118

5 NEUROBEHAVIOURNeurobehavioural studies encompass the effects of exposure to

ELF electromagnetic fields on the nervous system and its responses at differ-ent levels of organization. These include the direct stimulation of peripheraland central nerve tissue, perceptual effects resulting from sensory stimula-tion, and effects on central nervous system function. Effects on the latter canbe assessed both electrophysiologically by recording the electrical activity ofthe brain, and by tests of cognition, assessment of mood, and other studies.

The nervous system also has a central role in the control of otherbody systems, particularly the cardiovascular system, through direct nervouscontrol, and the endocrine system, through neural input into the pineal andpituitary glands. These glands in turn influence reproduction and develop-ment, and in a more general way, physiology and well-being.

The brain and nervous systems function by using electrical signals,and may therefore be considered particularly vulnerable to low frequencyEMFs and the resultant induced electric fields and currents. Substantial num-bers of laboratory experiments with volunteers and animals have investigatedthe possible consequences of exposure to weak EMFs on various aspects ofnervous system function, including cognitive, behavioural and neuroendo-crine responses. In addition, epidemiological studies have been carried outon the relationship between EMF exposure and both suicide and depression.

These studies have been reviewed by NRC (1997), NIEHS (1998),IARC (2002), ICNIRP (2003) and McKinlay et al. (2004). In particular,ICNIRP (2003) reviewed in detail some of the evidence summarized here.

In general, there are few effects for which the evidence is strong,and even the more robust field-induced responses seen in the laboratory stud-ies tend to be small in magnitude, subtle and transitory in nature (Crasson etal., 1999; Sienkiewicz et al., 1993).

5.1 Electrophysiological considerationsAn examination of the electrophysiological properties of the ner-

vous system, particularly the central nervous system (CNS: brain and spinalcord) gives an indication of its likely susceptibility to the electric fieldsinduced in the body by EMF exposure. Ion channels in cell membranes allowpassage of particular ionic species across the cell membrane in response tothe opening of a “gate” which is sensitive to the transmembrane voltage(Catterall, 1995; Hille & Anderson, 2001; Mathie, Kennard & Veale, 2003).It is well established that electric fields induced in the body either by directcontact with external electrodes, or by exposure to low frequency magneticfields, will, if of sufficient magnitude, excite nerve tissue through their inter-action with these voltage-gated ion channels. Sensitivity is therefore prima-rily to the transmembrane electric field and varies widely between differention channels (Hille & Anderson, 2001; Mathie, Kennard & Veale, 2003;Saunders & Jefferys, 2002). Many voltage-gated ion channels are associatedwith electrical excitability and electrical signalling. Such electrically excit-

119

able cells not only comprise neurons, glial and muscle cells, but also endo-crine cells of the anterior pituitary, adrenal medulla and pancreas, gametesand, with reservations, endothelial cells (Hille & Anderson, 2001).

All these cells generally express voltage-gated sodium and calciumchannels. Both of these ion channels are involved in electrical signaling andcalcium ions activate a number of crucial cellular processes including neu-rotransmitter release, excitation-contraction coupling in muscle cells andgene expression (Catterall, 2000; Hille & Anderson, 2001). Some ion chan-nels, for example voltage-gated potassium and chloride ion channels, alsoexist in other, non-excitable tissues such as those in the kidney and liver andshow slow electric potential changes but their voltage sensitivity is likely tobe lower (Begenisich & Melvin, 1998; Cahalan, Wulff & Chandy, 2001; Cat-terall, 2000; Jan & Jan, 1989; Nilius & Droogmans, 2001). Since voltage-gated ion channels in excitable cells are steeply sensitive to the transmem-brane electric potential, electric field strength in tissue is a more relevantparameter to relate to electrically excitable cell thresholds than current den-sity (Bailey et al., 1997; Blakemore & Trombley, 2003; Reilly, 2005; Shep-pard, Kavet & Renew, 2002). In fact, the relevant parameter in determiningthe transmembrane current and hence the excitability is the linear gradient inelectric field (Tranchina & Nicholson, 1986), which in turn relates to geo-metric parameters of the neuron, including the degree of bending of the axon.

Peripheral nerves comprise neurons whose cell bodies are locatedwithin the CNS with extended processes (axons) that lie outside the CNS.They conduct action potentials (impulses) towards (sensory nerves) or from(motor nerves) the spinal cord and nerve stimulation shows an all-or-nothingthreshold behaviour. Excitation results from a membrane depolarisation ofbetween 10–20 mV, corresponding to an electric field in tissue of 5–25V m-1 (McKinlay et al., 2004). Pulsed magnetic fields, where the rate ofchange of field induces large localised electric fields, can directly stimulateperipheral nerves and nerve fibres located within the brain (see below).

Cells of the central nervous system are considered to be sensitive toelectric fields induced in the body by exposure to ELF magnetic fields at lev-els that are below threshold for impulse initiation in nerve axons (Jefferys,1995; Jefferys et al., 2003; Saunders, 2003; Saunders & Jefferys, 2002).Such weak electric field interactions have been shown in experimental stud-ies mostly using isolated animal brain tissue to have physiological relevance.These interactions result from the extracellular voltage gradients generatedby the synchronous activity of a number of neurons, or from those generatedby applying pulsed or alternating currents directly through electrodes placedon either side of the tissue. Jefferys and colleagues (Jefferys, 1995; Jefferyset al., 2003) identified in vitro electric field thresholds of around 4–5 V m-1.Essentially, the extracellular gradient alters the potential difference acrossthe neuronal membrane with opposite polarities at either end of the neuron; atime-constant of a few tens (15–60) of milliseconds results from the capaci-tance of the neuronal membrane (Jefferys et al., 2003) and indicates a limitedfrequency response. Similar arguments concerning the limited frequency

120

response of weak electric field effects due to the long time-constants (25 ms)arising from cell membrane capacitance have been given by Reilly (2002)regarding phosphene data.

The CNS in vivo is likely to be more sensitive to induced low fre-quency electric fields and currents than are in vitro preparations (Saunders &Jefferys, 2002). Spontaneous activity is higher, and interacting groups or net-works of nerve cells exposed to weak electrical signals would be expected,on theoretical grounds, to show increased sensitivity through improved sig-nal-to-noise ratios compared with the response of individual cells (Adair,2001; Stering, 1998; Valberg, Kavet & Rafferty, 1997). Much of normal cog-nitive function of the brain depends on the collective activity of very largenumbers of neurons; neural networks are thought to have complex non-lineardynamics that can be very sensitive to small voltages applied diffusely acrossthe elements of the network (Adair, 2001; ICNIRP, 2003; Jefferys et al.,2003). Gluckman et al. (2001) placed the detection limit for network modula-tion in hippocampal slices by electric fields at around 100 mV m-1 . Recentexperimental work by Francis, Gluckman & Schiff (2003) confirms a neuralnetwork threshold of around 140 mV m-1, which the authors found was lowerthan single neuron thresholds, based on a limited number of measurements.A lower limit on neural network sensitivity to physiologically weak inducedelectric fields has elsewhere been considered on theoretical grounds to bearound 1 mV m-1 (Adair, Astumian & Weaver, 1998; Veyret, 2003). Thetime-course of the opening of the fastest voltage-gated ion channels can beless than 1 ms (Hille & Anderson, 2001), suggesting that effects at frequen-cies up to a few kilohertz should not be ruled out. Accommodation to aslowly changing stimulus resulting from slow inactivation of the sodiumchannels will raise thresholds at frequencies less than around 10 Hz.

Other electrically excitable tissues with the potential to show net-work behaviour include glial cells located within the CNS (e.g. Parpura etal., 1994), and the autonomic and enteric nervous systems (see Sukkar, El-Munshid & Ardawi, 2000), which comprise interconnected non-myelinatednerve cells and are distributed throughout the body and gut, respectively.These systems are involved in regulating the visceral or “housekeeping”functions of the body; for example, the autonomic nervous system isinvolved in the maintenance of blood pressure. Muscle cells also show elec-trical excitability; only cardiac muscle tissue has electrically interconnectedcells. However, Cooper, Garny & Kohl et al. (2003), in a review of cardiacion channel activity, conclude that weak internal electric fields much belowthe excitation threshold are unlikely to have any significant effect on cardiacphysiology. EMF effects on the heart could theoretically result from indirecteffects mediated via the autonomic nervous system and CNS (Sienkiewicz,2003). Effects on the endocrine system could potentially also be mediatedthis way, although the evidence from volunteer experiments indicates thatacute ELF magnetic field exposure up to 20 µT does not influence the circa-dian variation in circulating levels of the hormone melatonin (Warman et al.,2003b), nor other plasma hormone levels (ICNIRP, 2003).

121

5.2 Volunteer studiesAn electric charge is induced on the surface of a human (or other

living organism) exposed to a low frequency electric field that alternates inamplitude with the frequency of the applied field. The alternation of the sur-face charge with time induces an electric field and therefore current flowwithin the body; in addition, exposure to a low frequency magnetic fieldinduces circulating eddy currents and associated electric fields. If of suffi-cient magnitude, these induced electric fields and currents can interact withelectrically excitable nerve and muscle tissue. Generally, however, the sur-face charge effects of exposure to low frequency electric fields become pro-hibitive long before the internal electric fields become large enough to elicita response in the tissue.

