of each session is described by 13 free parameters. Three of these parameters ($R_0$, ${\alpha }_1$ and ${\alpha }_{\mathrm{\infty }}$) describe the learning process and do not vary with horizon or uncertainty condition. Ten of these parameters ($A$, $B$ and $\sigma$ in the different horizon and information conditions) describe the decision process. All parameters are estimated for each subject in each stimulation condition and the key analysis asks whether parameters change between vertex and RFPC stimulation.

Parameter	Horizon dependent?	Uncertainty dependent?	TMS dependent?
prior mean, $R_0$	no	no	yes
initial learning rate, ${\alpha }_1$	no	no	yes
asymptotic learning rate, ${\alpha }_{\mathrm{\infty }}$	no	no	yes
information bonus, $A$	yes	n/a	yes
spatial bias, $B$	yes	yes	yes
decision noise, $\sigma$	yes	yes	yes