Effects of excessive substance use on inhibitory control in patients with attention-deficit/hyperactivity disorder

Background: Patients with Attention-Deficit/Hyperactivity Disorder (ADHD) often develop comorbid Substance Use Disorder (SUD) [1]. Although literature indicates that both disorders are associated with abnormalities related to inhibitory control [2,3], combined ADHD and SUD effects in patients are yet to be examined. Objective: To investigate the effects of substance abuse on inhibitory control in patients with ADHD. Methods: Our sample consisted of ADHD groups with and without substance abuse (ADHD+SUD and SUD-only, respectively) and healthy controls (n=34 per group). Groups were matched with family history of SUD to eliminate SUD trait effects. Behavioral and functional MRI (fMRI) data in Goand Stop-trials (successful and failed) during a Stop-Signal task were collected. Behavioral data analysis consisted of one-way ANCOVAs for reaction times (RTs), intra-individual coefficient of variation (ICV) and stop-signal reaction times (SSRTs). The analysis of the fMRI data included preprocessing with ICA-AROMA and first-level analysis with a fixed-effects model (FSL FEAT). Group differences were examined with a mixed-effects model with 5000 random permutations (FSL randomize). Beta-values from significant clusters (t>2.3 & p<.05) were extracted and post-hoc comparisons with Bonferroni-Holm correction were performed. Results: Behavioral results showed that the ADHD groups demonstrated longer RTs and increased ICV than controls during Go-trials (RTs: p=.003 for ADHD-only, p<.001 for ADHD+SUD; ICV: p<.001 for ADHD-only, p<.001 for ADHD+SUD). ADHD+SUD also showed longer SSRTs than controls (p=.008) during Stop-trials. The fMRI data analysis revealed that, (i) during Go-trials, ADHD+SUD showed hypoactivation of OFC compared to ADHD-only (p<.001) and controls (p=.01), while ADHD-only showed hyperactivation of a cluster including cingulate and paracingulate gyri compared to the other two groups (p=.004 for ADHD+SUD, p<.001 for controls). (ii) During successful Stop-trials, ADHD+SUD showed decreased activation of paracingulate gyrus compared to ADHD-only (p=.014) and controls (p<.001). Moreover, ADHD-only activated the same cluster less than controls (p=.042). (iii) During failed Stop-trials, controls showed decreased activation of a cluster containing caudate nucleus and putamen (p<.001for ADHD-only, p<.001 for ADHD+SUD). Additionally, we found a cluster located in Angular, Supramarginal and MTG with decreased activation for the ADHD groups compared to controls (p<.001 for ADHD-only, p=.014 for ADHD+SUD) and further decreased activation for ADHD-only compared to ADHD+SUD (p=.006). Conclusion: During Go-trials, the ADHD groups showed worse behavioral performance than controls. Despite similar performance to each other, the patient groups were characterized by distinct neural activation patterns. ADHD+SUD showed neural deficits in cognitive control network that might be associated with the poor performance observed, while ADHD-only seemed to make use of the same network during Go-trials. We speculate that ADHDonly might rely on proactive control, partially compensating for deficits in reactive cognitive control. Additional analysis for this is needed. During successful Stop-trials, both ADHD groups showed deficits in cognitive control network, with more pronounced deficits for ADHD+SUD. This suggests a severity pattern for the ADHD groups, with greater impairments potentially associated with comorbid substance abuse. Finally, failed inhibitory control seems to be linked to suboptimal function of fronto-striatal network in controls, but different abnormalities seem to account for disinhibition in the other groups.