Characterizing physiological noise in the brainstem: passband SSFP vs. GRE-EPI

Introduction. Functional MRI of the brainstem is complicated by: 1) small grey matter nuclei, 2) severe distortions due to susceptibility boundaries and 3) high levels of physiological noise, all of which make conventional GRE-EPI BOLD problematic. SSFP fMRI has the potential to produce high-resolution images with reduced distortions [1-3], and recent work reported reduced physiological noise for SSFP in cortical regions [5]. SSFP may therefore be amenable to fMRI of the brainstem and other “hard to image” regions of the brain. However, this previous study [5] had several shortcomings if one is interested in brainstem fMRI: 1) SSFP was compared to a short TE/TR variant of GRE (i.e., spoiled gradient echo, or SPGR) rather than GRE-EPI, 2) no inferior regions were covered and 3) no information was provided about the relative contribution of respiratory and cardiac fluctuations. The present work addresses these issues by comparing different physiological noise components in passband SSFP, SPGR (i.e., GRE with acquisition matched to SSFP) and GRE-EPI.