A Fast Patch-Based Hankel Low-Rank Method for Magnetic Resonance Spectroscopy Reconstruction

Area of Science:

• Magnetic Resonance Spectroscopy (MRS)
• Computational Chemistry
• Structural Biology

Background:

• Sparse sampling accelerates multi-dimensional magnetic resonance spectroscopy (MRS) acquisition.
• Current low-rank reconstruction methods offer high-fidelity recovery but suffer from long reconstruction times.
• This limitation hinders the practical application of advanced MRS techniques.

Purpose of the Study:

• To develop a novel approach for significantly reducing reconstruction times in sparsely-sampled MRS.
• To address the computational bottleneck of existing low-rank reconstruction methods.
• To maintain high spectral reconstruction quality while improving efficiency.

Main Methods:

• Introduced a novel method to reduce the dimensionality of the low-rank Hankel-like matrix used in MRS reconstruction.
• Leveraged dimensionality reduction to decrease computational complexity.
• Validated the approach using both simulated and real-world MRS data.

Main Results:

• Achieved a substantial acceleration in reconstruction times, exceeding fourfold compared to conventional low-rank methods.
• Demonstrated that the proposed method does not compromise the quality of spectrum reconstructions.
• Successfully applied the technique to both simulated and experimental MRS data.

Conclusions:

• The novel dimensionality reduction technique offers a significant speed-up for sparsely-sampled MRS reconstruction.
• This advancement makes advanced MRS techniques more accessible and practical for chemistry and structural biology.
• The method provides a computationally efficient alternative without sacrificing spectral fidelity.