Table 2. Choice behavior in the Wheels of Fortune task for MS patients and healthy controls (regression analyses, panel logit with individual random effect).

	Healthy controls (n = 38)				MS patients (n = 72)
A
choice	Coeff	Std Err	z	p	Coeff	Std Err	z	p
EV	0.0338	0.0018	18.49	<0.001	0.0271	0.0012	22.26	<0.001
d	−0.0036	0.0008	−4.32	<0.001	0.0004	0.0006	0.67	.502
r	0.0035	0.0006	5.96	<0.001	0.0054	0.0004	12.19	<0.001
cst	0.3300	0.0763	4.32	<0.001	0.1954	0.0520	3.75	<0.001
	Log likelihood = −673.57; Wald chi2(2) = 552.77; prob>chi2 = 0.0000				Log likelihood = −1298.23; Wald chi2(2) = 1100.22; prob>chi2 = 0.0000
B
choice	Coeff	Std Err	z	p	Coeff	Std Err	z	p
EV	0.0297	0.0013	23.39	<0.001	0.0284	0.0009	32.33	<0.001
risk	0.0013	0.0008	1.57	0.116	−0.0029	0.0006	−4.58	<0.001
cst	0.1876	0.0678	2.77	0.006	0.0391	0.0480	0.81	0.415
	Log likelihood = −729.01; Wald chi2(2) = 555.41; prob>chi2 = 0.0000				Log likelihood = −1397.32; Wald chi2(2) = 1109.13; prob>chi2 = 0.0000

A. Model of choice integrating the effects of anticipating disappointment (d) and regret (r) in addition to the maximization of expected values (EV). Both MS patients and controls chose anticipating r and maximizing EV. MS patients did not choose anticipating d.

B. Model of choice integrating the effects of risk in addition to the maximization of expected values (e). Healthy controls were risk neutral while MS patients were risk averse.