Optimal shrinkage denoising breaks the noise floor in high-resolution diffusion MRI

Area of Science:

• Medical Imaging
• Biophysics
• Signal Processing

Background:

• Spatial resolution in diffusion magnetic resonance (MR) imaging is limited by noise.
• Weak signals from small voxels, especially with high diffusion sensitization, are obscured by noise due to the non-Gaussian nature of MR magnitude signals.

Purpose of the Study:

• To develop a method for noise suppression in diffusion MR imaging.
• To improve spatial resolution and enable precise characterization of tissue microstructure and white matter pathways.

Main Methods:

• Optimal shrinkage of singular values in complex-valued k-space data from multiple receiver channels.
• Exploration and comparison of different low-rank signal matrix recovery strategies.
• Integration with background phase removal techniques.

Main Results:

• Remarkable suppression of the noise floor.
• An 11-fold reduction in the noise floor achieved with the optimal strategy.
• Substantially improved imaging resolution without hardware upgrades or extended acquisition times.

Conclusions:

• The developed framework effectively reduces noise in diffusion MR imaging.
• This method enhances imaging resolution for detailed microstructural and white matter pathway analysis.
• The approach outperforms existing denoising methods and is cost-effective.