Pediatric Radiology (2021) 51:716–723

https://doi.org/10.1007/s00247-020-04894-9

MINISYMPOSIUM: PEDIATRIC MRI QUALITY AND SAFETY

Components of a magnetic resonance imaging system and their

relationship to safety and image quality
Suraj D. Serai 1,2 & Mai-Lan Ho 3 & Maddy Artunduaga 4,5 & Sherwin S. Chan 6,7 & Govind B. Chavhan 8,9

Received: 17 June 2020 / Revised: 9 September 2020 / Accepted: 19 October 2020 / Published online: 15 April 2021
# Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract
Magnetic resonance imaging (MRI) is a powerful diagnostic tool that can be optimized to display a wide range of clinical conditions.
An MRI system consists of four major components: a main magnet formed by superconducting coils, gradient coils, radiofrequency
(RF) coils, and computer systems. Each component has safety considerations. Unless carefully controlled, the MRI machine’s strong
static magnetic field could turn a ferromagnetic object into a harmful projectile or cause vertigo and headache. Switching magnetic
fields in the gradients evokes loud noises in the scanner, which can be mitigated by ear protection. Gradients also generate varying
magnetic fields that can cause peripheral nerve stimulation and muscle twitching. Magnetic fields produced by RF coils deposit energy
in the body and can cause tissue heating (with the potential to cause skin burns). In this review, we provide an overview of the
components of a typical clinical MRI scanner and its associated safety issues. We also discuss how the relationship between the
scanning parameters can be manipulated to improve image quality while ensuring a safe operational environment for the patients and
staff. Understanding the strengths and limitations of these parameters can enable users to choose optimal techniques for image
acquisition, apply them in clinical practice, and improve the diagnostic accuracy of an MRI examination.

Keywords Children . Gradients . Image quality . Magnetic resonance imaging . Radiofrequency coils . Safety . System
components

Introduction complex relationship (Fig. 1). Understanding the interplay of

key components of the MRI system is foundational to acquir-
Magnetic resonance imaging (MRI) is a comprehensive diag- ing a diagnostic-quality MR image. MR image acquisition
nostic modality with high soft-tissue contrast resolution and requires four basic steps: (1) place the patient in a uniform
superior spatial and temporal resolution that can obtain ana- magnetic field generated by the combination of the MRI com-
tomical, metabolic, chemical and physiological data from the ponents, (2) displace the equilibrium magnetization vector
body [1]. An MRI system is composed of four main compo- with an RF pulse, (3) collect the signal as the magnetization
nents: main magnet, gradient coils, radiofrequency (RF) coils vector returns to equilibrium and (4) convert the collected
and computer systems that control and interface the various signals into images using the computer’s signal processing
components. Each component interacts with the others in a algorithms.

* Suraj D. Serai 5
Children’s Health Medical Center, Dallas, TX, USA
[email protected]
6
Department of Radiology, University of Missouri-Kansas City
1
School of Medicine, Kansas City, MO, USA
Department of Radiology, Children’s Hospital of Philadelphia,
3401 Civic Center Blvd., Philadelphia, PA 19104, USA 7
Department of Radiology, Children’s Mercy Hospital,
2 Kansas City, MO, USA
Perelman School of Medicine at the University of Pennsylvania,
Philadelphia, PA, USA
8
3 Department of Diagnostic Imaging, The Hospital for Sick Children,
Department of Radiology, Nationwide Children’s Hospital, Toronto, ON, Canada
Columbus, OH, USA
4 9
Pediatric Radiology Division, Department of Radiology, Department of Medical Imaging, University of Toronto,
University of Texas Southwestern Medical Center, Dallas, TX, USA Toronto, ON, Canada
Pediatr Radiol (2021) 51:716–723 717

