Transdiagnostic Connectome-Based Prediction of Response Inhibition

Response inhibition (RI) deficits are a core feature across diagnostic categories of mental disorders. However, it remains unclear whether the brain networks underlying different forms of RI deficits are disorder-shared or disorder-specific, and how they interact with aberrant brain connectivity across disorders. Connectome-based predictive modeling (CPM) provides a novel approach for exploring the brain networks associated with RI abnormalities across diagnostic categories of mental disorders. Publicly available resting-state functional magnetic resonance imaging data from individuals with schizophrenia (n = 47), bipolar disorder (n = 47), and attention-deficit/hyperactivity disorder (n = 40), as well as healthy controls (n = 121), were utilized to construct whole-brain network predictive models for different forms of RI (action cancellation and action restraint). The brain networks of different forms of RI were further compared with abnormal brain networks in the diagnostic groups. Action restraint and action cancellation exhibited both shared and distinct brain networks. There was a dissociation in the relationship between the brain networks underlying different forms of RI and the aberrant connectivity patterns observed across diagnostic categories. Our models successfully predicted action restraint performance across diagnostic categories, whereas the model failed to effectively predict action cancellation due to the influence of disease-related aberrant connectivity on the brain networks underlying action cancellation. Nevertheless, the action cancellation model demonstrated generalizability to novel, healthy participants (n = 220) from an independent dataset. Our study clarifies the complex relationship between deficits in RI and the neuropathology of mental disorders and provides a foundation for more accurate cognitive assessment and targeted interventions. Our findings highlight the importance of refining RI constructs and emphasize the value of applying connectome methods to reveal cross-diagnostic neural mechanisms.