Abstract
PURPOSE: Accuracy investigation of volumetric navigators for motion correction, with emphasis on geometric EPI distortions at ultrahigh field. METHODS: High-resolution Dixon images were collected in different head positions and reconstructed to water, fat, T(2) *, and B(0) maps. Resolution reduction was performed, and the T(2) * and B(0) maps were used to apply effects of TE and EPI distortions to simulate various volumetric water and fat navigators. Registrations of the simulated navigators were compared with registrations of the original high-resolution images. RESULTS: Increased accuracy was observed with increased spatial resolution for non-EPI navigators. When using EPI, the distortions had a negative effect on registration accuracy, which was most noticeable for high-resolution navigators. Parallel imaging helped to alleviate those caveats to a certain extent, and 5-fold acceleration gave close to similar accuracy to non-EPI in most cases. Shortening the TE by partial Fourier sampling was shown to be mostly beneficial, except for water navigators with long readout durations. The EPI blip direction had an influence on navigator accuracy, and positive blip gradient polarities (yielding mostly image stretching frontally) typically gave the best accuracy for water navigators, whereas no clear recommendation could be made for fat navigators. Generally, fat EPI navigators had lower accuracy than water EPI navigators with otherwise similar parameters. CONCLUSIONS: Echo planar imaging has been widely used for MRI navigators, but the induced distortions reduce navigator accuracy at ultrahigh field. This study can help protocol optimization and guide the complex tradeoff between resolution and EPI acceleration in navigator parameter setup.