RF Pulse Design for Parallel Excitation in MRI.

Abstract

Parallel excitation in MRI uses localized coils driven by independent RF waveforms as a mechanism for spatially encoding RF energy deposition. Because localized coil (or sensitivity) encoding is imposed instantaneously, one can create shorter pulses by trading gradient encoding for sensitivity encoding. However, the parallel pulse design problem is complicated by the non-Fourier nature of sensitivity encoding and the potential for patient-dependent problem inputs, requiring pulses to be designed rapidly online. In this project, I investigate novel techniques for parallel RF pulse design, with a focus on fast and general methods. I first propose a model-based iterative small-tip-angle pulse design method that is facilitated by a linear Fourier analysis of small-tip-angle excitation. It allows the user to rapidly design pulses with compensation for non-idealities such as main field inhomogeneities. We show in simulations and experiments that it produces pulses of higher accuracy than competing methods. The non-linear large-tip-angle regime requires more complex pulse design methods. To address this problem, I also investigate two fast large-tip-angle pulse design methods. Both are formulated as a series of Bloch simulations interleaved with small-tip-angle pulse designs whose results sum to produce accurate large-tip-angle pulses. Small-tip-angle pulse designs use approximate linear models for the perturbations induced by adding a small-tip-angle pulse to a large-tip-angle pulse. The first method uses the Fourier small-tip-angle equation as a linear model. We demonstrate that it is fast, robust and simple to implement, but it has some drawbacks, such as the inability to control excitation phase, that are addressed by the second method. The second method is based on a novel analytical linearization of the Bloch equation about an RF pulse. While more complex than the first method, we show that it produces pulses of higher accuracy, and can be applied to a broader range of pulse design problems. Both methods produce large-tip-angle pulses of higher accuracy than small-tip-designed pulses that are scaled to produce large-tip-angles.