magnetic field in extreme circumstances (e.g., in case of a fire

or if a metallic object reaches the magnetic field and endangers
the patient or the equipment). The quench vents all the liquid
helium out of the room to the outside via a quench pipe sys-
tem, resulting in a complete loss of the magnetic field.
The strength of a magnetic field is measured in units of
gauss (G) and tesla (T). One tesla is equal to 10,000 G. The
main magnetic field of a 1.5-T magnet is about 30,000 times
the strength of the earth’s magnetic field (which is approxi-
mately 0.5 G). The strength of the magnetic field affects the
detected signal by increasing the energy difference between
the ground (equilibrium) and excited states. A higher magnet-
ic field strength increases longitudinal magnetization because
more protons align along the main axis of the magnetic field,
increasing the signal-to-noise ratio (SNR). This improved
SNR can be used to generate images with improved spatial
resolution and to perform fast imaging acquisition. The higher
magnetic field strength not only affects SNR but also affects
ferromagnetic objects that have the potential to become dead-
ly. Although the 3-T field is twice the strength of the 1.5-T
field, it is difficult to determine when the same object will
become a hazardous projectile in scanners of these strengths.
The stray magnetic field outside the bore of the magnet is
known as the fringe field. MRI systems are passively and
actively shielded to confine the fringe field within the scan
room. Magnetic field strength less than 5 G is considered
inconsequential to patient safety. In most systems, the 5-G
Fig. 1 MRI system components and their relationship. a, b Block
diagram (a) and connectivity (b) of a typical MRI system. A/D analog field is confined within the scan room, so the fringe field
to digital, RF radiofrequency doesn’t affect any external area.
Because the static magnetic field is always on, the most
fundamental safety issues are related to two aspects: human
Magnet exposure to the magnetic field, and the risk of a strong mag-
netic field potentially creating projectiles in the MRI suite.
The magnet is the heart of the MRI system. Almost all mag- Available data indicate that the strength of the magnetic field
nets used for clinical pediatric imaging are superconducting used in current clinical MR environments is free of deleterious
magnets of 1.5 tesla (T) (low field strength) or 3.0 T (high bio-effects. The International Electrotechnical Commission
field strength) [2–4]. A superconducting magnet consists of a (IEC) and the Food and Drug Administration (FDA) consider
series of coils wound on a cylindrical form within a bath of standard clinical-use MRI (1.5- and 3.0-T scanners) as a “non-
liquid helium enclosed in a cryostat. The niobium-titanium significant risk” device when operated within specified pa-
alloy coiled wires become superconducting at temperatures rameters and provided that appropriate safe practice guide-
below 10 Kelvin (−263 °C). This temperature is achieved by lines are followed diligently [5, 6].
bathing the system in liquid helium, which in turn is thermally Because a typical clinical MRI is on at full strength at all
insulated by the cryostat, minimizing helium boil-off. When times, safety conditions must constantly be in place to protect
the magnet is started for the first time, the helium in the cryo- any person or object near the MRI suite. As stated, one of the
stat cools the superconducting coils to −269 °C. Then, the greatest potential hazards around the magnet is the missile
current is passed to the coils by an external power source in effect. In many cases, these objects are inadvertently carried
a process called “ramping up.” As long as the magnet is into the MRI zone by hospital personnel who lack MRI train-
cooled properly, the current flowing through the ing. Examples include MRI-incompatible or MRI-conditional
superconducting coils can maintain the magnetic field devices like anesthesia carts and cleaning equipment. The
strength. When the liquid helium is released from the system, static magnetic field has been shown to produce diamagnetic
the magnet heats up and loses its superconducting properties. forces leading to temporary vertigo, dizziness and metallic
This process, called a “quench,” can result from technical tastes in the mouth, though this occurs mainly in experimental
failures or can be initiated by MRI staff to shut off the settings and especially at higher field strengths [7, 8].
718 Pediatr Radiol (2021) 51:716–723

