Magnetoreception in laboratory mice: sensitivity to extremely low-frequency fields exceeds 33 nT at 30 Hz.

0.44 MB·English

by  Prato, Frank S.

Magnetoreception in laboratory mice:
sensitivity to extremely low-frequency
fields exceeds 33 nT at 30 Hz
rsif.royalsocietypublishing.org
Frank S. Prato1,2,3, Dawn Desjardins-Holmes1, Lynn D. Keenliside1,
Janice M. DeMoor1, John A. Robertson1,3 and Alex W. Thomas1,2,3
1Bioelectromagnetics Group, ImagingProgram,Lawson Health ResearchInstitute, London,Ontario,
Research Canada N6A 4V2
2DepartmentofImaging, St. Joseph’s HealthCare, London, Ontario,Canada N6A 4V2
3DepartmentofMedicalBiophysics,Schulich School ofMedicineand Dentistry, Western University,London,
Citethisarticle:PratoFS,Desjardins-Holmes Ontario,Canada N6A 5C1
D, Keenliside LD, DeMoor JM, Robertson JA,
Thomas AW. 2013 Magnetoreception in Magnetoreception in the animal kingdom has focused primarily on behav-
ioural responses to the static geomagnetic field and the slow changes in
laboratory mice: sensitivity to extremely
its magnitude and direction as animals navigate/migrate. There has been
low-frequency fields exceeds 33nT at 30Hz.
relatively little attention given to the possibility that weak extremely
J R Soc Interface 10: 20121046.
low-frequency magnetic fields (wELFMF) may affect animal behaviour.
http://dx.doi.org/10.1098/rsif.2012.1046 Previously, we showed that changes in nociception underanambient mag-
netic field-shielded environment may be a good alternative biological
endpointtoorientationmeasurementsforinvestigationsintomagnetorecep-
tion.Hereweshowthatnociceptioninmiceisalteredbya30Hzfieldwith
Received: 19 December 2012 a peak amplitude more than 1000 times weaker than the static component
Accepted: 10 January 2013 of the geomagnetic field. When mice are exposed to an ambient magnetic
field-shielded environment 1h a day for five consecutive days, a strong
analgesic(i.e.antinociception)responseisinducedbyday5.Introductionof
a static field with an average magnitude of 44mT (spatial variability
of+3mT)marginallyaffectsthisresponse,whereasintroductionofa30Hz
Subject Areas:
time-varyingfieldasweakas33nThasastrongeffect,reducingtheanalgesic
biophysics effect by 60 per cent. Such sensitivity is surprisingly high. Any purported
detection mechanisms being considered will need to explain effects at
Keywords: suchwELFMF.
magnetic field sensitivity, ambient magnetic
field shielding, nociception, analgesia,
extremely low-frequency magnetic fields,
1. Introduction
geomagnetic static field
Theliteratureontheeffectsofgeomagneticfieldsonanimalorientationandhoming
hasfocusedprimarilyontheresponsetothestaticfieldratherthanthepotentialfor
a response to comparatively weak extremely low-frequency magnetic fields
Author for correspondence:
(wELFMF).AlthoughthereisawELFMF‘jitter’onthestaticfieldassociatedwith
Frank S. Prato
solaractivityandthunderstorms(e.g.Schumannresonances[1]),itisnormallyof
e-mail: [email protected] theorderofananotesla.Hence,itisunlikelythatsuchrandomweakfieldsaffect
animalorientationandhoming,exceptperhapsbybeingdisruptiveduringintense
solar activity and thunderstorms, when values can exceed 100nT. Of more rel-
evance is the observation that birds and other animals can orient using the
geomagneticfieldtoaprecisionofapproximately18[2–4].Thiswouldcorrespond
totheabilitytodetectwELFMFoftheorderoftensofnanotesla.
There have been a few attempts to establish a threshold for behavioural
effectsofwELFMF.Kirschvinketal.[5],usingafoodrewardexperiment,estab-
lishedthattheminimalfieldsdetectablebyhoneybeeswere4.3mTat10Hzand
430mTat60Hz.Nishimuraetal.[6]reportedthateffectsontailliftingbehav-
iourcouldbeobservedinaspeciesoflizardsat6and8Hzforapeakfieldof
2.6mT. In a study by Burda et al. [7], disruptive effects on alignment of rumi-
nants by a 50Hz field from high voltage electric power lines continued out
&
2013TheAuthors.PublishedbytheRoyalSocietyunderthetermsoftheCreativeCommonsAttribution
License http://creativecommons.org/licenses/by/3.0/, which permits unrestricted use, provided the original
authorandsourcearecredited.
toadistancebetween20and100m,wherethefieldstrengths fields, and (iii) the induction of analgesia may also be par- 2
arebetween0.1and1mT.Burgeretal.[8]reportedthata1mT tially attenuated by the introduction of a static field of
0.5Hzmagneticfieldaffectedc-Fosexpression,ameasureof geomagnetic amplitude (44mT+spatial variability of 7%). rs
neuronal activity, in the navigation circuit in Ansell’s mole- Earlier work [19] suggested an induced current mechanism if.ro
y
a
rats.Further,Ossenkoppetal.[9]observedthatopioid-related based on the observation that the size of the effect of the ls
o
behaviours in laboratory mice (CF-1) were affected on 17 introduced ELF magnetic fields was dependent on the pro- cie
December 1982 when there was a significant increase in the duct of the ELF frequency and the ELF magnitude. typ
u
degree of geomagnetic disturbance (from Ap index of 11 to However, how such weak fields could induce sufficient cur- blis
62 corresponding to approx. 120nT). There have also been rent remains a mystery [23]. In addition, other proposed hin
g
reports[10,11]thatanundergroundELFantennagenerating magnetoreception mechanisms, such as those based on rad- .o
0.1–0.5mT between 72 and 80Hz can disrupt bird orien- ical pairs [24] or arrays of iron particles [25–27], are also rg
tation, which supports an account of effects of geomagnetic theoretically challengedbysensitivities down to tens of nT. J
R
disturbances on orientation inbirds[12]. S
o
Althoughmostoftheworkonanimalnavigation,homing c
In
andorientationhasfocusedonnon-mammaliananimals,there 2. Material and methods te
havebeenmorerecentstudiesindicatingthatmammals,and rfac
e
specificallyrodents, canalso detectthegeomagneticfield.In 2.1. Animals 1
0
:
addition to the c-Fos work of Burger et al. [8], others have Adult male Swiss CD-1 mice (Charles River, St. Constant, 2
0
reported that modification of the direction of the static com- Quebec, Canada) two to four months old and weighing 25– 12
1
ponentofthegeomagneticfieldaffectslocationofbehaviours 35gwereused.Micewerehousedindividuallyinpolycarbonate 04
6
such asnest-makingand sleeping[13–16]. Inaddition,there cagesundera12L:12Dcycleat22+28C.Foodandwaterwere
isgrowingevidencethatlaboratoryrodents(Siberianhamster freelyavailable.
