A large international study led by Laurens van de Mortel and the ENIGMA-OCD working group investigated the possibility of predicting which patients with obsessive–compulsive disorder (OCD) benefit from cognitive behavioural therapy. Using machine learning, the researchers developed predictive models on both clinical information and brain imaging data from across the globe with the aim to move toward more personalised treatment strategies.
Findings showed that models based on routine clinical data were able to predict cognitive behavioural therapy outcome with about 69% accuracy. This suggests that information gathered in standard psychiatric assessments can provide guidance on how likely a patient is to benefit from cognitive behavioural therapy across a diverse world-wide patient population. In contrast, models trained on brain activity data alone performed only slightly better than chance. Moreover, combining brain data with clinical data did not improve predictions as compared to clinical data alone. These results suggest that, at least for now, brain scans do not add meaningful value in predicting treatment outcome and that standard clinical data remains the most reliable tool for predicting therapy outcomes in OCD.
These findings highlight the potential usefulness of predictive models and routine clinical data for guiding treatment decisions. The use of costly and complex neuroimaging data currently seems to add little predictive value in clinical care. Future research should investigate whether other types of (neuro)biological data offer more value, and aim to develop large and more harmonised datasets for the prediction of other treatments for OCD.
Laurens A. van de Mortel, Willem B. Bruin, Pino Alonso, Sara Bertolín, Jamie D. Feusner, Joyce Guo, Kristen Hagen, Bjarne Hansen, Anders Lillevik Thorsen, Ignacio Martínez-Zalacaín, Jose M. Menchón, Erika L. Nurmi, Joseph O’Neill, John C. Piacentini, Eva Real, Cinto Segalàs, Carles Soriano-Mas, Sophia I. Thomopoulos, Dan J. Stein, Paul M. Thompson, Odile A. van den Heuvel, Guido A. van Wingen,
Development and validation of a machine learning model to predict cognitive behavioral therapy outcome in obsessive-compulsive disorder using clinical and neuroimaging data. Journal of Affective Disorders, Volume 389, 2025, 119729, ISSN 0165-0327, https://doi.org/10.1016/j.jad.2025.119729.