5.2.1 Surface electric chargeThe surface electric charge can be perceived directly through the

induced vibration of body hair and tingling sensations in areas of the body,particularly the arms, in contact with clothing, and indirectly through sparkdischarges between a person and a conducting object within the field. In sev-eral studies carried out in the 1970’s and 1980’s (summarized by Reilly,1998a; 1999), the threshold for direct perception has shown wide individualvariation; 10% of the exposed subjects had detection thresholds of around 2–5 kV m-1 at 60 Hz, whereas 50% could detect fields of 7–20 kV m-1. Theseeffects were considered annoying by 5% of the test subjects exposed underlaboratory conditions above electric field strengths of about 15–20 kV m-1.In addition to showing a wide variation in individual sensitivity, theseresponses also vary with environmental conditions, particularly humidity;the studies referred to above, however, included both wet and dry exposureconditions.

It has been estimated that spark discharges would be painful to 7%of subjects who are well-insulated and who touch a grounded object within a5 kV m-1 field (Reilly, 1998a; Reilly, 1999) whereas they would be painful toabout 50% in a 10 kV m-1 field. Unpleasant spark discharges can also occurwhen a grounded person touches a large conductive object such as a largevehicle that is “well-insulated” from ground and is situated within a strongelectric field. Here, the threshold field strength required to induce such aneffect varies inversely with the size of the conductive object. In both cases,the presence in the well-insulated person or object of a conductive pathwayto ground would tend to mitigate the intensity of any effect (Reilly, 1998a;Reilly, 1999), as would the impedance to earth of the grounded object or per-son.

People can perceive electric currents directly applied to the bodythrough touching, for example, a conductive loop in which current is inducedby exposure to environmental electromagnetic fields. Thresholds for directlyapplied currents have also been characterised. At 50 to 60 Hz, the malemedian threshold for perception was between 0.36 mA (finger contact) and1.1 mA (grip contact), while pain occurred at 1.8 mA (finger contact).

122

Median thresholds for women were generally found to be two thirds of themale thresholds, while children were assumed to have median thresholds halfof male threshold values (WHO, 1993). There is also a wide variety in theindividual’s ability to detect currents, there is, for example, about one orderof magnitude difference in the perception threshold at the 0.5 percentile andthe 99.5 percentile at 50/60 Hz (Kornberg & Sagan, 1979). Generally, theability to detect fields or currents decreases with increasing frequency. Thishas been characterised for the perception of currents; the threshold is increas-ing by about two orders of magnitude at higher frequencies: 0.36 mA at 50/60 Hz, 4 mA at 10 kHz and 40 mA at 100 kHz (WHO, 1993).

A series of extensive studies on 50 Hz population thresholds inmore than 1000 people from all ages have recently been carried out by Leit-geb and colleagues. Leitgeb & Schröttner (2002) examined perceptionthresholds in 700 people aged between 16 and 60 years, approximately halfof them women. This study was recently extended to include 240 childrenaged 9–16 years, and about 20 people aged 61 years or more (Leitgeb, Schro-ettner & Cech, 2005). In both studies, electric current was applied to the fore-arm using pre-gelled electrodes, and considerable care was taken to rule outsubjective bias.

A summary of the studies on perception of electric currents directlyapplied to the body is given in Table 33.

Leitgeb, Schroettner & Cech (2005) note that the median perceptionthreshold for the population is 268 µA, almost 50% lower than the presentlimit of 500 μA recommended by the IEC (1994). They also note that whilstthe median threshold for women is approximately two thirds of the malethreshold values, children aged between 9 and 16 do not exhibit as a high asensitivity as had been assumed.

An issue with perception levels is that they really depend on the siteof application of the current (cheek and inner forearm being very sensitive)

Table 33. 50 Hz electric current perception values (Iw) for different perception probabilities (p) for men, women and the general population a

Iw (µA)

p (%) Men Women Children Population

90 602 506 453 553

50 313 242 252 268

10 137 93 112 111

5 106 68 78 78

0.5 53 24 35 32a Source: Leitgeb & Schroettner, 2002; Leitgeb, Schroettner & Cech, 2005.

123

and the area of application of the current (i.e. current density). The lattermakes the comparison of current values difficult (Reilly, 1998a).

5.2.2 Nerve stimulationLarge, rapidly changing, pulsed magnetic fields used in various

specialised medical applications such as magnetic resonance imaging (MRI)and transcranial magnetic stimulation (TMS) can induce electric fields largeenough to stimulate nervous tissue in humans. Minimum, orientation-depen-dent stimulus thresholds for large diameter (20 µm) myelinated nerve axonshave been estimated to be approximately 6 V m-1 at frequencies up to about1–3 kHz (Reilly, 1998a; Reilly, 1999). In addition, accommodation to aslowly changing stimulus resulting from slow inactivation of sodium chan-nels will raise thresholds at low frequencies. In MRI, nerve stimulation is anunwanted side effect of a procedure used to derive cross-sectional images ofthe body for clinical diagnosis (see Shellock, 2001). Threshold rates ofchange of the switched gradient magnetic fields used in MRI for perception,discomfort and pain resulting from peripheral nerve stimulation are exten-sively reviewed by Nyenhuis et al. (2001). Generally, median, minimumthreshold rates of change of magnetic field (during periods of < 1 ms) forperception were 15–25 µT s-1 depending on orientation and showed consid-erable individual variation (Bourland, Nyenhuis & Schaefer, 1999). Thesevalues were somewhat lower than previously estimated by Reilly (1998a;1999), possibly due to the constriction of eddy current flow by high imped-ance tissue such as bone (Nyenhuis et al., 2001). Thresholds rose as the pulsewidth of the current induced by the switched gradient field decreased; themedian pulse width (the chronaxie) corresponding to a doubling of the mini-mum threshold (the rheobase) ranged between 360 and 380 µs but againshowing considerable individual variation (Bourland, Nyenhuis & Schaefer,1999). Numerical calculations of the electric field induced by pulses in the84 subjects tested by Nyenhuis et al. (2001) have been used to estimate themedian threshold for peripheral nerve stimulation at 60 Hz as 48 mT (Bailey& Nyenhuis, 2005). Furthermore, Nyenhuis et al. (2001)using data frommeasurements on human volunteers estimated a rheobase electric field of 2.2V m-1 in tissue.

In TMS, parts of the brain are deliberately stimulated in order toproduce a transient, functional impairment for use in the study of cognitiveprocesses (see Reilly, 1998a; Ueno, 1999; Walsh, Ashbridge & Cowey,1998). Furthermore, in TMS, brief, localised, suprathreshold stimuli aregiven, typically by discharging a capacitor through a coil situated over thesurface of the head, in order to stimulate neurons in a small volume (a fewcubic centimetres) of underlying cortical tissue (Reilly, 1998a). The inducedcurrent causes the neurons within that volume to depolarise synchronously,followed by a period of inhibition (Fitzpatrick & Rothman, 2000). When thepulsed field is applied to a part of the brain thought to be necessary for theperformance of a cognitive task, the resulting depolarisation interferes withthe ability to perform the task. In principle then, TMS provides cognitiveneuroscientists with the capability to induce highly specific, temporally and

124

spatially precise interruptions in cognitive processing – sometimes known as“virtual lesions”. Reilly (1998a) noted induced electric field thresholds to beof the order of 20 V m-1. However, Walsh & Cowey (1998) cited typicalrates of change of magnetic field of 30 kT s-1 over a 100 µs period transientlyinducing an electric field of 500 V m-1 in brain tissue.

People are likely to show variations in sensitivity to induced elec-tric fields. In particular, epileptic syndromes are characterised by increasedneuronal excitability and synchronicity (Engelborghs, D'Hooge & De Deyn,2000); seizures arise from an excessively synchronous and sustained dis-charge of a group of neurons (Engelborghs, D'Hooge & De Deyn, 2000; Jef-ferys, 1994). TMS is widely used, apparently without adverse effects.However, repetitive TMS has been observed to trigger epileptic seizure insome susceptible subjects (Fitzpatrick & Rothman, 2000; Wassermann,1998). These authors also reported short- to medium-term memory impair-ments and noted the possibility of long-term cognitive effects from alteredsynaptic activity or neurotransmitter balance. Contraindications for TMS useagreed at an international workshop on repetitive TMS safety (Wassermann,1998) include epilepsy, a family history of seizure, the use of tricyclic anti-depressants, neuroleptic agents and other drugs that lower seizure threshold.Serious heart disease and increased intracranial pressure have also been sug-gested as contraindications due to the potential complications that would beintroduced by seizure.