Therefore, it is important to move patients in and out of the pain. In the setting of standard clinical practice, peripheral
magnet bore slowly to reduce these transient effects, especial- nerve stimulation is not considered to be a significant danger
ly with higher-field systems. in MRI. However, at levels an order of magnitude beyond the
threshold, there is a theoretical potential for stimulation of the
heart and diaphragm. Because of this, the IEC and FDA have
Gradients used data on magnetic and electric nerve stimulation to define
safe peripheral nerve stimulation exposure thresholds based
The primary function of gradient coils, also simply called on age, weight and size [5]. Modern clinical-use MRI systems
“gradients,” is to allow spatial encoding of the MR signal. are designed to calculate dB/dt values for patients and permit
Gradients produce an additional magnetic field that varies in scanning only within FDA-specified safety standards.
its strength along its direction and is superimposed on the Another safety concern associated with gradients is noise.
main magnetic field. Producing an image from MR signals MRI scanners can generate loud noises when pulses of current
requires (1) choosing a specific slice within the patient’s body are pushed through the gradient coil during the spatial
to be examined and (2) determining the voxels to be designat- encoding of the MRI signal [9]. Because the gradient coil is
ed within the slice. MRI has three sets of magnetic field gra- placed inside a strong magnetic field, a pulsed Lorentz force is
dients: X, Y and Z (Fig. 2). For a standard clinical MRI, a induced, vibrating the coil structure, which in turn generates a
magnetic field gradient along the z-axis is the slice selection compression wave in the air perceived as the scanner noise
gradient. The Y-field gradient produces phase encoding with- [9]. The FDA limits the sound in the scanner to a maximum of
in the slice. The X-field gradient produces frequency encoding 140 dB. In currently used MRI systems, noise levels of 70–
within the slice. 110 dB are reported, depending on the acquisition method
Safety considerations for gradient coils are related to rapid [10, 11]. Higher magnetic fields, larger gradient amplitudes
switching of the time-varying gradient magnetic field (dB/dt). and faster gradient switching all increase noise. This causes
This can potentially induce electric fields in humans, leading discomfort and anxiety in patients, especially children, and
to peripheral nerve stimulation, which, if it exceeds a certain presents a severe obstacle in MRI studies [12]. Discomfort
threshold, can cause skin tingling, muscle twitching and even and ear damage can be prevented if noise levels are reduced

Fig. 2 Field gradients. a–c

Diagrams illustrate the field
gradients generated in each of the
X (a), Y (b) and Z (c) directions
Pediatr Radiol (2021) 51:716–723 719

