PPP1R21-Related neurodevelopmental disorder: Phenotypic delineation, variant spectrum, and pathophysiological mechanisms.

Cognitive-behavioral therapy (CBT) is currently the first-line treatment for internalizing disorders including obsessive-compulsive disorder. Although it has been shown to be highly effective for OCD treatment overall, up to half of patients do not experience remission from their disorder or even clinically relevant symptom relief. This inconsistent response to treatment has motivated research into predicting which patients will benefit from psychotherapy and who is at risk of non-response (for an overview of CBT outcome studies in psychiatric disorders, see). Researchers have tried to identify OCD patients at risk of non-response based on sets of demographic and clinical variables alone. Overall OCD symptom severity and individual items measuring OCD symptoms have allowed for prediction of treatment outcome. However, predictive variables beyond symptom severity, as well as predictive performance vary considerably across studies. Which clinical variables constitute robust and informative predictors for OCD treatment beyond symptom severity measures thus continues to pose an open question. In addition to clinical features, machine learning (ML) studies have tried to boost predictive performance by including neuroimaging data as a source of information on endophenotypes. Indeed, the presence of endophenotypes in OCD is supported by some electroencephalographic (EEG) studies (e.g.). Findings from functional magnetic resonance imaging (fMRI) suggest connectivity alterations in OCD (for a review see), including aberrant connectivity in default-mode and fronto-parietal networks. One study examined graph-theoretic measures of dynamic functional connectivity in OCD. These measures can concisely characterize brain connectivity profiles and provide global summary features as a sparse set of predictors. The study demonstrated higher modularity and impaired functional flexibility in the OCD patient group.