[17] and C57Bl/6J mice [18]) can be taught behaviours that
canbeaffectedwhenexposedtoalteredstaticmagneticfield 2.2. Assessment of nociception
components of the geomagnetic field. However, except for
Nociception(painsensitivity)wasmeasuredasthelatencyofafoot
the work by Burger et al., all of these studies ignored the
lifting/licktoanaversivethermalstimulus(hotplatetest;modelHP
non-static magnetic field components, i.e. they failed to
AccuScan Instruments, Inc., Columbus, OH) at 50+0.58C. The
report and account for the wELFMF that could have been maximum individual latency observed was 45s; hence, all mice
presentowingtogeomagneticdisturbancesoranthropogenic were removed from the heated surfacebefore the cut-off timeof
sourcesof50or60Hzmagneticfields.Inthisregard,itisof 60ssetbytheAnimalUseSubcommittee.
interestthatOliveriusovaetal.[16]pointoutthatorientation
of mole-rats in a rotated geomagnetic field exhibit much
2.3. Magnetic exposure conditions
higher scatter of bearings as compared with orientation in
Extremely low-frequency magnetic fields inside the mu-metal
the‘natural’magneticfield.Couldthisbecausedbya‘natural’
boxes(0.20–0.35mTstatic,lessthan0.001mT60Hz)wereattenu-
wELFMFthatwasnotrotated?Hence,wehaveproposedthat
atedmorethan50times[28]incomparisonwithambientfields
thresholdstothesensingofstaticandwELFMFshouldbecar-
(fromstaticto125Hz).Themagnitudeofthemagneticfieldasa
riedoutundermagneticfieldshieldingconditionswheresuch
function of frequency in the animal housing room and for the
fields can be carefully introduced without confounds of
testing room has been previously reported [19]. The static field
exposuretoambientwELFMF. was approximately 45mT in magnitude, and the ELF spectrum
Hereweshow,usingalaboratorymousemodelinwhich shows peaks at 60, 120, 180 and 240Hz, which combined have
magnetic sensitivity may be driven by the same mechanism an amplitude of approximately 0.15mT. The mu-metal boxes
as that for bird orientation and homing [19,20], that sensi- are identical to the one described by Koziak et al. [29]. The
tivity to tens of nanotesla does occur. This is consistent boxes (33(cid:2)38(cid:2)20cm) were made of 1.6mm (1/16 in) thick
with the concept that animals capable of using the geomag- mu-metal (Magnetic Shield Corp., Bensenville, IL or Amuneal
Manufacturing Corp., Philadelphia, PA), with magnetically
netic field for orientation and homing perceive the spatial
shielded holes (diameter of 2.5cm) at each of the four corners
variation in the static field as an induced wELFMF as the
(1cm off the sides) of both the base and top surfaces. A mu-
animal traverses that spatial variation. Alternatively, the
metal cylinder (2.5cm high) surrounded each hole, shielding
wELFMF variation in a projected constant field sensed by a
theambientmagneticfieldandlight.Themu-metalboxeswere
‘fixed’receptorcouldalsobeperceivedastheanimalchanges
laminatedinsidewithblackopaquepolyethylene,asitisimper-
its orientationbyadegree or two. vioustovirtuallyallsolvents. Individual micewereplacedin a
We have established that laboratory-bred mice, when 26(cid:2)16(cid:2)12cm clean transparent polycarbonate cage and cov-
exposed to an ambient magnetic field-shielded environment ered by a clear polycarbonate (Lexan) top with ventilation
for 1h per day for five consecutive days, develop a strong holes (diameter of 8mm). This cage was inserted into the
antinociceptive response by the fifth day [21]. This response mu-metalbox.Individualmicearefreetomovewithinthepoly-
is probablyopioid-mediated and is sensitive to the intensity carbonatecageinthemu-metalbox.Ashambox,identicaltothe
dimensions of the mu-metal boxes, was constructed of opaque
and wavelength of light [20], similar to some aspects of
fibreglassmaterialandwasalsolinedwithopaquepolyethylene.
bird navigation [22]. Using this experimental protocol, we
The sham box had no effect on the ambient magnetic field. In
estimated in a preliminary pilot experiment in 20 animals
addition,aboxmadeofstainlesssteel(AmunealManufacturing
that nociceptive behaviour effects may be detected for
Corp) was added that did not attenuate magnetic fields below
30Hz fields possibly as low as 30nT [19]. Here, we show
250Hzbymorethan3percentbutwasexpectedtosignificantly
that (i) the effect at 30Hz, 30nT is robust, because it has attenuate the ambient ELF electric field. This stainless steel
beenconfirmedin120animals,(ii)theinductionofanalgesia box was identical in appearance and similar in weight to the
isdefinitelynotrelatedtotheeliminationofambientelectric mu-metalboxes.
