Mohan Lal Jayatilake1, Christoph Juchem 2, Michael Mullen1, Gregor Adriany1, Robin de Graaf2, and Michael Garwood1
1Center for Magnetic Resonance Research, University of Minnesota, Minnesota, MN, United States, 2Department of Radiology and Biomedical Imaging, Yale University School of Medicine, New Haven, CT, United States
Synopsis
A highly uniform magnetic
field (B0) is typically required to generate MR images. In the original STEREO
(for STEering REsonance over the Object) method, spatial variations in B0 are compensated
by adjusting pulse amplitudes and frequencies in a temporal manner. Here we present
a novel design of a
limited set of well-defined multi-coil (MC) arrays that can optimize magnetic
field distortions across the object of interest. Introduction
Standard approaches to generating MR images
require a highly uniform magnetic field (B0) and thus necessitate the use of
magnets that are physically large relative to the object to be imaged, leading
to the high cost of magnet construction and siting. Previously we
proposed an approach called STEREO (for STEering REsonance over the Object) as
a means to generate images using magnets with drastically reduced B0 uniformity
[1]. Unlike standard Fourier imaging techniques, STEREO exploits full spatiotemporal
encoding, and images are reconstructed by a non-standard approach (e.g., by
solving the inverse problem). In the original STEREO method, spatial variations
in B0 and radiofrequency field (B1) are compensated by adjusting pulse
amplitudes and frequencies in a temporal manner. In the present study, we
investigated whether it is possible to use the dynamic multi-coil technique (DYNAMITE)
[2] as a means to create spatiotemporal excitation for STEREO. We aimed to design a limited set of well-defined MC
arrays able to strongly reduce the magnetic field distortions across the object
of interest
Materials and Methods
After
selection of center positions (x0C,
y0C, z0C) of the current carrying circular coils on the
cylindrical surface (Fig 1), the z-component of the
magnetic field generated by the circular coils (BZ) per unit current was then derived analytically by
use of the Biot-Savart law, which is adopted from Juchem’s work [3].
$$B_{Z}(x,y,z)=
\frac{\mu_0.n}{2\pi}.\sum_c\left(\frac{(y-y_{0C})}{\sqrt{((x-x_{0C})^2 +
(z-z_{0C})^2 )[(r + \sqrt{(x-x_{0C})^2 + (z-z_{0C})^2})^2 +(y - y_{0C})^2)}}
\right)$$
$$\times \left((-K(\kappa^2) +\frac{(r^2 + (x-x_{0C})^2 + (y-y_{0C})^2
+ (z-z_{0C})^2)}{[(r - \sqrt{(x-x_{0C})^2 + (z-z_{0C})^2})^2 +(y -
y_{0C})^2)]}.E(\kappa^2)\right)$$ (1)
Here μ0 is the magnetic field
constant, n is the
number of turns, r is the
radius of the coil, c is the coil
number, and x, y, and z are the Cartesian coordinates describing three-dimensional space
[4]. K(κ2) and E(κ2) refer to the complete elliptical
integrals of the first and second kind, respectively. Equivalent values in
Hertz were derived using the Larmor equation for protons (1H). A simulated, uncorrected magnetic induction field, ΔHZ, was generated within the 256x256x256mm FOV and
64x64x64 matrix. Assuming constant coil currents, ΔHZ was then projected into the magnetic field basis, BZ,
to obtain the corresponding current values, IC,
in each individual coil to compensate ΔHZ,
within a spherical region of interest (ROI) with a 10-mm radius. The estimation
of IC (see Eq.(2)) was
based on the Least-squares fitting based on the
Levenberg-Marquardt method. The current values in each coil is,
$$I_{C} = [B_{Z}(x,y,z)^{T}. B_{Z}(x,y,z)]^{-1}. B_{Z}(x,y,z)^{T}. \triangle H_{Z}(x,y,z)$$ (2)
The
compensation of field homogeneity was extended by moving a ROI along the spiral
trajectory (Fig 2A), in clock-wise direction at 4000 different time points
during the period of 16 ms in synchrony with traversing shimmed region. A
uniform B1 pulse of 5.9 μT was introduced
simultaneously with modulated coil currents for duration of 16 ms. The Bloch equations were carried out to predict
the spatiotemporal excitation profile of the B1 pulse. The maximum transverse
magnetization (Figure 2B) and maximum flip angle were evaluated during the
whole B1 pulse for each pixel. Further, by introducing a linearly
varying B1 pulse (Figure 2D), the Bloch simulations were carried out to
evaluate the maximum transverse magnetization, MXY (Figure 2C) [5]. Potential relaxation effects were
not considered in the simulations.
Results
Fig 2(A) shows selected spherical ROIs along the
spiral trajectory.
Fig 2(B) illustrates the maximum
MXY produced in the 64x64 matrix in the presence of
a uniform 5.9 uT B1 field. It can be seen that uniform excitation along the
traversed spiral trajectory is achieved. Above a threshold B1 amplitude, the
method is insensitive to B1 field inhomogeneity, as demonstrated in Fig 2(C)
which shows the
MXY produced by the linearly varying B1
shown in Fig2(D).
Discussion
In this work, simulations have provided
proof-of-concept for a unique way of exciting spins uniformly by dynamically
driving the coil currents in a multi-coil array with DYNAMITE while applying RF
irradiation. The next step will involve experimental validation of this
spatiotemporal excitation as a means to accomplish MR imaging in extreme field
inhomogeneity.
Acknowledgements
NIH grant R24 MH105998. References
[1] Snyder et al., Magn. Reson. Med. 72,49–58 (2014).; [2] Juchem et al., J Magn Reson.204:281-9 (2010).; [3] Juchem et al., Magn. Reson. Med. 63,171–180 (2010).; [4] Schill et al., IEEE Trans Magn; 39:961–967 (2003).; [5] Jang et al., Magn. Reson. Med. (2015).