5.2.3 Retinal functionThe effects of exposure to weak low frequency magnetic fields on

human retinal function are well established. Exposure of the head to mag-netic flux densities above about 5 mT at 20 Hz, rising to about 15 mT at 50Hz, will reliably induce faint flickering visual sensations called magneticphosphenes (Attwell, 2003; Sienkiewicz, Saunders & Kowalczuk, 1991;Taki, Suzuki & Wake, 2003). It is generally agreed that these phosphenesresult from the interaction of the induced electric current with electricallysensitive cells in the retina. Several lines of evidence suggest the productionof phosphenes by a weak induced electric field does not involve the initialtransduction of light into an electrical signal. Firstly, the amplification of theinitial signal generated by the absorption of light takes place primarilythrough an intracellular “second-messenger cascade” of metabolic reactionsprior to any change in ion channel conductivity (Hille & Anderson, 2001).Secondly, the phosphene threshold appears unaffected by “dark” adaptationto low light levels (Carpenter, 1972). In addition, phosphenes have beeninduced in a patient with retinitis pigmentosa, a degenerative illness prima-rily affecting the pigment epithelium and photoreceptors (Lövsund et al.,1980).

There is good reason to view retinal circuitry as an appropriatemodel for induced electric field effects on CNS neuronal circuitry in general(Attwell, 2003). Firstly, the retina displays all the processes present in otherCNS areas, such as graded voltage signalling and action potentials, and has asimilar biochemistry. Secondly, in contrast to more subtle cognitive effects,

125

phosphenes represent a direct and reproducible perception of field interac-tion. A clear distinction can be made in this context between the detection ofa normal visual stimulus and the abnormal induction of a visual signal bynon-visual means (Saunders, 2003); the latter suggests the possibility ofdirect effects on cognitive processes elsewhere in the CNS.

Thresholds for electrically induced phosphenes have been esti-mated to be about 10–14 mA m-2 at 20 Hz (Adrian, 1977; Carstensen, 1985).A similar value (10 mA m-2 at 20 Hz), based on studies of magneticallyinduced phosphenes, has been derived by Wake et al. (1998). The equivalentelectric field threshold can be estimated as around 100–140 mV m-1 using atissue conductivity for brain tissue of about 0.1 S m-1 (Gabriel, Gabriel &Corthout, 1996). More recently, Reilly (2002) has calculated an approximate20 Hz electric field threshold in the retina of 53 mV m-1 for phosphene pro-duction. A similar value (60 mV m-1) has been reported elsewhere (see Saun-ders, 2003). Subsequently, however, Taki et al. (2003) indicated thatcalculations of phosphene thresholds suggested that electrophosphenethresholds were around 100 mV m-1, whereas magnetophosphene thresholdswere around 10 mV m-1 at 20 Hz.

More detailed calculation by Attwell (2003) based on neuroanatom-ical and physiological considerations, suggests that the phosphene electricfield threshold in the extracellular fluid of the retina is in the range 10–60mV m-1 at 20 Hz. There is however, considerable uncertainty attached tothese values. In addition, the extrapolation of values in the extracellular fluidto those appropriate for whole tissue, as used in most dosimetric models, iscomplex, depending critically on the extracellular volume and other factors.With regard to the frequency response, Reilly (2002) suggests that the nar-row frequency response is the result of relatively long membrane time con-stants of around 25 ms. However, at present, the exact mechanismunderlying phosphene induction is unknown. It is not clear whether the nar-row frequency response is due to intrinsic physiological properties of the ret-inal neurons, as suggested by Reilly (2002) above and by Attwell (2003)considering active amplification process in the retinal neuron synaptic termi-nals, or is the result of central processing of the visual signal (Saunders,2003; Saunders & Jefferys, 2002). This issue can only be resolved throughfurther investigation.

5.2.4 Brain electrical activity Since the first suggestion that occupational exposure to EMFs

resulted in clinical changes in the electroencephalogram (EEG) was pub-lished in 1966 (Asanova & Rakov, 1966; 1972), various studies have investi-gated if exposure to magnetic fields can affect the electrical activity of thebrain. Such methods can provide useful diagnostic information regarding thefunctional state of the brain, not only from recordings of the spontaneousactivity at rest but also from recording the sensory functions and subsequentcognitive processes evoked in response to specific stimuli (evoked or event-related potentials, ERPs). Nevertheless, neurophysiological studies usingmagnetic fields need to be performed with much care and attention since

126

they can be prone to many potential sources of error and artefact (NIEHS,1998). Changes in arousal and attention of volunteers, in particular, can sub-stantially affect the outcome of these studies.

Various studies have investigated the effects of magnetic fields onbrain activity by analysing the spectral power of the main frequency bands ofthe EEG (Bell et al., 1992; Bell et al., 1991; Bell, Marino & Chesson, 1994a;Bell, Marino & Chesson, 1994b; Gamberale et al., 1989; Heusser, Tellschaft& Thoss, 1997; Lyskov et al., 1993b; Lyskov et al., 1993a; Marino, Bell &Chesson, 1996; Schienle et al., 1996; Silny, 1986). These studies have used awide variety of experimental designs and exposure conditions, as well ashealthy volunteers and patients with neurological conditions, and thus aredifficult to compare and evaluate. Despite some scattered field-dependentchanges, most notably in the alpha frequency band, and with intermittentexposure perhaps more effective than continuous exposure, these studieshave produced inconsistent and sometimes contradictory results.

A difficulty with interpretation of the EEG in individuals at rest isthat the intra-individual variability is very high. The variability of ERPs ismuch lower, resulting in better reproducibility, and other studies have inves-tigated the effects of magnetic fields and combined electric and magneticfields on these potentials within the EEG waveform. There are some differ-ences between studies, but generally, the early components of the evokedresponse corresponding to sensory function do not appear affected by expo-sure (Graham & Cook, 1999; Lyskov et al., 1993b). In contrast, large andsustained changes on a later component of the waveform representing stimu-lus detection may be engendered by exposure at 60 mT (Silny, 1984; 1985;1986), with lesser effects occurring using fields of 1.26 mT (Lyskov et al.,1993b), and nothing below 30 µT (Graham & Cook, 1999). Finally, exposureduring the performance of some discrimination and attention tasks mayaffect the late major components of the EEG which are believed to reflectcognitive processes involved with stimulus evaluation and decision making(Cook et al., 1992; Crasson et al., 1999; Graham et al., 1994), although Cras-son and Legros (2005) were unable to replicate the effects they reported pre-viously. There also is some evidence that task difficulty and fieldintermittency may be important experimental variables. However, all thesesubtle effects are not well defined, and some inconsistencies between studiesrequire additional investigation and explanation.

A summary of studies on changes in brain electrical activity whileawake is given in Table 34.

127

Tabl

e 34

. Bra

in e

lect

rical

act

ivity

Test

a

Expo

sure

Res

pons

eC

omm

ents

Aut

hors

Dur

ing

expo

sure

to m

agne

tic fi

elds

up

to 1

0 H

z

Stan

dard

EE

G: C

3,4;

P3,

4; O

1,2

; FF

T at

10

Hz,

no

reco

rds

durin

g ex

posu

re6

fem

ale

and

4 m

ale

volu

ntee

rs

10 H

z10

0 µT

10 m

in

Imm

edia

tely

afte

r exp

osur

e th

e sp

ectra

l pow

er o

f the

bra

in a

ctiv

ity

was

low

er th

an b

efor

e ex

posu

re a

nd

10 m

in a

fterw

ards

, but

onl

y at

the

occi

pita

l ele

ctro

des

was

this

diff

er-

ence

sig

nific

ant.

Dat

a fo

r ind

ivid

ual v

olun

teer

s ar

e no

t pre

sent

ed, a

nd th

ere

is

no in

form

atio

n co

ncer

ning

the

rate

of r

espo

nder

s.

Bel

l, M

arin

o &

Che

sson

, 19

94b

Stan

dard

EE

G: C

3,4;

P3,

4; O

1,2

; FF

T at

1.5

Hz

or 1

0 H

z ba

nd

13 h

ealth

y su

bjec

ts a

nd 6

pat

ient

s

5 or

10

Hz

20, 4

0 µT

2 s

on, 5

s o

ff

In e

ach

pers

on, t

he m

agne

tic fi

eld

alte

red

the

brai

n ac

tivity

at t

he fr

e-qu

ency

of s

timul

atio

n, b

ut n

o sy

s-te

mat

ic c

hang

es o

f bra

in a

ctiv

ity.

The

stre

ngth

of t

he e

ffect

was

pr

opor

tiona

l nei

ther

to fr

e-qu

ency

nor

to fi

eld

stre

ngth

.

Bel

l, M

arin

o &

Che

sson

, 19

94a

Stan

dard

EE

G: C

3,4;

P3,

4; O

1,2

; fre

-qu

ency

spe

ctru

m a

naly

sis

exce

pt

for <

2.5

Hz

and

9-11

Hz.

13 v

olun

teer

s, 6

pat

ient

s

1.5

or 1

0 H

z80

µT

2 s

on, 5

s o

ff

ICO

S (i

ntra

-sub

ject

com

paris

on o

f st

imul

us a

nd n

on-s

timul

us s

tate

) was

al

tere

d by

ELF

exp

osur

e in

58%

of

the

subj

ects

.

