Magnetic R~SOIWZCS? Imaging. Vol. 10, pp. 689-694. Printed in the USA. All rights reserved.
1992 Copyright 0
0730-725X/92 $5.00 + .OO 1592 Pergamon Press Ltd.
l Original Contribution
BIO-EFFECTS OF HIGH MAGNETIC FIELDS: A STUDY USING A SIMPLE ANIMAL MODEL JEREMY WEISS, * RICHARD C. HERRICK,*~ KATHERINE H. TABER, *t CHARLES CONTANT,$ AND GORDON A. PLISHKER*$ *Baylor College of Medicine, Magnetic Resonance Center, The Woodlands, TX 77380, USA Departments of tRadiology, SNeurosurgery, and $Neurology, Baylor College of Medicine, Houston, TX 77030, USA The desire to do clinical imaging and spectroscopy at magnetic field strengths greater than 2 Tesla (T) necessitates investigation of possible bioeffects at these high fields. A simple T-maze was utilized to evaluate the aversive effects of exposure to three levels of static magnetic field (0,1.5, and 4 T). The right arm of the maze extended into the center of a 30-cm horizontal bore magnet, while the left arm extended into a mock magnet bore with the same dimensions. The self-shielded design of the magnet reduces the fringe field to zero within 1 m of the bore, placing the start box of the maze outside the 5-G line of the magnet. Each rat performed a total of ten trials at each level of magnetic field strength. A follow-up subset was run at 4 T with the maze reversed. At 0 T, the rats entered the magnet freely. No significant differences from the control were observed at 1.5 T. At 4 T, however, in 97% of the trials the rats would not enter the magnet. In the maze-reversed subset a majority of the rats turned toward the magnet, indicating that they had learned an aversive response from the previous trials at 4 T. However, in only 4 decisions out of 58 did the rats actually enter the magnet. Eighteen decisions to turn around were made at the edge of the magnet in a region of strong field gradients (up to 13 T/m) and a field strength up to 1.75 T. We propose that the aversive response is most likely due to magnetic induction effects caused by motion in a strong magnetic field gradient. Keywords:
Magnetic resonance; Biosafety;
Static magnetic field effects.
There is a growing concern over the bioeffects of exposure to high magnetic fields, fueled in part by reports that technicians working around 4-T magnets sometimes experience dizziness, nausea, vertigo, and visual phosphenes.’ In addition, it has been known for some time that high magnetic field pulses can be used to directly stimulate nerves and brain.2 Recently, direct muscle stimulation has been reported during an MRI scan with magnetic field time rate of change (dB/dt) of 61 T/sec.3 Experimental demonstration of the effects of exposure to the magnetic fields used in magnetic resonance imaging (MRI) on spatial orientation performance have been difficult to obtain. No effects have been found at very low (0.15 T) or intermediate (1.5 T) magnetic field strength on spatial memory.4 However, short term exposure (3-4 hr) at 1.6 T increased investigatory behavior in rats, while exposure to 0.3 T depressed investigatory behavior. 5 The large fringe field associated with the higher
Adult male Long Evans Hooded rats (50-57 days of age, 200-225 g, n = 32) were housed in groups of four under standard conditions. Within each group, the rats were distinctly marked with iodine tincture to maintain individual identification. Prior to a testing period, each group was transported to the testing site (“2 miles). Four rats were eliminated from the study due to lack of cooperation. A simple T maze was constructed from clear acrylic,
RECENED g/24/91; ACCEPTED3/16/92. Address correspondence to Dr. Richard Herrick, Baylor
Magnetic Resonance Center, 9450 Grogan’s Mill Rd., Suite 110, The Woodlands, TX 77380, USA.