Asetoffourcoilswasplacedinsideallboxesexceptthefibre- levels in all boxes were below the sensitivityof the spectropho- 3
glass sham box. The coils were made of 150 turns of AWG #28 tometer (LightSpex; McMahan Research Laboratories, Chapel
magnetwire(0.32mmdiameterwithathinlayerofenamelinsula- Hill,NC)andcertainlybelow2.0(cid:2)1016photonss21m22. rs
tion), and were put on plastic rectangular formers with mean if.ro
dimensionsof30(cid:2)17.8cm.TheDCresistancewasapproximately ya
31Voneachformer.Thefourcoilswerespaced9.7cmapartina 2.5. Experimental procedures lsoc
ie
Merritt-likeconfigurationforatotallengthof29cmoncentre.The Three experiments were undertaken: the first to confirm the ty
p
coilswereconnectedtoeachchannelofafour-channelcontrolled effects of re-introduction of a 65nT, 30Hz field; the second to u
b
currentamplifier.Tominimizetheintroductionofanelectricfield, establish that a 33nT, 30Hz field was above the detection lis
h
the coils were electrically shielded with conductive silver paint thresholdandthethirdtoinvestigatetheeffectoftheintroduc- in
g
andcopperfoil.Connectionleadswereroutedthroughthemagne- tion of a static field of the same average magnitude as the .o
rg
ticallyshieldedholesatthetoprearcornerofthemu-metalenclosure ambient geomagnetic static field, but with a spatial variability
andconnectedtothecontrolledcurrentamplifier.Eachcoilcarries of +7% over the volume in which the mice could traverse. J
approximately3.787(cid:2)1024Apeaktodelivera500nTpeakfield. Note that all stated values of ELF magnetic field amplitude are RS
o
The coils are nominally 31V each (single wound). Calculated peakvaluesofthesinusoidalfields. c
power is (3.787(cid:2)1024)2(cid:2)31¼4.446(cid:2)1026W per coil or Inte
1.778(cid:2)1025W for the four coils. For sinusoidal operation, this rfa
valueismultipliedby0.707givinganapparentpowerof1.257(cid:2) 2.5.1. Experiment 1 ce
1
1025WRMS.Itisunlikelythatheatingisaconfoundinour30Hz Forfiveconsecutivedays,eachmouse(N¼240)wasexposedto 0:
2
experiments because the maximum ELF field used in these oneoffourconditions(for1h)andtestedpre-andpostexposure 0
1
experimentswas65nT.Incontrast,thepowerneededtoproduce (hotplatelatencyinseconds).Thetimeofdayoftheexposurewas 21
0
the44mTstaticfield(approx.0.1W)isequivalenttoaboutathird randomized between 09.00 and 15.00. The exposure conditions 46
ofthemetabolicrateofa30gmouse(approx.0.34W).Notethat usedweresham(fibreglassenclosurewithnoELFMFshielding),
within these coils, the magnetic field varied by +7% over the stainlesssteelenclosurewithoutactivationofinsertedcoils,posi-
volumeoccupiedbythemousecage. tive control (mu-metal enclosure providing static and ELFMF
Thesignalgeneratorwasanin-housearbitraryfunctiongen- shieldingwithoutactivationofinsertedcoils)andfivemu-metal
erator configured to produce either 64-step waveforms for enclosureswithactivecoilsat30Hzsinusoidalwithamagnetic
frequenciesabove51Hzor512-stepwaveformsforlowerfrequen- field amplitude at 65nT. Note that there were 30 animals per
cies.UsingaTektronixInc.(Beaverton,OR)TDS3032oscilloscope exposurecondition;however,animalswereexposedsingly.
(FFT Log scale Hanning window), the second harmonic was
measured at 256dB, the third at 249dB and the fourth at
2.5.2. Experiment 2
256dB of the fundamental frequencies for a 100Hz 64-step
sinusoidalpattern.The6.4kHzfundamentalofthe64stepsfora This was identical to experiment 1 except that four of the five
100Hzsinusoidalsignalwasalsomeasured.Thesecondharmonic experimental exposureboxesweresetat33nT,withthefifthat
is41dBbelowthe100Hzfundamental,andharmonicsareseen 65nT. As in experiment 1, there were 30 animals per exposure
up to the 20th at about 272dB (just visible above noise). Note conditionforatotalof240animals.
that,fortheintroductionofastaticfield,arippleatELFfrequencies
wouldbeattenuatedbymorethan70dB,asthe60Hzrippleon 2.5.3. Experiment 3
the+15Vamplifierisapproximately1percentandtheoutput
Thisonlyusedfourboxes.Thefibreglasssham,thestainlesssteel
opampshaveatypicalpowersupplyrejectionratioof69dB.
controlandthemu-metalpositivecontrolwereidenticaltothose
Amplitude calibration was done by selecting sets of current
in experiments 1 and 2.The fourth boxwasmu-metal with the
limitingresistorsandadjustingtheamplifiergaincontroltoachieve
coil energized to produce an average horizontal static field of
adesiredfieldstrength,asmeasuredbyaBartington3-axisfluxgate
44mT with a variation of +7% over the volume occupied by
magnetometer(MAG-03MS1000,BartingtonInstruments,Oxford,
themousecagesetwithineachbox.Notetherewere40animals
UK) with afull-scalewaveform pattern.Subsequent lower fields
perexposureconditionforatotalof160.
wereobtained bychangingthesoftware:dividing thewaveform
In order to keep the experiments double-blinded regarding
pattern amplitudes by the required factor and uploading as
exposure conditions, the signal generators did not display the
needed.Fieldstrengthwasmeasuredaspeakvalues.Theaccuracy
amplitudeorpatternselectionandonlyL.D.K.(atechnicalman-
ofourMAG-03MS1000magnetometerwascheckedbycomparison
agerseparatefromtheactualexperiments)knewtheconditions.
with our MAG-03MC 500 unit for which we have a calibrated
WhileD.D.H.carriedouttheexperiments,A.W.T.completedthe
source (MAG-03MC-CU). No drift in the accuracy of the unit
analysispriortoL.D.K.disclosingtheexposureconditions.
couldbedetected.
