Effects of GSM mobile phone on the central nervous system not ...
R e s e a r c h
Effects of GSMon the Ce
Evaluation of nine
alleged effects of electro-
magnetic fields caused by
mobile phones operating
via the actually applied
GSM standard (Global
Standard for Mobile
Phones) on the Central
Nervous System are
analysed based upon a
number of studies pub-
lished in recent years
The studies mentioned above used the von Klitzing (1992)
electroencephalogram (EEG) as a tool for CETECOM/von Klitzing (1994)
measuring such interaction. During EEG von Klitzing (1995)
measurements electrodes placed at the head Reiser et al. (1995)
skin transmit electric signals partially pro- Spittler et al. (1997)
duced by processes occurring within the Hietanen et al. (1997)
brain. As the active brain segments are Röschke, Mann (1997)
located at different parts of the head, the Thuroczy et al. (1997)
signals are deducted by 21 different elec- Krafzyk et al. (1999)
trodes distributed in a given pattern. Since In 1992, during experiments von Klitz-
the low voltages recorded during EEG ing observed alterations in his test sub-
measurements are easily altered either by jects occurring in the range of alpha waves.
various technical influences or by the par- In 1995 he exposed 17 test subjects to an
ticipant’s behavior, these examinations electromagnetic field for 3 x 15 min, each
require a particularly careful handling ex- exposure being followed by a 15-min
cluding all possible disturbances. break. Field strengths were several points
In the following, we will deal with nine below the field strengths in other studies.
studies on this topic published in recent Von Klitzing detected a decreased activity
years: in the alpha wave range (8-12 Hz) and
26 NEWS l e t t e r 1 . 01
bile Nervous System not proven
simultaneously, an increase of activity at trode was placed nearest to the field source. tervals there was a break of ten min dur-
lower frequencies (in the theta and gam- However, when the two measurement phas- ing which the field was switched off. This
ma range) during exposure. However, there es were evaluated separately the result experimental strategy offered the possi-
is no information available on the order of could not be confirmed as significant. For bility to record individual alterations in
exposure and control intervals. As a sta- the 900-MHz GSM field the result was a single subjects as well as across all test
tistical evaluation of results also is lack- significant increase of the beta activity subjects. Under the conditions of this ex-
ing, it is not possible to assess the signifi- during the time period after the end of perimental set-up no alterations of the EEG
cance of the study results. exposure. At other EEG measurement caused by the field of the GSM phone could
Following the first study of von Klitzing points no alterations of potentials could be detected.
from 1992, in 1994 an EEG measurement be detected. However, the effects can not In the same year, however, Thuroczy et
in three participants was carried out by be excluded by the statistical approach al. presented different findings. The scien-
CETECOM and von Klitzing. In two of three selected by the authors (sign-test for paired tists conducted two-channel EEG meas-
participants von Klitzing again detected samples). urements in 76 persons seperately screen-
EEG alterations possibly caused by elec- Hietanen et al. examined several mobile ing the side of the head exposed to the
tromagnetic fields. CETECOM as well eval- phones including GSM 900 and GSM 1800 electromagnetic field and the side of the
uated the data comparing results with von devices looking at possible effects on the head without exposure. The experiment
Klitzing ultimately reaching the conclu- EEG. All devices operated at maximum followed a firm pattern of exposures and
sion that von Klitzing’s findings resulted transmission power exposing test partici- sham exposures. By means of this experi-
from methodical or technical impairments. pants for 20 min to the electromagnetic mental design alterations of the EEG’s al-
As only three participants were involved field. Sham exposures, too, were part of pha waves could be proven. However, the
statistical evaluation is lacking. Interest- the study. Only at exposure to a mobile authors did not compare all exposure and
ingly, one of the participants allegedly fell radio device alterations in the delta wave sham exposure intervals. Further – ss is
asleep during measurements. As sleep as a range were found; however, this effect can seen in the figures – there were observed
rule causes EEG alterations, the obtained not be seen as significant. Measurements increases as well as decreases of the alpha
results are highly doubtful, regarding the of delta waves prove especially critical, activity, i.e. a uniform alteration of alpha
issue at hand. since there may occur strong disturbances waves is not recognisable. Therefore, an
In 1995, Reiser and his co-workers too caused by transpiration at measurement evaluation of results seems rather diffi-
looked for effects of electromagnetic fields point areas. All other applied devices cult, particularly as the exact conditions
on the EEG. Their investigation was de- showed no impact on the participants’ EEG. of controls were not available.
