This appendix discusses possible physiological effects in human subjects from 20 T exposures and related MRI experiments. There have been temporary sensory effects noted by human subjects at lower fields, and extrapolation of these to much higher fields requires evaluations of the biophysics and additional experimentation with tissues, animals, and human subjects. The committee presents a short history of the health effects controversy followed by physiological and perception observations and then a synopsis of the physics behind these experimental observations of effects of static, radiofrequency fields, and switched magnetic field gradients.
SCOPE OF THE PROBLEM
Investigations of health effects from magnetic field exposures have been placed into three categories: static field effects, switched gradient effects, and radio-frequency (RF) heating effects. Experiments commenced in earnest 30 years ago, when the promise of widespread use of magnetic resonance imaging of humans became clear from the 1973 initial experiments demonstrating the imaging potentials by Lauterbur (Lauterber, 1973). Health effects studies include effects on molecular enzyme kinetics of genotoxic experiments up to 10 T and whole human body exposures up to 9.4 T. The conclusions from these investigations are that there are no harmful effects from static fields up to 9.4 T. There are known nerve stimulations from induced E-fields following rapid gradient switching with a threshold near 6 V/m. The RF thermal effects are dependent on calculated specific absorbed power. Though the health effects observed are reversible and not considered harmful, there
is no question that high static magnetic fields do have a temporary physiological effect on the central nervous system, as detailed in the next section on static field effects.
The issue before the committee is—What are the potential effects of fields that are two times those to which human beings have been exposed, particularly when the theoretical expectation involves a force or energy dependence on the B-field squared? The health effects emphasis is on static field effect potentials, because for nuclei other than protons, the RF frequencies are in the range already experienced, and with respect to imaging gradients, no major differences in the gradient-switched field rates are expected. But for protons, the anatomical distribution of power deposition and dielectric resonance effects need investigation for imaging and spectroscopy at 20 T. The relative permittivity of tissues at frequencies above 400 MHz is known, and both theoretical and experimental work can explore the potentials for hazards. Proton magnetic resonance functional imaging becomes important at 20 T, where RF penetration (1/e) at 860 MHz of 6-10 mm allows examination of the human cortex. Next, the committee gives a short history of inquiry into static and RF fields and then synopsizes the observations made using human subjects at fields above 4 T as well as animal behavior results.
STATIC FIELD EFFECTS: HISTORICAL PERSPECTIVE AND SCIENTIFIC LITERATURE
Public awareness of potential problems from electromagnetic fields dates from the publication of a statistical relationship between childhood leukemia and the presumed increases in magnetic fields in their homes. The magnetic fields were those related to electric power line inputs, and the method of quantification was the configuration of wires from the main high-voltage lines to the transformers to the homes. These were codified by a “wire code” for the homes in a particular neighborhood. Magnetic fields that epidemiologists were investigating are 60 Hz sinusoidal fields with only generally less than 3 μT amplitude associated with ac power lines. Though there was no compelling evidence that the wire codes and actual fields in the homes were strongly correlated, the controversy persisted for a number of years in the 1980s. These concerns and related concerns about electric fields near homes from overhead power lines led to a NRC committee evaluation and report. That committee concluded that there are no known human health effects from oscillating magnetic fields with an amplitude in the range to be expected in public places (NRC, 1997). An exception is the physiological effect of increased bone healing from the induced electric fields associated with exposure of bone fractures to oscillating magnetic fields of low amplitude (Bassett et al., 1974). Nearly 300 literature citations comprise this report on cell, animal, human, and epidemiology studies. Concerns about industrial operations (e.g., induction heating, aluminum plants)
and high-energy physics accelerators, where fields of 0.05 to 1.0 T are experienced by technicians and scientists for a few hours each day, led to an epidemiology study, which found no increased incidence of cancers. But this, like many other studies, had limited power even though almost 1,600 individuals were evaluated (Budinger et al., 1986). Epidemiology studies that endeavor to show effects must limit the claims to statements such as this: The results show that there was not an increase in prevalence of disease X beyond a factor of Y, where Y in the case of static fields is 2 for cardiovascular effects and 3 for cancer. Because disease incidences are low, a proper study to show, say, a 1.5-fold increase in incidence or measurable change in prevalence requires large groups (100,000 for most diseases) and, for diseases of very low prevalence, such as brain tumors and some leukemias, millions of subjects.