Mar

ino,

Bel

l &

Che

sson

, 19

96

Stan

dard

EE

G, O

1,2

spec

tral a

naly

-si

s of

thet

a (3

.5-7

.5 H

z), a

lpha

(7.5

-12

.5H

z) a

nd b

eta

band

s (1

2.5-

25

Hz)

25 fe

mal

e an

d 36

mal

e vo

lunt

eers

3 H

z10

0 µT

pp20

min

one

expo

sure

and

one

con

trol

sess

ion

Sig

nific

ant c

hang

es in

thet

a an

d be

ta

frequ

ency

ban

ds a

fter e

xpos

ure

rela

-tiv

e to

con

trols

, int

erpr

eted

as

slig

htly

pro

noun

ced

redu

ctio

n of

al

ertn

ess

durin

g ex

posu

re.

Exp

osur

e an

d co

ntro

l ses

-si

ons

on d

iffer

ent d

ays,

the

two

sess

ion

days

wer

e no

t tre

ated

as

a d

oubl

e bl

ind

stud

y.

Heu

sser

, Te

llsch

aft &

Th

oss,

199

7

Dur

ing

expo

sure

to m

agne

tic fi

elds

bet

wee

n 45

and

60

Hz

Stan

dard

EE

G26

exp

erie

nced

pow

er u

tility

line

-m

en

50 H

z ex

posu

re d

urin

g w

ork-

day

aver

age

expo

sure

23

µTon

e da

y liv

e, o

ne d

ay s

ham

No

chan

ges

in a

lpha

EE

G, n

or e

vi-

denc

e of

EE

G a

bnor

mal

ities

.In

terv

entio

n st

udy,

not

labo

ra-

tory

.G

ambe

rale

et

al.,

198

9

128

Tabl

e 34

. Con

tinue

d

Stan

dard

EE

G (1

0-20

sys

tem

): C

3,4;

P3,

4; O

1,2

; FFT

at 1

-18.

5 H

z in

0.5

Hz

step

s3

fem

ale,

11

mal

e vo

lunt

eers

60 H

z25

or 5

0 µT

2 s

on, 8

s o

ff, fi

rst 2

s u

sed

as c

ontro

l

No

syst

emat

ic e

ffect

s fo

r fre

quen

cy

band

s an

d ac

tivity

-pow

er in

tens

ities

. In

50

% o

f vol

unte

ers

dim

inis

hed

EE

G p

ower

was

obs

erve

d as

a

resp

onse

to th

e fie

ld.

Bel

l et a

l.,

1991

Stan

dard

EE

G: C

3,4;

P3,

4; O

1,2

; FF

T at

1-1

8.5

Hz

in 0

.5 H

z st

eps

10 h

ealth

y vo

lunt

eers

and

10

neu-

rolo

gica

l pat

ient

s

60 H

zB

DC

: 78

µT, B

AC

: 78

µT, s

in-

gle

and

com

bine

d2

s on

, 5 s

off,

firs

t 2 s

use

d as

con

trol

19 o

ut o

f 20

pers

ons

resp

onde

d to

th

e fie

lds:

ove

rall

35%

to B

DC

; 70%

of

the

patie

nts

and

80%

of t

he v

olun

-te

ers

to B

AC

, res

pons

e to

BA

C w

as

not d

iffer

ent f

rom

the

resp

onse

s to

th

e co

mbi

natio

n B

AC

+ B

DC

. Fie

ld-

indu

ced

incr

ease

and

dec

reas

e of

br

ain

activ

ity, n

o sy

stem

atic

cha

nges

w

ere

obse

rved

for t

he h

emis

pher

es

or a

ctiv

ity lo

ci.

EE

G c

hang

es w

ere

low

er in

pa

tient

s th

an in

hea

lthy

sub-

ject

s.

Bel

l et a

l.,

1992

Stan

dard

EE

G s

pect

ral a

naly

sis

6 fe

mal

e an

d 8

mal

e vo

lunt

eers

45 H

z1.

26 m

T1s

on,

1s

off c

ycle

ove

r 15

min

, one

exp

osur

e an

d on

e co

ntro

l ses

sion

Sig

nific

ant i

ncre

ase

of th

e po

wer

val

-ue

s of

alp

ha a

nd b

eta

band

s af

ter

expo

sure

, no

chan

ges

in d

elta

- and

th

eta-

band

s.

Lysk

ov e

t al

., 19

93a

Stan

dard

EE

G s

pect

ral a

naly

sis.

be

fore

and

afte

r exp

osur

e11

fem

ale

and

9 m

ale

volu

ntee

rs

45 H

z1.

26 m

T10

per

sons

: 1 h

con

tinuo

us

field

, 10

pers

ons:

1s

on/o

ff in

term

itten

t fie

ld fo

r 1 h

One

exp

osur

e an

d on

e co

n-tro

l ses

sion

Sev

eral

sta

tistic

ally

sig

nific

ant

chan

ges;

incr

ease

of a

lpha

act

ivity

du

ring

inte

rmitt

ent e

xpos

ure

and

decr

ease

of d

elta

act

ivity

. Inc

reas

e of

be

ta w

aves

in fr

onta

l but

not

in o

ccip

-ita

l der

ivat

ions

.

Inho

mog

eneo

us re

sults

, con

-si

sten

t with

incr

ease

d re

lax-

atio

n. D

oubl

e bl

ind

stud

y.

Lysk

ov e

t al

., 19

93b

129

Tabl

e 34

. Con

tinue

d

Dur

ing

expo

sure

to m

agne

tic fi

elds

at h

ighe

r fre

quen

cies

Stan

dard

EE

G, F

3,4,

P3,

4,O

1,2,

trad

i-tio

nal f

requ

ency

ban

ds 0

.1 -

30 H

z,

psyc

holo

gica

l par

amet

ers

and

ques

tionn

aire

s26

fem

ale

and

26 m

ale

volu

ntee

rs

Sphe

rics

sim

ulat

ion:

10

kH

z50

0 µs

dur

atio

n, ra

ndom

in

terv

als

betw

een

50 a

nd 1

50

ms

38 A

m-1

pea

k va

lue

elec

tric

field

shi

elde

dex

posu

re 1

0 m

in a

nd c

ontro

l se

ssio

n

Sig

nific

ant r

educ

tion

of th

e po

wer

on

ly in

the

alph

a fre

quen

cy b

and

(8-

13 H

z) in

par

ieta

l and

occ

ipita

l der

i-va

tions

, whe

n an

alys

ing

sub-

grou

ps

only

in 1

0 - 1

0.75

Hz.

P /

O re

cord

-in

gs s

how

sig

nific

ant r

educ

tions

in

the

volta

ge p

ower

.

Fact

ors

such

as

phys

ical

con

-di

tion

and

neur

otic

s w

ere

con-

side

red

as m

edia

tors

of

sphe

rics

effe

ctiv

enes

s.

Sch

ienl

e et

al

., 19

96

Evok

ed p

oten

tials

afte

r exp

osur

e to

EM

F

Visu

al e

voke

d po

tent

ials

100

subj

ects

5 –

50 H

z pu

lsed

mag

netic

fie

ldup

to 1

00 m

T

Pha

se re

vers

al o

f com

pone

nts

of th

e vi

sual

evo

ked

pote

ntia

l at 6

0 m

T.

Very

inte

nse

field

s.S

ilny,

198

4;

1985

; 198

6

Aud

itory

evo

ked

pote

ntia

ls6

fem

ale

and

8 m

ale

volu

ntee

rs45

Hz

1.26

mT

1s o

n, 1

s o

ff, 1

5 m

inon

e ex

posu

re a

nd o

ne c

ontro

l se

ssio

n

N10

0 co

mpo

nent

s w

ere

shor

ter,

ampl

itude

s w

ere

redu

ced.

Ly

skov

et

al.,

1993

a

Aud

itory

evo

ked

pote

ntia

ls11

fem

ale

and

9 m

ale

volu

ntee

rs45

Hz

1.26

mT

10 p

erso

ns: 1

h c

ontin

uous

fie

ld10

per

sons

: 1 h

1 s

on/

off

inte

rmitt

ent f

ield

one

expo

sure

and

one

con

trol

sess

ion

Not

affe

cted

.Ly

skov

et

al.,

1993

b

130

Tabl

e 34

. Con

tinue

d

Aud

itory

, vis

ual a

nd s

omat

osen

sory

ev

oked

pot

entia

ls b

efor

e, d

urin

g an

d af

ter e

xpos

ure

36 (m

ale

and

fem

ale)

sub

ject

s

60 H

z 14

.1 o

r 28.

3 µT

45 m

in

No

effe

ct e

xcep

t a re

duce

d am

pli-

tude

of t

he s

omat

osen

sory

evo

ked

pote

ntia

l in

the

low

er e

xpos

ure

grou

p.