to less than 80 dB. This can be accomplished through sound- Radiofrequency coils
attenuating materials like earplugs or noise-cancelling head-
phones. New methods of reducing noise are now becoming Radiofrequency coils are used to send RF pulses and receive
clinically available, such as “quiet gradient coils,” in which the signal back from the patient’s body. They are one of the
either the net Lorentz force is compensated between current most important determinants of SNR and signal uniformity
pathways or reduced gradient slew rate is applied [13–15]. [10]. The energy transmitted by RF pulses flips magnetization
The new reduced-noise quiet gradient acquisition methods away from the z-axis [5]. When the RF pulse tilts magnetiza-
use a radial center-out sampling scheme in which endpoints tion by 90° away from the z-axis, it is known as a 90° pulse.
of each spoke follow a spiral path in time (Fig. 3). The slower Tilted by 180°, it is called 180° RF pulse. RF coils can be
gradient direction changes in this sampling strategy result in transmitter, receiver or combined transmitter–receiver [10,
less noise. These techniques decrease sound intensity by 35– 11]. The magnetic field produced by RF coils (B1) is perpen-
40% [13]. Another advantage of spiral scanning is that the dicular to the main magnetic field (B0). Two basic types of RF
image is acquired faster than when using conventional k- coils are volume and surface coils. Volume coils cover large
space sampling techniques. The disadvantages of this method fields of view (FOVs). The main volume coil of the scanner,
of acquisition are a relative loss of image sharpness and in- typically a transmitter, is located in the bore as the innermost
creased streaking artifacts, which might be perceived as in- ring of the magnet (this is also called a body coil). Surface
crease in image noise or reduced SNR [16]. Another disad- coils are placed directly on the limited anatomy of interest and
vantage of quiet gradient techniques is longer acquisition time have small FOVs. The SNR improves as the imaging area is
as compared with the same sequence with higher slew rates. closer to the surface coil and decreases with distance away
The lower sound level from spiral acquisition techniques from the coil surface. The advantages of volume coils (large
could benefit pediatric, claustrophobic and non-cooperative coverage and homogeneous signal) and of surface coils (high
patients. signal) are combined in a phased-array coil. This coil is made
of multiple surface coils and attached to multiple receiver
channels; depending on the number of these elements,
phased-array coils range from 4 to greater than 64 channels.
Because they can spatially localize signal, these coils are used
for parallel imaging.
Radiofrequency transmit coils deposit energy into the body
in the form of heat by creating electric fields that interact with
electrically conductive tissues [5]. The amount of heat that can
occur in a patient is quantified by the specific absorption rate
(SAR), categorized as global and local [5]. Global SAR is
easily estimated by the MRI scanner computer’s power mon-
itors. Local SAR is difficult to estimate because it is influ-
enced by the local electromagnetic field (which varies in re-
sponse to field inhomogeneity) and by the distribution of the
conductive tissues [5]. As per IEC, the typical standard limi-
tation for global SAR in MRI is a maximum of 4 watts per
kilogram (W/kg) [5]. SAR is 4 times greater in a 3.0-T MRI
than in a 1.5-T MRI [5]. One strategy to minimize SAR is to
reduce the RF pulse power emitted by the transmission coil —
at the expense of longer acquisition times and changes in
contrast [5]. Another reason to reduce RF pulse power is to
lower the risk of skin burns and heating or interfering with
implanted devices [7].
Burns, often of second degree, are the most frequently
encountered adverse event with MRI [7]. The risk of
burns increases with higher SAR, but RF energy modifi-
cations can prevent raising core and local temperatures.
This is particularly important in patients with reduced
Fig. 3 Quiet gradient acquisition. a, b Schematic of a typical “silent”
pulse sequence (a) and acquisition of k-space (b) with spiral trajectory.
thermoregulation (e.g., obese), communication deficits or
Gx X-gradient, Gy Y-gradient, Gz Z-gradient, kx k-space along x-axis, ky implanted devices such as those with elongated shapes
k-space along y-axis, TE echo time (e.g., leads, electrodes), which are prone to becoming
720 Pediatr Radiol (2021) 51:716–723

heated [7]. Another risk of overheating relates to the Signal-to-noise ratio

shape of conductive materials. Conductive loops, which
can be circular, U-shaped or S-shaped, carry a much Signal-to-noise ratio measures the objective image quality in
higher risk of heating by electrical current induction from MRI protocols. Higher SNR typically indicates an image of
magnetic fields [7]. Burns can result from a bare cable greater clinical utility, whereas images with low SNR often
touching the patient’s skin or metallic implants, or from appear grainy. In MRI, the SNR, scan time and image resolu-
improper patient positioning (e.g., causing a loop with the tion are interdependent (Fig. 4). The signal detected by the RF
limbs) [7]. Conducting loops in tissue through the extrem- coil is dependent on several protocol parameters (Fig. 5).
ities and trunk can also be a rare source of burns. This can “Noise” is a general term describing pixel-to-pixel variations
be avoided by preventing skin-to-skin contact in the ex- in signal. The noise in a local region of interest can be de-
tremities. Careful extremity positioning with or without scribed by the standard deviation of the pixel signal values in
material between the extremities to break up the tissue that region. The SNR is equal to the ratio of the average signal
loops can mitigate this rare event. RF pulses can also intensity over the standard deviation of the noise. Image qual-
induce voltage and interrupt the pacing system of active ity can be improved by increasing SNR through a variety of
devices [7]. factors that may or may not be within the operator’s control,
including magnetic field strength, voxel size and number of
averages (Table 1). Increasing SNR by increasing the number
of excitations/averages (NEX/NSA) will directly affect the
Computer systems length of scan time (Fig. 6). The other way to increase the
SNR is by increasing slice thickness or voxel size at the ex-
The multiple computer systems embedded in an MRI scanner pense of reducing the spatial resolution, but with no change in
have a range of functions. Primarily they (1) control the RF the length of scan time.
and gradient pulses, (2) collect the data and (3) process and Other parameters affecting the SNR are the sequence used,
display the generated image. These computer systems work echo time (TE), repetition time (TR) and the flip angle. Longer
synergistically to control the RF and gradient systems, sample repetition times increase SNR but result in loss of the T1
the incoming signal at discrete time intervals, and process the effect. Conversely, the SNR decreases as the TE increases.
data to produce recognizable images. Automated post- With a short TE, the T2 contrast is lost. For this reason, the
processing analysis software can distinguish between frequen- option of shortening TE to improve SNR is available only for
cy domains to process data. The post-processing computer T1-weighted sequences.
analysis can be carried out in the frequency domain (i.e. fol- There is a threshold at which increasing SNR confers al-
lowing Fourier transformation of the free induction decay) or most no image-quality advantage to the observer, and any
alternatively in the time domain (i.e. by direct fitting of the excess of SNR might best be used as currency to pay for
free induction decay). The data are then converted to improvements in image quality, to increase spatial resolution,
displayed images. or to decrease scan time [17].

