MULTI-ECHO RECOVERY WITH FIELD INHOMOGENEITY COMPENSATION USING STRUCTURED LOW-RANK MATRIX COMPLETION

Area of Science:

• Magnetic Resonance Imaging
• Neuroimaging
• Image Reconstruction

Background:

• Echo-planar imaging (EPI) is crucial for functional MRI but suffers from geometric distortions due to magnetic field inhomogeneities.
• These distortions limit spatial resolution and affect functional activation sensitivity, especially with single echo times.
• Current acceleration techniques have limitations in acquiring multiple echo times within a practical repetition time (TR).

Purpose of the Study:

• To introduce a novel rosette-based acquisition scheme and a structured low-rank reconstruction algorithm.
• To overcome geometric distortions and improve functional activation sensitivity in EPI.
• To enable distortion-free imaging with enhanced spatial resolution.

Main Methods:

• Development of a rosette-based k-space acquisition trajectory.
• Implementation of a structured low-rank reconstruction algorithm tailored for EPI data.
• Exploitation of the time series' exponential structure for simultaneous echo recovery.

Main Results:

• Successful recovery of distortion-free images from multi-echo EPI data.
• Demonstrated improvement in functional activation sensitivity across different brain regions.
• Potential for higher spatial resolution imaging compared to conventional EPI.

Conclusions:

• The proposed rosette-based scheme and reconstruction offer a promising solution for EPI distortions.
• This method enhances the reliability and sensitivity of functional MRI studies.
• It paves the way for more accurate and higher-resolution neuroimaging.