3rd order DSU pre-emphasis calibration using spatio-temporal field monitoring

Purpose/Introduction: The promising advantages resulting from a transition to ultra-high-field in MR applications, such as higher SNR, spectral resolution and stronger BOLD-effect, come along with problems, like B 1 - and B 0-inhomogeneities, among others. An auspicious approach to cope with the latter is a slice-wise updated higher order shim (DSU), during a running sequence [1]. Fast switching of shim currents causes eddy currents within the shim coils themselves and in surrounding structures, which in turn give rise to severe field distortions. Therefore a careful pre-emphasis-calibration is necessary. This work presents an iterative pre-emphasis-calibration for a 3rd-order DSU on a 7T human MR system, using a 3rd-order field camera [2,3] to monitor the spatio-temporal field-evolution, which is faster and more precise compared to previously shown FASTERMAP algorithms [4,5]. 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 and a DSU unit (Resonance Research Inc., Billerica, USA). The field camera is based on 16 NMR probes equally distributed on a sphere with a diameter of 20 cm. This setup allows for observing all 16 field components at microsecond resolution. The time dependent field distributions were recorded, while switching currents in each shim coil sepa- rately. The eddy current induced decays were fitted by a tri-exponential fit in MATLAB, to calculate the time constants that were used to initialize the iterative pre-emphasis approach. After adjusting capacitances within the DSU unit to set time ranges, the time constants and amplitudes for the preemphasis currents could be refined via a software interface. In this way eddy current compensation in the switched coil itself, as well as cross terms to Z0 was implemented. Due to power limitations some of the 3rd-order pre-emphasis terms could not be fully optimized, and also cross terms to 1st order terms could not be addressed yet. Results: Figures (1) and (2) show the time dependence of some shim fields during and after switching. As shown, the optimized pre-emphasis reduces the eddy currents to a minimum and the deviation from the desired pulse shape is less than 2 after 2ms. Although the pre-emphasis calibration is not perfect, yet (fig. (2c)), the results of in vivo measurements show a significant improvement over measurements acquired with a global shim (fig. (3a,d))