Frequencycalibrationwasachievedbyselectionofappropri-
ate waveform patterns and numerical calculation of the point 2.6. Data analysis
latency. Output frequency was then confirmed with both a
Foreachofthethreeexperiments,thedatawereanalysedusingan
Fluke 179 Multimeter (Fluke Electronics Canada, Mississauga,
omnibusmixeddesignanalysisofvariance(ANOVA;IBMSPSS
ON,Canada)andaTektronixTDS3032oscilloscope.
Statistics v. 19.0, USA) investigating the relationships between
Acoustic properties inside the test enclosures were checked
withaBru¨el&Kjær(DK-2850Nærum,Denmark)one-inchcon- day(5,repeated),pre-versuspostexposure(2,repeated)andcon-
dition(8,independent,forexperiments1and2;4,independent,for
densermicrophone(model4131)andpreamplifier(model2801).
experiment3).ParametricposthocanalysisusingTukey’shonestly
OutputwasmeasuredwiththeoscilloscopeusingFFTLogscale
significantdifference(HSD)examinedthedifferencesbetweensets
Hanning window and showed no detected signal outside of
ofdata.Furtherparametricposthocanalysis(single-tailedpaired
normalambientbackground.
t-test)comparedpost-exposurelatencytimesondays4and5for
eachofthefive30Hzexposuresinexperiments1and2withthe
2.4. Lighting conditions
positivecontrol(0nT)andwiththeshamcontrol(fibreglassbox)
Becausewehaveshownthattheinducedanalgesiacanbeattenu- fortherespectiveexperiments.Inexperiment3,days4and5of
ated/abolishedbysimultaneouslightexposure,carewastakento the 44mT exposed animals were similarly compared with the
eliminate light intensity and wavelength as confounders. Light positivecontrol(0nT)andtheshamcontrol(fibreglassbox).
30 4
rs
25 if.ro
y
a
ls
o
c
20 iety
p
u
latency (s) 15 blishing.org
10
J
R
S
o
c
5 In
te
rfa
c
e
0 1
0
:
sham s-steel mu-metal 65 nT 65 nT 65 nT 65 nT 65 nT 20
control control control 30 Hz 30 Hz 30 Hz 30 Hz 30 Hz 12
1
0
condition 4
6
Figure 1. Pre- (open circles) and post-exposure (filled circles) latencies for 30Hz, 65nT as well as fibreglass sham, stainless steel control and 0nT (mu-metal
positivecontrol).Thefiveentriesforeachoftheeightconditionsrefertothefiveexperimentaldays.Errorbarscorrespondto+standarderrorofthemean(s.e.m.).
Wheres.e.m.barsarenotevident,theyfallwithinthesymbols.TheANOVAanalysisshowedasignificantthree-wayinteractionbetweenday(5,repeated)bypre–
post (2, repeated) by condition (8, independent) (F ¼5.30, p,0.001,h2¼0.14).
2,28
30
25
20
s)
y (
c
n 15
e
at
l
10
5
0
sham s-steel mu-metal 33nT 33nT 33nT 33nT 65nT

control control control 30Hz 30Hz 30Hz 30Hz 30Hz

condition
Figure2.Pre-(opencircles)andpost-exposure(filledcircles)latenciesfor30Hz,65nTandfour30Hz,33nTexposuresaswellasafibreglasssham,astainless
steel control and a 0 nT mu-metal positive control. The five entries for each of the eight conditions refer to the five experimental days. Error bars correspond
to+standard error of the mean (s.e.m.). Where s.e.m. bars are not evident, they fall within the symbols. The ANOVA analysis showed a significant three-
way interaction between day (5, repeated) by pre–post (2, repeated) by condition (8, independent) (F ¼4.00, p,0.001,h2¼0.11).
2,28
A third analysis was performed to determine the fractional 3. Results
reductioninanalgesiathatcouldbeattributedtothesinusoidal
or static magnetic fields introduced into the mu-metal box. Figures1–3showtheresultsofexperiments1,2and3.Forall
Thiswasdonebycalculatingthesumofthedifferencesbetween three experiments, the ANOVA analysis (details in the
post- and pre-latencies for days 3, 4 and 5 for each of the respective figure captions) showed a highly significant
five groups in experiments 1 and 2 and for the single exposure three-way interaction. Post hoc Tukey’s HSD analyses of
group in experiment 3. This sum was subtracted from the homogeneity for sets of data across conditions are shown in
sum for the respective positive control for each experiment. tables1–3forexperiments1–3,respectively.Table4provides
This difference was then divided by the summed latencies
the significance values calculated for the post hoc compari-
for the respective positive control. The result corresponds to
sons for days 4 and 5 and the fractional decreases summed
the fractional decrease in induced analgesia for days 3, 4 and
overdays 3, 4and 5in theinduced analgesia.
5 caused by the introduced magnetic field. This provided
For experiment 1, the introduction of the 65nT, 30Hz
a total of 11 outcomes, six at 65nT, four at 33nT and one for
exposure reduced the induction of analgesia observed on
thestaticfield.
