Modeling essential connections in obsessive-compulsive disorder patients using functional MRI.

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Obsessive-compulsive disorder (OCD) diagnosis relies on subjective neuropsychological assessments.
• Current diagnostic methods for OCD can be influenced by patient and clinician bias.
• There is a need for objective, quantitative methods for OCD diagnosis.

Purpose of the Study:

• To develop and validate a classification model for OCD diagnosis using functional magnetic resonance imaging (fMRI).
• To identify brain functional connectivity (FC) variations associated with OCD.
• To explore the role of specific brain regions, particularly the cerebellum, in OCD.

Main Methods:

• Functional connectivity (FC) matrices were computed from resting-state fMRI data.
• A novel Riemann Kernel principal component analysis (PCA) model was used for feature extraction, preserving topological information.
• Hierarchical features were classified using an ensemble XGBoost algorithm.

Main Results:

• The proposed classification model achieved an accuracy of 91.8% in distinguishing OCD patients from healthy controls.
• The cortico-striatal-thalamic-cortical (CSTC) circuit and cerebellum were identified as key regions associated with OCD.
• Patients with OCD exhibited decreased posterior cerebellar connections compared to controls.

Conclusions:

• The study presents an efficient fMRI-based method for characterizing OCD patients.
• The findings offer a new perspective on OCD-related brain features, emphasizing the cerebellum's role.
• This research may provide novel insights into the neurobiological underpinnings and genetic etiology of OCD.