Improved parallel magnetic resonance imaging reconstruction with multiple variable density sampling.

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging
• Image Reconstruction

Background:

• Parallel Magnetic Resonance Imaging (pMRI) methods like GRAPPA reconstruct undersampled k-space data.
• Traditional pMRI methods require auto-calibration signal (ACS) lines sampled at k-space center.
• Standard sampling involves full sampling in the center and uniform undersampling in outer k-space.

Purpose of the Study:

• To propose a novel reconstruction method using multiple variable density sampling (MVDS).
• To improve image quality in pMRI using fewer ACS lines and multiple reduction factors.
• To reduce noise and artifacts in reconstructed MR images.

Main Methods:

• Implemented a novel multiple variable density sampling (MVDS) pattern.
• Kept auto-calibration signal (ACS) lines unchanged.
• Applied multiple reduction factors: smaller near ACS, larger at the outermost k-space.

Main Results:

• The MVDS method achieved lower reconstruction error, especially with fewer ACS lines.
• NL-GRAPPA reconstruction of MVDS-sampled data resulted in reduced noise and artifacts.
• The proposed method demonstrated superior performance compared to traditional sampling patterns.

Conclusions:

• The MVDS sampling strategy significantly improves pMRI image quality.
• This method is particularly effective in low ACS line scenarios.
• MVDS offers a promising alternative for efficient and high-quality MR image reconstruction.