Multi-site classification within subgroups, split according to medication status, resulted in fair performance even after accounting for correlated clinical variables (i.e., disease duration) through additional splits. Evidence from rodent studies suggests that serotonin reuptake inhibitors (SRIs) mediate neuroplasticity in various cortical and subcortical structures through glio-genesis and neuro-genesis40–42. However, little is known about how these findings might translate to humans and what the effects of long-term medication use are43. The few longitudinal studies with small samples suggest that SRI treatment normalizes brain volumes. One study reported significantly larger thalamic volumes in twenty-one treatment-naïve pediatric patients compared to HC and that these differences decreased following paroxetine treatment44. Another study reported that smaller putamen volume in treatment-naïve patients was no longer detectable in the thirteen patients that were treated with fluoxetine45. Nonetheless, it remains unclear whether these structural changes are related to medication use or to symptom improvement. In contrast to these normalizing effects of SRIs, the previous univariate meta-analyses and mega-analyses of the ENIGMA-OCD study found significantly thinner cortices in medicated adult OCD patients and smaller cortical surface area in medicated
