High-Accelerated Parallel Imaging With the Inherent Local Feature in PE-xSPEN MRI.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Image Reconstruction
• Accelerated Imaging Techniques

Background:

• Undersampling in MRI's phase-encoding (PE) dimension causes aliasing, creating overlapping object replicas.
• Conventional methods struggle to differentiate these aliased signals, especially without multi-channel coils.

Purpose of the Study:

• To develop a highly accelerated, high-resolution imaging technique using PE-xSPEN.
• To leverage spatiotemporal encoding for distinguishing spatial signal contributions before k-space acquisition.
• To enable differentiation of aliased localized information in k-space.

Main Methods:

• PE-xSPEN utilizes modulated hyperbolic phase, coupling the y and z spatial dimensions.
• This encoding ensures signals at a given k-space location originate predominantly from local voxels.
• Aliased spatial replicas remain distinguishable post-inverse Fourier transform, even without multi-channel coils.

Main Results:

• The PE-xSPEN GRAPPA framework demonstrated reduced residual aliasing compared to traditional Fourier-encoded signals.
• Performance gains were validated through MRI simulations, phantom tests, and in vivo brain imaging.
• The method enhances unfolding efficiency when combined with parallel imaging.

Conclusions:

• The PE-xSPEN GRAPPA framework effectively controls aliasing artifacts using hyperbolic phase modulation.
• It generates k_z-dependent coil responses, minimizing artifacts and noise amplification.
• This approach offers improved reconstruction for accelerated, high-resolution MRI.