Improving GRAPPA reconstruction by frequency discrimination in the ACS lines

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Image Reconstruction

Background:

• GRAPPA is a parallel imaging technique that reconstructs MR images from undersampled k-space data.
• GRAPPA utilizes autocalibrated (ACS) lines, sampled at the Nyquist rate, to estimate reconstruction weights.
• ACS lines contain both low-frequency (DC component) and higher-frequency information, potentially leading to reconstruction artifacts.

Purpose of the Study:

• To propose a method for reducing GRAPPA reconstruction artifacts by selectively using ACS lines.
• To improve the accuracy of GRAPPA reconstruction weights by discriminating low-frequency spectrum components.

Main Methods:

• The proposed method involves excluding data points around the DC component within the ACS lines.
• A simple approach is to eliminate a central square window from the k-space ACS region.
• More sophisticated methods for selecting ACS coefficients can also be employed.

Main Results:

• Empirical testing on real multi-coil MRI data demonstrates significant enhancement rates.
• The method maintains GRAPPA's computational complexity and reduces the g-factor.
• Reconstruction accuracy improved by up to 35% with 32 ACS lines and an acceleration factor of 3.
• The approach showed further accuracy improvements when combined with other reconstruction techniques.

Conclusions:

• The proposed method enhances GRAPPA coefficient accuracy, leading to superior final image reconstruction.
• This technique is fully compatible with the original GRAPPA algorithm and other optimization methods.
• The method is easily implementable in commercial MRI scanning software.