field strength magnets has made standard radial and T maze testing impossible in the past. The recent introduction of a high field (up to 4.5 T) self-shielded magnet with virtually no fringe field has made such studies possible. In the present study, the behavior of rats in a standard T maze is evaluated. These experiments are designed to simulate the exposure of a human patient entering a magnet for diagnostic purposes. METHODS
689
690
Magnetic Resonance Imaging 0 Volume 10, Number 4, 1992
30-cm bore, iron-shielded magnet that was rampable from 0 to 4.5 T. The shielded design causes the magnetic field strength to fall off rapidly with distance from the magnet. The 5-G line is 1 m from the front of the magnet (see Fig. 1). Prior to each trial, the maze was cleaned with alcohol to eliminate any olfactory cues. The rat was deposited into the start box between the two arms of the maze and allowed to move freely. At all times during the trial the observer stood behind the start box between the two arms of the maze. The clear acrylic top of the maze was in place during all trials to keep the rat in the maze and to help reduce ambient noise. A decision was recorded when the head of the rat crossed into the next region. Decisions were recorded on paper in all cases. In addition, several trials were recorded on videotape. The trial was considered complete either when the rat made a Decision 5 (going to the end of one arm of the maze) or when 4 min had elapsed. At the end of each trial the rat was returned to its cage. Each rat performed two sets of five trials during a testing session. All four rats completed their first set, then the procedure began again. Each group was returned to its home cage at the end of the second set. Experiments were performed at 0, 1.5, and 4 T. Following the completion of all experimental trials at all three field strengths, a subset of four randomly selected rats performed five trials at 4 T in the same T maze with
and the outside painted with opaque white enamel. Each arm was 183 cm in length, 10 cm wide, and 10 cm deep. The top of the maze was removable and made of clear acrylic to facilitate cleaning and observation. Each arm of the maze was divided into five lengths, which will hereafter be called decisions. Decision 1 was 45.5 cm in length. Decisions 2-4 were 30.5 cm in length each; Decision 4 extended 15 cm inside the bore of the magnet (see Fig. 1). Decision 5 extended to the center of the magnet (46 cm). When in position the right arm of the maze (A) extended into the center of the magnet, while the left arm of the maze (B) extended into the center of the sham magnet (a set of gradient coils of the exact inner diameter of the magnet). Both the magnet and the sham were covered with red construction paper to eliminate visual cues. The magnet was situated in a very large common work area with uniform temperature, bright uniform overhead lighting, and a relatively high level of ambient noise. The light and tactile conditions experienced in the two arms of the maze were identical. The primary noise source in the environment was from gas venting from the magnet’s cryogenic system. This resulted in a constant white noise, which tended to drown out other intermittent noise in the environment. Noise amplitude from this source was constant during control and exposure. All experiments were performed utilizing a Baymed
4.0
3.0
z
_o_
1ST
+
4T
. 4A
3A
2A
:lA : 5 Gauss Line
Decision
Points
Fig. 1. The two axial field plots at 1.5 and 4 T taken from the center 1.5 T at magnet center and 6.8 G for 4 T at magnet center.
of the magnet.
At 91.5 cm, the field is 6.2 G
for
Bio-effects of high magnetic fields 0 J.
the arms reversed, so that the B arm led into the magnet. Upon encountering an aversive stimulus, it is expected that the rat would stop, rather than continue down the entire length of the maze. Thus, an increase in decisions is predicted at the point at which the magnetic field becomes an aversive stimulus. A Pearson Chi-square analysis was performed to evaluate statistical significance.
WEISS ET AL.
Table 1. Tabulation of outcomes collapsed to decision to turn into the magnet, away from the magnet, or no decision for control, 1.5 T, and 4 T trials
from magnet
Away
280
240
240
q
.2 .’ A
F? .9 $ P
160
‘ii 2 5 % e! LA
154 (55.00V0)
254 (90.71%)
(5.&o)
(6&o)
111 (39.64%)
109 (38.93%)
(2.5&o)
n Control q 1.5 Tesla
200 1.5Tesla
160 (57.14%)
4.0 T
magnet), region -5 (in the mock magnet) or no decision (0) was made. These data and Vo of total outcomes for each field strength are shown in Table 1. There was no difference found between 0 T and 1.5 T in arm preference. At 4 T, however, arm B (away from magnet) was clearly preferred (JI < .OOl). The results were unchanged when individual rats were maintained as a separate variable in the statistical analysis. When the arms of the magnet were reversed, the number of Decision 4 in the B arm (magnet arm) increased dramatically, while the number of Decision 5 decreased (see Fig. 3). The rats were observed to travel
280
200
1.5 T
&70)
magnet
RESULTS
t
ControI
No decision Toward
A partial tabulation of decisions is shown for both the magnet arm A (see Fig. 2A) and the sham arm B (see Fig. 2B) for all magnetic field strengths tested. For each trial, only the decisions indicating the furthest extent of travel down each arm are tabulated, giving at most two decisions per trial. One decision is tabulated if the rat stays in just one arm of the maze. The frequency of each decision is similar at 0 T (control) and 1.5 T when arm A and B are compared. At 4 T, however, there is a clear increase in Decision 5 in B and a decrease in Decision 5 in A. A Pearson Chi-square statistical analysis was performed after collapsing the results of each trial such that each trial was allowed only one outcome; thus, the rat ended up in region 5 (in the
691
4.0 Tesla 160
6 120
80
40
5 s z-
120
u!?!
80
40
1
3
2
Decision (A)
4 Points
5
1
2
4
3 Decision
5
Points
(B)
Fig. 2. (A) Represents the direction toward the magnet (arm A). (B) Represents the direction away from the magnet (arm B). Frequency of the decisions made is plotted on the y-axis, while distance is represented by the x-axis. Only the decisions indicating furthest extent of travel down each arm are tabulated. One decision is tabulated if the rat stays in just one arm of the maze.