signed as a cross study during which each Consequently, the authors claim that mo- A similar EEG-based cross study was
participant was examined various times bile radio device emissions have no disad- published in 1997 by Röschke and Mann.
and simultaneously served as a control. vantageous effects on the human brain. The EEGs of 34 participants again was re-
Besides, a double-blind set-up was select- In 1997 Spittler et al. exposed 52 test corded by means of two-channel technique.
ed, i.e. neither the scientist nor the patici- subjects to the radiation of a GSM 900 The study subjects were exposed and sham
pant knew when the exposure to an elec- mobile phone of a 217-Hz pulse frequen- exposed to the electromagnetic field of a
tromagnetic field began. cy. The 52 subjects participating in this GSM 900 mobile phone in random se-
Evaluation of the 147-MHz field tests experiments were divided into two groups. quence. Each interval had a duration of
showed an increase of the alpha and beta While the members of one group were ex- 10 min, each interval was followed by a
activity at a measurement point at the back posed for ten min to the electromagnetic break of 30 min. Ultimately, Röschke and
of the head when exposure and end phase field, the subjects of the other group were Mann could not detect any alteration of
are taken together. The measurement elec- only sham exposed. Between exposure in- the EEG caused by electromagnetic fields.
NEWS l e t t e r 1 . 01 27
R e s e a r c h
Taken from the study of Eulitz et al. (1998): Curve of the spectral short-term power at standard
(‘frequent’) and aimed stimuli (‘infrequent’) in the frequency range of 19-31 Hz measured at two
measurement points located symmetrically towards the head middle line (‘left’/’right’). The
radiation source was located to the left. As to the aimed stimuli (presentation at 0 ms) a
difference is seen dependent of the exposure (‘with PEMF’/ ‘without PEMF’) being significant
from 260 ms onwards.
Taken from the study of Krause et al. (2000): Curve of the spectral short-term power in the
frequency range of 4-6 Hz at encoding and retrieval of words during a memory experiment.
Approximately from 300 ms onwards after word presentation (at 0 ms) the spectrum shows
differences dependent of exposure (‘EMF on’/’EMF off’).
28 NEWS l e t t e r 1 . 01
Krafzyk et al. conducted a sophisticated phones can have an impact on the human night prior to the experiment at the labo-
test series publishing their results in 1999. sleep. Apart from EEG measurements also ratory to adaptate to the new sleeping en-
As a method, they selected a placebo-con- the electrocardiogram (ECG), the electro- vironment.
trolled cross study including a complex myogram (EMG) and the electro-oculogram In 1998 Wagner et al. planned to repeat
EEG measurement carried out over a time (EOG) provided additional physiologically the results of the Mann study from the
period of three and a half hours for each relevant quantities being of importance for year 1996 in a study of their own. Howev-
participant. Here, evoked potentials as well characterising sleep phases. The measure- er, they reduced the power flux density of
as MAEP (medium latency evoked poten- ments were continuously recorded 0.5 mW/cm² to 0.02 mW/cm². They did
tials) were observed; further, the P-300 throughout the night. not succeed in replicating the results of
experiment was applied. During this ex- In 1996 Mann and Röschke exposed test the Mann study, though at least a similar
periment participants have to identify cer- persons to a GSM field of an assessed power trend of parameters was recognisable.
tain sounds out of a whole sound sequence. flux density of 0.05 mW/cm². This experi- These results led to the assumption that
As an electromagnetic field source mobile ment was conducted as a double-blind an effect of the GSM field might depend
phones of the GSM 900 and GSM 1800 study including sham exposures. To adapt on its strength. To look into this assump-
standard were used. As the scientists claim to the unusual environment the partici- tion in 1999 Röschke conducted a third
in their evaluation, during the EEG partic- pants had to spend the night before the experiment with a comparable experimen-
ipants showed no effects that could be ex- measurements at the sleep laboratory. Dur- tal set-up but a significantly increased
plained by the irradiated electromagnetic ing statistical evaluation (ANOVA = stand- power flux density of 0.2 mW/cm². Thus,
fields. ard analysis of variance) a shortened phase his workgroup addressed the issue of the
Thus, it must be emphasized that none of sleep onset (reduced sleep onset laten- relation between results and the dose of
of the studies presented here gives evi- cy) and a dimished percentage of REM- the electromagnetic field. However, this
dence of interaction between GSM fields phase were registered when the electro- test could neither detect any effect; even
and comprehensive brain functions. This magnetic field was switched on. The EEG the trend of parameters determined by
result is supported by the fact that the of each participant was separately evalu- Wagner was not found during this third
electromagnetic signals of a GSM mobile ated following the measurements of the experiment.
