Abstract
Purpose/Introduction: The use of bSSFP to detect BOLD was
introduced in 2001 [1], and since then, different variants of bSSFP have been investigated [2]. At 9.4T, the original stopband method [1], which is based on BOLD-related frequency and T2-changes, is too instable, as breathing or small motion induces frequency shifts of 20–100 Hz. But even for passband bSSFP, a very short TR is mandatory at 9.4T to minimize banding artifacts, especially for
whole-brain coverage. The goal of this study is to compare BOLDrelated signal contributions in 3D FLASH and bSSFP acquisitions.
Subjects and Methods: All experiments were performed at 9.4T on two healthy volunteers with IRB approval. A custom-built head coil [3] was used for signal transmission/reception. Slab-selective bSSFP and FLASH experiments were performed at 1 mm resolution with two slab thicknesses (18 mm and 36 mm). The acquisition time was
kept constant by adjusting the slice oversampling factor (100 and 0 in the 18 mm and 36 mm experiments, respectively). Other parameters: TR = 3.8 ms, TE = 1.9 ms, FA = 12, GRAPPA R = 3, 6/8 partial Fourier (POCS reconstruction), volumeTA = 3.25 s. In addition, whole brain experiments with nonselective excitation were performed at 1.2 mm resolution (TR = 2.6 ms, TE = 1.3 ms, FA = 12, CAIPIRINHA [6] R = 2 9 4D=2, 6/8 partial Fourier, volumeTA = 3 s). The fMRI paradigm consisted of a flickering radial checker-board which was presented ten times in alternating six frames off- and on-periods. For
analysis, the data were processed with FSL FEAT [7].
Results: Despite the short TE, strong activation is observed in the slab-selective FLASH experiments, whereas activation in the nonselective FLASH experiment is only visible in large vessels. In contrast, strong BOLD activation can be found in all three bSSFP experiments (Fig. 1).
Discussion/Conclusion: The FLASH acquisitions show a significant inflow effect for small slabs, and almost no activation for whole-brain excitation. This indicates that the T2*-related contributions to BOLD are negligible for FLASH, and thus also for bSSFP. From this can be concluded that the observed BOLD activation in slab-selective bSSFP
is related to inflow, T2 and diffusion effects, whereas whole-brain bSSFP shows a pure T2 and diffusion contrast similar to a spin-echo. This is also supported by Monte-Carlo simulations and thus suggests a much more confined effect to small vessels as compared to T2*-
weighted acquisitions. This also indicates a reduced sensitivity to larger draining veins and thus higher spatial specificity.