Magnetic Resonance Imaging 0 Volume 10, Number 4, 1992
692
:
1
2
3
4
5
Decision Polnts
Fig. 3. Results of the random subset run. The direction has now been changed so that the magnet is now on the left (arm B) and the mock magnet is on the right (arm A). Tabulation is the same as for Fig. 2.
down the B arm, halt suddenly at Decision 4, and reverse direction. Several animals were seen to spread out all four limbs and back up the maze several steps before turning around. A Pearson Chi-square statistical analysis was performed on the 4-T data in Figs. 2 and 3. Chi-square was computed to be 60.09 with nine degrees of freedom. The increase in decision 4B shown in Fig. 3 over decisions in Fig. 2 proved to be statistically significant, with p < .OOOl. DISCUSSION
The results from 0 T (control) and 1.5 T indicate the rats exhibited no strong preference for either arm (A-magnet, B - nonmagnet). At 4 T there was a significant increase in selection of arm B. This clear preference for the “nonmagnet” arm of the maze suggests that an aversive stimulus was associated with the magnet at 4 T that was not present at 1.5 T. When the arms of the maze were reversed, the frequency of Decisions 3 and 4 in arm B (now the magnet arm) increased, suggesting that the rats had learned to travel toward the left in the initial experiments. After moving some distance down arm B toward the magnet the rats were observed to halt and reverse direction. They then proceeded all the way to the end of the other arm, resulting in a high number for Decision 5 in arm A. Thus,
the experimental results strongly suggest that some aspect of the environment around the magnet when at 4 T was quite aversive to the rats. A puzzling feature of the 4-T data is that the rats are making their decision to turn back in a region of relatively low static magnetic field. At decision point 5A the magnitude of the field is not much above 1.5 T, which in the previous trial caused no behavioral changes. That is, the subjects entered the center of the magnet, in which region they were experiencing a 1.5-T static magnetic field, with no evidence of aversion. The coarse data sampling does not allow us to make a firm statement about the exact magnetic field magnitude when the rats turned back, but most of the rats turned around before reaching decision point 5A. We propose that this behavior may be related to the magnetic field gradient in the area. If we examine the gradient of the field, that is, the rate of change along the z axis, we see that when the magnet is at 1.5 T the maximum gradient the rat is exposed to is approximately 0.07 T/m. When the magnet is at 4 T, however, the gradient at the mouth of the magnet is already above 0.05 T/m and rapidly climbs to 13 T/m in the area spanned by decision box 4A. These observations suggest that the magnetic field gradient as well as the absolute field magnitude may be a factor in the aversive stimulus. It is possible that this area of strong magnetic field gradient would be aversive even without motion, but it is not possible in these experiments to test for this condition independently. A stationary rat could interact with the magnetic field through either permanent or induced magnetic dipole moments in the tissues of the rat. A homogeneous field exerts a torque on the dipole, while a magnetic field gradient exerts a translational force. All tissue exhibits a weak induced diamagnetic dipole moment; this is not a very likely mechanism because of the weakness of the effect as well as the fact that the magnetic dipole moment of protein molecules would not change significantly with conformational change. Paramagnetic dipole moments, such as found in deoxyhemoglobin, myoglobin, cytochromes, and iron-sulfur proteins, would have at least one or two orders of magnitude greater interaction with the magnetic field than a purely diamagnetic protein molecule. It is difficult to postulate how an interaction with any of these molecules could result in a stimulus to the rat’s nervous system. Ferromagnetic deposits, in the form of magnetite particles or the iron-storage protein ferritin, would have an interaction several orders of magnitude stronger than paramagnetism. In fact magnetotactic bacteria have been identified that use this method to navigate along the earth’s magnetic field. Such a stimulus mechanism
Bio-effects of high magnetic fields 0 J. WEISSET AL.
sophisticated setup. Qualitative observations during the trial and on review of the videotape showed no obvious changes in the rate or pattern of travel down the maze between control and exposure. Table 2 shows estimates of dB/dt and resultant E field induced by these two types of motion assuming a velocity along the z axis of 0.2 m/set, a rotational velocity of 1 rad/sec, and a diameter for the rat’s head of 3 cm. Note that the Btrans in the gradient field at 4 T already exceeds the estimate of E,,, in the center of the magnet at 1.5 T. If the aversive stimulus was the electric field with a threshold on the order of 1 x 10m2 V/m (dB/dt = 1.3 T/set), it would explain why the rats turned away in an area of relatively low static magnetic field but high field gradient. This level is more than an order of magnitude below the thresholds reported for peripheral nerve stimulation (~61 T/set). Thus, direct nerve stimulation is not a likely mechanism for the aversive response. A threshold for magneto-phosphenes has been estimated at a dB/dt of approximately 1.3 T/sec6 dB/dt for translational movement in the gradient at 4 T are above this value, indicating that this could be a possible mechanism for the aversive response. There is also research showing that living organisms are sensitive to extremely weak electric fields and in the case of pulsating fields, may be able to detect fields below the thermal noise level.’ Proteins, such as the ion channel proteins in membranes, have many polar groups that affect their function and it is believed that electric fields can modulate their behavior. Electric fields are thought to interact directly with these proteins through these polar groups in their side chains, inducing an “electro-conformational” change. The noise level in membranes has been estimated to be on the order of 3 x lop3 to 3 x 10e2 V/m, which is on the order of our estimate for induced electric fields in the rat. Effects have been reported with field exposures as low as 2 x lop3 V/cm.* Although these studies are of the effects of time-varying electric fields, they are still relevant because of the time-varying nature of the electric fields induced in the rat. We can thus establish at
