Predicting cognitive-behavioral therapy outcomes in obsessive-compulsive disorder from inhibitory control neural activity: A mega-analysis and machine learning study from the ENIGMA-OCD consortium

Objective Cognitive behavioral therapy (CBT) is an effective treatment for obsessive-compulsive disorder (OCD), yet it remains difficult to predict who will respond. This study investigates associations between neural activity during inhibitory control tasks and CBT outcomes, and its value as a predictive marker of CBT outcome. Methods Using fMRI data from individuals performing an inhibitory control task across five ENIGMA-OCD samples (n=130, age range=8-57, 54% female), mass-univariate associations were analyzed between response inhibition and error processing activity and three CBT outcomes: response, remission, and change in symptom severity. Random forest classifier and support vector machine models using leave-one-sample-out and 5-fold cross-validation were used to predict CBT response and remission from fMRI activity and clinical data. Results Remission after CBT was associated with weaker activity in default mode regions during response inhibition and in the right supramarginal gyrus during error processing. Greater symptom reduction was linked to weaker pre-treatment activity across frontoparietal, dorsal attention, visual, and subcortical regions during response inhibition, and stronger default mode activity during error processing. These group-level effects were partially confirmed with machine learning, which predicted CBT response using inhibitory and error-related activity with modest AUC values of 0.66-0.67. Prediction of CBT remission or using other data types did not exceed chance-level. Conclusion Wide-spread weaker activity during response inhibition and stronger activity in default mode regions during error processing pre-treatment appear beneficial for CBT. Despite modestly predictive AUC-values, these findings cannot yet be translated into clinically relevant individual markers of CBT outcome.