Dou

ble-

blin

d, c

ount

erba

l-an

ced

stud

y.G

raha

m &

C

ook,

199

9

Even

t-rel

ated

pot

entia

ls a

fter e

xpos

ure

to E

MF

Ele

ctro

des

Cz,

P3,

4 fo

r eve

nt-

rela

ted

pote

ntia

ls (P

300)

, fol

low

ing

audi

tory

or v

isua

l stim

uli i

n th

e O

dd-

ball

task

dur

ing

expo

sure

30 m

ale

volu

ntee

rs

60 H

z9

kV m

-1, 2

0 µT

18 e

xpos

ed a

nd s

ham

-ex

pose

d ov

er fo

ur 6

-h s

es-

sion

s, 1

2 ex

pose

d in

all

ses-

sion

s

Am

plitu

de o

f the

aud

itory

P30

0 w

as

incr

ease

d. V

isua

l ER

Ps

wer

e no

t af

fect

ed.

Effe

cts

on a

udito

ry E

RP

com

-po

nent

s w

ere

grea

test

soo

n af

ter a

ctiv

atio

n of

fiel

d an

d af

ter s

witc

hing

off

at th

e en

d of

th

e se

ssio

n.

Coo

k et

al.,

19

92

Eve

nt-r

elat

ed b

rain

pot

entia

ls

(N20

0-P

300)

follo

win

g au

dito

ry

stim

uli i

n th

e O

ddba

ll ta

sk d

urin

g ex

posu

re54

mal

e su

bjec

ts

3 m

atch

ed g

roup

s of

18

men

ea

ch, t

wo

6-h

sess

ions

, exp

o-su

re o

r sha

m, 6

0 H

z:a)

6 k

V m

-1, 1

0 µT

b) 9

kV

m-1

, 20

µT

c) 1

2 kV

m-1

, 30

µT

Sig

nific

ant i

ncre

ases

of P

300

late

ncy

in g

roup

b),

but d

ecre

ase

durin

g sh

am e

xpos

ure.

N20

0-P

300

com

pone

nt c

om-

plex

alte

red

in a

ll gr

oups

, ord

er

of e

xpos

ure

did

not a

ffect

re

sults

. Dou

ble

blin

d, c

oun-

terb

alan

ced

stud

y.

Gra

ham

et

al.,

1994

Eve

nt-r

elat

ed p

oten

tials

dur

ing

per-

form

ance

of t

he O

ddba

ll ta

sk, t

he

dich

otic

list

enin

g ta

sk a

nd th

e C

NV

pa

radi

gm a

fter e

xpos

ure

21 m

ale

subj

ects

50 H

z10

0 µT

30 m

in, c

ontin

uous

or i

nter

-m

itten

t he

ad o

nly

Diff

eren

ces

in E

RP

ampl

itude

s w

ere

seen

dur

ing

the

dich

otic

list

enin

g ta

sk.

Som

e ef

fect

s w

ere

inco

nsis

tent

be

twee

n tri

als.

Dou

ble-

blin

d st

udie

s.

Cra

sson

et

al.,

1999

Eve

nt-r

elat

ed p

oten

tials

dur

ing

per-

form

ance

of t

he O

ddba

ll ta

sk, t

he

dich

otic

list

enin

g ta

sk a

nd th

e C

NV

pa

radi

gm a

fter e

xpos

ure

18 m

ale

subj

ects

50 H

z10

0 µT

30 m

in, c

ontin

uous

or i

nter

-m

itten

t he

ad o

nly

No

effe

cts

in E

RP

am

plitu

des

wer

e se

en d

urin

g th

e di

chot

ic li

sten

ing

task

or i

n ot

her m

easu

res

of p

erfo

r-m

ance

.

Rep

licat

ion

and

exte

nsio

n of

ab

ove

stud

y by

the

sam

e gr

oup.

Cra

sson

&

Legr

os,

2005

a C

, F, O

& P

repr

esen

t sta

ndar

d E

EG

reco

rdin

g el

ectro

de p

ositi

ons;

FFT

= F

ast F

ourie

r Tra

nsfo

rm.

131

5.2.5 SleepSleep is a complex biological process controlled by the central ner-

vous system and is necessary for general health and well-being. The possibil-ity that EMFs may exert a detrimental effect on sleep has been examined intwo studies. Using the EEG to assess sleep parameters, Åkerstedt et al.(1999) reported that continuous exposure of healthy volunteers to 50 Hz at 1µT at night caused disturbances in sleep. In this study, total sleep time, sleepefficiency, slow-wave sleep (stage III and IV), and slow-wave activity weresignificantly reduced by exposure, as was subjective depth of sleep. Graham& Cook (1999) reported that intermittent, but not continuous, exposure to 60Hz, 28 µT magnetic fields at night resulted in less total sleep time, reducedsleep efficiency, increased time in stage II sleep, decreased time in rapid eyemovement (REM) sleep and increased latency to first REM period. Consis-tent with a pattern of poor and broken sleep, volunteers exposed to the inter-mittent field also reported sleeping less well and feeling less rested in themorning.

A comparison between these two studies is made difficult becauseof the differences in the exposure levels used, 1 µT (Åkerstedt et al., 1999)vs. 28 µT (Graham & Cook, 1999) and also of other differences in thedesign. As to the results, in the Åkerstedt study, results were apparentlyobtained by low-level continuous exposure, whereas the Graham study failedto elicit such results by continuous exposure, but did produce similar resultswith intermittent exposures. Further studies with similar designs are neededbefore any conclusions can be drawn.

A summary of studies on brain electrical activity during sleep isgiven in Table 35.

Table 35. Brain electrical activity during sleep

Test Exposure Response Comments Authors

Sleep EEGs, conventional recordings8 female and 10 male healthy volunteers

50 Hz1 µTone night (23:00-07:00) with field on, one night with field off

Significantly reduced slow wave activity and slow wave sleep. Also tendency for reduced total sleep time, sleep efficiency, REM sleep (not statistically signifi-cant).

Absolute values were within the normal variability; the observed changes are far from clinical sig-nificance. Blind study, balanced design.

Åkerstedt et al., 1999

Sleep EEG, 3 nights (23:00-07:00), Cz, C4, Oz 24 male volun-teers

60 Hz28.3 µT, circularly polarised8 sham-exposed controls, 7 sub-jects exposed to continuous fields, 9 to intermittent 1 h on, 1 h off, 15 s on/off cycle

Intermittent exposure to magnetic fields pro-duced significant distur-bances in nocturnal sleep EEGs in 6 of 9 per-sons: decreased sleep efficiency, altered sleep architecture, suppres-sion of REM sleep, lower well-feeling of several subjects in the morning.

No effect was seen during con-tinuous field expo-sure relative to sham-exposed controls. Double-blind, counter-balanced study.

Graham & Cook, 1999

132

5.2.6 Cognitive effectsDespite the potential importance of field-induced effects on atten-

tion, vigilance, memory and other information processing functions, rela-tively few studies have looked for evidence of changes in cognitive abilityduring or after exposure to low frequency EMFs. These have been reviewedby NIEHS (1998), Cook, Thomas & Prato (2002), Bailey (2001), Crasson(2003) and ICNIRP (2003). While few field-dependent changes have beenobserved, it is important to consider that this type of study may be particu-larly susceptible to various environmental and individual factors which mayincrease the variance of the experimental endpoint and decrease the power todetect a small effect. This may be particularly important, since any field-dependent effects are likely to be small with fields at environmental levels(Sienkiewicz et al., 1993; Whittington, Podd & Rapley, 1996).

The effects of acute exposure to magnetic fields on simple andchoice reaction time have been investigated in several recent studies using awide range of magnetic flux densities (20 µT – 1.26 mT) and experimentalconditions. Some studies did not find any field-dependent effects (Gam-berale et al., 1989; Kurokawa et al., 2003b; Lyskov et al., 1993b; Lyskov etal., 1993a; Podd et al., 2002; Podd et al., 1995), although modest effects onspeed (Crasson et al., 1999; Graham et al., 1994; Whittington, Podd & Rap-ley, 1996) and accuracy during task performance (Cook et al., 1992;Kazantzis, Podd & Whittington, 1998; Preece, Wesnes & Iwi, 1998) havebeen reported. However, Crasson & Legros (2005) were unable to replicatethese observations. These data also suggest that effects may depend on thedifficulty of the task (Kazantzis, Podd & Whittington, 1998; Whittington,Podd & Rapley, 1996) and that exposure may attenuate the usual improve-ment seen with practice in reaction time (Lyskov et al., 1993b; Lyskov et al.,1993a; Stollery, 1986)

A few studies have reported subtle field-dependent changes in othercognitive functions, including memory and attention. Using a battery of neu-ropsychological tests, Preece, Wesnes & Iwi (1998) found that exposure to a50 Hz magnetic field at 0.6 mT decreased accuracy in the performance ofnumerical working memory task and decreased sensitivity of the perfor-mance in a word recognition task. Similarly Keetley et al. (2001) investi-gated the effects of exposure to 28 µT, 50 Hz fields using a series ofcognitive tests. A significant decrease in performance was seen with oneworking memory task (the trail-making test, part B) that involves visual-motor tracking and information processing within the prefrontal and parietalareas of the cortex. Podd et al. (2002) reported delayed deficits in the perfor-mance of a recognition memory task following exposure to a 50 Hz field at100 µT. Trimmel & Schweiger (1998) investigated the effects of acute expo-sure to 50 Hz magnetic fields at 1 mT. The fields were produced using apower transformer, and volunteers were exposed in the presence of a 45 dBsound pressure level noise. Compared with a no-field, no-noise condition andnoise alone (generated using a tape recording) significant reductions in visualattention, perception and verbal memory performance were observed during

133

Tabl

e 36

. Cog

nitiv

e ef

fect

s

Test

Expo

sure

Res

pons

eC

omm

ents

Aut

hors

Rea

ctio

n tim

e, v

igila

nce,

mem

ory

and

perc

eptio

n sp

eed

test

ed b

efor

e an

d af

ter e

ach

day

26 e

xper

ienc

ed p

ower

util

ity li

nem

en.