in the rat, however, would be expected to come into play at much lower field strengths than 1.5 T, where no aversive stimulus was observed. We suggest that it is more likely that the aversive effect is exerted via magnetic induction effects resulting from the rat’s motion in the magnetic field. According to Faraday’s law a change in magnetic flux through a loop of radius r will result in an induced electric field around the loop given by: r dB
E=--2z where E is the induced electric field in volts/meter, r is the radius of the loop in meters, and dB/dt is the time rate of change of magnetic field in T/set. In MRI this effect is usually thought of in the context of gradient switches or eddy currents in the organism from radiofrequency (RF) irradiation. An induction effect can also arise, however, from translational movement of the organism in a field gradient or rotational motion in the field. For translational motion in the z direction, dB/dt is given by: dB = grad, Bv, dt
where, v, is the velocity along the z axis and grad,B is the field gradient along z. For rotational motion, maximum dB/dt is given by: dB = -BQ dt
where Q is rotational velocity in rad/sec. To estimate electric fields induced in the rat, the motion is approximated by a translational motion along the z axis and a rotational motion of the head due to head wagging. Actual rate of travel of the rats in the maze is difficult to assess without a much more
Table 2. Estimates
1.5 T center 1.5 T 4A/SA
boundary boundary
electric
Field
Gradient
U-1
(T/m)
1.5
4 T center 4 T 4A/5A
of dB/dt and induced
0
field in the rat’s head as a result of motion dB&,,, (T/set)
dB/dt,,,
0
1.5
6
1.2
0.6
4
0
0
4
2.6
1.75
13
in the magnetic
field
(T/set)
0.6 1.75
693
0 9 x 10-3 0 1.95 x 10-2
1.1 x 10-2 4.5 x 10-3 3 x 10-2 1.3 x 10-2
Magnetic Resonance Imaging 0 Volume
694
least the plausibility of a time-varying electric field as a cause of an aversive stimulus. Electric fields in a conductive medium also give rise to currents which may have a biological effect. One proposed mechanism for the vertigo experienced by human subjects at 4 T is magneto-hydrodynamic pressure on the cupola within the vestibular mechanism.’ A minimum pressure threshold for sensations has been estimated at 1.25 x lo-’ Pa.9 Magnetohydrodynamic pressure is given by: P = baflB’/p
For humans, assuming a = .003 m, b = .00015m, Q = 10.5 radians/set, and B = 4 T, P was estimated to be 1.5 x 10e4 Pa, showing that stimulation of the cupola is quite plausible. A similar expression can be derived for linear motion in a gradient. For a canal in a plane perpendicular to the field: P = bav,B grad, B/p
.
Magnetohydrodynamic pressure in this case would be equal at all points around the canal and thus would not be expected to deflect the cupola. Magneto-hydrodynamic pressure does not appear to be a likely mechanism for an aversive stimulus arising from a field gradient. CONCLUSION The results of this study indicate that there is a measurable aversive effect at 4 T on the T maze behavior of rats, but no effect at 1.5 T. The induced aversion is most likely a result of magnetic induction effects arising from movement through an area of magnetic field gradient.
10, Number 4, 1992
Acknowledgments-This work was supported by a grant from the Kleberg Foundation. The authors gratefully acknowledge Aviva Ferragi and Jay Porter for providing magnetic field measurements presented in this manuscript.
REFERENCES 1. Schenck, J. Health and physiological effects of human exposure to whole body four tesla magnetic fields during MRI. In: R.L. Magin, R.P. Liburdy, B. Persson (Eds). Biological Effects and Safety Aspects of Nuclear Magnetic Resonance Imaging and Spectroscopy. New York Academy of Sciences Workshop, May 15-17, 1991. new method for 2. Hallett, M.; Cohen, L. Magnetism-a stimulation of nerve and brain. JAMA 264:538-541; 1989. 3. Cohen, M.S.; Weisskoff, R.M.; Rzedzian, R.R.; Kantor, H.L. Sensory stimulation by time-varying magnetic fields. Magn. Res. Med. 14:409-414; 1990. 4. Inms, N.K.; Ossenkopp, K.P.; Prato, ES.; Sestini, E. Behavioral effects of exposure to nuclear magnetic resonance imaging: II. Spatial memory tests. Magn. Rex Imaging 4:281; 1986. 5. Smirnova, N.P. Behavior in rats in “open field” following the action of a magnetic field. Zhurnal Uysskei Ner-
voi 32~72; 1982. 6. Anderson, L.E. Biological response of animals to timevarying magnetic fields. In: R.L. Magin, R.P. Liburdy, B. Persson (Eds). Biological Effects and Safety Aspects of Nuclear Magnetic Resonance Imaging and Spectroscopy. New York Academy of Sciences Workshop, May 15-17, 1991. 7. Weaver, J.C.; Astumin, B.D. The response of living cells to very weak electric fields: The thermal noise limit. Science 247~459-462; 1990. 8. McLeod, K.J.; Lee, R.C.; Ehrlich, H.P. Frequency dependence of electric field modulation of fibroblast protein synthesis. Science 236:1465-1469; 1987. range of 9. Oman, C.M.; Young, L.R. The physiological pressure difference and cupula deflections in the human semicircular canal. Acta Otolaryngol. 74~324-331; 1972.