When nuclear magnetic resonance (NMR) was found to have major applications to human health but involved whole body exposure to fields more than 10,000 times Earth’s magnetic field, the scientific community began extensive investigations on cells, animals, and humans, with exposures approaching 10 T. The most extensive compendium of studies was published by the Advisory Group on Non-ionizing Radiation as a document of the Health Protection Agency of Great Britain (Health Protection Agency, 2008). The focus of that investigation is on static fields. But because the MRI and fMRI procedure involves use of rapidly changing magnetic fields to acquire spatial information and RF oscillating fields to achieve the resonance condition, the evaluation of safety of MRI and spectroscopy must include the slew rate of magnetic field changes and the RF power. These three aspects are discussed below.
A first concern surrounding very high magnetic fields is the danger from projectiles, including tools, metal furniture, and other ferromagnetic objects that can become projectiles (Schenck, 1992; Schenck, 2000). A life-threatening danger is to personnel with metallic implants, including pacemakers and wound clips. But other than the fact that the forces will be greater when the field and the field gradient become larger, one does not expect increased hazards from forces at 20 T. The current screening of individuals before entering the magnet is an adequate safeguard from harm from pacemakers and other implanted ferromagnetic objects.
Field Effects on Diamagnetic Materials (Water and Tissues)
The force for saturated ferromagnetic objects is proportional to the gradient, and this can be as high as 10-40 T/m depending on the magnet and the position of the subject, but there is no force on ferromagnetic objects in a homogeneous field. For diamagnetic and paramagnetic materials the force is proportional to the
product of the B-field and the gradient of the field. Even in a homogeneous field the turning torque can exist depending on the magnetic susceptibility anisotropy. Though the forces on diamagnetic materials are smaller than those on ferromagnetic materials, they are sufficient to levitate frogs in the fringe fields of a 16 T spectrometer (Berry and Geim, 1997). The levitation effect on diamagnetic materials, including water, ethanol, wood, plastic, acetone, and graphite, was noted in 1991 (Beaugnon and Tourier, 1991). One can expect a liquid surface profile to change significantly from a level surface in very high magnetic fields. For example, the surface of pure water (diamagnetic susceptibility of -9.031 × 10-6) will decrease at the field center and rise at the edges of a magnetic field of 10 T in a 200 mm diameter, 1,300 mm long magnet (Hirota et al., 1995). The amplitude of the surface profile is almost 40 mm. The change in heights from one part of the field to another is the result of the conservation of potential energy.
Whether this phenomenon affects the inner ear sense of gravity direction is not known but will be discussed further in the next section.
Magnetic Field Effects on Vestibular Apparati of Fish, Birds, and Mammals
A wide variety of experimental observations implicate the vestibular apparatus for the variety of symptoms and signs manifested by animals and human subjects in high magnetic fields as well as in the fringe fields of high-field magnets, where forces can cause small but physiologically significant relative tissue motion. Observed symptoms and signs include avoidance by animals of high fields and field gradients (Weiss et al., 1992; Houpt et al., 2007), animal head tilt while in a homogeneous field (Houpt et al., 2003), and turning behavior of animals after exiting high magnetic fields (Houpt et al., 2003). Human subjects have definite symptoms of nausea, vertigo, nystagmatism, and some reversible decline in cognitive function at fields of 4 T, 7 T, 8 T, and 9.4 T (Schenck, 1992; Patel et al., 2008; Theysohn et al., 2008; Kangarlu et al., 1999; Glover et al., 2007; van Nierop et al., 2012). Effects on the vestibular system are believed to underlie these phenomena.