50 H

z ex

posu

re d

urin

g w

orkd

ayav

erag

e ex

posu

re 2

3 µT

one

day

live,

one

day

sha

m

No

diffe

renc

e in

per

for-

man

ce b

etw

een

expo

sed

and

non-

expo

sed

days

.

Inte

rven

tion

stud

y, n

ot

labo

rato

ry.

Gam

bera

le e

t al

., 19

89

Rea

ctio

n tim

e (R

T) a

nd ta

rget

-del

e-tio

n te

st (T

DT)

6 fe

mal

e an

d 8

mal

e vo

lunt

eers

45 H

z1.

26 m

T1

s on

, 1 s

off

cycl

e, 1

5 m

inon

e ex

posu

re a

nd o

ne c

ontro

l ses

sion

No

sign

ifica

nt d

iffer

ence

s fo

r R

T, T

DT

not a

ffect

ed.

Lysk

ov e

t al.,

19

93a

Rea

ctio

n tim

e (R

T)11

fem

ale

and

9 m

ale

volu

ntee

rs45

Hz

1.26

mT

10 p

erso

ns: 1

h c

ontin

uous

fiel

d10

per

sons

: 1 h

our 1

s o

n/of

f int

erm

it-te

nt fi

eld

one

expo

sure

and

one

con

trol s

essi

on

RT

not d

irect

ly a

ffect

ed.

Lear

ning

to p

erfo

rm th

e R

T te

st (d

ecre

ase

of R

T in

re

peat

ed tr

ials

) affe

cted

by

exp

osur

e .

Lysk

ov e

t al.,

19

93b

Rea

ctio

n tim

e to

ligh

t fla

shed

at v

ari-

able

inte

rval

s du

ring

expo

-sur

e12

sub

ject

s (e

xpt 1

) and

24

subj

ects

(e

xpt 2

), m

ale

and

fem

ale

Exp

erim

ent 1

:10

.1 o

r 0.2

Hz

1.1

mT

300

sE

xper

imen

t 2:

0.2

or 4

3 H

z1.

8m

T 30

0 s

No

effe

cts

foun

d.E

xper

imen

t 2 d

esig

ned

to

test

for p

ossi

ble

para

met

-ric

reso

nanc

e th

eory

. D

oubl

e bl

ind

stud

ies.

Pod

d et

al.,

19

95

Rea

ctio

n tim

e, a

ccur

acy

and

mem

ory

reco

gniti

on60

Hz

100

µT1

s on

, 1 s

off

for 1

1 m

in

Effe

ct o

n m

emor

y, n

ot o

n re

actio

n tim

e or

acc

urac

y.R

esul

ts d

iffer

ent f

rom

pre

-vi

ous

stud

ies

(Whi

tting

ton

et a

l., 1

996)

Pod

d et

al.,

20

02

134

Tabl

e 36

. Con

tinue

d

Rea

ctio

n tim

e, a

ccur

acy,

tim

e pe

rcep

-tio

n an

d vi

sual

per

cept

ion

12 m

ale

and

8 fe

mal

e su

bjec

ts

50 H

z22

µT

circ

ular

ly p

olar

ised

with

har

mon

-ic

s an

d re

petit

ive

trans

ient

s up

to 1

00

µT 55 m

in

No

effe

cts.

Kur

okaw

a et

al

., 20

03b

Rea

ctio

n tim

e (R

T), a

ttent

ion,

diff

er-

entia

l rei

nfor

cem

ent o

f low

resp

onse

ra

te (D

RL)

54 m

ale

volu

ntee

rs

3 m

atch

ed g

roup

s of

18

men

eac

h, tw

o 6-

h se

ssio

ns, e

xpos

ure

or s

ham

, 60

Hz:

a) 6

kV

m-1

, 10

µTb)

9 k

V m

-1, 2

0 µT

c) 1

2 kV

m-1

, 30

µT

Slo

wer

reac

tion

time

in O

dd-

ball t

ask

and

low

er a

ccur

acy

of D

RL

in g

roup

a) o

nly.

No

effe

ct o

n ot

her m

ea-

sure

s or

in o

ther

exp

osur

e gr

oups

. Dou

ble

blin

d,

coun

terb

alan

ced

stud

y.

Gra

ham

et a

l.,

1994

A v

isua

l dur

atio

n-di

scrim

inat

ion

task

w

ith 3

leve

ls o

f diff

icul

ty10

0 m

ale

and

fem

ale

subj

ects

50 H

z 10

0 µT

inte

rmitt

ent

9 m

in

Dec

reas

ed re

actio

n tim

e fo

r th

e ha

rdes

t lev

el o

f per

for-

man

ce.

A re

laxe

d si

gnifi

canc

e le

vel (

0.15

) was

use

d.

Dou

ble-

blin

d, c

ount

er-b

al-

ance

d st

udy.

Whi

tting

ton,

P

odd

& R

ap-

ley,

199

6

Rey

Aud

itory

Ver

bal L

earn

ing

test

(w

ith d

elay

ed re

call)

and

Dig

it Sp

an

Task

21 m

ale

subj

ects

.

50 H

z10

0 µT

con

tinuo

us o

r int

erm

itten

t30

min

head

onl

y

No

effe

cts

(rep

orte

d in

dis

-cu

ssio

n).

Dou

ble-

blin

d st

udie

s.C

rass

on e

t al.,

19

99

Cho

ice

seria

l rea

ctio

n tim

e ta

sk, t

ime

estim

atio

n ta

sk, i

nter

val p

rodu

ctio

n ta

sk, v

igila

nce

task

, dig

it sp

an m

em-

ory

task

and

Wilk

inso

n A

dditi

on ta

sk

durin

g ex

posu

re54

mal

e su

bjec

ts

60 H

z9

kV m

-1 a

nd 2

0 µT

2 x

3 h

/ day

for 4

day

s

Few

er e

rror

s in

cho

ice

reac

-tio

n tim

e ta

sk. N

o ef

fect

s on

re

actio

n tim

e, m

emor

y or

vi

gila

nce.

Dou

ble-

blin

d, c

ount

erba

l-an

ced

stud

y.C

ook

et a

l.,

1992

135

Tabl

e 36

. Con

tinue

d

A v

isua

l dur

atio

n-di

scrim

inat

ion

task

w

ith 3

leve

ls o

f diff

icul

ty.

40 m

ale

and

59 fe

mal

e su

bjec

ts

50 H

z 10

0 µT

inte

rmitt

ent

7.9

min

Impr

oved

acc

urac

y fo

r the

ha

rdes

t lev

el o

f per

for-

man

ce.

A re

laxe

d si

gnifi

canc

e le

vel (

0.3)

was

use

d. D

ou-

ble-

blin

d, c

ount

er-b

al-

ance

d st

udy.

Kaz

antz

is,

Pod

d &

Whi

t-tin

gton

, 199

8

Imm

edia

te w

ord

reca

ll, re

actio

n tim

e,

digi

t vig

ilanc

e ta

sk, c

hoic

e re

actio

n tim

e, s

patia

l wor

king

mem

ory,

nu

mer

ic w

orki

ng m

emor

y, d

elay

ed

wor

d re

call

and

reco

gniti

on a

nd p

ic-

ture

reco

gniti

on d

urin

g ex

posu

re.

16 (m

ale

and

fem

ale)

sub

ject

s

50 H

z or

sta

tic m

agne

tic fi

elds

at 0

.6

mT

appl

ied

to th

e he

ad. D

urat

ion

not

spec

ified

. Cur

rent

den

sity

in h

ead

esti-

mat

ed a

s 2–

6 m

A m

-2

Red

uced

acc

urac

y of

wor

d an

d nu

mbe

r rec

all a

nd p

er-

form

ance

of c

hoic

e re

actio

n tim

e ta

sk.

Ran

dom

ised

blin

d cr

oss-

over

des

ign.

Pre

ece,

W

esne

s &

Iwi,

1998

Dur

atio

n D

iscr

imin

atio

n Ta

sk a

nd

Stro

op C

olou

r Wor

d te

st.