Contrast-to-noise ratio

Determinants of magnetic resonance image Contrast-to-noise ratio is a measure of how distinguishable

quality two structures are from each other. In clinical MRI, the CNR
between two structures is the difference in SNR between those
Several MRI parameters can be modified to produce the best two relevant tissue types (CNR = SNR1 – SNR2). CNR is
possible image quality while maintaining a reasonable data controlled by the same factors that affect SNR. Improving
acquisition time. These factors include SNR, contrast-to- CNR clarifies the distinction between two clinical areas of
noise ratio (CNR), spatial resolution, scan time, and number interest. If the CNR is too low, the details of the structure
of excitations/acquisitions (NEX/NSA). The four basic com- might be obscured by image noise.
ponents of the MR system have direct or indirect effects on the
SNR, scan time and imaging quality. Higher field strength Spatial resolution and scan time
provides more signal that can be used either for getting better
SNR or reducing the scan time. The gradient performance In MRI, the resolution is determined by the number of voxels
affects spatial resolution and imaging speed. RF coils are in a specified FOV. The higher the image resolution, the better
one of the most important determinants of the MR signal the small pathologies can be diagnosed. Spatial resolution is
and SNR. Phased-array coils provide better signal and SNR proportional to the number of voxels. Voxel size can be cal-
as compared to volume coils. culated by dividing the FOV by the matrix size (e.g., FOV:
Pediatr Radiol (2021) 51:716–723 721

Table 1 Relationship between an

imaging factor and its effect on Imaging factor Effect on MR image
MR image
Increase magnetic field strength Increase in SNR, increase in CNR
Increase in field of view Increase in SNR, decrease in spatial resolution
Increase in voxel/matrix size Increase in SNR, increase in partial volume effects,
decrease in spatial resolution
Increase TR (repetition time) Increase in SNR, increase in maximum allowable slices,
increase in scan time, decrease in T1-W contrast,
increase in T2-W contrast
Increase TE (echo time) Decrease in SNR, decrease in number of allowable slices,
decrease in T1-W contrast
Increase TI (inversion time) Short TI: null fat (STIR); long TI: null CSF (FLAIR)
Increase flip angle Increase in SNR
Increase slice thickness Increase in anatomical coverage, decrease in scan time,
increase in partial volume artifact, decrease in spatial resolution
Increase in slice distance/gap Increase in anatomical coverage, decrease in scan time,
decrease in SNR, increase in risk of missing pathology
Increase in matrix size (frequency) Increase in spatial resolution, no change in scan time, increase in SNR
Increase in matrix size (phase) Increase in spatial resolution, increase in scan time, decrease in SNR
Increase averaging (NEX) Increase in scan time, increase in SNR, decrease in motion artifacts
Increase in receiver bandwidth Decrease in SNR, no change in scan time, decrease in
chemical shift artifact
Increase in transmitter bandwidth Increase in SNR, decrease in spatial resolution
Utilization of partial k-space Decrease in SNR, decrease in scan time