30 Table 2. Post hoc Tukey’s HSD analysis of homogeneity for sets of data 5
across conditions for experiment 2. (Means for groups in homogeneous
25 subsets are displayed.The error term is m.s.e.¼3.490). rs
if.ro
y
20 a
s) subset lso
atency ( 15 ecoxnpdoistuioren n 1 2 3 cietypub
l 10 lis
h
s-steel control 30 14.3267 in
g
5 .o
33nT, 30Hz 30 14.9400 14.9400 rg
0 sham control 30 15.2700 15.2700 J
sham s-steel mu-metal static R
control control control 44 mT 33nT, 30Hz 30 15.5167 15.5167 So
c
condition 33nT, 30Hz 30 15.7467 15.7467 In
te
Figure3.Pre-(opencircles)andpost-exposure(filledcircles)latenciesfora 33nT, 30Hz 30 15.7633 15.7633 rfa
c
44mTstaticfieldexposureaswellasafibreglasssham,astainlesssteelcon- 65nT, 30Hz 30 15.8233 15.8233 e
1
trol and a 0nT mu-metal positive control. The five entries for each of the 0:
mu-metal 30 17.2967 2
four conditions refer to the five experimental days. Error bars correspond 0
1
to +standard errorof the mean (s.e.m.). Where SEM bars are not evident, control 21
0
4
theyfallwithinthesymbols.TheANOVAanalysisshowedasignificantthree- significance 0.063 0.599 0.051 6
way interaction between day (5, repeated) by pre–post (2, repeated) by
condition (4, independent) (F ¼5.98, p,0.001,h2¼0.13).
2,38
Table 3. Post hoc Tukey’s HSD analysis of homogeneity for sets of data
Table 1. Post hoc Tukey’s HSD analysis of homogeneity for sets of data across conditions for experiment 3. (Means for groups in homogeneous
across conditions for experiment 1. (Means for groups in homogeneous subsets aredisplayed.The errorterm is m.s.e.¼6.413).
subsets aredisplayed.The errorterm is m.s.e.¼2.200).
subset
subset exposure
exposure condition n 1 2 3
condition n 1 2 3
s-steel 40 13.3250
s-steel control 30 13.6600 control
sham control 30 15.0300 sham control 40 14.5225 14.5225
65nT, 30Hz 30 15.1967 static 44mT 40 15.9600 15.9600
65nT, 30Hz 30 15.5867 mu-metal 40 16.8975
65nT, 30Hz 30 15.7133 control
65nT, 30Hz 30 15.8233 significance 0.153 0.058 0.351
65nT, 30Hz 30 16.0933 16.0933
mu-metal 30 17.0700
control
4. Discussion
significance 1.000 0.106 0.180
Theresultsofexperiment1(figure1)indicateanattenuation
ininducedanalgesiaof58percentforanexposureof65nTat
30Hz (peak amplitude frequency product of 1950nT-Hz)
the third to fifth days of testing versus that in the positive (table 4). This is consistent with fractional attenuation
controlby58+3%.Forexperiment2,33nT,30Hzreduced values of 0.44 for 1500nT-Hz (at 30Hz) and 0.60 for
the induction of analgesia seen in the positive control by 3000nT-Hz (at 30Hz) previously reported by our group
60+8%.Inexperiment3,thestaticfieldreducedtheinduc- [19].Resultsfromexperiment2(figure2)indicateanattenu-
tion of analgesia on the third to fifth days of testing versus ation in induced analgesia of 60 per cent foran exposure of
that in the positive control by 30.5 per cent. As expected in 990nT-Hz (table 4), which seems high based on previous
all three experiments, the sham exposure did not induce work. However, this recent work is based on an experiment
analgesia. The stainless steel box, anticipated to shield only that included a total of 120mice, whereas the previous
the ELF electric fields, also did not induce analgesia in any work only had results for 20mice. In figure 2, relatively
of theexperiments (figures1–3). largeeffectsareseenat33nT,30Hz,whichstronglysuggests
Thelatencytimespost-exposureondays4and5weresig- the threshold of detection is below 1000nT-Hz! Additional
nificantly different (p,0.05) as compared with the mu- experimentsfurtherreducingpeakfieldareneededtotitrate
metalcontroland theshamcontrolforallthreeexperiments the threshold at 30Hz and to determine any dependence of
(see table 4). thatthreshold withfrequency.
Table 4. Single-tailed posthoc significance (p)values comparing post-exposure latency times on days 4and 5in magnetic field-exposed mice tothe respective 6
positive (0nT) and sham (fibreglass) controls, and theinduced fractional attenuationin induced analgesia(exp, experiment; freq,frequency; ampl, amplitude).
rs
statistical significance (p-value) if.roy
a
ls
o
c
ie
at4 at4 at5 at5 ty
p
days days days days u
peak b
lis
exp. freq. ampl. fractional attenuation h
in
no. (hz) (nT) nT-Hz to 0nT to sham to 0nT to Sham in induced analgesia g.o
rg
exp. 1 30 65 1950 ,0.0001 ,0.0001 ,0.0001 ,0.0001 0.449
J
30 65 1950 ,0.0001 0.0010 ,0.0001 ,0.0001 0.605 R
S
o
30 65 1950 ,0.0001 0.0040 ,0.0001 ,0.0001 0.613 c
In
30 65 1950 ,0.0001 0.0020 ,0.0001 ,0.0001 0.655 terfa
c
30 65 1950 ,0.0001 0.0005 ,0.0001 ,0.0001 0.585 e
1
0
exp. 2 30 65 1950 ,0.0001 ,0.0001 ,0.0001 ,0.0001 0.554 :
2
0
30 33 990 0.0004 ,0.0001 ,0.0001 ,0.0001 0.361 12
1
0
30 33 990 ,0.0001 0.0002 ,0.0001 ,0.0001 0.631 46
30 33 990 ,0.0001 0.0200 ,0.0001 0.0002 0.752
30 33 990 ,0.0001 0.0003 ,0.0001 ,0.0001 0.646
exp. 3 0 44000 0 0.0320 ,0.0001 0.0260 ,0.0001 0.305
Experiment3(figure3)suggestsforthefirsttimethatthis is difficult to understand how these mechanisms could
experimental protocol may detect effects of exposure to a explaineffectsofwELFMFof33nTpeakat30Hz.Ourearlier
static field with a variation of +3mT. Previous experiments work suggested effects consistent with an induced current.