18 m

ale

subj

ects

50 H

z10

0 µT

con

tinuo

us o

r int

erm

itten

t30

min

head

onl

y

No

effe

ct o

n re

actio

n tim

e an

d pe

rform

ance

acc

urac

y.D

oubl

e bl

ind

with

cou

nter

-ba

lanc

ed e

xpos

ure

orde

r.C

rass

on &

Le

gros

, 200

5

Syn

tact

ic a

nd s

eman

tic v

erba

l rea

-so

ning

task

s, 5

-cho

ise

seria

l rea

ctio

n tim

e ta

sk, a

nd v

isua

l sea

rch

task

s du

ring

expo

sure

76 m

ale

subj

ects

50 H

z cu

rren

t50

0 µA

dire

ctly

app

lied

to h

ead

and

shou

lder

s5.

5 h

/ day

for 2

day

s

Incr

ease

d la

tenc

y in

syn

tac-

tic re

ason

ing

task

. P

ossi

ble

diffe

renc

es

betw

een

grou

ps. D

oubl

e-bl

ind

proc

edur

es w

ith

cros

s ov

er d

esig

n.

Stol

lery

, 198

6;

1987

Rey

Aud

itory

Ver

bal L

earn

ing

test

; D

igit

Span

Mem

ory

Task

; Dig

it S

ym-

bol S

ubst

itutio

n te

st; S

peed

of C

om-

preh

ensi

on T

est a

nd T

rail

Mak

ing

Test

30 s

ubje

cts,

bot

h se

xes

50 H

z28

µT

50 m

inve

rbal

and

writ

ten

test

s ad

min

iste

red

20 m

in fr

om e

xpos

ure

onse

t

Mos

t res

ults

indi

cate

d no

ef

fect

, but

dat

a su

gges

tive

of d

etrim

enta

l effe

ct o

n sh

ort-t

erm

lear

ning

and

ex

ecut

ive

func

tioni

ng.

Dou

ble-

blin

d cr

oss-

over

de

sign

.K

eetle

y et

al.,

20

01

Visu

al d

iscr

imin

atio

n, p

erce

ptio

n, v

er-

bal m

emor

y an

d m

ood

and

sym

ptom

ch

eckl

ist

66 (m

ale

and

fem

ale)

sub

ject

s

50 H

z1

mT

45dB

noi

se c

ompa

red

to n

oise

alo

ne

Sig

nific

ant r

educ

tion

in

visu

al a

ttent

ion,

per

cept

ion

and

verb

al m

emor

y pe

rfor-

man

ce.

Dou

ble

blin

d st

udie

s.Tr

imm

el &

S

chw

eige

r, 19

98

136

field exposure. The presence of the noise during exposure, however, compli-cates interpretation of this study.

Generally, while electrophysiological considerations suggest thatthe central nervous system is potentially susceptible to induced electricfields; cognitive studies have not revealed any clear, unambiguous finding.There is a need for a harmonisation of methodological procedures used indifferent laboratories, and for dose-response relationships to be investigated.The studies on various cognitive effects from ELF field exposure are summa-rized in Table 36.

5.2.7 HypersensitivityIt has been suggested that some individuals display increased sensi-

tivity to levels of EMFs well below recommended restrictions on exposure.People self-reporting hypersensitivity may experience a wide range of severeand debilitating symptoms, including sleep disturbances, general fatigue, dif-ficulty in concentrating, dizziness, and eyestrain. In extreme forms, everydayliving may become problematical. A number of skin problems such aseczema and sensations of itching and burning have also been reported, espe-cially on the face, and, although there may be no specific symptom profile(see Hillert et al., 2002), increased sensitivity to chemical and other factorsoften occurs (Levallois et al., 2002). The responses to EMFs are reported tooccur at field strengths orders of magnitude below those required for conven-tional perception of the field (Silny, 1999). These data have been reviewedby Bergqvist & Vogel (1997) and more recently by Levallois (2002),ICNIRP (2003) and Rubin et al. (2005).

In contrast to anecdotal reports, the evidence from double-blindprovocation studies (Andersson et al., 1996; Flodin, Seneby & Tegenfeldt,2000; Lonne-Rahm et al., 2000; Lyskov, Sandström & Hansson Mild,2001b; Swanbeck & Bleeker, 1989) indicate that neither healthy volunteersnor self-reporting hypersensitive individuals can reliably distinguish fieldexposure from sham-exposure. In addition, subjective symptoms and circu-lating levels of stress-related hormones and inflammatory mediators couldnot be related to field exposure. Similar results were reported in a survey ofoffice workers (Arnetz, 1997). In studies reported by Keisu (1996) and byToomingas (1996), the outcome of tests on an individual was used therapeu-tically in the medical handling of the patient. In none of these series wasthere any reproducible association between exposure and symptoms. Furthertest series have been performed in Sweden, the UK and in Germany, includ-ing an unsuccessful repetition of the Rea et al. (1991) study (see below), butthese have not been published in a peer reviewed form. For a review, seeBergqvist & Vogel (1997). These results are consistent with the view thathypersensitivity to EMFs is a psychosomatic syndrome, suggested by Gothe,Odoni & Nilsson (1995).

Not all studies dismiss the possibility of EMF hypersensitivity,however. Two studies have reported weak positive field discrimination(Mueller, Krueger & Schierz, 2002; Rea et al., 1991) and another two studies

137

reported subtle differences in heart rate, visual evoked potentials, electroret-inogram amplitudes and electrodermal activity between normal and hyper-sensitive volunteers (Lyskov, Sandström & Hansson Mild, 2001a; Sandströmet al., 1997). The study by Rea et al. (1991) has, however, been criticised onseveral methodological grounds (ICNIRP, 2003): the selection of individu-als, the exposure situation and whether the test was blind or not. There issome morphological evidence to suggest that the numbers and distribution ofmast cells in the dermis of the skin on the face may be increased in individu-als displaying hypersensitive reactions (Gangi & Johansson, 2000; Johans-son et al., 1994; Johansson, Hilliges & Han, 1996). Increased responsivenesswas attributed to changes in the expression of histamine and somatostatinand other inflammatory peptides. Similar effects in the dermis have also beenreported following provocation tests to VDU-type fields in normal, healthyvolunteers (Johansson et al., 2001).

EMF hypersensitivity was addressed by the World Health Organi-zation (WHO) at a workshop held in Prague in October 2004 (WHO, 2005).It was proposed that this hypersensitivity, which has multiple recurrentsymptoms and is associated with diverse environmental factors tolerated bythe majority of people, should be termed “idiopathic environmental intoler-ance (IEI) with attribution to EMF”. The workshop concluded that IEI incor-porates a number of disorders sharing similar nonspecific symptoms thatadversely affect people and cause disruptions in their occupational, social,and personal functioning. These symptoms are not explained by any knownmedical, psychiatric or psychological disorder, and the term IEI has no med-ical diagnostic value. IEI individuals cannot detect EMF exposure any moreaccurately than non-IEI individuals, and well-controlled and conducted dou-ble-blind studies have consistently shown that their symptoms are not relatedto EMF exposure per se. A summary of hypersensitivity studies is given inTable 37.

5.2.8 Mood and alertnessThe possible impact of EMFs on mood and arousal has also been

assessed in double-blind studies in which volunteers completed mood check-lists before and after exposure. No field-dependent effects have beenreported using a range of field conditions (Cook et al., 1992; Crasson et al.,1999; Crasson & Legros, 2005; Graham et al., 1994). However, in contrastStollery (1986) reported decreased arousal in one of two participating groupsof subjects when mild (500 µA) 50 Hz electric current was passed throughthe head, upper arms, and feet. This was done to simulate the internal electricfields generated by exposure to an external electric field strength of 36 kVm-1. Also Stevens (2001) reported that exposure to a 20 Hz, 50 µT magneticfield increased positive affective responses displayed to visual stimuli com-pared with sham-exposure. Arousal, as measured by skin conductance, gavevariable results. Table 38 summarizes the studies on effects of ELF fieldexposure on mood and alertness.

138

Tabl

e 37

. Hyp

erse

nsiti

vity

Test

Expo

sure

Res

pons

eC

omm

ents

Aut

hors

Ski

n sy

mpt

oms

30 p

atie

nts

VD

U:

stat

ic e

lect

ric fi

eld

0,2

and

30 k

V

m-1

ELF

mag

netic

fiel

d: 5

0 an

d 80

0 nT dB

/dt:

23 a

nd 3

35 m

T s-

1

No

resp

onse

rela

ted

to e

xpos

ure.

Hea

t, re

dden

ing,

itch

-in

g, s

tingi

ng, o

edem

a in

exp

osed

and

sha

m

expo

sed

situ

atio

ns.

Sw

anbe

ck

& B

leek

er,

1989

Per

cept

ion

and

sym

ptom

s17

pat

ient

s Fi

elds

from

VD

U, p

re-te

sted

as

caus

ing

sym

ptom

s in

ope

n pr

ov-

ocat

ion

prio

r to

doub

le b

lind

ses-

sion

s. In

shi

elde

d la

bora

tory

.

16 in

divi

dual

s fa

iled

to d

etec

t (g

uess

) pre

senc

e of

the

field

s,

sym

ptom

s w

ere

rela

ted

to g

uess

es,

not t

o th

e fie

lds.

