I. Saha, J. Moore, S. Sengupta, S. Pradhan, J. M. Joers, and J. C. Gore Vanderbilt University Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, Department of Radiology and Radiological Sciences, Vanderbilt University, Nashville, TN, United States, Department of Physics and Astronomy, Vanderbilt University, Nashville, TN, United States, Department of Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, Department of Radiology, Children’s Hospital of Wisconsin, Milwaukee, WI, United States, The Medical College of Wisconsin, Milwaukee, WI, United States

Introduction Accurate field mapping is integral to the success of ultra-high field MRI. Knowledge of the spatial variations in transmitted RF fields (B1) allows for the calibration of RF pulse amplitude, guides the design of adiabatic and other B1-insensitive pulses, and enables real-time pulse designs that capitalize on parallel transmission technologies (e.g., spectral-spatial excitations and RF shimming). Maps of the receive field (B1) permit post-proccessing image intensity corrections and facilitate the design of RF coils for high-field applications. Static field variations (ΔB0) maps are used in an array of applications from RF pulse calibration to post-processing image distortion correction. One of the challenges currently facing high-field MRI research is the fast acquisition and calculation of all such field map data without sacrificing accuracy of the measurements. Here, we demonstrate a workflow for measuring/calculating all relevant fields at 3×3×5 mm resolution through the entire human brain in less than three minutes at 7 T. This approach differs from previously reported fast mapping protocols in that a single-shot EPI readout is employed such that a multi flip-angle fitting technique can be utilized to calculate the RF fields without the time penalty associated with long TR, non EPI sequences.

Fig. 3: Images acquired at 7 T with a GRE sequence (TR = 5s) using sinc excitations (left) and the composite pulse excitations from Fig. 2 (middle). Line profiles (right) reflect signals from the two pulse types after adjustment for the slice profile differences reported in Fig. 2. Results for both 45° (top row) and 90° (bottom row) pulses are included. Fig. 2: Slice profiles as measured in a mineral oil phantom for a Gaussian-modulated sinc pulse (left), and the 45° (middle) and 90° (right) optimized pulses from Fig. 1. B1-insensitive slice-selective pulses constructed from optimized non-selective composite waveforms

sub-pulses, each with 32 μs duration; (b) optimized phase modulation waveform; (c) simulated values of transverse magnetization for the optimized pulse over a range of relative B1 magnitudes (horizontal axis) and frequency offsets (vertical axis); (d) simulated frequency profiles for various relative B1 magnitudes. A slice-selective B1-insensitive composite pulse design for improved excitation uniformity at 7 T

Purpose At high magnetic fields, the wavelengths of the RF pulses transmitted in MRI are comparable to the size of the human torso/head [1], which gives rise to inhomogeneous B1 fields. Variations of the RF field can be reduced (for example) by the use of tailored slice-selective pulses optimized for the desired excitation profile [2] based on spokes or an optimized composite pulse with slice selection [3]. A refocusing pulse is often needed in various imaging scenarios. Solutions currently used in use involve adiabatic refocusing pulses, or sequences of sinc-pulses susceptible to field inhomogeneities. A design of a composite refocusing pulse suitable for use in human imaging at 7T is presented here. With the assumption that it is preceded by a slice-selective excitation, the refocusing solution is immune to inhomogeneities within a predefined universal parameter space of B1 and ΔB0 values (UPS).

Fig. 2 Results for three HS pulses: a high bandwidth pulse (row 1), a reduced bandwidth pulse (row 2), and an optimized pulse with reduced bandwidth (row 3). Flip-angle maps are calculated on the grid of targeted static and RF field variations (column 1) and within an axial slice of the human brain (column 2). Frequency profiles for B1/ B1,nom intensities of 0.35, 0.50, 0.75, and 1.00 are shown in column 3 with AM and FM waveforms in column 4. Hyperbolic secant parameter optimization for non-selective inversion at 7 T