hadsuggestedthattheinductionofanalgesiawasnotassoci- However, the formulation of such a mechanism based on
atedwiththeshieldingofthestaticgeomagneticfield.But,in induced currents faces formidable challenges associated
that earlier work [28], the Plexiglass box was placed within with background tissue noise [23]. Further, the results
threeorthogonallynestedcoilsthatzeroedthestaticambient reportedherecastdoubtonourearlierhypothesisthateffects
field but not the ELF field, and the attenuation of stress- nearthethresholdvaryastheamplitudefrequencyproduct.
inducedanalgesiawasinvestigatedforasingle2hexposure. With respect to iron particles, Kirschvink et al. [5] pro-
Note that these square Helmholtz-like coils were approxi- posed that the detection mechanism for magnetoreception
mately2mindiameterandhencehadrelativelylittlespatial in bees is due to iron particles that hold a permanent mag-
variationoverthedimensionscorrespondingtothespaceavail- netic dipole. Experiments done by Walker & Bitterman [30]
able for mouse movement. Herewe have re-introduced only suggest that the median threshold sensitivity to a static
the magnitude of the local static ambient field. Note that an field is 260nT but they report one bee with a sensitivity
effect of 30.5 per cent attenuation in induced analgesia is of down to 26nT. Kirschvink et al. [5] explored bee sensitivity
interest.Inpreviouswork[19],weshowedthatthemaximum to a 10 and 60Hz field and determined thresholds of 4.3
attenuation in induced analgesia through the introduction and 430mT, respectively. This is clearly far in excess of
ofELFmagneticfieldswas70.7percent.Thecombinationof 33nT. However, their experimental set-up requires the bees
theeffectsofthestaticfield(30.5%)andtheELFfield(70.7%) to travel through a strong magnetic field gradient under
equalsapproximately100percentattenuationand,therefore, ambient magnetic field conditions. Polk [26] and Adair [25]
wespeculatethatthismayexplaintheentirephenomenon. estimated athreshold of 5mTat 60Hz fora single magnetic
There is a major potential confound to the interpretation magnetite particle, which seems low compared with the
of experiment 3. It is entirely possible that the mice were results of the bee work. In addition, arrays of iron particles
exposed to an apparent ELF magnetic field owing to move- could increase sensitivity and one-dimensional arrays are
ment of the mice within the spatial variation in this static clearly used and effective in magnetotactic bacteria [31]. If
field, i.e. variations as large as+3mT. If this turns out to we estimate an improvement in signal to noise proportional
be correct, then the effect produced by this ‘apparent’ ELF to the square root of the number of particles in the array,
is at a higheramplitude but much lower frequency thanthe then approximately 10000 would be needed to allow
33nT,30Hzexposures.Experimentstoresolvethisquestion athresholdof50nT,ratherthan5mT.Inaddition,ifthepar-
shouldbeundertakenwhere(i)thegradientisreduced,and/ ticles were too small or too large to hold a permanent field,
or (ii) the gradient is increased while maintaining the same then an array of such iron magnetite particles could also
averagestaticfield of 44mT. explain effectsthat have been reported that do not discrimi-
An effect of Earth-strength static magnetic fields on lab- nate between the Earth’s magnetic poles. Note that the very
oratory mice could involve at least two mechanisms important recent workof Wu & Dickman [27] demonstrates
currentlybeingdebatedintheanimalorientation/hominglit- convincingly that there is a neurosubstrate for magnetic
erature: a radical pair [24] and/or arrays of iron particles field detection including polarity, direction and magnitude
made of magnetite [27]. However, as presently modelled, it of a static field. But their conjecture that the detector is a
permanent magnetic particle is simply based on speculation theELFfrequency,theELFamplitudeandtherelativeorien- 7
driven by alternate neuronal firing when the static field is tation of the ELF field vector, and that of a ‘superimposed’
rotated 1808. In their paper they report a static magnetic staticfield[36,37].Itisofinterestthatthesetheorieswereheav- rs
field threshold of 20mT. Again, experiments are conducted ily criticized [38,39] in that the oscillating metal ion in the if.ro
y
a
under ambient wELFMF conditions. It would be important potential well could not achieve coherence long enough to ls
o
to see the effect on neuronal firing in (i) a shielded environ- detect anELFMF. Binhi [40] has recast this model to address cie
mentand (ii) underwELFMFin ashieldedenvironment. these issues and has fit theoretical curves to the data from a typ
u
Withrespecttothefreeradicalbiophysicaldetectionmech- numberofthecyclotronresonancesinvestigators(seeespecially blis
anism,themainexperimentalindicationsarelightsensitivity/ figs 2.11, 4.32, 4.45). Also, Lednev’s work should not be dis- hin
g
dependencyandaninclinationcompass(meaningtheinability counted given his prediction, supported by experimental .o
to detect the Earth’s magnetic field polarity). However, an data, that effects are expected under zeroed static field con- rg
alternative explanation is an array of non-magnetic particles ditions [36,37]. Note that similar criticisms are currently an J
R
and the need for light to trigger an essential post-detection issuewithrespecttotheradicalpairmechanism[41].However, S
o
event.Theattractivenesstosomeisthatrequirementsofcoher- theevidenceassociatedwithorientation,migrationandhoming c
In
enceandentanglementofafreeradicalpairwouldusherina has become so convincing in recent years that some are te
quantumbiologyeffect [24,32,33].But,inarecenttheoretical suggestingthatquantumcoherenceofthespinstatesmustbe rfac
e
analysis,Hogbenetal.[34]suggestthatcoherenceandentan- preservedevenunder‘warmandwet’conditions[32].Perhaps, 1
0
:
glement are, in fact, not needed. The authors argue that theprogressbeingmadeinsolid-statephysicswithrespectto 2
0
liftingthisrequirement‘offersnewandmoreflexibleopportu- spinquantummemory[33]mayprovidecluesastohowspin 12
1
nitiesforthedesignofbiologicallyinspiredmagneticcompass coherencelifetimecanbeextended.Suchanextensionwould 04
6
sensors’. However, the radical pair mechanism is certainly imply isolation from the ‘thermal bath’, but would also need
challenged to explain thresholds of 33nT at 30Hz without to address issues of frequency preference and the very low
prolongedcoherenceandentanglement.SensitivitytoanELF detectionthresholdestablishedherein.Abetterestablishment
frequencymayinfactbeapropertyofa‘downstream’biologi- of the sensitivity threshold of wELFMF and dependence on
caleventthatdetermineswhetherornottheinitialdetection frequencywouldhelpvalidatethesemodelsandallowacalcu-
results in a behavioural/physiological response. This would lationoftheneededradicalpaircoherencetime.