The vestibular apparatus is the size of a children’s marble imbedded in the bony structure of the skull and comprises the inner ear. The vestibular organ is the master sensor for balance and motion, acting in coordination with the brain and spinal cord and using the visual system and body proprioception. The structure shown in Figure F.1 has three units: three orthogonally oriented circular canals (sacs) filled with fluid to sense angular motion (semicircular canals); two linear force detectors for gravity and linear acceleration (utrical and saccule); and a spiral structure that detects sound pressure waves over a wide band of frequencies (cochlea). All three systems use the deflection of hair cells to stimulate nerve impulses. Inertial forces of the endolymph fluid in the semicircular canals result in the bending of a structure at the base of each semicircular canal; multiple calcium
FIGURE F.1 Vestibular apparatus is about 20 mm in width and height. The semicircular canals sense angular motion, the two ball-like structures sense gravity and linear acceleration, and the spiral is the sound pressure transducer. The inset shows the otoliths embedded in a gel matrix whose motion stimulates the hair cells of the utricle and saccule. SOURCE: Courtesy of Thomas Budinger, University of California, Berkeley.
carbonate particles (10-100 μm) embedded in a gelatin platform move hair cells in response to gravity and linear motion in the utrical and saccule; and the motion of fluid in response to sound pressure waves stimulates hair cells lining the channels of the cochlea. The calcium carbonate particles (known as otoliths) in birds and fish contain some magnetite and thus are thought to play a role in navigation as well as balance (Harada et al., 2001). Definitive experiments that implicate the vestibular system in animal behavior showed no effects if the vestibular system is ablated (Carson et al., 2009).
There is no agreement regarding the predominant mechanism between Lorentz forces associated with movement of ions in a magnetic field (Roberts et al., 2011; Antunes et al., 2012), nerve conduction effects from changes in magnetic field (Glover et al., 2007), or magnetic field torque associated with susceptibility anisotropy of diamagnetic tissues (Budinger, 1981). Sensors in the utricule and saccule are hair cells driven by a small structure containing particles composed of mostly calcium carbonate (otoliths). Sensors in the semicircular canals are cells that distort from relative motion of fluid and the cell surface. Pressures of 2 mPa are sufficient to stimulate nystagmus (Kassemi et al., 2005).
Anticipated Human Responses to 20 T Magnet Exposures
The actual force on diamagnetic tissues of the body is dependent on field, field gradient, free space permeability, and susceptibility of the tissues. The susceptibility of tissues is in the range of that of diamagnetic water, ca. –10-6, but some tissues
such as those containing ferritin protein can have a paramagnetic susceptibility, which means that the forces on these tissues will be in the opposite direction to those of diamagnetic susceptibility (Schenck, 1996). The magnetization of tissues is given by χB0/m0, where χ is the susceptibility and B0 is the magnetic field. The presence of heterogeneous susceptibility is manifest in the contrast in images that show the associated phase differences.
The fringe field of a 20 T magnet will have a product of field and field gradient of 17 T2/m based on extrapolated values from measurements made for a 7 T fringe field as shown in Figure 2 of van Nierop et al. (2012). Our extrapolated values are 4.5 T/m and a field of 3.75 T at 0.5 m from the bore of 20 T magnet. We can compare this value to that at which rodents demonstrate avoidance on entering the 14.1 T magnet. From Figure 1 of Houpt et al. (2007), it was calculated that the field and gradient at the avoidance level were 2 T and 35 T/m, respectively, thus, the product is 70 T2/m. This force threshold is much greater than that estimated for a 20 T fringe field, but within the bore of a 20 T magnet the gradients could be much greater depending on the magnet design. If the maximum gradient on entering the magnet is 15 T/m when the local field is 10 T, the value of field times field gradient will be 150 T2/m. The dimensions T2/m when divided by the permeability and modified by susceptibility of the specific tissue give the volume force in N/m3.
