Joint fractional segmentation and multi-tensor estimation in diffusion MRI.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Biomedical Engineering

Background:

• Accurate segmentation of white matter tracts is crucial for understanding brain connectivity.
• Estimating diffusion models in regions with crossing fibers and partial voluming presents significant challenges.
• Current diffusion-weighted imaging (DWI) methods struggle with complex white matter architectures.

Purpose of the Study:

• To develop a novel Bayesian approach for fractional segmentation of white matter tracts.
• To simultaneously estimate multi-tensor diffusion models in the brain.
• To improve the reliability of diffusion measurements in challenging white matter regions.

Main Methods:

• A Bayesian framework incorporating spatially correlated tensor priors within white matter tracts.
• Fractional segmentation allowing multiple tensor compartments per voxel.
• Simultaneous estimation of tract weights and diffusion tensors.

Main Results:

• Reliable estimation of multiple tensor compartments in fiber crossing regions, even with low angular resolution DWI.
• Reduced effects of partial voluming by distinguishing diffusion compartments associated with each tract.
• Accurate recovery of ground truth volume fractions and tensor compartments on synthetic data.
• Improved segmentation and diffusion measurement statistics on real data with crossing tracts.

Conclusions:

• The proposed Bayesian method offers a robust solution for fractional white matter tract segmentation and multi-tensor modeling.
• This approach enhances the accuracy and reliability of diffusion MRI analysis, particularly in complex white matter structures.
• The method has the potential to improve diagnostic capabilities and our understanding of neurological conditions.