Neuroimaging Markers for Studying Gulf-War Illness: Single-Subject Level Analytical Method Based on Machine Learning.

Area of Science:

• Neuroimaging
• Veterans' Health
• Biomarker Discovery

Background:

• Gulf War illness (GWI) affects ~250,000 veterans with chronic, persistent symptoms.
• GWI veterans experience poorer health and more chronic conditions than non-deployed peers.
• Objective biomarkers are needed for accurate GWI diagnosis and management.

Purpose of the Study:

• To identify objective neuroimaging biomarkers for Gulf War illness diagnosis.
• To evaluate the effectiveness of T1W-MRI, DTI, and NDI in detecting GWI characteristics.
• To develop a machine learning model for GWI classification using neuroimaging data.

Main Methods:

• Applied T1-weighted MRI, diffusion tensor imaging (DTI), and neurite density imaging (NDI).
• Utilized group-level statistical comparisons and single-subject machine learning analysis.
• Trained a classifier using white matter NDI features for GWI case identification.

Main Results:

• Neurite density imaging (NDI) demonstrated high sensitivity in characterizing GWI.
• The NDI-based classifier achieved 90% accuracy and 0.941 F-score in cross-validation.
• NDI measures are sensitive to microstructural and macrostructural brain changes in GWI veterans.

Conclusions:

• NDI shows significant promise as a diagnostic tool for Gulf War illness.
• Neuroimaging, particularly NDI, can reveal brain alterations associated with GWI.
• These findings support NDI's value for improving GWI diagnosis and treatment studies.