eliminatetheneedoftheradicalpairtobesensitivetomagnetic
fields over a number of milliseconds, and hence the The animals were housed, and all experiments were conducted,
according to the Canadian Council on Animal Care guidelines as
dependence on a prolonged coherence and entanglement
approved by the Animal Use Subcommittee of the University of
timewouldnolongerberequired.
WesternOntario(protocolno.2008-083).
In the non-orientation bioelectromagnetics literature,
there has been a data-supported theory that ion–protein
ThisworkwassupportedinpartbyaCanadianInstitutesofHealth
complexes can detect wELFMF [35]. The associated initial Research operating grant (FRN # MOP 43874). We would like to
transduction theories require specific resonances between thankShelaghRossforassistanceinpreparingthismanuscript.
References
1. Polk C. 1982Schumann resonances. In CRC 7. BurdaH,BegallS,CervenyJ,NeefJ,NemecP.2009 12. Moore FR.1977Geomagnetic disturbanceand the
HandbookofAtmospherics, vol. 1(ed.H Volland), Extremely low-frequencyelectromagnetic fields orientation ofnocturnally migrating birds. Science
pp.111.Boca Raton,FL: CRC Press. disruptmagneticalignmentofruminants.Proc.Natl 196, 682–684.(doi:10.1126/science.854743)
2. Wiltschko W, Wiltschko R.1996Magnetic Acad.Sci.USA106,5708–5713.(doi:10.1073/pnas. 13. HollandRA, Thorup K,Vonhof MJ, Cochran WW,
orientation in birds. J.Exp.Biol.199,29–38. 0811194106) Wikelski M. 2006Navigation:batorientationusing
3. IoaleP,TeyssedreA.1989Pigeonhoming:effectsof 8. Burger T,Lucova M, MoritzRE, Oelschlager HH, Earth’s magneticfield. Nature 444, 702.(doi:10.
magneticdisturbances before release oninitial Druga R,Burda H,Wiltschko W, WiltschkoR, 1038/444702a)
orientation. Ethol. Ecol. Evol.1, 65–80. (doi:10. Nemec P.2010Changing and shielded magnetic 14. Burda H,Marhold S, Westenberger T, Wiltschko R,
1080/08927014.1989.9525531) fields suppressc-Fosexpression inthe navigation WiltschkoW.1990Magneticcompassorientationin
4. Walcott C.1978Anomaliesin theEarth’s circuit: input from the magnetosensorysystem thesubterranean rodentCryptomys hottentotus
magneticfield increasethescatterofpigeon’s contributestothe internalrepresentation ofspace (Bathyergidae). Experientia46,528–530.(doi:10.
vanishing bearings. In Animalmigration, in asubterranean rodent.J.R. Soc. Interface7, 1007/BF01954256)
navigation andhoming(ed.K Schmidt-Koenig, 1275–1292. (doi:10.1098/rsif.2009.0551) 15. KimchiT, Terkel J. 2001Magnetic compass
WT Keeton),pp.143–151.Heidelberg, 9. Ossenkopp KP, KavaliersM, HirstM. 1983 orientationintheblindmoleratSpalaxehrenbergi.
Germany: Springer. Reduced nocturnal morphineanalgesia in J.Exp.Biol.204,751–758.
