Biomarkers of Suicidal Ideation in Depression: Insights From VMHC Analysis and Machine Learning.

Background: Suicidal ideation (SI) is a major cause of death in patients with major depressive disorder (MDD). Although current clinical tools can assess suicide risk, objective neurobiological markers based on research remain lacking. Clinical evidence suggests that resting-state functional magnetic resonance imaging (rs-fMRI) studies utilizing voxel-mirrored homotopic connectivity (VMHC) analysis can uncover the neural mechanisms underlying mental disorders. This study explores differences in interhemispheric connectivity between MDD patients with and without SI, aiming to identify imaging biomarkers for suicide risk. Methods: This study included 48 SI patients and 44 non-SI patients. VMHC values were calculated to assess interhemispheric functional connectivity. Brain regions with significant differences between the groups were identified. A support vector machine (SVM) model was applied to evaluate the utility of VMHC values in distinguishing SI patients from non-SI patients with MDD. Results: Patients with suicidal ideation exhibited significantly increased VMHC values in the superior frontal gyrus, putamen, inferior temporal gyrus, and cerebellum compared to those without suicidal ideation. The SVM model achieved an accuracy of 77.2%, sensitivity of 83.3%, specificity of 70.5%, and an area under the curve (AUC) of 0.81. When combining VMHC values from multiple brain regions, classification accuracy improved to 86.8%. Conclusion: MDD patients with SI exhibit abnormal interhemispheric connectivity, with VMHC abnormalities in specific brain regions serving as potential biomarkers for suicide risk. The integration of machine learning and neuroimaging highlights the clinical relevance of VMHC as a tool for early detection and targeted intervention in suicide prevention.