XQ-SR: Joint x-q space super-resolution with application to infant diffusion MRI

Area of Science:

• Neuroimaging
• Medical Physics
• Biomedical Engineering

Background:

• Diffusion MRI (DMRI) is crucial for studying brain development and disorders.
• Low spatio-angular resolution in DMRI limits detailed structural analysis, especially in infant brains.
• Existing super-resolution (SR) methods often address spatial or wavevector domains separately.

Purpose of the Study:

• To develop a novel framework for joint super-resolution (SR) in both spatial (x-space) and wavevector (q-space) domains for DMRI.
• To improve the resolution and quality of DMRI data, particularly for challenging infant brain imaging.
• To overcome limitations of existing SR techniques by enhancing both spatial and angular resolution concurrently.

Main Methods:

• Established signal relationships in x-q space using robust neighborhood matching.
• Utilized these relationships to regularize the inverse problem of recovering high-resolution DMRI data.
• Developed a joint super-resolution framework applicable to post-acquisition data enhancement.

Main Results:

• Demonstrated successful high-resolution data recovery from low-resolution DMRI.
• Achieved remarkably improved image quality in synthetic, adult, and infant DMRI datasets.
• Validated the effectiveness of joint spatial and wavevector domain SR.

Conclusions:

• The proposed joint SR framework effectively enhances DMRI resolution in both spatial and wavevector domains.
• This method offers a significant improvement for analyzing subtle brain structures, particularly in infant neuroimaging.
• The technique provides a valuable tool for advancing quantitative analysis in DMRI without extending scan durations.