Disrupted pathways from frontal-parietal cortex to basal ganglia and cerebellum in patients with unmedicated obsessive compulsive disorder as observed by whole-brain resting-state effective connectivity analysis - a small sample pilot study

Objective To date, a systematic characterization of abnormalities in resting-state effective connectivity (rsEC) in obsessive-compulsive disorder (OCD) is lacking. The present study aimed to systematically characterize whole-brain rsEC in OCD patients as compared to healthy controls. Methods Using resting-state fMRI data of 50 unmedicated patients with OCD and 50 healthy participants, we constructed whole-brain rsEC networks using Granger causality analysis followed by univariate and multivariate comparisons between patients and controls. Similar analyses were performed for resting-state functional connectivity (rsFC) networks to examine how rsFC and rsEC differentially capture abnormal brain connectivity in OCD. Results Univariate comparisons identified 10 rsEC networks that were significantly disrupted in patients, and which were mainly associated with frontal-parietal cortex, basal ganglia, and cerebellum. Conversely, abnormal rsFC networks were widely distributed throughout the whole brain. Multivariate pattern analysis revealed a classification accuracy as high as 80.5% for distinguishing patients from controls using combined whole-brain rsEC and rsFC. Conclusions The results of the present study suggest disrupted communication of information from frontal-parietal cortex to basal ganglia and cerebellum in OCD patients. Using combined whole-brain rsEC and rsFC, multivariate pattern analysis revealed a classification accuracy as high as 80.5% for distinguishing patients from controls. The alterations observed in OCD patients could aid in identifying treatment mechanisms for OCD.

Automated summary

The study revealed significant disruptions in 10 rsEC networks in OCD patients, mainly involving frontal-parietal cortex, basal ganglia, and cerebellum. Conversely, abnormal rsFC networks were distributed throughout the entire brain. Multivariate pattern analysis showed a classification accuracy as high as 80.5% for distinguishing patients from controls using combined whole-brain rsEC and rsFC.