Goal 1: Monetary Reward Neural Effect
The SPM analyses of the monetary main effect (45¢ or 1¢ > 0¢) in all subjects revealed activations in 25 regions comprising 9 clusters which included the right and left lateral OFC, the lateral and ventromedial (including the anterior cingulate cortex) PFC, the mesencephalon, thalamus, and cerebellum (but also the occipital lobe), all bilaterally ( and ). However, consistent with our first a priori hypothesis, direct group analyses revealed that the activations in the comparison subjects but not the cocaine abusers were driving these results ( and ).
Figure 1 SPM results of the monetary reward effect (45¢ or 1¢ > 0¢) in all subjects (A: comparison subjects, N=13, and cocaine abusers, N=16) and in direct group comparisons (B: comparison subjects > cocaine abusers). Statistical (more ...)
The complementary ROI analyses revealed a significant monetary main effect in six of these clusters (, clusters in boldface, and ). Further, a money by group interaction was significant in the left OFC (); indeed, an overall test of coincidence of the study groups’ regression lines was statistically significant (F=3.49(2,80), p < 0.05), indicating different lines of best fit (from lowest to highest monetary reward) in this ROI as a function of group. All other significant results are marked in and further described in Discussion.
Figure 2 Average BOLD signals in the ROIs (see and , regions in boldface) located at the left orbitofrontal cortex (A: OFC), prefrontal cortex (B: PFC, mean signal), right mesencephalon (C: MSN), and left cerebellum (D: CBL) as a function of monetary (more ...)
We examined the effect of age, urine status, and cigarette smoking by conducting correlations or t-tests with all 9 clusters’ responses to absolute or relative monetary reward (45¢, 1¢, or 0¢ > baseline, 45¢>0¢, and 1¢>0¢) as the dependent variables (3 covariates × 9 ROIs × 5 reward conditions = 135 analyses); even with a lenient Bonferroni correction (p<0.01), there were no significant correlations between age and any of these ROIs, nor did monetary responses in these ROIs differ as a function of urine status or cigarette smoking history, in each of the study groups or in the complete sample.
Goal 2: State Motivation
All three behavioral measures of state motivation (45¢ > 0¢) were significantly intercorrelated in the comparison subjects (variables 1–3 in , lower half) but not in the cocaine abusers (, upper half). In the cocaine subjects there was instead a correlation between the differential RT with MPQ self-control, such that the faster the RT to the higher monetary reward, the more the self-reported trait control. Again, age, urine status, or cigarette smoking did not affect these results.
Table 2 Correlations between selected dependent study variables. Differential (change) scores were calculated between the high monetary reward and the neutral cue (45¢ - 0¢) for all three state motivation (variables 1–3) and also for the (more ...)
Goal 3: Brain-Behavioral Associations
The differential signal change (45¢ > 0¢) in the lateral PFC correlated significantly with motivation at both the state (differential RT) and trait (MPQ achievement) levels and also with trait self-control but only in the cocaine abusers (variable 8 with variables 1, 4–5, , upper half). Voxel based correlations in the cocaine abusers confirmed the involvement of the lateral PFC in RT (sA), achievement (sB), and self-control (sC and , inserted map). Selected drug use variables (s) did not correlate with the differential BOLD response in the lateral PFC or with MPQ self-control; further, this correlation () remained unchanged after controlling with partial correlations for age, urine status, and smoking history (see ).
Figure 3 A: Correlation between the lateral PFC and inhibitory control in 16 cocaine abusers. Scatterplot shows association between the BOLD signal change for monetary reward as compared to the neutral cue (45¢ > 0¢) in the lateral PFC (more ...)
In the comparison subjects there was a significant correlation between the lateral PFC and the right OFC (45¢ > baseline for both) in both voxel based (whole brain) (sD and , inserted map) and ROI (, linear regression) analyses. These correlations were not significant in the cocaine abusers.