Seung-Kyun Lee1, Kishore V. Mogatadakala2, Dominic Graziani1, Jean-Baptiste Mathieu2, Thomas K.-F. Foo1, and Matt A. Bernstein3
1GE Global Research, Niskayuna, NY, United States, 2GE Healthcare, Florence, SC, United States, 3Mayo Clinic, Rochester, MN, United States
Synopsis
Head orientation dependence of the peripheral nerve stimulation (PNS)
thresholds and the induced electric fields of a high-performance, asymmetric
head-only gradient coil were studied experimentally and by numerical simulation.
In the experiment, the gradient field direction was fixed and the subject head
was rotated in the transverse plane. The subject-reported PNS thresholds nearly
doubled when the head's anterior-posterior direction was parallel to the
gradient compared to when the head was approximately perpendicular to the
gradient. Human-body-model simulation suggested that the orientation dependence
may be primarily due to locally concentrated electric fields in the corrugated
regions of the face.Introduction
A compact, asymmetric head-only gradient coil with improved cooling and
electromagnetic design has been shown to enable fast, high-quality imaging at
slew rates up to 700 T/m/s [1-5] with reduced PNS compared to a whole-body
gradient coil. Whereas the PNS advantage of a compact neuroimaging gradient
coil is well-documented [6,7], the dependence of the PNS thresholds on the gradient
direction in the head has been only sporadically reported in the literature [7,8]
with limited quantitative explanation. In particular, the stronger PNS from a
left-right (L/R) gradient coil as compared to an anterior-posterior (A/P) coil
has been attributed to the difference in the cross sectional areas in the
sagittal vs coronal mid-plane in the head [7]. However, this does not seem to
fully explain a marked suppression of A/P-gradient PNS observed consistently on
many subjects [1,4,5] with different head aspect ratios. Here we present an
experiment in which the PNS thresholds of an asymmetric head-only gradient coil
were measured as a subject rotated his head while the gradient direction was
fixed in space. A numerical simulation probed electric field (E-field) hot
spots at different head orientations.
Methods
The
PNS measurement was conducted on a healthy male volunteer under an IRB-approved
protocol using a compact head-only gradient coil described in [1-3] inside a
whole-body 3T scanner (GE MR750W, Fig 1a). A train of bipolar trapezoidal
gradient pulses (Fig. 2), with 1 ms flat-top time, a variable zero-to-peak
amplitude ΔG and a variable zero-to-peak rise-time
Δt,
were applied to both the X (horizontal) and the Y (vertical) coils
simultaneously, creating a 45°-oblique gradient field. The subject reported
the onset of sensation as ΔG was incremented at one of three
rise-times. The maximum ΔG was limited to 85 mT/m. The
experiment was repeated as the subject rotated his head about the Z-axis by three
different angles. After each measurement at a given angle, the subject was
imaged with a 3-plane localizer (FSE, TR/TE/FOV/slice = 706 ms/79 ms/28 cm/8
mm) to record the head position in the gradient coil (Fig. 1b).
Numerical
simulation was performed on a human body model (electrical conductivity = 0.2
S/m) in Comsol Multiphysics (Burlington, MA, USA) (Fig. 3a,b). The gradient
coil (X) was modeled after the electromagnetic design of the head-only gradient
coil, and was driven with a 1 kHz sine-wave with 660A peak current. The induced
E-fields at different regions of the head were recorded as the human model was
rotated around the Z axis with respect to the gradient coil in 5° steps.
Results
Figure (1c) shows the PNS thresholds at different angles and rise-times.
The thresholds when the head was approximately parallel to the gradient field (θ = 10°) was nearly twice the
thresholds at θ = 75° for the two shortest
rise-times. The difference in thresholds between θ = 44° and θ = 75° was not as dramatic
as the difference between θ = 44° and θ = 10°, indicating
relatively sharp drop in PNS as the head aligns with the gradient direction. Figure
(3c) shows the orientation dependence of the simulated E-field magnitudes. The
E-field hot-spot occurs in the nose-bridge area for the PNS-sensitive
orientation (i.e., when the head and the gradient are orthogonal, θ = ±90°), which is
consistent with the PNS locations reported by a majority of subjects in earlier
studies [1,4]. The concentrated E-field in this area exhibits a significant
drop when the head aligns with the gradient field (Fig. 3c).
Discussion and Conclusion
The
PNS thresholds of a compact asymmetric head gradient coil depend strongly on
the relative orientation of the head with respect to the gradient field in the
transverse plane [7,8]. Unlike in previous studies we obtained the experimental
data by rotating the head while the gradient direction was fixed in space, thereby
eliminating any possibility that different X, Y gradient coil designs
contribute to the dependence. A human-body-model simulation suggested that the
strong orientation dependence of the E-fields concentrated in certain corrugated
areas of the face, e.g., the bridge of the nose, could be the main contributor
to the observed PNS characteristics of a compact head gradient coil. Extension
of the present work to include other degrees of freedom in head
position/orientation and more sophisticated head models [9] is currently
planned. The high PNS thresholds of the head gradient coil in the A/P direction
can be exploited for high-speed Cartesian readout such as EPI.
Acknowledgements
This work was supported in part by the NIH grant 5R01EB010065. The views
herein do not necessarily represent those of NIH.References
[1] Mathieu et al, ISMRM 2015, #1019. [2] Huston et al, ISMRM 2015,
#971. [3] Huston et al, ASNR 2015, #O-232. [4] Lee et al, ISMRM 2014, #310. [5]
Lee et al, Magn Reson Med (accepted). [6] Chronik and Rutt, Magn Reson Med
46:386 (2001). [7] Schmitt et al, Echo-planar imaging: theory, technique, and
application. Springer (1998), p212. [8] Goodrich et al, Concepts Magn Reson
44B:66 (2015). [9] Feldman et al, Magn Reson Med 68:1973 (2012).