Abstract
Purpose/Introduction: The application of ultra-high static magnetic field strengths to functional MRI (fMRI) yields important advantages, such as higher SNR and stronger BOLD-contrast. However, the increasing field strengths
come along with problems, e.g. B0 and B1 inhomogeneities, as well. To address the B0 inhomogeneity dynamic shim updating (DSU) with an excellent eddy-current compensation (ECC) proved to be a promising approach [1].
This work presents the application of 3rd order DSU with full ECC to echo planar imaging (EPI), which is the basis for BOLD-contrast fMRI.
Subjects and Methods: All measurements were performed on a 7T Philips Achieva whole-body system (Philips Healthcare,Cleveland,USA), equipped with a full set of 3rd order shim coils. A DSU LoadGo Unit (ResonanceR
esearchInc.,Billerica,USA) was connected directly to the higher order shim amplifiers. The linear shim terms are driven by the gradient amplifiers, which could not be directly assessed by the LoadGo Unit. Therefore an in-house
build gradient driver box has been constructed, which sums up the voltages from the LoadGo Unit and the driving voltages for the gradient amplifiers.
As fast switching of the shims induces eddy-currents in the shim coils and the surrounding conductive structures, which lead to severe long-lasting B0 field distortions, a careful pre-emphasis calibration of the system is necessary.
This has been performed using a spatio-temporal field-monitoring approach, as described previously [2]. In order to actively compensate for eddy-currents, the LoadGo unit overshoots the nominal driving voltage on a short timescale.
To maintain the systems functionality, the shim strengths of the 2nd and 3rd order shims therefore had to be reduced to 73 to 90 and 43 to 53 of their maximum values, respectively.
Results: In vivo multi-slice EPIs of the brain (TE/TR=28ms/3000ms,EPI factor=99) have been acquired in two healthy volunteers (fig.(2)), once using a static global shim, determined by a vendor pre-implemented routine closely related to FASTESTMAP, and once using DSU where the shim is updated 5ms before the excitation of each slice (fig.(1)).
Discussion/Conclusion: The setup and calibration used enabled for very fast switching of shims, with a settling of residual eddy-currents within less than 2ms. Therefore, switching of the shims within a running sequence only 5ms
before the excitation was possible, contrary to more than 60ms as reported in earlier publications [3,4], allowing for slice-wise dynamic updated shimmed EPIs and a significant gain in image quality, without the cost of prolonged scan time.
