Multiscale white matter fiber tract coregistration: a new feature-based approach to align diffusion tensor data.

Area of Science:

• Medical Imaging
• Neuroimaging
• Computational Anatomy

Background:

• Diffusion Tensor Imaging (DTI) is crucial for visualizing white matter (WM) structural connectivity.
• Accurate coregistration of DTI data is essential for robust analysis and comparison across subjects.
• Existing voxel-based methods can be sensitive to noise and may not fully capture complex fiber pathway structures.

Purpose of the Study:

• To develop an automatic, multiscale, feature-based rigid-body coregistration technique for DTI.
• To utilize the local curvature (kappa) and torsion (tau) of WM fiber pathways as features for registration.
• To evaluate the accuracy, precision, and feasibility of the proposed technique for human brain DTI data.

Main Methods:

• A novel coregistration technique based on the (kappa, tau)-space of WM fiber pathways.
• Mean Squared Difference (MSD) in (kappa, tau)-space as the similarity measure.
• Incorporation of a scale-space representation for multiscale robustness and Principal Component Analysis (PCA) for transformation parameter calculation.

Main Results:

• The proposed automatic multiscale technique demonstrated high accuracy and precision in simulations on synthetic DTI data.
• The method inherently supports region of interest (ROI) coregistration and handles both global and local transformations.
• Comparison with a voxel-based approach on in vivo human brain DTI data showed the feasibility and advantages of the fiber pathway-based technique.

Conclusions:

• The presented automatic multiscale feature-based coregistration technique offers a robust and accurate method for aligning DTI data.
• Leveraging white matter fiber pathway geometry in (kappa, tau)-space provides a powerful alternative to voxel-based registration.
• This technique shows significant potential for improving the analysis of neuroimaging data in clinical and research settings.