5. Kirschvink J, Padmanabha S, Boyce C,Oglesby J. mice following ageomagnetic disturbance. 16. OliveriusovaL,NemecP,KralovaZ,SedlacekF.2012
1997Measurement ofthethreshold sensitivityof Neurosci. Lett. 40,321–325.(doi:10.1016/ Magnetic compassorientation in twostrictly
honeybeestoweak, extremely low- 0304-3940(83)90059-9) subterranean rodents:learned orspecies-specific
frequency magneticfields. J.Exp.Biol.200, 10. Southern WE. 1975Orientation ofgull chicks innate directional preference?J. Exp.Biol.215,
1363–1368. exposedtoprojectSanguine’selectromagneticfield. 3649–3654.(doi:10.1242/jeb.069625)
6. NishimuraT,OkanoH, TadaH,NishimuraE, Science 189,143–145.(doi:10.1126/science. 17. DeutschlanderME,FreakeMJ,BorlandC,PhillipsJB,
Sugimoto K,Mohri K,Fukushima M. 2010Lizards 1138371) MaddenRC,AndersonLE,WilsonBW.2003Learned
respondtoan extremely low-frequency 11. Larkin RP, Sutherland PJ. 1977Migratingbirds magneticcompassorientation bythe Siberian
electromagneticfield.J.Exp.Biol.213,1985–1990. respond toProjectSeafarer’selectromagnetic field. hamster Phodopus sungorus. Anim.Behav.65,
(doi:10.1242/jeb.031609) Science 195,777–779. 779–786.(doi:10.1006/anbe.2003.2111)
18. MuheimR,Edgar NM,SloanKA, PhillipsJB. 2006 25. AdairRK.1994Constraintsofthermalnoiseonthe 34. HogbenHJ, Biskup T,HorePJ. 2012Entanglement 8
Magnetic compassorientation in C57BL/6Jmice. effects ofweak60-Hzmagnetic fields actingon andsourcesofmagneticanisotropyinradicalpair-
Learn Behav. 34, 366–373.(doi:10.3758/ biological magnetite.Proc. Natl Acad. Sci.USA91, basedavianmagnetoreceptors.Phys.Rev.Lett.109, rs
BF03193201) 2925–2929. (doi:10.1073/pnas.91.8.2925) 220501.(doi:10.1103/PhysRevLett.109.220501) if.ro
y
19. PratoFS, Desjardins-HolmesD, KeenlisideLD, 26. PolkC. 1994Effectsofextremely-low-frequency 35. Binhi VN.2000Amplitude and frequency als
o
DeMoorJM,RobertsonJA,StodilkaRZ,ThomasAW. magnetic fields onbiologicalmagnetite. dissociation spectraofion-protein complexes cie
2011Thedetection threshold forextremely low Bioelectromagnetics 15, 261–270.(doi:10.1002/ rotatinginmagneticfields.Bioelectromagnetics21, typ
u
frequency magneticfields may be below bem.2250150308) 34–45.(doi:10.1002/(SICI)1521-186X b
lis
1000nT-Hzin mice.Bioelectromagnetics32, 27. WuLQ, Dickman JD.2012Neural correlates ofa (200001)21:1,34::AID-BEM6.3.0.CO;2-8) hin
g
561–569. (doi:10.1002/bem.20661) magnetic sense. Science 336, 1054–1057. 36. Lednev VV. 1996Bioeffectsofweak combined, .o
20. PratoFS, Desjardins-HolmesD, KeenlisideLD, (doi:10.1126/science.1216567) constantandvariablemagneticfields.Biofizika41, rg
McKay JC,Robertson JA, Thomas AW. 2009Light 28. Choleris E,DelSeppiaC, Thomas AW, Luschi P, 241–252. J
R
altersnociceptiveeffectsofmagneticfieldshielding GhioneG,MoranGR,PratoFS.2002Shielding,but 37. Lednev VV, SrebnitskayaLK, Il’yasova YN, S
o
in mice:intensityand wavelength considerations. notzeroingoftheambientmagneticfieldreduces RozhdestvenskayaZY, KlimovAA, Belova NA,Tiras c
In
J.R. Soc. Interface6, 17–28. (doi:10.1098/rsif. stress-induced analgesia inmice. Proc. R.Soc. B KP. 1997Magnetic parametric resonancein te
2008.0156) 269,193–201.(doi:10.1098/rspb.2001.1866) biosystems:experimental verification ofthe rfac
e
21. PratoFS, Robertson JA,Desjardins D, HenselJ, 29. KoziakAM,DesjardinsD,KeenlisideLD,ThomasAW, predictions ofatheoryusing regenerating 1
0
:
Thomas AW. 2005Daily repeated magneticfield PratoFS.2006Light altersnociceptiveeffectsof planarians Dugesiatigrina asatest system. 2
0
shielding inducesanalgesia in CD-1mice. magnetic fieldshielding. Bioelectromagnetics27, Biophysics 41, 825–835. 12
1
Bioelectromagnetics 26,109–117.(doi:10.1002/ 10–15. (doi:10.1002/bem.20170) 38. Engstrom S. 1996Dynamic properties ofLednev’s 04
6
bem.20056) 30. WalkerME,BittermanME.1989Honeybeescanbe parametric resonancemechanism.
22. WiltschkoR,StapputK,ThalauP,WiltschkoW.2010 trained torespond toverysmall changesin Bioelectromagnetics17, 58–70. (doi:10.1002/
Directionalorientation ofbirds bythemagnetic geomagnetic fieldintensity.J.Exp. Biol.145, (SICI)1521-186X(1996)17:1,58::AID-BEM8.3.0.
fieldunder differentlight conditions. J.R. Soc. 489–494. CO;2-5)
Interface7(Suppl.2),S163–S167.(doi:10.1098/rsif. 31. BlakemoreRP.1975Magnetotacticbacteria.Science 39. AdairRK. 1998A physical analysis oftheion
2009.0367.focus) 190,377–379.(doi:10.1126/science.170679) parametric resonancemodel.Bioelectromagnetics
23. Adair RK. 1991Constraintsonbiological effectsof 32. Gauger EM,Rieper E,MortonJJL, Benjamin SC, 19,181–191.(doi:10.1002/(SICI)1521-
weakextremely-low-frequencyelectromagnetic Vedral V. 2011Sustained quantumcoherence 186X(1998)19:3,181::AID-BEM6.3.0.CO;2-Y)
fields. Phys.Rev. A 43, 1039–1048.(doi:10.1103/ and entanglementin the avian compass. Phys. 40. BinhiVN.2002Magnetobiology:underlyingphysical
PhysRevA.43.1039) Rev.Lett. 106,040503. (doi:10.1103/PhysRevLett. problems, pp.473.Bath, UK: Academice Press.
24. Mouritsen H,HorePJ. 2012Themagneticretina: 106.040503) 41. AdairRK.1999Effectsofveryweakmagneticfields
light-dependent and trigeminal magnetoreception 33. BoehmeC,McCameyDR.2012Physics.Nuclear-spin onradicalpairreformation.Bioelectromagnetics20,
in migratorybirds. Curr.Opin. Neurobiol. 22, quantummemorypoisedtotakethelead.Science 255–263.(doi:10.1002/(SICI)1521-186X
343–352. (doi:10.1016/j.conb.2012.01.005) 336,1239–1240. (doi:10.1126/science.1223439) (1999)20:4,255::AID-BEM6.3.0.CO;2-W)