In SDI, drug dependence was measured using the computerized Composite International Diagnostic Interview (CIDI)-Substance Abuse Module (SAM) (CIDI-SAM) (17). CIDI-SAM is a structured interview designed for trained, lay interviewers, and has been shown to have good test-retest and inter-rater reliability(18). For each drug, symptom count and date of last use were recorded. CIDI-SAM was not given to controls. Performance data on the modified gambling task were available for 34 (15 controls, 19 SDI) of the 39 subjects. We used a modification of the Iowa Gambling Task (IGT) adapted for an fMRI experiment (16). Details of the task have been previously described (15). There were 80 trials for which the subject chose “Play or Pass” and these were divided into 2 time blocks, early and late. The number of times an individual chose to play “bad” decks on early compared to late trials was totaled. Repeated measures analysis of variance (rmANOVA) using IQ, education, and age as covariates was performed in SPSS analyzing for effects of group by time interaction.
