Franck Girard, PhD
Valentine L. Marcar, DPhil
Franciszek Hennel, PhD
Ernst Martin, MD
Index terms:
Magnetic resonance (MR), pulse
sequences
Magnetic resonance (MR), technology
Radiology 2000; 216:900 –902
Abbreviation:
RARE = rapid acquisition with
relaxation enhancement
1
From the Department of Magnetic
Resonance, University Children’s Hospi-
tal Zurich, Steinwiestrasse 75, CH-8032
Zurich, Switzerland (F.G., V.L.M., E.M.)
and the Commissariat a L’Energie Atom-
ique—Service Hospitalier Fre ´de ´ric Joliot,
Orsay, France (F.H.). Received January 4,
1999; revision requested February 23;
final revision received December 28; ac-
cepted January 12, 2000. Supported in
part by grant 31-50734.97 from the
Swiss National Science Foundation and
grant BBW: 95.0314 from the Swiss Fed-
eral Office for Education and Science.
Address correspondence to F.G. (e-mail:
frank.girard@kispi_unizh.ch).
© RSNA, 2000
Author contributions:
Guarantor of integrity of entire study,
F.G.; study concepts, F.H.; study design,
F.G.; definition of intellectual content,
F.G.; literature research, F.H.; clinical
studies, F.G., V.L.M.; experimental stud-
ies, F.G., V.L.M.; data acquisition, F.G.,
V.L.M., E.M.; data analysis, F.G.; manu-
script preparation and editing, F.G.;
manuscript review, V.L.M.
Anatomic MR Images
Obtained with Silent
Sequences
1
The authors evaluated silent mag-
netic resonance (MR) imaging se-
quences for their suitability in provid-
ing high-spatial-resolution anatomic
images that are of sufficient quality
to be useful in a clinical setting. The
authors compared the images ob-
tained with a silent rapid acquisition
with relaxation enhancement (RARE)
sequence to its standard counterpart
with respect to signal-to-noise ratio,
distribution of gray level, and spatial
resolution. No real differences were
observed between the standard and
the silent RARE MR images. Anatomic
images were also acquired with a silent
spin-echo sequence. Acoustic noise
levels with the silent sequences were at
least 22 dB (A-weighted scale) lower
than those with standard sequences,
without loss of image quality.
Magnetic resonance (MR) images are pro-
duced by applying radio-frequency pulses
together with magnetic field gradients.
The rapid switching of these gradients
creates an induced Lorentz force, which
in turn causes vibrations in the gradient
coil structure. These vibrations are per-
ceived as loud acoustic noise (1). Previous
studies showed that levels around 80 dB
on the A-weighted scale could be reached
with standard rapid acquisition with re-
laxation enhancement (RARE) sequences
(2,3), and levels up to 115 dB are not
uncommon with echo-planar imaging
sequences (4). These acoustic noise levels
occasionally approach the limits set by
almost all safety boards, implying that
ear protection should be used by patients
and staff working near the magnet tube
to avoid long-term hearing loss (5). The
gradient acoustic noise during an MR im-
aging examination is not only a major
source of discomfort, it is the principle rea-
son for the waking of children who had to
be sedated, which invariably means the ex-
amination must be abandoned. For anx-
ious patients (6), those who experience
claustrophobia, and children, this acoustic
noise is an additional distress factor.
In some instances, the gradient acous-
tic noise also represents an obstacle to
performing specific functional measure-
ments, particularly those involving audi-
tory stimuli (7–9). Finally, the acoustic
noise level generated by a sequence can
be expected to increase with increasing
field strength (1). Reduction in the acous-
tic noise level will become an important
issue when field strength as high as 8 T is
used for imaging purposes.
Few useful solutions have been offered
to overcome this problem. Acoustic noise-
cancellation devices have been designed
(10,11) and specific gradient coils have
been constructed (12,13), but these ap-
proaches are costly. From the point of view
of adapting imaging sequences, little work
has been done. One approach along these
lines was the application of a reduced slew
rate in functional MR imaging (14); another
was the replacement of gradient pulsing
with a mechanically rotated coil (15). The
latter solution is currently restricted to use
in the spin-echo sequence and may not be
implemented with standard systems, espe-
cially since use of a rotating coil may be
considered a potential safety hazard.
An approach to reduce the acoustic
noise level is use of “silent” MR imaging
sequences (1), which are based on the
linear acoustic response of the gradient
system (16) (ie, the imager acoustic noise
spectrum is given by the product of the
spectrum of the gradient waveform and
the frequency response function of the
system). Silent sequences make use of
“soft” gradient waveforms that contain
no frequencies for which the frequency
response function is high. These meth-
ods are implemented on the basis of the
three rules of Hennel et al (1), that is,
sinusoidal gradient slopes, maximum slope
duration, and minimum number of slopes.
We acquired images with silent RARE,
900