CNR contrast-to-noise ratio, CSF cerebrospinal fluid, FLAIR fluid-attenuated inversion recovery, NEX number of
excitations, SNR signal-to-noise ratio, STIR short tau inversion recovery

320 mm, matrix: 256×256; voxel size: 320/256=1.25 mm). and resolution are manipulated to achieve the desired pixel size.
There are two spatial resolution parameters used in MRI for For a three-dimensional (3-D) image, scan time can be manip-
the production of a two-dimensional (2-D) image: the frequen- ulated by using slice thickness as well as by the frequency and
cy matrix and the phase matrix. The frequency matrix does not phase matrixes. In a 3-D acquisition, the third dimension is also
affect scan time, whereas the phase matrix size is proportional phase-encoded and therefore directly affects scan time.
to scan time. Thus, a phase matrix of 256 takes twice as long
to acquire as a phase matrix of 128. In general, phase matrix

Fig. 5 Diagram shows the interplay of protocol parameters directly

Fig. 4 Diagram illustrates that the signal-to-noise ratio (SNR), scan time affecting signal in MRI. FOV field of view, NEX number of excitations,
and image resolution are interdependent in a complex relationship RF radiofrequency, TE echo time, TR repetition time
722 Pediatr Radiol (2021) 51:716–723

Fig. 6 Formula for MRI

acquisition time. Sec seconds, TR
repetition time

Number of excitations (NEX) or number of acquisi- Aggressive parallel imaging can also decrease scan time,
tions (NSA) though it reduces SNR and can result in image artifacts if
applied too strongly.
The number of averages (i.e. signal averaging) is also referred
to as the number of excitations (NEX) or the number of ac-
quisitions (NSA). NEX is one of the most direct methods used Summary
to control SNR. In principle, each imaging cycle is repeated
and the results are averaged to form the final image. The An MRI scanner is composed of four major components: main
approximate amount of improvement in SNR is calculated magnet, gradient coils, RF coils and computer system, all of
as the square root of the number of excitations. Using multiple which pose risks to patient and operator safety. The main
averages can minimize artifacts from respiratory motion. magnet is associated with projectiles. The gradient coils are
Signal from motion artifacts varies more than the stationary loud and can cause peripheral nerve stimulation. The RF coils
tissue signal, and their intensity can be reduced by destructive can cause burns and re-program some implanted devices. An
interference. Therefore, if scan time is not a concern, increas- ideal MRI examination balances scan time and image quality.
ing the NEX/NSA is an easy way to boost the SNR. The This balance can be positively affected by new hardware, new
disadvantage of signal averaging is that it increases the total sequences and optimized sequence parameters.
acquisition time in proportion to the number of repetitions.
Higher NEX/NSA increases SNR but does not affect contrast. Compliance with ethical standards
Phased-array coils allow two or more coil elements to collect
imaging data simultaneously. The superior SNR and spatial Conflicts of interest Sherwin Chan has a research grant and honorarium
from Jazz Pharmaceuticals.
resolution of phased-array coils also decreases the number of
signal averages, thereby decreasing scan time.