phone already are strongly dampened by different sleep phases. Apart from a sig- In 1999 Borbély et al. exposed their par-
head tissue. At a tissue depth of some cm nificant increase of the EEG power during ticipants during a placebo-controlled dou-
the dampening of a GSM 900 field amounts REM phase no effects were detected being ble-blind cross study to a 900-MHz GSM
to more than 90 per cent. As to a GSM possibly caused by the electromagnetic signal with different modulation frequen-
1800 field, the dampening effect is even field. cies. In this set-up, the frequency mixture
stronger. Thus, theoretically only outer The results of this study gave consider- of 2, 8, 217 and 1736 Hz was meant to
brain regions may be influenced by the able cause for discussion. However, we simulate the simultaneous activity of a base
electromagnetic field. Since the EEG dur- must bear in mind that the observed ef- station and a mobile phone. A three-di-
ing waking state is controlled by the tha- fects may either be caused by the GSM pole antenna was installed in the vicinity
lamus being located deep inside the brain, carrier signal or by modulation frequency. of the test person’s head to simulate the
an impact of GSM mobile phone electro- In 1998 a study of Hinrichs et al. showed near-field situation. The specific absorp-
magnetic fields on the EEG because of this that a GSM 1800 field has no effect at all tion rate (SAR) lay at approximately 1 W/
natural position alone is very unlikely. on sleep relevant parameters. In a double- kg. During one night the baseline of each
Penetration depth of these fields is not blind study participants were exposed to a participant was recorded by tests, in the
sufficient to cause alterations. GSM 1800 field with the mobile phone following night measurements were made.
antenna being mounted in a distance of Each participant had to absolve two of
Impact of electromagnetic approximately 1.4 m to the head of the these sessions. Here too, exposure and
fields caused by GSM mobile test subject. This set-up was meant to sim- sham exposure occurred in a random pat-
phones on sleep phases ulate conditions in the vicinity of a fully tern.
Between 1996 and 1999 several studies active base station. Exposures and sham The results showed an increased activi-
were carried out aimed to decide wether exposures took place in four subsequent ty during the non-REM sleep phases in the
electromagnetic fields of GSM mobile nights. Here too, participants spent the alpha band (7-14 Hz). This trend was es-
NEWS l e t t e r 1 . 01 29
pecially pronounced during the first non- experiment could be the consequence of posures and exposures were performed; as
REM sleep phase. Further, a shortened the fact that participants did not spend a field source a GSM mobile phone was
waking phase was observed. The authors their adaptation night in the measurement used. Information on frequency, field
report that in particular test persons who room. strength and SAR values is not available.
had been the first to be sham exposed In his study, Borbély used highly com- All measurements were replicated after two
more often woke during the first night. plex pulse signals. Lyskov (1993) reported weeks. Of particular notice is that the test
They explained this phenomenon rather EEG alterations caused by a 45-Hz field. persons were not ‘blinded’ and therefore
by the fact that these participants spent However, the field alone was not respon- knew when the field was activated. No
adaptation night in a room other than the sible for the alterations. Lyskov observed significant effects were observed.
the test room; allegedly, the electromag- these EEG alterations only under condi- The studies until now performed
netic fields they were exposed to did not tions when the field repeatedly was could not confirm any effect of GSM fields
lead to this effect. switched on and off in a 1-second inter- on evoked potentials. Therefore, an im-
The contradictory results from the stud- val. Therefore, it can not be excluded that pact of GSM fields on brain activity and
ies of Mann, Wagner and Röschke do not Borbély’s results are caused by the com- functions so far seems to be unlikely. How-
provide clear evidence of interaction be- plex signal structure. Thus, a replication ever, one must bear in mind that there are
tween GSM fields and sleeping behavior, of the Borbély study under standardised no results available on motor-sensorially
too, though the studies’ experimental strat- conditions seems reasonable and neces- evoked potentials. Existing studies should
egies are similar. A dose-dependent effect sary. be revised considering this aspect, too.
