Shimming and MRS

Adjustment of the static magnetic field homogeneity, commonly known as the B0 shimming or simply shimming, is essential for magnetic resonance spectroscopy because it determines the spectral resolution which is critical for reliable metabolite quantification. Inhomogeneities in the B0 field, resulting primarily from susceptibility differences between air and tissue, are scaled with the B0 field and become highly non-linear at ultra-high magnetic fields. Although efficient minimization of the B0 inhomogeneity in the whole human brain is extremely difficult especially at high fields, successful results were recently achieved at 7T by using shim coils up to the 3 order [1]. The B0 field is usually most distorted in prefrontal regions due to its close proximity to the sinuses. Different techniques have been proposed in combination with active shimming to minimize strong local inhomogeneities including localized shim coils [2,3], diamagnetic and paramagnetic passive shims [4-6] and dynamic shimming [7]. In general, successful shimming requires efficient shimming methods for mapping the B0 field variations over the region of interest as well as a shim system (coils and drivers) which is strong enough to compensate these field gradients. Methods developed for B0 field mapping can be grouped into two categories; methods based on 3D B0 mapping [8-10] and B0 mapping along projections [11-15]. In both types of shimming techniques, information about the B0 field variation is calculated from phase differences acquired during the evolution of the magnetization in a non-homogeneous field. The precision of the B0 field mapping depends on the duration of the evolution time and the spatial resolution of the mapping. Longer evolution times increase sensitivity to small B0 changes and allow a fine adjustment of shims. However, longer evolution times result in a substantial signal loss and severe phase unwrapping problems when the B0 field homogeneity is poor, typically in the beginning of the shimming process. Therefore, methods utilizing multiple evolution times are preferable [13,8,9]. The strength of linear shims (X, Y, Z) is not a problem, because powerful gradient coils and gradient amplifiers are preferentially used for 1 order shim corrections in MR scanners. Therefore, limitations in the shim strength are more an issue with higher-order shim coils, usually the strength of the 2 order shim system. The maximum strength of a shim coil depends on its inductance and the maximum current of the corresponding shim driver. The higher shim coil inductance requires less current to generate a specific field, however, this may result in stronger coupling to the gradient coil. Which means that the gradient switching can negatively influence the stability of the shim current. Shim coils with lower inductance may exhibit decreased coupling with the gradient coil but require more current to generate a given field, which can result in difficulties with sufficient heat extraction from the shim coil. MR spectroscopic imaging (MRSI) requires a highly homogeneous B0 field within a large region of the brain, typically across a whole slice through the brain or possibly within the entire brain volume. Due to substantial B0 field distortions at high magnetic fields, adjustment of an acceptable B0 homogeneity in large volumes requires shim corrections up to the 3 order and probably some 4-order terms, such as Z4, might be necessary [8]. Despite all this effort, complete elimination of small local deviations from the B0 field uniformity is very difficult. On the other hand, single voxel MRS requires adjustment of the B0 field homogeneity in a relatively small volume, which is technically much easier. In addition, the B0 field distortions can be very well approximated using only 1 and 2 order shim terms, which is the most common shim coil configuration on MR scanners. High levels of the B0 uniformity can be achieved in small