SLfRank: Shinnar-Le-Roux Pulse Design With Reduced Energy and Accurate Phase Profiles Using Rank Factorization.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Pulse Sequence Design
• Biomedical Engineering

Background:

• The Shinnar-Le-Roux (SLR) algorithm is a standard method for designing frequency-selective pulses in MRI.
• The sequential solution in the original SLR algorithm leads to sub-optimal pulse designs due to bi-linear coupling.
• Existing methods often result in suboptimal pulse energy and phase profile accuracy.

Purpose of the Study:

• To enhance the Shinnar-Le-Roux (SLR) algorithm for designing frequency-selective pulses.
• To reduce the energy requirements of designed pulses.
• To improve the accuracy of phase profiles in generated pulses.

Main Methods:

• Implemented a convex relaxation technique to address the bi-linear coupling in Cayley-Klein (CK) parameters.
• Modified the design process of the SLR algorithm by jointly optimizing polynomial pairs instead of sequentially.
• Utilized numerical experiments to validate the performance of the improved algorithm.

Main Results:

• The proposed method generates pulses with significantly lower energy, up to 26% reduction compared to the original SLR algorithm.
• Pulses designed with the improved algorithm exhibit more accurate phase profiles, particularly linear phase profiles for slice excitation.
• Numerical experiments indicate that the new approach achieves near-globally optimal pulse designs in practice.

Conclusions:

• The enhanced SLR algorithm offers a more efficient and accurate method for designing frequency-selective pulses in MRI.
• Convex relaxation provides an effective solution to the bi-linear coupling problem in CK parameter design.
• The improved pulse design leads to reduced energy deposition and enhanced signal fidelity in MRI applications.