Table 2.

Univariate and multivariate standardized coefficients (95% CIs) predicting the expected-value signal in vmPFC

Dependent: Q in vmPFC	Coefficient (univariate)	Coefficient (multivariate)
Age	−0.32 (−0.61 to −0.03, P = 0.030)	0.30 (−0.17 to 0.76, P = 0.211)
D1-R in NAcc	0.41 (0.12 to 0.69, P = 0.006)	0.41 (−0.00 to 0.82, P = 0.052)
vmPFC-accumbens FA	0.48 (0.22 to 0.75, P = 0.001)	0.49 (0.15 to 0.83, P = 0.006)

Both D1-R BP_ND and FA in the connection between NAcc and vmPFC independently predicted the strength of this value signal. It should be noted that we present these effects as independent here, based on the previously reported significant relationship between BP_ND in NAcc and the expected-value signal in vmPFC. However, this relationship exceeded the threshold for significance when entered into the regression with age and FA in NAcc-vmPFC. Univariate coefficients represent the result of bivariate correlations, whereas multivariate coefficients represent multiple regression coefficients of a model including all predictors.