1992 Copyright 0
0730-725X/92 $5.00 + .OO 1592 Pergamon Press Ltd.
l Original Contribution
BIO-EFFECTS OF HIGH MAGNETIC FIELDS: A STUDY USING A SIMPLE ANIMAL MODEL JEREMY WEISS, * RICHARD C. HERRICK,*~ KATHERINE H. TABER, *t CHARLES CONTANT,$ AND GORDON A. PLISHKER*$ *Baylor College of Medicine, Magnetic Resonance Center, The Woodlands, TX 77380, USA Departments of tRadiology, SNeurosurgery, and $Neurology, Baylor College of Medicine, Houston, TX 77030, USA The desire to do clinical imaging and spectroscopy at magnetic field strengths greater than 2 Tesla (T) necessitates investigation of possible bioeffects at these high fields. A simple T-maze was utilized to evaluate the aversive effects of exposure to three levels of static magnetic field (0,1.5, and 4 T). The right arm of the maze extended into the center of a 30-cm horizontal bore magnet, while the left arm extended into a mock magnet bore with the same dimensions. The self-shielded design of the magnet reduces the fringe field to zero within 1 m of the bore, placing the start box of the maze outside the 5-G line of the magnet. Each rat performed a total of ten trials at each level of magnetic field strength. A follow-up subset was run at 4 T with the maze reversed. At 0 T, the rats entered the magnet freely. No significant differences from the control were observed at 1.5 T. At 4 T, however, in 97% of the trials the rats would not enter the magnet. In the maze-reversed subset a majority of the rats turned toward the magnet, indicating that they had learned an aversive response from the previous trials at 4 T. However, in only 4 decisions out of 58 did the rats actually enter the magnet. Eighteen decisions to turn around were made at the edge of the magnet in a region of strong field gradients (up to 13 T/m) and a field strength up to 1.75 T. We propose that the aversive response is most likely due to magnetic induction effects caused by motion in a strong magnetic field gradient. Keywords:
Magnetic resonance; Biosafety;
Static magnetic field effects.
There is a growing concern over the bioeffects of exposure to high magnetic fields, fueled in part by reports that technicians working around 4-T magnets sometimes experience dizziness, nausea, vertigo, and visual phosphenes.’ In addition, it has been known for some time that high magnetic field pulses can be used to directly stimulate nerves and brain.2 Recently, direct muscle stimulation has been reported during an MRI scan with magnetic field time rate of change (dB/dt) of 61 T/sec.3 Experimental demonstration of the effects of exposure to the magnetic fields used in magnetic resonance imaging (MRI) on spatial orientation performance have been difficult to obtain. No effects have been found at very low (0.15 T) or intermediate (1.5 T) magnetic field strength on spatial memory.4 However, short term exposure (3-4 hr) at 1.6 T increased investigatory behavior in rats, while exposure to 0.3 T depressed investigatory behavior. 5 The large fringe field associated with the higher
Adult male Long Evans Hooded rats (50-57 days of age, 200-225 g, n = 32) were housed in groups of four under standard conditions. Within each group, the rats were distinctly marked with iodine tincture to maintain individual identification. Prior to a testing period, each group was transported to the testing site (“2 miles). Four rats were eliminated from the study due to lack of cooperation. A simple T maze was constructed from clear acrylic,
RECENED g/24/91; ACCEPTED3/16/92. Address correspondence to Dr. Richard Herrick, Baylor
Magnetic Resonance Center, 9450 Grogan’s Mill Rd., Suite 110, The Woodlands, TX 77380, USA.