Animal experiments as well as recent human subject exposures have shown the effects are temporary. But the neurological network of which the vestibular system is part has a tremendous plasticity, so that damage to the vestibular system might be hidden through adaptation. Animal exposure studies in addition to tissue exposure studies and mathematical simulations will be necessary in order to assure the safety of expected exposure periods. As already shown, pressure thresholds for activation of parts of the vestibular apparatus are in the range of 2 mPa, and adjacent tissues (e.g., ferritin-loaded brain tissue vs. normal tissue) are expected to have a 200-fold increase in susceptibility (extrapolated from Schenck, 1992).
Turning Torque on Macromolecules and Large Molecular Assemblages
In addition to high magnetic field effects from susceptibility differences between tissues, there is an important effect of high fields on molecules or tissue elements with large susceptibility differences between major and minor axes of molecular assemblages (e.g., retinal rods, chloroplasts, platelets). The energy is given by
where V is the volume of the unit, χl and χr are the susceptibility for the long and radial directions, and θ is the angle between the radial axis and the B-field. It is well known that, under conditions of negligible viscosity, fields of 1 T will orient
chloroplasts, blood platelets, retinal rods, and even large macromolecules. The potential for unwanted effects on molecular systems in animals and human beings requires investigation at fields of 20 T because the turning torque scales as B-field squared. The orientation effect on molecular assemblages and macromolecules has been demonstrated in studies of stimulating bone formation to grow in specified directions (Kotani et al., 2002). The origin of diamagnetic anisotropy in proteins and polypeptides is attributed to diamagnetic anisotropy of the planar peptide bonds (Worcester, 1978).
Lorentz Force Movement of Conducting Nerves
The Lorentz force, per unit volume, on conducting nerves in a magnetic field is given by
F = J × B
where J is the current density. The maximum displacement, as modeled by Roth and Baser (2009), is given by (JB/4μ) a2ln(b/a), where the current density J is 10 A/m2, the field is 20 T, and μ is the tissue shear modulus of 10 kPa. The nerve radius a is assumed to be 2 mm and that of the surrounding tissue return ion flow b is 25 mm. The calculated displacement is only 0.05 μm.
Effects of High Magnetic Fields on Nerve Conduction Speed
The Lorentz effect of force on moving ions in a magnetic field can slow nerve conduction velocity. The motion of sodium and potassium ions during nerve conduction can be pictured as small current loops along the axis of the conducting nerve fiber. If a field is applied at right angles to the nerve fiber, one can expect the ion current paths will be distorted. An ion of charge e in an electric field, E, and a magnetic field, B, will have forces Fe = eE and Fm = eVdB, respectively. Here, Vd is the drift velocity of the ions. Wikswo (1980) shows Vd is between 0.033 m/s and 6.6 × 10-5 m/s depending on the values chosen for estimation. This suggests that a 10 percent change in nerve conduction velocity will occur at 24 T. Verification of this can be done now at fields available at NHFML.
Cardiovascular Electrical Signal Artifacts
Electrocardiogram (ECG) signals at high field show artifacts associated with external wire conductors moving in the external field, and these temporal changes in recorded voltages are not associated with a physiological effect. There are two additional effects of importance as they do induce E-fields within the body that
are proportional to field strength. With each heartbeat as much as about 70 ml of blood is ejected into the aorta with a cross section of 17 mm. The initial velocity is 0.3 m/s. In a static magnetic field the potential across the aorta is about 1 V (i.e., 20 T × 0.3 m/s × 0.17 m). This potential will be detected by the ECG leads and give a signal that will occur mostly at the T wave, but a complex wave form occurs due to the varying directions of blood flow through the chambers of the heart and main vessels of the thorax (Tenforde et al., 1983). A result is the magnetocardiogram.
