A Diagnosis Model of Typhoon-Related Post-Traumatic Stress Disorder Based on Fixel-Based Analysis in Machine Learning

Area of Science:

• Neuroimaging
• Psychiatry
• Machine Learning

Background:

• Post-traumatic stress disorder (PTSD) is a common mental health condition following trauma, with environmental disasters like super typhoons being significant triggers.
• Previous research indicates white matter alterations in individuals with PTSD.
• Fixel-based analysis (FBA) is an advanced diffusion MRI technique for detailed white matter microstructure assessment.

Purpose of the Study:

• To investigate the utility of FBA as an imaging biomarker for PTSD in super typhoon survivors.
• To reduce subjective bias in PTSD diagnosis by leveraging objective imaging data.
• To explore the combination of FBA with machine learning for enhanced PTSD identification.

Main Methods:

• Diffusion MRI data were acquired from three groups: PTSD (n=27), trauma-exposed controls (TEC, n=33), and healthy controls (HC, n=30).
• Fixel-based analysis (FBA) was employed to assess white matter integrity using metrics like fiber density (FD), fiber cross-section (FC), and fiber density-cross section (FDC).
• Machine learning models were developed and validated using five-fold cross-validation to differentiate between PTSD, TEC, and HC groups.

Main Results:

• PTSD patients exhibited increased FD in the right frontopontine tract and right middle longitudinal fascicle compared to HC.
• Elevated FDC values were observed in the bilateral frontopontine tract and left thalamo-premotor tract in the PTSD group.
• Machine learning models demonstrated high performance in distinguishing PTSD from TEC (accuracy=0.89, AUC=0.95) and all three groups (macro-averaged F1-score=0.99).

Conclusions:

• Typhoon survivors with PTSD show distinct structural alterations in brain white matter, detectable via FBA.
• The integration of FBA with machine learning provides a promising approach for objective PTSD diagnosis.
• These findings highlight potential imaging biomarkers for PTSD, improving understanding of its microstructural correlates.