Accelerated MR diffusion tensor imaging using distributed compressed sensing.

Area of Science:

• Medical Imaging
• Biophysics
• Signal Processing

Background:

• Diffusion Tensor Imaging (DTI) acquisition is time-consuming, often requiring minutes to hours.
• Accelerated DTI acquisition is crucial for clinical and research applications.
• Exploiting image priors can improve reconstruction speed and quality.

Purpose of the Study:

• To investigate the feasibility and efficacy of distributed compressed sensing for accelerating DTI.
• To leverage joint sparsity in diffusion-weighted images for faster DTI acquisition.
• To evaluate the performance of distributed compressed sensing in reconstructing DTI data.

Main Methods:

• Retrospective undersampling of fully sampled DTI datasets (simulated and experimental heart) with acceleration factors 2-6.
• Reconstruction of diffusion-weighted images using l2-l1 norm minimization.
• Evaluation of reconstruction accuracy using root-mean-square error and DTI index maps.

Main Results:

• Distributed compressed sensing demonstrated superiority over basic compressed sensing in simulations.
• Reconstruction accuracy was dependent on signal-to-noise ratio and acceleration factors.
• Accurate DTI indices (fractional anisotropy, mean diffusivities, eigenvector orientation) were achieved at acceleration factors up to 4.

Conclusions:

• Distributed compressed sensing effectively accelerates DTI acquisition.
• This technique shows potential for practical reduction of DTI scan times.
• Distributed compressed sensing offers a viable solution for faster DTI data acquisition.