References
Scan time
1. Grover VP, Tognarelli JM, Crossey MM et al (2015) Magnetic
resonance imaging: principles and techniques: lessons for clini-
In clinical MRI, scan time is computed as: scan time = TR ×
cians. J Clin Exp Hepatol 5:246–255
number of averages × number of phase-encoding steps. In 2. Serai SD, Hu HH, Ahmad R et al (2020) Newly developed methods
addition to increasing workflow and throughput, appropriate for reducing motion artifacts in pediatric abdominal MRI: tips and
scan time per sequence and minimal total study scan time is pearls. AJR Am J Roentgenol 214:1042–1053
crucial to patient compliance, especially when it comes to 3. Serai SD, Jones BV, Podberesky DJ, Coley B (2013) Is it time for a
dedicated pediatric MRI ACR accreditation program? J Am Coll
young children. Even when scan time is not a major issue, Radiol 10:274–278
shorter scans improve patient cooperation, reduce the oppor- 4. Serai SD, Rigsby CK, Kan HJ et al (2018) Inclusion of pediatric-
tunity for movement-related artifacts, and decrease sedation specific indications and procedures in the new ACR MRI accredi-
and anesthesia time, if applicable. If the scan time for a par- tation program. J Am Coll Radiol 15:1022–1026
ticular sequence or study is too long, it can be decreased by 5. International Electrotechnical Commission (2010) Medical electri-
cal equipment — part 2–33: particular requirements for the basic
optimizing the study protocol. Still, decreasing scan time af- safety and essential performance of magnetic resonance equipment
fects image quality and should be balanced appropriately. One for medical diagnosis. IEC 60601–2-33. U.S. Food and Drug
of the easiest ways is to reduce TR, but this affects image Administration website. https://www.accessdata.fda.gov/scripts/
contrast. The next easiest way is to decrease the number of cdrh/cfdocs/cfStandards/detail.cfm?standard__identification_no=
37168. Accessed 29 Sep 2020
phase-encoding steps by using a rectangular FOV, half- 6. U.S. Department of Health and Human Services, Food and Drug
Fourier technique, or reduction in phase matrix. However, Administration (2014) Criteria for significant risk investigations of
these three methods result in increase of image noise. magnetic resonance diagnostic devices — guidance for industry and
Pediatr Radiol (2021) 51:716–723 723

Food and Drug Administration staff. Online document. https://www. 13. Alibek S, Vogel M, Sun W et al (2014) Acoustic noise reduction in
fda.gov/media/71385/download. Accessed 28 Sep 2020 MRI using silent scan: an initial experience. Diagn Interv Radiol
7. Hartwig V, Giovannetti G, Vanello N et al (2009) Biological effects 20:360–363
and safety in magnetic resonance imaging: a review. Int J Environ 14. Fuelkell P, Langner S, Friedrich N et al (2018) Software-based
Res Public Health 6:1778–1798 noise reduction in cranial magnetic resonance imaging: influence
8. Franco G, Perduri R, Murolo A (2008) Health effects of occupa- on image quality. PLoS One 13:e0206196
tional exposure to static magnetic fields used in magnetic resonance 15. Corcuera-Solano I, Doshi A, Pawha PS et al (2015) Quiet
imaging: a review. Med Lav 99:16–28 PROPELLER MRI techniques match the quality of conventional
9. Mansfield P, Glover PM, Beaumont J (1998) Sound generation in PROPELLER brain imaging techniques. AJNR Am J Neuroradiol
gradient coil structures for MRI. Magn Reson Med 39:539–550 36:1124–1127
10. McJury M, Blug A, Joerger C et al (1994) Short communication: 16. Zaccagnino E, Devincent C, Leelakanok N et al (2019) Assessment
acoustic noise levels during magnetic resonance imaging scanning of quiet T2 weighted PROPELLER sequence in pediatric abdomi-
at 1.5 T. Br J Radiol 67:413–415 nal imaging. Clin Imaging 53:12–16
17. McRobbie DW, Moore EA, Graves MJ, Prince MR (2007) MRI from
11. McJury MJ (1995) Acoustic noise levels generated during high
picture to proton, 2nd edn. Cambridge University Press, New York
field MR imaging. Clin Radiol 50:331–334
12. Quirk ME, Letendre AJ, Ciottone RA, Lingley JF (1989) Anxiety
in patients undergoing MR imaging. Radiology 170:463–466 Publisher’s note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.