Dong Chen1,2, Folkmar Bornemann1, Mika W. Vogel2, Laura I. Sacolick2, Guido Kudielka2, Yudong Zhu3
1Center for Mathematical Science, Technical University of Munich, Munich, Germany; 2Imaging Technologies, GE Global Research Europe, Munich, Germany; 3New York University Langone Medical Center, New York, NY, USA
Parallel RF transmit offers additional degrees of freedom in excitation pulse design. One application of that is accelerating multidimensional selective excitation by under-sampling the excitation k space. Where to ideally under-sample the k space with optimal sparsity is determined by the target profile, B1 maps and error tolerance, which is known a priori. We propose an adaptive k space sparsifying method, which exploits this prior-knowledge using a greedy-wise algorithm that simultaneously sparsify over all Tx coils and significantly reduces the pulse duration. The method was validated with Bloch simulations and parallel transmit phantom imaging experiments.