can only be assumed as long as there would Some uncertainty remains as long as
be an extremely non-linear relation be- the results of the Borbély study are not Cognitive studies with
tween dose and effect. Even at the highest refuted by a replication study. Neverthe- event-related potentials (ERP)
power flux density (0.02 mW/cm²) no ef- less, we may assume that there is no inter- and event-related
fects could be proven. This agrees with the action between GSM fields and human desynchronisations/synchro-
results of Hinrichs et al. (1998) for the sleeping behavior. nisations (ERD/ERS)
GSM 1800 field, though pulse frequency Another component of Krafczyk’s 1998
and field parameters (near-field/far-field) Impact of electromagnetic study were tests of the so-called P-300
are different. fields of GSM mobile phones experiment. During this experiment par-
Only Paschke et al. (1996) and Reite et on evoked potentials (EP) ticipants have to identify certain sounds
al. (1994) presented results from which an In their 1998 study (see EEG) Krafczyk being part of a standard sound sequence.
impact of electromagnetic fields on sleep et al. conducted tests with evoked poten- 300 milliseconds after the stimulus the EEG
can be derived. However, the applied field tials. The participants were visually stim- can register the corresponding response.
differs very much from the GSM field ap- ulated by a chess-board pattern changing Based upon the ANOVA evaluation, Krafc-
plied of the above mentioned studies, since eight times per second. At this fequency zyk could not detect any EEG deviation
the authors treated persons suffering from there was a distinctive EEG alteration. caused by the GSM field neither in ampli-
sleep disorders with a complex modulated However, the evoked potential was not tude nor in time.
high-frequency field of 27 MHz. Paschke changed by the magnetic field. This result In a study from the year 1998, Freude et
and Reite report shortened sleep onset determined by means of the ‘ANOVA meth- al. detected a possible impact of GSM fields
phases and an earlier beginning of deep od’ was significant. Also the somewhat on physiological processes. They exam-
sleep phases. delayed evoked potentials (medium laten- ined the so-called readiness potential, for
The results of Borbély et al. (1999), too, cy evoked potentials [MAEP]) were not in- example occurring immediately prior to
show an alleged impact of electromagnet- fluenced by the GSM field. an involuntary movement. In a double-
ic fields on sleeping behavior. However, In 1998, Urban et al. examined 20 test blind study test persons had to perform
the depicted effects only were significant persons in a cross study using Krafczyk’s two tasks:
because of the huge deviations between changing chess-board pattern to visually • mouse-click in self-selected time in-
the first sham exposure and the following produce potentials. The frequency of the tervals (easy task) and
basic value and exposure measurements. visual stimulus was 1.5 Hz, the simulation • observe a marker moving clockwise
Moreover, the erroneous results of this had a duration of five minutes. Sham ex- and stop it by mouse-click after three spins
30 NEWS l e t t e r 1 . 01
R e s e a r c h
precisely at the 12- o’-clock- position occurring with a delay of 260 to 380 mil- were presented and of the phase of word
(complex task). liseconds. Participants hearing the stand- identification showed that there are EMF-
During experiment, a multi-channel EEG ard sounds showed no alterations. Re- related ERD/ERS alterations. These altera-
was recorded. The electromagnetic field sponse times as well showed no impact of tions concern the tetha band (4-8 Hz) and
was either switched on (three to five min) the electromagnetic field. Eulitz’ results the alpha band (8-12 Hz). Higher frequen-
or off. During easy task performance the demonstrate that the GSM field, though cy bands were not examined. The quota of
EEG showed no change caused by the GSM having an influence on the background right answers under different test condi-
field. In contrast, the EEGs recorded dur- EEG during task performance, does have tions was similar. Event-related potentials
ing complex task performance showed a no effect on event-related activities or be- (ERP) showed no differences.
different result. A part of the measure- havior. Krause et al. limited their tests to altera-
ment points showed a small, but still sig- In a recent study from 2000, Krause et tions within the lower bands of the EEG
nificant decrease of amplitudes was regis- al. examined the impact of electromag- up to approximately 12 Hz. Eulitz et al.