Another time-varying signal is that associated with breathing, where now the induced voltage is proportional to the rate of change of area orthogonal to the field direction. At 20 T, breathing at 15 per minute and a chest expansion of 2 cm, this is only 1.5 mV (3.14 cm2/4 s × 20 T × 10-4m2/cm2). This potential will appear as a rise and fall of the ECG in synchrony with breathing. None of these effects is of physiological significance.
A flowing conductor in a magnetic field will experience charge separation, and this E-field when acted upon by the magnetic field will result in a force counter to the flow of the fluid. Flowing mercury can be stopped by magnetic fields. The question arose years ago regarding the retarding force that might increase peripheral resistance, causing a rise in blood pressure if human subjects are exposed to high fields. Theoretical studies published before 1990 concluded the magnetohydrodynamic effect would be prohibitive. The correct theory for magnetohydrodynamic effects has shown that earlier literature did not take into account viscous forces and all of the induced magnetic field (Keltner et al., 1990). This theory and experimental work at 4.7 T showed the magnetohydrodynamic effect is not significant at physiological flows in the aorta where the largest effect is expected.
SPECIFIC ABSORBED POWER AND RAPIDLY CHANGING GRADIENTS
Major safety issues that underpin limiting guidelines for magnetic resonance imaging and spectroscopy of human subjects are the RF heating effects, whose metric is the specific absorbed power. A second limitation on the pulse sequences is the magnitude of induced electric fields from rapidly changing magnetic fields. Whereas the known thresholds from the experiences over the last 40 years are not expected to be exceeded at 20 T, these thresholds will limit the power density and therefore the depth of penetration at the higher proton frequencies required for 20 T. The physiological thresholds of nerve stimulation will limit the use of some desired pulse sequences.
Specific Absorbed Power
The oscillating magnetic fields for the radio-frequency pulses used in imaging and spectroscopy induce oscillating electric fields in accord with Faraday induction. The E field is proportional to the frequency and the conducting body loop. The average electric field is E/√2, so that the absorbed power, SAR, by a mass of tissue is as follows:
SAR = E2σ/2ρ
where E is the magnitude of the E-field, s is the conductivity, and σ is the density.
An increase in field leads to an increase in the magnetic resonance frequency for a given spin. In turn the increase in frequency leads to an increase in induced E-field, and SAR is expected to increase with field. The conductivity of tissues also increases with frequency (a factor of 2 can be expected between 300 MHz (7 T) and 852 MHz (20 T) While this is the case for proton MRI, fMRI, and MRS, and will be the case for other spins, but these other spins (see Table 4.1 in Chapter 4) have frequency requirements many times lower than the resonant frequency of protons for a given field. The scientific community (e.g., Tang and Ibrahim, 2007) and regulatory advisors involved in safety are experienced with frequencies below 300 MHz and as one can see from Table 4.1, most of the nuclei comprising the human body have resonant frequencies below 300 MHz.
Induced E-Fields from Time-Varying Gradients
An E-field of 6 V/m induced by 60 T/s near a 30 cm diameter body part will cause a sensation of an electric shock (Budinger et al., 1991). The governing physics is the Maxwell-Faraday equation, which equates the electric field to the diameter of a loop defining the body being exposed to a rapid change in the magnetic field.
Volt/meter = − dB/dt × r/2
where dB/dt is the rate of the magnetic field gradient change associated with MRI pulse sequences and r is the object radius.
Visual sensations known as magnetophosphenes can be induced at about 2 T/s. This phenomenon has been studied since visual phosphenes were noted by d’Arsonval (1896), who moved the magnetic field source near the eyes. The mechanism is not as simple as suggested by equation (Carson et al., 2009). Because the threshold for visual sensations is dependent on reaching a field of at least 10 mT with a rise time of approximately 2 ms and a repetition rate less than 30 per second (Lövsund et al., 1980). These dependencies can be understood if the mechanism is a
mechanical distortion of retinal components from susceptibility anisotropy. Direct electric voltage application to the head can induce phosphenes, but the current densities of 17 μA/cm2 are much greater than the 2 μA/cm2 that is inferred from magnetic field changes. The emphasis on the mechanisms for phosphene observations is relevant to understanding other sensory phenomena that might manifest at 20 T, where larger fringe fields and B-fields are expected than experienced in the last 40 years of investigations.
