Region and volume dependencies in spectral line width assessed by 1H 2D MR chemical shift imaging in the monkey brain at 7 T.

High magnetic fields increase the sensitivity and spectral dispersion in magnetic resonance spectroscopy (MRS). In contrast, spectral peaks are broadened in vivo at higher field strength due to stronger susceptibility-induced effects. Strategies to minimize the spectral line width are therefore of critical importance. In the present study, 1H 2D chemical shift imaging at short echo times was performed in the macaque monkey brain at 7 T. Large brain coverage was obtained at high spatial resolution with voxel sizes down to 50 microl being able to quantify up to nine metabolites in vivo with good reliability. Measured line widths of metabolites decreased from 14.2 to 7.6 Hz with voxel volumes of 3.14 ml to 50 microl (at increased spatial resolution). The line width distribution of the metabolites (7.6+/-1.6 Hz, ranging from 5.5 to 10 Hz) was considerably smaller compared to that of water (10.6+/-2.4 Hz) and was also smaller than reported in 1H MRS at 7 T in the human brain. Our study showed that even in well-shimmed areas assumed to have minimal macroscopic susceptibility variations, spectral line widths are tissue-specific exhibiting considerable regional variation. Therefore, an overall improvement of a gross spectral line width--directly correlated with improved spectral quality--can only be achieved when voxel volumes are significantly reduced. Our line width optimization was sufficient to permit clear glutamate (Glu)-glutamine separation, yielding distinct Glu maps for brain areas including regions of greatly different Glu concentration (e.g., ventricles vs. surrounding tissue).