Tract-based probability densities of diffusivity measures in DT-MRI

Area of Science:

• Neuroimaging
• Biomarker Discovery
• Statistical Modeling

Background:

• Diffusion Tensor Imaging (DTI) provides insights into white matter microstructure.
• Biomarkers are crucial for diagnosing and understanding neurological conditions like Vascular Cognitive Impairment (VCI).
• Probability density functions (PDFs) offer a robust way to characterize complex data distributions.

Purpose of the Study:

• To evaluate the utility of probability density functions of DTI diffusivity measures as biomarkers.
• To assess the effectiveness of these PDFs in differentiating between VCI patients and healthy controls.
• To explore the application of PDFs in classifying individual subjects.

Main Methods:

• Estimation of univariate and bivariate probability densities for DTI measures (FA, MD, RD, AD) and tract arc length.
• Analysis focused on transcallosal fibers in the brain.
• Application of estimated densities for hypothesis testing and subject classification.

Main Results:

• Estimated densities and derived metrics like entropy successfully detected group differences between VCI patients and controls.
• High statistical power was achieved in identifying group disparities.
• Low classification errors were obtained when distinguishing between the two groups.

Conclusions:

• Probability density functions of DTI diffusivity measures are effective biomarkers for VCI.
• These PDF-based approaches offer powerful tools for hypothesis testing and subject classification in neurological research.
• The findings support the use of advanced statistical modeling in neuroimaging for clinical applications.