tered. Amazingly, this effect particularly netic fields on memory. Test persons had detected alterations within the beta band,
occurred at head spots located opposite to to perform memory tasks when a GSM i.e. in the range between 12.5 Hz and 30
the exposed area. The authors dealt with 900 mobile phone was switched on. Dur- Hz. Thus, the results of these studies can
the question whether these effects were ing each experiment round in total 92 text not be compared to each other.
caused by alterations of the cortical passages each containing five verbs were In contrast to the Krause study, the ef-
stimilus susceptibility, as claimed by Adey read via headset to the test persons. The fects detected by Eulitz et al. and Freude
in 1979. However, this assumption can not test persons had to decide whether the last et al. mainly occurred in one brain hemi-
be verified since the relation between word of a given passage also was part of sphere. Nevertheless, in a new study not
events at cellular level and the EEG at the first four words of the previous pas- yet published Krause et al. could confirm
head surface so far is not known well sage. The decisions had to be communi- their earlier results.
enough. cated via a push-button. During the tests Summarising the results of the different
In 1998, Eulitz et al. conducted a study a multi-channel EEG was recorded. Evalu- studies we must conclude that so far there
of combined ERP and ERD/ERS tests. Here, ation showed alterations of the ERD/ERS is no proof of electromagnetic fields show-
participants too had to perform a hearing in the lower frequency range of alpha ing a systematical effect on the electric
task being tied to a second task. At hear- waves (8-10 Hz). A differentiated analysis signals of the brain being caused by stim-
ing a certain test sound in a sequence of of the time interval in which the words ulus processing. Though ERD/ERS altera-
1000-Hz sounds (standard sounds), they
had to push a button as fast as possible.
The sounds they had to identify occurred
in a random distribution. Event-related
potentials from both tasks (sound identifi-
cation; pushing button) as well as ERD/
ERS spectra were separately evulated for
each task. Response times and number of
right and wrong responses were recorded,
too. Two of the four total experiments took
place at a GSM 900 field; each course had
a duration of 15 minutes.
The evaluation of measurement results
based upon the ANOVA method showed
significant deviations for the ERD spectra
under field exposure. Alterations were es-
pecially observed at the side of the head
pointing to the field source in the range
between 19 Hz and 31 Hz (beta band) and
NEWS l e t t e r 1 . 01 31
tions show that the specific cortical area Summary
may be influenced by electromagnetic Despite partially contradictory results
fields during stimulus processing, there are not any of the presented studies did prove
neither correspondences in the frequency a dependence between EEG structure and
band nor concerning the relation between exposure to electromagnetic fields. There-
the exposed hemisphere and the hemi- fore, the hypothesis that electromagnetic
sphere where effects occurred. fields have no effect on the waking EEG
Random effects can only be definitely seems to be highly probable.
excluded by further studies. The observed In view of the sleeping EEG three out of
effects did occur when the test persons five studies show no evidence of an im-
had to perform complex tasks involving pact of the GSM field on sleep. In spite of
several brain regions - meaning that elec- that we should not totally exclude the pos-
tromagnetic fields could either have an sibility that GSM fields have certain
influence on communication between the small effects on sleep.
involved brain regions or on intellectual Concerning the area of evoked poten-
capacity being challenged by the corre- tials an interaction with GSM fields can be
sponding task. excluded. Insofar, all studies confirm the
Tasks of gradually increasing difficulty null hypothesis.
could contribute new insight concerning Only field effects on the ERD/ERS oc-
this issue. Also the fact that the applied curing during easy task performance could
electromagnetic fields because of tissue be established. However, one may specu-
dampening only reach the top brain re- late that pulsed high-frequency fields dis-
gions should be considered when defining turb intracortical communication trans-
tasks for new experiments particularly ad- ported by high-frequency brain currents
dressing these peripheral regions. Ultimate- or that they decrease intellectual capacity.
ly, all observed effects lie within a varia-
tion width also being expected from other Hermann Wygoda, scientific journalist
potential altering influences. Here, we have (this article is a summary of a study presented
by Prof. Dr. Hermann Hinrichs, University
to mention the use of drugs affecting the
central nervous system as well as decreas-
ing concentration levels or the circadiane
rhythm. A health risk can not be derived
from the effects detected until now.
32 NEWS l e t t e r 1 . 01