field strength magnets has made standard radial and T maze testing impossible in the past. The recent introduction of a high field (up to 4.5 T) self-shielded magnet with virtually no fringe field has made such studies possible. In the present study, the behavior of rats in a standard T maze is evaluated. These experiments are designed to simulate the exposure of a human patient entering a magnet for diagnostic purposes. METHODS
689
690
Magnetic Resonance Imaging 0 Volume 10, Number 4, 1992
30-cm bore, iron-shielded magnet that was rampable from 0 to 4.5 T. The shielded design causes the magnetic field strength to fall off rapidly with distance from the magnet. The 5-G line is 1 m from the front of the magnet (see Fig. 1). Prior to each trial, the maze was cleaned with alcohol to eliminate any olfactory cues. The rat was deposited into the start box between the two arms of the maze and allowed to move freely. At all times during the trial the observer stood behind the start box between the two arms of the maze. The clear acrylic top of the maze was in place during all trials to keep the rat in the maze and to help reduce ambient noise. A decision was recorded when the head of the rat crossed into the next region. Decisions were recorded on paper in all cases. In addition, several trials were recorded on videotape. The trial was considered complete either when the rat made a Decision 5 (going to the end of one arm of the maze) or when 4 min had elapsed. At the end of each trial the rat was returned to its cage. Each rat performed two sets of five trials during a testing session. All four rats completed their first set, then the procedure began again. Each group was returned to its home cage at the end of the second set. Experiments were performed at 0, 1.5, and 4 T. Following the completion of all experimental trials at all three field strengths, a subset of four randomly selected rats performed five trials at 4 T in the same T maze with
and the outside painted with opaque white enamel. Each arm was 183 cm in length, 10 cm wide, and 10 cm deep. The top of the maze was removable and made of clear acrylic to facilitate cleaning and observation. Each arm of the maze was divided into five lengths, which will hereafter be called decisions. Decision 1 was 45.5 cm in length. Decisions 2-4 were 30.5 cm in length each; Decision 4 extended 15 cm inside the bore of the magnet (see Fig. 1). Decision 5 extended to the center of the magnet (46 cm). When in position the right arm of the maze (A) extended into the center of the magnet, while the left arm of the maze (B) extended into the center of the sham magnet (a set of gradient coils of the exact inner diameter of the magnet). Both the magnet and the sham were covered with red construction paper to eliminate visual cues. The magnet was situated in a very large common work area with uniform temperature, bright uniform overhead lighting, and a relatively high level of ambient noise. The light and tactile conditions experienced in the two arms of the maze were identical. The primary noise source in the environment was from gas venting from the magnet’s cryogenic system. This resulted in a constant white noise, which tended to drown out other intermittent noise in the environment. Noise amplitude from this source was constant during control and exposure. All experiments were performed utilizing a Baymed
4.0
3.0
z
_o_
1ST
+
4T
. 4A
3A
2A
:lA : 5 Gauss Line
Decision
Points
Fig. 1. The two axial field plots at 1.5 and 4 T taken from the center 1.5 T at magnet center and 6.8 G for 4 T at magnet center.
of the magnet.
At 91.5 cm, the field is 6.2 G
for
Bio-effects of high magnetic fields 0 J.
the arms reversed, so that the B arm led into the magnet. Upon encountering an aversive stimulus, it is expected that the rat would stop, rather than continue down the entire length of the maze. Thus, an increase in decisions is predicted at the point at which the magnetic field becomes an aversive stimulus. A Pearson Chi-square analysis was performed to evaluate statistical significance.
WEISS ET AL.
Table 1. Tabulation of outcomes collapsed to decision to turn into the magnet, away from the magnet, or no decision for control, 1.5 T, and 4 T trials
from magnet
Away
280
240
240
q
.2 .’ A
F? .9 $ P
160
‘ii 2 5 % e! LA
154 (55.00V0)
254 (90.71%)
(5.&o)
(6&o)
111 (39.64%)
109 (38.93%)
(2.5&o)
n Control q 1.5 Tesla
200 1.5Tesla
160 (57.14%)
4.0 T
magnet), region -5 (in the mock magnet) or no decision (0) was made. These data and Vo of total outcomes for each field strength are shown in Table 1. There was no difference found between 0 T and 1.5 T in arm preference. At 4 T, however, arm B (away from magnet) was clearly preferred (JI < .OOl). The results were unchanged when individual rats were maintained as a separate variable in the statistical analysis. When the arms of the magnet were reversed, the number of Decision 4 in the B arm (magnet arm) increased dramatically, while the number of Decision 5 decreased (see Fig. 3). The rats were observed to travel
280
200
1.5 T
&70)
magnet
RESULTS
t
ControI
No decision Toward
A partial tabulation of decisions is shown for both the magnet arm A (see Fig. 2A) and the sham arm B (see Fig. 2B) for all magnetic field strengths tested. For each trial, only the decisions indicating the furthest extent of travel down each arm are tabulated, giving at most two decisions per trial. One decision is tabulated if the rat stays in just one arm of the maze. The frequency of each decision is similar at 0 T (control) and 1.5 T when arm A and B are compared. At 4 T, however, there is a clear increase in Decision 5 in B and a decrease in Decision 5 in A. A Pearson Chi-square statistical analysis was performed after collapsing the results of each trial such that each trial was allowed only one outcome; thus, the rat ended up in region 5 (in the
691
4.0 Tesla 160
6 120
80
40
5 s z-
120
u!?!
80
40
1
3
2
Decision (A)
4 Points
5
1
2
4
3 Decision
5
Points
(B)
Fig. 2. (A) Represents the direction toward the magnet (arm A). (B) Represents the direction away from the magnet (arm B). Frequency of the decisions made is plotted on the y-axis, while distance is represented by the x-axis. Only the decisions indicating furthest extent of travel down each arm are tabulated. One decision is tabulated if the rat stays in just one arm of the maze.
