FMRI using high flip-angle alternating steady state balanced SSFP supported by Monte Carlo studies

Introduction The need exists for a functional MRI (fMRI) technique capable of artifact-free whole brain coverage, with good blood oxygenation level dependent (BOLD) contrast to noise ratio (CNR), and high temporal resolution. Conventional gradient echo (GRE) BOLD fMRI suffers from signal dropout in regions of magnetic field inhomogeneity, limiting coverage, while spin echo fMRI has reduced BOLD CNR and lower temporal resolution than GRE. Recently, passband balanced SSFP (pbSSFP) has attracted interest as a functional imaging technique because it can provide artifact-free whole brain coverage with good BOLD CNR [1]. However, an experimental implementation of pbSSFP fMRI with good temporal resolution has yet to be developed. PbSSFP images contain characteristic low signal banding artifacts at specific off-resonance frequencies, with locations that can be shifted by altering the RF phase cycling increment. Hence two pbSSFP images with complementary RF phase cycling increments, termed ‘on-resonance’ and ‘off-resonance’ acquisitions, can be combined, via maximum intensity projection (MIP), to produce a single artifact-free image. Alternating between onand off-resonance acquisitions induces transient oscillations in the MRI signal that necessitate a delay in image acquisition to avoid phase encode ghosting. Current whole brain pbSSFP fMRI implementations avoid these oscillations by acquiring a time course of on-resonance images followed by a time course of off-resonance images. However, the artifact free images thus produced consist of two images acquired several minutes apart, resulting in poor temporal resolution. To achieve temporal resolution typical of GRE based fMRI (15s), the paired onand off-resonance images must be acquired sequentially by alternating between the two steady states. Further investigation is needed to determine if alternating steady state pbSSFP (altSSFP) acquisitions can maintain BOLD contrast and signal stability while providing the temporal resolution typical of GRE fMRI. This study used a Monte Carlo model to investigate BOLD signal and contrast characteristics in altSSFP, with the goal of optimizing acquisition parameters. Simulations of RF catalyzed sequential on and off-resonance acquisitions were performed at 4T, with volume times (defined as Tvol, the time to collect both images) varying from 1-5s. Simulation results showed that BOLD contrast decreases as Tvol is decreased in altSSFP acquisitions. Interestingly, the use of higher flip angles (45-60°) than those typically used for artifact suppression in brain tissue (25-30°) produced larger and more uniform BOLD contrast with off-resonance conditions. This was observed for both altSSFP and conventional pbSSFP simulations. In summary, simulations suggest that high flip angle (45-60°) altSSFP acquisitions preserve up to 90% of the BOLD contrast observed in conventional pbSSFP acquisitions while providing good temporal resolution (2-3s volume times). This demonstrates the potential of altSSFP for fast, high CNR, whole brain fMRI acquisitions.