New problems are not expected to arise with pulse sequences needed for imaging at high fields. Power requirements for RF transmit and receive coils and requirements for the gradient coils involved in applications at 20 T will not present serious engineering problems if the frequencies are restricted to less than 400 MHz. Going beyond low gamma spins to proton imaging and spectroscopy at 20 T, 852 MHz, will be a challenge.
Antunes, A., P.M. Glover, Y. Li, O.S. Mian, and B.L. Day. 2012. Magnetic field effects on the vestibular system: Calculation of the pressure on the cupula due to ionic current-induced Lorentz force. Physics in Medicine and Biology 57:4477.
Bassett, C.A.L., R.J. Pawluk, and R.O. Becker. 1974. Augmentation of bone repair by inductively coupled electromagnetic fields. Science 184:575.
Beaugnon, E., and R. Tournier. 1991. Levitation of organic materials. Nature 349:470.
Berry, M.V., and A.K. Geim. 1997. Of flying frogs and levitrons. European Journal of Physics 18:307.
Budinger, T.F. 1981. Nuclear magnetic resonance (NMR) in vitro studies: Known thresholds for health effects. Journal of Computer Assisted Tomography 5:800.
Budinger, T.F., K.S. Bristol, D.A. Carpenter, P.A. Garbutt, P.D.C. Wong, and C. Yen. 1986. Epidemiological study of static magnetic field health effects. Progress in Physics and Molecular Biology 87:241.
Budinger, T.F., H. Fischer, D. Hentschel, H.-E. Reinfelder, and F. Schmitt. 1991. Physiological effects of fast-oscillating magnetic field gradients. Journal of Computer Assisted Tomography 15:909.
Carson, A.M., B.S. Kwon, J.C. Smith, and T.A. Houpt. 2009. Labyrinthectomy abolishes the behavioral and neural response of rats to a high-strength static magnetic field. Physiology and Behavior 97:36.
d’Arsonval, M.A. 1896. Dispositifs pour la mesure des courants alternatifs à toutes fréquences. Comptes Rendus des Séances de la Société de Biologie et de ses Filiales (Paris) 3:451.
Glover, P.M., I. Cavin, W. Qian, R. Bowtell, and P.A. Gowland. 2007. Magnetic-field-induced vertigo: A theoretical and experimental investigation. Bioelectromagnetics 28:349.
Harada, Y., M. Taniguchi, H. Namatame, and A. Iida. 2001. Magnetic materials in otoliths of bird and fish lagena and their function. Acta Oto-Laryngologica 121:590.
Health Protection Agency. 2008. Static Magnetic Fields Report of the Independent Advisory Group on Non-ionizing Radiation. Radiation, Chemical and Environmental Hazards 6.
Hirota, N., T. Homma, H. Sugawara, K. Kitazawa, M. Iwasaka, S. Ueno, H. Yokoi, Y. Kakudate, S. Fujiwara, and M. Kawamura. 1995. Rise and fall of surface level of water solutions under high magnetic field. Japanese Journal of Applied Physics 34: 991.
Houpt, T.A., D.M. Pittman, J.M. Barranco, E.H. Brooks, and J.C. Smith. 2003. Behavioral effects of high strength magnetic fields on rats. Journal of Neurosience 23:1498.
Houpt, T.A., J.A. Cassell, C. Riccardi, M.D. DenBleyker, A. Hood, and J.C. Smith. 2007. Rats avoid high magnetic fields: Dependence on an intact vestibular system. Physiology and Behavior 92:741.
