Estimation of diffusion properties in crossing fiber bundles

Area of Science:

• Neuroimaging
• Diffusion MRI
• Computational Neuroscience

Background:

• Modeling diffusion in crossing fibers remains a challenge in diffusion MRI.
• Accurate estimation of diffusion parameters is crucial for understanding white matter microstructure.

Purpose of the Study:

• To develop an optimization framework for selecting dual tensor models and diffusion weighting parameters (b-values).
• To precisely and accurately estimate diffusion shape and orientation parameters in complex white matter regions.

Main Methods:

• Utilized the Cramér-Rao lower bound (CRLB) to assess parameter variance.
• Performed Monte Carlo simulations to evaluate model performance.
• Employed Levenberg-Marquardt optimization with a Rician noise model.

Main Results:

• Constraining axial diffusion (λ∥) and modeling isotropic fraction (fiso) improved parameter estimation.
• Fractional Anisotropy (FA) of both tensors could be independently estimated with ~9% precision (SNR=25).
• A two-element b-vector (b = [1.0, 3.5] x 10³ mm⁻²s) proved sufficient for heterogeneous tissue analysis with low error.

Conclusions:

• The proposed framework enables accurate and precise estimation of diffusion parameters in crossing fibers.
• This method allows for consistent estimation of FA profiles along crossing tracts.
• Establishes fundamental limits for comparative studies analyzing white matter structures.