Magnetic Resonance Imaging 0 Volume 10, Number 4, 1992
692
:
1
2
3
4
5
Decision Polnts
Fig. 3. Results of the random subset run. The direction has now been changed so that the magnet is now on the left (arm B) and the mock magnet is on the right (arm A). Tabulation is the same as for Fig. 2.
down the B arm, halt suddenly at Decision 4, and reverse direction. Several animals were seen to spread out all four limbs and back up the maze several steps before turning around. A Pearson Chi-square statistical analysis was performed on the 4-T data in Figs. 2 and 3. Chi-square was computed to be 60.09 with nine degrees of freedom. The increase in decision 4B shown in Fig. 3 over decisions in Fig. 2 proved to be statistically significant, with p < .OOOl. DISCUSSION
The results from 0 T (control) and 1.5 T indicate the rats exhibited no strong preference for either arm (A-magnet, B - nonmagnet). At 4 T there was a significant increase in selection of arm B. This clear preference for the “nonmagnet” arm of the maze suggests that an aversive stimulus was associated with the magnet at 4 T that was not present at 1.5 T. When the arms of the maze were reversed, the frequency of Decisions 3 and 4 in arm B (now the magnet arm) increased, suggesting that the rats had learned to travel toward the left in the initial experiments. After moving some distance down arm B toward the magnet the rats were observed to halt and reverse direction. They then proceeded all the way to the end of the other arm, resulting in a high number for Decision 5 in arm A. Thus,
the experimental results strongly suggest that some aspect of the environment around the magnet when at 4 T was quite aversive to the rats. A puzzling feature of the 4-T data is that the rats are making their decision to turn back in a region of relatively low static magnetic field. At decision point 5A the magnitude of the field is not much above 1.5 T, which in the previous trial caused no behavioral changes. That is, the subjects entered the center of the magnet, in which region they were experiencing a 1.5-T static magnetic field, with no evidence of aversion. The coarse data sampling does not allow us to make a firm statement about the exact magnetic field magnitude when the rats turned back, but most of the rats turned around before reaching decision point 5A. We propose that this behavior may be related to the magnetic field gradient in the area. If we examine the gradient of the field, that is, the rate of change along the z axis, we see that when the magnet is at 1.5 T the maximum gradient the rat is exposed to is approximately 0.07 T/m. When the magnet is at 4 T, however, the gradient at the mouth of the magnet is already above 0.05 T/m and rapidly climbs to 13 T/m in the area spanned by decision box 4A. These observations suggest that the magnetic field gradient as well as the absolute field magnitude may be a factor in the aversive stimulus. It is possible that this area of strong magnetic field gradient would be aversive even without motion, but it is not possible in these experiments to test for this condition independently. A stationary rat could interact with the magnetic field through either permanent or induced magnetic dipole moments in the tissues of the rat. A homogeneous field exerts a torque on the dipole, while a magnetic field gradient exerts a translational force. All tissue exhibits a weak induced diamagnetic dipole moment; this is not a very likely mechanism because of the weakness of the effect as well as the fact that the magnetic dipole moment of protein molecules would not change significantly with conformational change. Paramagnetic dipole moments, such as found in deoxyhemoglobin, myoglobin, cytochromes, and iron-sulfur proteins, would have at least one or two orders of magnitude greater interaction with the magnetic field than a purely diamagnetic protein molecule. It is difficult to postulate how an interaction with any of these molecules could result in a stimulus to the rat’s nervous system. Ferromagnetic deposits, in the form of magnetite particles or the iron-storage protein ferritin, would have an interaction several orders of magnitude stronger than paramagnetism. In fact magnetotactic bacteria have been identified that use this method to navigate along the earth’s magnetic field. Such a stimulus mechanism
Bio-effects of high magnetic fields 0 J. WEISSET AL.
sophisticated setup. Qualitative observations during the trial and on review of the videotape showed no obvious changes in the rate or pattern of travel down the maze between control and exposure. Table 2 shows estimates of dB/dt and resultant E field induced by these two types of motion assuming a velocity along the z axis of 0.2 m/set, a rotational velocity of 1 rad/sec, and a diameter for the rat’s head of 3 cm. Note that the Btrans in the gradient field at 4 T already exceeds the estimate of E,,, in the center of the magnet at 1.5 T. If the aversive stimulus was the electric field with a threshold on the order of 1 x 10m2 V/m (dB/dt = 1.3 T/set), it would explain why the rats turned away in an area of relatively low static magnetic field but high field gradient. This level is more than an order of magnitude below the thresholds reported for peripheral nerve stimulation (~61 T/set). Thus, direct nerve stimulation is not a likely mechanism for the aversive response. A threshold for magneto-phosphenes has been estimated at a dB/dt of approximately 1.3 T/sec6 dB/dt for translational movement in the gradient at 4 T are above this value, indicating that this could be a possible mechanism for the aversive response. There is also research showing that living organisms are sensitive to extremely weak electric fields and in the case of pulsating fields, may be able to detect fields below the thermal noise level.’ Proteins, such as the ion channel proteins in membranes, have many polar groups that affect their function and it is believed that electric fields can modulate their behavior. Electric fields are thought to interact directly with these proteins through these polar groups in their side chains, inducing an “electro-conformational” change. The noise level in membranes has been estimated to be on the order of 3 x lop3 to 3 x 10e2 V/m, which is on the order of our estimate for induced electric fields in the rat. Effects have been reported with field exposures as low as 2 x lop3 V/cm.* Although these studies are of the effects of time-varying electric fields, they are still relevant because of the time-varying nature of the electric fields induced in the rat. We can thus establish at
