Labeling of ambiguous subvoxel fibre bundle configurations in high angular resolution diffusion MRI

Area of Science:

• Neuroimaging
• Diffusion MRI analysis
• Computational neuroscience

Background:

• High angular resolution diffusion imaging (HARDI) methods in MRI estimate dominant fiber orientations within voxels.
• Complex subvoxel fiber structures can lead to ambiguous orientation distribution functions, limiting tractography accuracy.

Purpose of the Study:

• To disambiguate complex subvoxel fiber tract configurations.
• To improve the performance of diffusion MRI fiber tractography.

Main Methods:

• Extension of a curve inference method using differential geometric estimates.
• Distinguishing between curving and fanning fiber bundles based on local neighborhood analysis.
• Modeling underlying fiber bundles using inferred curves instead of just orientations.

Main Results:

• The method successfully labels voxels into categories representing different fiber geometries.
• Distinguishes between distinct fiber geometries with similar voxel-wise orientation distributions.
• Experimental results show improved performance of a fiber-tracking algorithm.

Conclusions:

• The curve inference method enhances the accuracy of diffusion MRI by resolving subvoxel ambiguity.
• This approach improves fiber tractography by providing a more precise representation of neural pathways.
• Enables more reliable mapping of brain connectivity through improved tract reconstruction.