Control analyses for the influence of verbal IQ differences

As the ADHD groups differed in verbal IQ (WST) from controls we calculated ANCOVAs including verbal IQ as a covariate.

Differences between treated and untreated subjects

During the anticipation phase the main effect of condition was still significant in the gain condition (left: F(1,31)=19,82, FWE-corrected-p=.002, Tal: -21,8,-8); right: F(1,31)=19,98, FWE-corrected-p=.002, Tal: 15,5,-10) and trendwise in loss condition (F(1,31)=7,65, FWE-corrected-p=.073, Tal: -18,8,-11).

During gain anticipation, MPH subjects showed significantly reduced activity compared to controls in the BA 46 (T=4.33, p=.000; Tal: 42,38,4; ANCOVA: group effect: F(1,29)=9.38, uncorrected-p=.001, Tal: 42,38,4). However, there were no differences in the BA 45 between controls and drug-naïve subjects.

During loss anticipation, the effect of group in the middle frontal gyrus (BA 10) was not significant, but MPH-treated subjects with childhood ADHD showed again less activity compared to controls (T=3.61, p<.001; Tal: 33,35,7), but did not differ from drug-naïve ADHD subjects.

During outcome of loss-avoidance, controlling for IQ did not change our results in the insula (left: F(1,29)=11.95, uncorrected-p=.000, Tal: -36,11,-6); right: F(1,29)=10,53, uncorrected-p=.000, Tal: 36,8,-6). The activation in the right and left insula was significantly higher in subjects with MPH-treatment than in drug-naïve subjects (T>3.50, uncorrected-p=.001). After controlling for IQ, we found no effect in the precentral gyrus.

Differences between remitted and actual diagnosed subjects

After controlling for IQ we still revealed a significant effect of group during gain anticipation in the putamen (left: F(1,29)= 11.31, uncorrected-p=.000, Tal: -21,-2,11, k=16 ), indicating a higher neural response during gain anticipation in controls compared to subjects with adult ADHD (T=4.31, Tal: -21,-2,11, uncorrected p=.000), while remitted subjects with childhood ADHD did not differ from controls. Post-hoc t-tests revealed a higher neural response in the left putamen in remitted subjects compared to subjects with ongoing ADHD (T=3.81, uncorrected p=.000, Tal: -27,3,11).

Control analyses for the influence of drug abuse

Exploratively, we calculated ANOVAs using group as between-subject factor (Controls (n= 12) vs. childhood ADHD-drug-naïve (n=10) vs. childhood ADHD-MPH (n=8)) excluding subjects with a history of drug abuse.

Differences between treated and untreated subjects

During the anticipation phase the main effect of condition was still significant in the gain condition (left: F(1,27)=19,05, FWE-corrected-p=.003, Tal: -18,8,-8); right: F(1,27)=19,28, FWE-corrected-p=.002, Tal: 15,5,-10) and significant in the loss condition (left: F(1,27)=18,78, FWE-corrected-p=.003, Tal: -18,11,-6; right: F(1,27)=19,66; FWE-corrected-p=.002, Tal: 12,11,-11 ).

During gain anticipation, MPH-treated subjects showed significantly reduced activity compared to controls in the BA 45 (T=3.80, p=.000; Tal: -50,26,4); ANCOVA: group effect: F(1,27)=9.74, uncorrected-p=.000, Tal: -50,26,4). Even if not significant in the ANOVA, reduced activation in the BA 45 (T=4.04, p=.000, Tal: -50,29,4) in drug-naïve subjects compared to controls was observed.

During loss anticipation, the effect of group in the middle frontal gyrus (BA 10) was significant (F(1,27)=11.15, uncorrected-p=.000, Tal: 30,38,9, k=7), but in this analysis MPH-treated subjects with childhood ADHD not drug-naïve subjects, showed less activity compared to controls (T=4.65, uncorrected-p=.000, Tal:33,38,9) and drug-naïve subjects (T=3.50, p<.001; Tal: 30,38,12, k=2).

During outcome of loss-avoidance: After exclusion of subjects with drug abuse, we did not observe an effect in the insula and the precentral gyrus.

1