Fast and accurate reconstruction of HARDI data using compressed sensing

Area of Science:

• Neuroimaging
• Medical Physics
• Biomedical Engineering

Background:

• Brain disorders can degrade white matter integrity and connectivity.
• Diffusion MRI (dMRI) quantifies water diffusion changes in white matter, reflecting neural tract integrity.
• Long scanning times, especially for High Angular Resolution Diffusion Imaging (HARDI), limit dMRI's clinical use.

Purpose of the Study:

• To reduce HARDI scanning times using compressed sensing.
• To maintain or improve the accuracy of HARDI data estimation.
• To facilitate wider clinical integration of dMRI.

Main Methods:

• Applied compressed sensing theory to reduce diffusion-encoding gradients in HARDI.
• Compared the proposed method against alternative approaches.
• Quantified accuracy using mean squared error.

Main Results:

• Compressed sensing significantly reduced the number of required diffusion gradients.
• The proposed method achieved more accurate HARDI data estimation compared to alternatives.
• Mean squared error was lower with the compressed sensing approach.

Conclusions:

• Compressed sensing is a viable technique to shorten dMRI/HARDI acquisition times.
• This approach enhances the practicality of dMRI for clinical applications.
• Reduced scanning times can accelerate the integration of dMRI in diagnosing brain disorders.