Related Experiment Video
Updated: Feb 24, 2026
17:06
Co-analysis of Brain Structure and Function using fMRI and Diffusion-weighted Imaging
Published on: November 8, 2012
27.1K
Sparse Bayesian Inference of White Matter Fiber Orientations from Compressed Multi-resolution Diffusion MRI.
Pramod Kumar Pisharady1, Julio M Duarte-Carvajalino1, Stamatios N Sotiropoulos2
1CMRR, Radiology, University of Minnesota, Minneapolis, Minnesota, USA.
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|August 29, 2017
Summary
This study introduces a sparse Bayesian algorithm to improve fiber orientation estimation from compressed diffusion MRI data. The method enhances accuracy and robustness, even with under-sampled data.
Area of Science:
- Neuroimaging
- Biomedical Engineering
- Computational Neuroscience
Background:
- Diffusion MRI provides insights into tissue microstructure.
- Estimating crossing fiber orientations is crucial for accurate brain mapping.
- Current methods struggle with under-sampled or compressed diffusion MRI data.
Purpose of the Study:
- To develop a sparse Bayesian algorithm for robust fiber orientation and volume fraction estimation from compressed diffusion MRI.
- To enhance the RubiX algorithm by incorporating sparsity for improved reconstruction from under-sampled data.
Main Methods:
- A sparse Bayesian approach is proposed, modeling high-resolution data with parametric spherical deconvolution and low-resolution data with spatial partial volume representation.
- A dictionary is created using exponential decay components, with dictionary weights representing fiber volume fractions.
- Sparsity priors are introduced to exploit the inherent sparsity of fiber orientations in the data.
Main Results:
- The proposed method demonstrates improved accuracy in estimating fiber orientations.
- A significant decrease in the uncertainty of fiber orientation estimates was observed.
- Robustness in fiber orientation estimation was achieved, particularly for under-sampled diffusion MRI data.
Conclusions:
- The sparse Bayesian algorithm effectively estimates fiber orientations and volume fractions from compressed diffusion MRI.
- Sparsity incorporation enhances the RubiX algorithm's capability to handle under-sampled data.
- This approach offers a more accurate and reliable method for microstructural tissue parameter extraction in diffusion MRI.