in the rat, however, would be expected to come into play at much lower field strengths than 1.5 T, where no aversive stimulus was observed. We suggest that it is more likely that the aversive effect is exerted via magnetic induction effects resulting from the rat’s motion in the magnetic field. According to Faraday’s law a change in magnetic flux through a loop of radius r will result in an induced electric field around the loop given by: r dB
E=--2z where E is the induced electric field in volts/meter, r is the radius of the loop in meters, and dB/dt is the time rate of change of magnetic field in T/set. In MRI this effect is usually thought of in the context of gradient switches or eddy currents in the organism from radiofrequency (RF) irradiation. An induction effect can also arise, however, from translational movement of the organism in a field gradient or rotational motion in the field. For translational motion in the z direction, dB/dt is given by: dB = grad, Bv, dt
where, v, is the velocity along the z axis and grad,B is the field gradient along z. For rotational motion, maximum dB/dt is given by: dB = -BQ dt
where Q is rotational velocity in rad/sec. To estimate electric fields induced in the rat, the motion is approximated by a translational motion along the z axis and a rotational motion of the head due to head wagging. Actual rate of travel of the rats in the maze is difficult to assess without a much more
Table 2. Estimates
1.5 T center 1.5 T 4A/SA
boundary boundary
electric
Field
Gradient
U-1
(T/m)
1.5
4 T center 4 T 4A/5A
of dB/dt and induced
0
field in the rat’s head as a result of motion dB&,,, (T/set)
dB/dt,,,
0
1.5
6
1.2
0.6
4
0
0
4
2.6
1.75
13
in the magnetic
field
(T/set)
0.6 1.75
693
0 9 x 10-3 0 1.95 x 10-2
1.1 x 10-2 4.5 x 10-3 3 x 10-2 1.3 x 10-2
Magnetic Resonance Imaging 0 Volume
694
least the plausibility of a time-varying electric field as a cause of an aversive stimulus. Electric fields in a conductive medium also give rise to currents which may have a biological effect. One proposed mechanism for the vertigo experienced by human subjects at 4 T is magneto-hydrodynamic pressure on the cupola within the vestibular mechanism.’ A minimum pressure threshold for sensations has been estimated at 1.25 x lo-’ Pa.9 Magnetohydrodynamic pressure is given by: P = baflB’/p
For humans, assuming a = .003 m, b = .00015m, Q = 10.5 radians/set, and B = 4 T, P was estimated to be 1.5 x 10e4 Pa, showing that stimulation of the cupola is quite plausible. A similar expression can be derived for linear motion in a gradient. For a canal in a plane perpendicular to the field: P = bav,B grad, B/p
.
Magnetohydrodynamic pressure in this case would be equal at all points around the canal and thus would not be expected to deflect the cupola. Magneto-hydrodynamic pressure does not appear to be a likely mechanism for an aversive stimulus arising from a field gradient. CONCLUSION The results of this study indicate that there is a measurable aversive effect at 4 T on the T maze behavior of rats, but no effect at 1.5 T. The induced aversion is most likely a result of magnetic induction effects arising from movement through an area of magnetic field gradient.
10, Number 4, 1992
Acknowledgments-This work was supported by a grant from the Kleberg Foundation. The authors gratefully acknowledge Aviva Ferragi and Jay Porter for providing magnetic field measurements presented in this manuscript.
REFERENCES 1. Schenck, J. Health and physiological effects of human exposure to whole body four tesla magnetic fields during MRI. In: R.L. Magin, R.P. Liburdy, B. Persson (Eds). Biological Effects and Safety Aspects of Nuclear Magnetic Resonance Imaging and Spectroscopy. New York Academy of Sciences Workshop, May 15-17, 1991. new method for 2. Hallett, M.; Cohen, L. Magnetism-a stimulation of nerve and brain. JAMA 264:538-541; 1989. 3. Cohen, M.S.; Weisskoff, R.M.; Rzedzian, R.R.; Kantor, H.L. Sensory stimulation by time-varying magnetic fields. Magn. Res. Med. 14:409-414; 1990. 4. Inms, N.K.; Ossenkopp, K.P.; Prato, ES.; Sestini, E. Behavioral effects of exposure to nuclear magnetic resonance imaging: II. Spatial memory tests. Magn. Rex Imaging 4:281; 1986. 5. Smirnova, N.P. Behavior in rats in “open field” following the action of a magnetic field. Zhurnal Uysskei Ner-
voi 32~72; 1982. 6. Anderson, L.E. Biological response of animals to timevarying magnetic fields. In: R.L. Magin, R.P. Liburdy, B. Persson (Eds). Biological Effects and Safety Aspects of Nuclear Magnetic Resonance Imaging and Spectroscopy. New York Academy of Sciences Workshop, May 15-17, 1991. 7. Weaver, J.C.; Astumin, B.D. The response of living cells to very weak electric fields: The thermal noise limit. Science 247~459-462; 1990. 8. McLeod, K.J.; Lee, R.C.; Ehrlich, H.P. Frequency dependence of electric field modulation of fibroblast protein synthesis. Science 236:1465-1469; 1987. range of 9. Oman, C.M.; Young, L.R. The physiological pressure difference and cupula deflections in the human semicircular canal. Acta Otolaryngol. 74~324-331; 1972.
