Pseudo-random center placement O-space imaging for improved incoherence compressed sensing parallel MRI

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Image Reconstruction Algorithms

Background:

• Nonlinear spatial encoding magnetic (SEM) field strategies, like O-space imaging, can introduce artifacts in accelerated MRI scans.
• Compressed sensing (CS) enables image reconstruction from undersampled data by leveraging signal sparsity.

Purpose of the Study:

• To develop and evaluate a pseudo-random center placement (CP) O-space CS approach for enhanced MRI.
• To optimize image reconstruction by modulating SEM fields and improving incoherence.

Main Methods:

• Investigated the relationship between the incoherence parameter and CS validity.
• Optimized O-space acquisition for CS by perturbing Z(2) strength.
• Demonstrated the method on a 3T human scanner using an eight-channel receiver array.

Main Results:

• Pseudo-random CP O-space imaging demonstrated improved incoherence between sensing and sparse domains.
• Images reconstructed using pseudo-random CP O-space showed reduced mean squared error compared to linear SEM field methods.

Conclusions:

• Pseudo-random CP O-space imaging with a nonlinear SEM field is effective for CS-based reconstruction.
• This method significantly reduces image mean squared error at high acceleration for 2D slices compared to linear encoding.