To the best of our knowledge, this is the first report on IGT performance in children and adolescents with OCD. The observed poor IGT performance suggests impaired decision-making function in OCD patients. Previous studies have described poor IGT performance in adult OCD patients [8–11] and one study demonstrated impaired IGT performance in patients with prominent hoarding symptoms , both of which are in line with our current findings regarding childhood-onset OCD. However, Nielen et al.  did not find significant differences in IGT performance between OCD patients and healthy volunteers.
Electrophysiological recordings of the OFC have revealed that orbitofrontal neurons play an important role in processing the motivational value of rewarding outcomes [30–32]. This fits very well with the fact that IGT performance is specifically sensitive to OFC activity. In the current study, children with OCD performed similar to controls in the early stages of the IGT, but selected more disadvantageous cards towards the end of the task. The persistent selection of cards from disadvantageous decks in IGT is a behavior also observed in patients with damage to the OFC 
. Taken together, OFC dysfunction may be associated with the pathophysiology of childhood OCD.
The present findings that IGT performance correlated significantly with the CY-BOCS score during the period with the most severe symptoms but not with the CY-BOCS score at the time of the assessment suggest that IGT performance may be associated with the potential severity of OCD rather than the state of present symptom exacerbation. We infer that the poor performance on IGT in antidepressant users may be because antidepressants were more likely to be chosen as a treatment for those with severe symptoms during the initial clinical assessment. However, an alternative explanation may be that the negative effects of antidepressants on cognitive performance result in poor IGT performance. Further studies are needed to elucidate what factors influence IGT performance in children with OCD.
The current study showed no significant differences in performance on the WCST between OCD patients and controls. WCST performance is specifically related to function of the DLPFC , a region that has no direct influence on the orbitofrontal-striatal-thalamic circuit. Therefore, it is plausible that WCST performance is not as impaired as IGT performance in OCD.
The only previous study which examined WCST performance in children with OCD  reported reduced WCST scores in children with OCD compared to healthy controls, with effect sizes of d = 1.24 for total errors and d = 1.30 for categories completed. Our sample size had 97% power to detect an effect size of 1.2 at an alpha level of 0.05 (calculated by G*Power 3.1.3 ). Therefore, our results indicate that any potential difference in WCST performance between OCD patients and controls in the current study was smaller than that derived by Shin et al. . We infer that the current inconsistency with the results of Shin et al. may be because patients and controls were not matched for IQ in their study.
There are numerous studies on WCST performance in adult OCD patients; however, results are inconsistent across studies. Several groups reported a reduced performance in OCD patients [9, 36–38], while others reported no differences compared with healthy controls [39–42]. The inconsistencies across these studies may be attributable, in part, to the comorbidity of PDD or ADHD in some of the OCD patients. Children with high-functioning autistic disorder or ADHD display poor performance on the WCST . To better understand and clarify these findings, any comorbidity with PDD or ADHD, the age of onset and duration of OCD, the intelligence of the subjects, and the age of assessment should all be taken into account.
There are several limitations to the current study. First of all, the number of participants was too small to draw definitive conclusions. Secondly, some children with OCD were on medication, which may have affected the results. Finally, we did not use neuroimaging methods such as fMRI to confirm our inference that the function of the OFC was associated with the results obtained in our study.
In summary, children with OCD displayed poor performance on IGT but did not differ on WCST scores compared to controls. These findings are in line with the hypothesis that OCD is associated with dysfunction of orbitofrontal-striatal-thalamic circuitry. However, large-scale studies combined with fMRI are needed to confirm the role of the orbitofrontal cortex in the pathophysiology of childhood OCD.