Anatomical Biomarkers for Adolescent Major Depressive Disorder from Diffusion Weighted Imaging using SVM Classifier

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Psychiatry

Background:

• Adolescent Major Depressive Disorder (MDD) is a significant mental health concern with severe long-term consequences.
• Identifying reliable biomarkers for MDD in adolescents is crucial for early diagnosis and effective treatment.
• Current diagnostic methods may not fully capture the neurobiological underpinnings of adolescent MDD.

Purpose of the Study:

• To explore anatomical brain features using machine learning to identify biomarkers for adolescent Major Depressive Disorder (MDD).
• To distinguish between MDD patients and healthy subjects based on neuroimaging data.
• To evaluate the classification performance of the proposed machine learning methodology.

Main Methods:

• Utilized diffusion tensor imaging (DTI) to measure anatomical connectivity between brain regions.
• Applied topological measurements from anatomical brain networks.
• Employed p-value based filtering and minimum redundancy maximum relevance (mRMR) for feature selection.
• Performed leave-one-out cross-validation for performance evaluation.

Main Results:

• Achieved an accuracy of 78%, with 90.39% sensitivity and 79.66% precision in classifying 79 subjects.
• Identified key distinguishing features including network centrality and participation coefficients in specific brain regions (e.g., right lingual gyrus, right lateral occipital sulcus).
• Highlighted the role of altered connectivity and network topology in adolescent MDD.

Conclusions:

• Machine learning analysis of anatomical brain networks can effectively identify biomarkers for adolescent MDD.
• Specific network measures derived from DTI data show promise in distinguishing MDD patients.
• This approach offers a potential avenue for developing objective diagnostic tools for adolescent MDD.