And

erss

on

et a

l., 1

996

Per

cept

ion

and

sym

ptom

s15

pat

ient

s an

d 26

con

trols

Fiel

ds fr

om V

DU

s an

d ot

her

obje

cts.

Sub

ject

s te

sted

in th

eir

norm

al h

ome

envi

ronm

ent,

usin

g a

varie

ty o

f dev

ices

.

15 in

divi

dual

s fa

iled

to d

etec

t pre

s-en

ce o

f the

fiel

ds, s

ympt

oms

wer

e no

t rel

ated

to th

e fie

lds.

Flod

in,

Sen

eby

&

Tege

nfel

dt,

2000

Pro

voca

tion

stud

y of

stre

ss h

orm

one

lev-

els,

ski

n bi

opsi

es a

nd fa

cial

ski

n se

nsa-

tions

24 p

atie

nts

and

12 c

ontro

ls

VD

Us:

5 H

z–2

kHz:

12

V m

-1

198

nT

2 kH

z–40

0 kH

z: 1

0 V

m-1

18

nT30

min

/ w

eek

for 4

wee

ks

Non

e of

the

test

par

amet

ers

diffe

red

betw

een

expo

sed

and

sham

ex

pose

d co

nditi

ons,

but

ski

n sy

mp-

tom

s ap

pear

ed in

the

open

pro

voca

-tio

n te

sts.

Dou

ble-

blin

d st

udy.

Lonn

e-R

ahm

et a

l.,

2000

139

Tabl

e 37

. Con

tinue

d

EE

G, v

isua

l evo

ked

pote

ntia

ls, e

lect

ro-

derm

al a

ctiv

ity, E

CG

and

blo

od p

ress

ure.

20 p

atie

nts

and

20 c

ontro

l sub

ject

s

60 H

zin

term

itten

t 15

sec

on/o

ff cy

cle

at

10 T

mag

netic

fiel

d ex

posu

re a

nd

sham

exp

osur

e ap

plie

d ra

ndom

ly

durin

g a

40 m

in p

erio

d

Mag

netic

fiel

d ex

posu

re h

ad n

o ef

fect

on

any

of th

e pa

ram

eter

s ex

amin

ed.

Pat

ient

s re

porti

ng E

HS

di

ffere

d fro

m c

ontro

l su

bjec

ts in

bas

elin

e va

lues

.

Lysk

ov,

San

dströ

m

& H

anss

on

Mild

, 200

1b

Gen

eral

hea

lth s

urve

y of

133

offi

ce

empl

oyee

s. E

xplo

rato

ry s

tudy

of s

kin

dise

ase,

offi

ce e

rgon

omic

s an

d ai

r qua

l-ity

in 3

offi

ce w

orke

rs re

porti

ng E

MF

hype

rsen

titiv

ity c

ompa

red

to 5

con

trols

VD

Us:

5 H

z–2

kHz

~ 10

–15

V m

-1

100–

150

nT

10%

(13)

of g

ener

al s

taff

repo

rted

EM

F hy

pers

ensi

tivity

; no

diffe

r-en

ces

in s

kin

sym

ptom

s be

twee

n E

MF

hype

rsen

sitiv

es a

nd c

ontro

ls

in e

xplo

rato

ry s

tudy

.

The

auth

ors

wer

e no

t ab

le to

attr

ibut

e E

MF

hype

rsen

sitiv

ity to

any

pa

rticu

lar e

nviro

nmen

-ta

l fac

tor.

Arn

etz,

19

97

Per

cept

ion

and

sym

ptom

s in

one

fem

ale

patie

nt10

dou

ble-

blin

d te

sts

Fiel

ds fr

om V

DU

Th

e di

scom

fort

the

patie

nt e

xper

i-en

ced

had

no c

orre

latio

n to

whe

ther

or

not

the

mon

itor a

ctua

lly w

as o

n.

The

patie

nt re

cons

id-

ered

her

ow

n pe

rcep

-tio

n of

the

illne

ss, a

nd

in ti

me

the

sym

ptom

s re

cede

d co

mpl

etel

y.

Kei

su, 1

996

Per

cept

ion

and

sym

ptom

s in

one

pat

ient

50 H

z 34

or 1

00 µ

T1

or 1

0 s

repe

ated

Pos

itive

resp

onse

whe

n hu

mm

ing

of th

e co

ils a

udib

le, d

isap

pear

ed

whe

n “c

amou

flage

d” b

y m

aski

ng

nois

e.

. To

omin

gas,

19

96

Sym

ptom

s an

d ph

ysio

logi

cal r

eact

ions

100

subj

ects

Low

leve

l mag

netic

fiel

ds (<

1

µT) a

t var

ying

freq

uenc

ies

(0.1

H

z–5

MH

z), i

n sh

ield

ed la

bora

-to

ry.

16 o

ut o

f 100

indi

vidu

als

reac

ted

repe

ated

ly to

fiel

ds b

y se

vera

l pa

ram

eter

s (s

ympt

oms,

pup

il di

am-

eter

cha

nges

etc

.).

Not

sur

e w

heth

er fu

lly

blin

d st

udy.

Rea

et a

l.,

1991

140

Tabl

e 37

. Con

tinue

d

EM

F pe

rcep

tion

49 s

ubje

cts

with

EH

S a

nd 1

4 co

ntro

ls50

Hz

100

V m

-1

6 µ

Tra

ndom

ly p

rese

nted

as

2 m

in

bloc

k of

exp

osur

e / s

ham

exp

o-su

re

Per

cept

ion

by 7

sub

ject

s, b

ut n

o di

f-fe

renc

e in

per

cept

ion

betw

een

sub-

ject

s w

ith o

r with

out s

elf-r

epor

ted

EH

S.

Mue

ller,

Kru

eger

&

Sch

ierz

, 20

02

Ele

ctro

card

iogr

am, v

isua

l evo

ked

pote

n-tia

ls (V

EP)

and

ele

ctro

retin

ogra

ms

10 s

ubje

cts

repo

rting

EM

F hy

pers

ensi

v-ity

and

10

cont

rols

Exp

osur

e to

flic

kerin

g lig

ht a

t be

twee

n 25

and

75

flash

es p

er

seco

nd. N

o E

MF

expo

sure

Hig

her V

EP

ampl

itude

s in

EM

F hy

pers

ensi

tive

patie

nts.

Diff

eren

ces

betw

een

mea

n ag

e of

pat

ient

s an

d co

ntro

ls (3

7 vs

47

year

).

San

d-st

röm

et

al.,

199

7

Sel

f-rep

orte

d sy

mpt

oms,

blo

od p

ress

ure,

he

art r

ate,

(ski

n) e

lect

rode

rmal

act

ivity

, E

EG

s an

d vi

sual

evo

ked

pote

ntia

ls10

sub

ject

s re

porti

ng h

yper

sens

itivi

ty

and

10 c

ontro

ls

No

EM

F ex

posu

reD

iffer

ence

s be

twee

n pa

tient

s an

d co

ntro

ls re

gard

ing

self-

repo

rted

sym

ptom

s, h

eart

rate

, ele

ctric

al

activ

ity o

f the

ski

n, a

nd v

isua

l ev

oked

pot

entia

l am

plitu

des.

Lysk

ov,

San

d-st

röm

&

Han

sson

M

ild, 2

001a

Imm

uno-

fluor

esce

nt s

tain

ing

of m

ast

cells

from

ski

n bi

opsi

es13

hea

lthy

subj

ects

VD

U (T

V o

r PC

) exp

osur

e fo

r 2

or 4

hIn

crea

se in

num

ber o

f mas

t cel

ls in

pa

pilla

ry a

nd re

ticul

ar d

erm

is in

5

subj

ects

.

Joha

nsso

n et

al.,

200

1

141

Tabl

e 38

. Moo

d an

d al

ertn

ess

Test

Expo

sure

Res

pons

eC

omm

ents

Aut

hors

Moo

d A

djec

tive

Che

cklis

t bef

ore

and

afte

r ex

posu

re; S

tanf

ord

Sle

epin

ess

Sca

le

befo

re d

urin

g an

d af

ter e

xpos

ure

30 m

ale

subj

ects

60 H

z9

kV m

-1 a

nd 2

0 µT

2 x

3 h

/ day

for 4

day

s

No

effe

ct.

Dou

ble-

blin

d, c

ount

erba

lanc

ed

stud

y.C

ook

et

al.,

1992

Ale

rtnes

s R

atin

g S

cale

, Moo

d A

djec

tive

Che

cklis

t bef

ore

and

afte

r exp

osur

e54

mal

e su

bjec

ts

60 H

z6

kV m

-1 a

nd 1

0 µT

9 kV

m-1

and

20

µT12

kV

m-1

and

30

µT2

x 3

h / d

ay fo

r 4 d

ays

No

effe

ct.

Dou

ble-

blin

d, c

ount

erba

lanc

ed

stud

y.G

raha

m e

t al

., 19

94

Stat

e