Shimming and MRSI

INTRODUCTION With the advent of high field 3T and ultrahigh field 7T MR systems, significant improvements in a variety of brain imaging modalities were anticipated. Although improvements in SNR have been demonstrated at 3T and 7T [1, 2], limitations due to increased static field (B0) inhomogeneity have also been noted. One modality particularly affected by poor B0 homogeneity is magnetic resonance spectroscopic imaging, (MRSI). Although excellent B0 homogeneity is a requisite for MRSI, the hardware requirements in terms of shim strength and shim order necessary for 7T is controversial, with the majority of systems delivered having only 2 order shims. The arguments forwarded against the efficacy of higher order shims include: 1) the required strengths for higher order terms are sufficiently large as to make them unattainable with conventional shim technology and 2) the effects of higher order terms across individual voxels is minimal due to the small voxel sizes used in MRSI studies. The first issue, which is characterized by the global homogeneity, B Global 0 , and calculated over the entire ROI is a significant factor in determining the performance of frequency selective pulses, such as those generally used for water suppression or spectral editing. The second issue is characterized by the local homogeneity, B Local 0 , which is measured over individual voxels and impacts spectral resolution within a voxel. This latter effect can significantly impact the number of pixels which provide useable spectra. Therefore, in this presentation we will: 1) evaluate the role and requirements for 3 order shims for SI studies at 7T for both B Global 0 and B Local 0 ; 2) demonstrate high resolution spectroscopic imaging from a variety of brain regions using 1-3 order shimming and 3) evaluate the potential role for 4 order shimming.