Effect of Regularization Parameter and Scan Time on Crossing Fibers with Constrained Compressed Sensing

Area of Science:

• Medical Imaging
• Neuroimaging
• Biophysics

Background:

• Diffusion Tensor Imaging (DTI) calculates a single fiber orientation per voxel, limiting its ability to resolve complex structures.
• Crossing Fiber Angular Resolution of Intra-voxel structure (CFARI) is a novel compressed sensing approach for resolving intra-voxel fiber architecture.
• Understanding CFARI's parameter dependencies is crucial for optimizing its performance.

Purpose of the Study:

• To investigate the performance of CFARI under varying signal-to-noise ratios (SNR) and regularization parameters.
• To elucidate the tradeoffs associated with scan parameters and algorithm choices in CFARI.
• To identify optimal strategies for enhancing CFARI's accuracy in resolving crossing fibers.

Main Methods:

• Simulations were conducted using diffusion-weighted magnetic resonance data.
• CFARI performance was evaluated as a function of SNR and the regularization parameter (beta).
• Experiments utilized two random tensors with varying diffusivities and eigenvalues but identical fractional anisotropy.

Main Results:

• Acquiring repeated measurements improves CFARI performance for a fixed scan time.
• A spatially variable, data-adaptive regularization parameter significantly stabilizes CFARI results.
• The optimal selection of the regularization parameter beta is data-dependent and spatially varying.

Conclusions:

• CFARI shows promise for resolving complex intra-voxel fiber structures in diffusion imaging.
• Optimizing CFARI requires careful consideration of SNR, regularization, and acquisition strategies.
• Data-adaptive regularization is key to robustly stabilizing CFARI results in complex white matter regions.