Resolution of crossing fibers with constrained compressed sensing using diffusion tensor MRI.

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Diffusion tensor imaging (DTI) is a standard neuroimaging technique for mapping tissue microstructure and brain connectivity.
• The DTI tensor model is limited in regions with crossing fibers, failing to represent multiple intra-voxel orientations.
• Existing methods to address crossing fibers often require extensive diffusion MRI data acquisition, hindering clinical application.

Purpose of the Study:

• To develop a compressed sensing technique for resolving crossing fibers in diffusion MRI.
• To enable accurate analysis of brain connectivity using routinely acquired clinical diffusion MRI data.
• To overcome the limitations of the conventional tensor model in complex white matter regions.

Main Methods:

• A compressed sensing approach was employed, assuming data can be sparsely represented by a linear combination of tensors with varying orientations.
• A fast hierarchical compressed sensing algorithm was developed for efficient computation of optimal orientations.
• A novel metric was introduced for comparing estimated fiber orientations.

Main Results:

• The proposed compressed sensing method successfully resolved crossing fibers using conventional diffusion MRI data (30 directions, b=700 s/mm²).
• Performance was validated through both simulations and in vivo imaging.
• The technique achieved comparable or superior results to standard q-ball imaging, which requires significantly more data.

Conclusions:

• Compressed sensing offers a viable solution for resolving crossing fibers in diffusion MRI with rapid, clinically feasible data acquisition.
• This method enhances the utility of diffusion MRI for studying brain connectivity, particularly in complex white matter tracts.
• The technique holds promise for improving diagnostic capabilities and research in neurological disorders affecting white matter integrity.