Table S1.

Mean parameter estimates and fit information for the three models, separately for each treatment group (SD in parentheses).

		Model 1	Model 2	Model 3
β	Placebo	0.095 (0.028)	0.137 (0.042)	0.197 (0.058)
	Reboxetine	0.105 (0.039)	0.152 (0.081)	0.245 (0.129)
	Citalopram	0.093 (0.035)	0.135 (0.053)	0.157 (0.061)
$\widehat{\lambda }$	Placebo	−	0.73	0.70
	Reboxetine	−	0.73	0.65
	Citalopram	−	0.85	0.84
$\widehat{\theta }$	Placebo	−	45.9	45.6
	Reboxetine	−	45.6	45.3
	Citalopram	−	49.7	49.5
$\widehat{\kappa }$	Placebo	0.93	1.07	−
	Reboxetine	1.03	1.17	−
	Citalopram	0.86	1.01	−
−LL	Placebo	4380	3789	3821
	Reboxetine	4415	3751	3780
	Citalopram	4349	3858	3901
BIC	Placebo	8913	7757	7804
	Reboxetine	8994	7691	7732
	Citalopram	8842	7885	7954
p	Placebo	<0.001	>0.999	<0.001
	Reboxetine	<0.001	>0.999	<0.001
	Citalopram	<0.001	>0.999	<0.001

Model 2 provided the best fit to the data.

Note: Model 1, mean pay-off estimation without decay; Model 2, mean pay-off estimation with decay; Model 3, pay-off distribution estimation with decay; -LL, negative log likelihood (smaller values indicate better fit); BIC, Bayesian information criterion; p, BIC model weight.