Kangarlu, A., R.E. Burgess, H. Zhu, T. Nakayama, R.L. Hamlin, A.M. Abduljalil, and P.-M.L. Robitaille. 1999. Cognitive, cardiac and physiological safety studies in ultra high field magnetic resonance imaging. Journal of Magnetic Resonance Imaging 17:1407.
Kassemi, M., D. Deserranno, and J.G. Oas. 2005. Fluid-structural interactions in the inner ear. Computers and Structures 83:181.
Keltner, J.R., M.S. Roos, P.R. Brakeman, and T.F. Budinger. 1990. Magnetohydrodynamics of blood flow. Magnetic Resonance in Medicine 16:139.
Kotani, H., M. Iwasaka, A. Curtis, and S. Ueno. 2000. Magnetic orientation of collagen and bone mixture. Journal of Applied Physics 87:6191.
Lauterbur, P.C. 1973. Image formation by induced local interactions: Example employing magnetic resonance. Nature 242:190.
Lövsund P., P.Å Öberg, and S.E.G. Nilsson. 1980: Magnetophosphenes: A quantitative analysis of thresholds. Medical and Biological Engineering and Computing 18:326.
National Research Council. 1997. Possible Health Effects of Exposure to Residential Electric and Magnetic Fields. National Academy Press, Washington, D.C.
van Nierop, L.E., P. Slottje, M.J.E. van Zandvoort, F. de Vocht, and H. Kromhout. 2012. Effects of magnetic stray fields from a 7 Tesla MRI scanner on neurocognition: A double-blind randomized crossover study. Occupational and Environmental Medicine 69:759.
Patel, M., R.A. Williamsom, S. Dorevitch, and S. Buchanan. 2008. Pilot study investigating the effect of the static magnetic field from a 9.4T MRI on the vestibular system. Journal of Occupational and Environmental Medicine 50:576.
Roberts, D.C., V. Marcelli, J.S. Gillen, J.P. Carey, C.C. Della Santina, and D.S. Zee. 2011. MRI magnetic field stimulates rotational sensors of the brain. Current Biology 21:1635.
Roth, B.J., and P.J. Basser. 2009. Mechanical model of neural tissue displacement during Lorentz effect imaging. Magnetic Resonance in Medicine 61:59.
Schenck, J. 1992. Health and physiological effects of human exposure to whole body four-tesla magnetic field during MRI. Annals of the New York Academy of Science 649:285.
Schenck, J.F. 1996. The role of magnetic susceptibility in magnetic resonance imaging: MRI magnetic compatibility of the first and second kinds. Medical Physics 23:815.
Schenck, J. 2000. Safety of strong, static magnetic fields. Journal of Magnetic Resonance Imaging 12:2.
Tang, L., and T.S. Ibrahim. 2007. On the radio-frequency power requirements of human MRI. Progress in Electromagnetics Research Symposium (PIERS) Online 3:886.
Tenforde, T.S., C.T. Gaffey, B.R. Moyer, and T.F. Budinger. 1983. Cardiovascular alterations in Macaca monkeys exposed to stationary magnetic fields: Experimental observations and theoretical analysis. Bioelectromagnetics 4:1.
Theysohn, J.M., S. Maderwald, O. Kraff, C. Moenninghoff, M.E. Ladd, and S.C. Ladd. 2008. Subjective acceptance of 7 Tesla MRI for human imaging. Magma 21:63.
Weiss, J., R.C. Herrick, K.H. Taber, C. Contant, and G.A. Plishker. 1992. Bio-effects of high magnetic fields: A study using a simple animal model. Magnetic Resonance Imaging 10:689.
Wikswo, J.P. 1980. An estimate of the steady magnetic field strength required to influence nerve conduction. IEEE Transactions on Biomedical Engineering 27:722.
Worcester, D.L. 1978. Structural origins of diamagnetic anisotropy in proteins. Proceedings of the National Academy of Sciences U.S.A. 75:5475.