Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/11005068
Effectsof7Hz-modulated450MHz
electromagneticradiationonhuman
performanceinvisualmemorytasks
ArticleinInternationalJournalofRadiationBiology·November2002
DOI:10.1080/09553000210153934·Source:PubMed
CITATIONS
READS
50
57
5authors,including:
JaanusLass
ViiuTuulik
TallinnUniversityofTechnology
TallinnUniversityofTechnology
76PUBLICATIONS574CITATIONS
37PUBLICATIONS331CITATIONS
SEEPROFILE
SEEPROFILE
RainFerenets
HiieHinrikus
TallinnUniversityofTechnology
TallinnUniversityofTechnology
13PUBLICATIONS280CITATIONS
78PUBLICATIONS502CITATIONS
SEEPROFILE
Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,
lettingyouaccessandreadthemimmediately.
SEEPROFILE
Availablefrom:HiieHinrikus
Retrievedon:19September2016
int. j. radiat. biol 2002, vol. 78, no. 10, 937± 944
EVects of 7 Hz-modulated 450 MHz electromagnetic radiation on
human performance in visual memory tasks
J. LASS†*, V. TUULIK†, R. FERENETS†, R. RIISALO‡ and H. HINRIKUS†
(Received 7 January 2002; accepted 2 May 2002)
Abstract.
Purpose: The aim was to examine low-level 7 Hz-modulated
450 MHz radiation eVects on human performance in visually
presented neuropsychological tasks associated with attention and
short-term memory.
Materials and methods: A homogeneous group of 100 subjects (37
female, 63 male) were randomly assigned to either the exposed
Õ
(10–20 min, 0.158 mW cm 2 ) or the sham-exposed group. A
battery of three diVerent tests measured attention and shortterm memory. Task 1 involved alternately selecting black digits
from 1 to 25 in ascending order and white digits from 24 to 1
in descending order. The time spent on the task and the number
of errors were recorded and analysed. Task 2 involved viewing
a picture of 12 objects during 3 s, followed by a list of 24 words.
The subject was required to select words representing previously
presented objects. In task 3, an array of letters in 10 rows (60 in
each row) was presented, and the subject was required to identify
all examples of a particular two-letter combination.
Results: The results of tasks 1 and 3 showed a signiŽ cant increase
in variances of errors ( p <0.05) in the exposed versus the shamexposed group. The results of task 2 indicated a signiŽ cant
decrease in errors ( p <0.05) in the exposed group.
Conclusions : The data provide additional evidence that acute lowlevel exposure to microwaves modulated at 7 Hz can aVect
cognitive processes such as attention and short-term memory.
1. Introduction
Man-made electromagnetic Ž elds (EMF) are often
substantially stronger than the natural Ž eld. Therefore, many studies have focused on the biological
eVects of EMF. Non-linear and non-thermal biological eVects of non-ionizing Ž elds have been investigated for many years ( Johnson and Guy 1972, Bawin
et al. 1973, Chizhenkova 1988, Vander Vorst and
Duhamel 1996, Hinrikus and Riipulk 1996, Lu et al.
2000). It has been found that the EMF of digital
radio telephone handsets may aVect the brain during
sleep and cause changes in the spectral power of the
electroencephalogram (EEG), and increase the duration of slow wave sleep (Borbély et al. 1999). It has
also been reported that a low-level 50 Hz EMF may
have an in uence on event-related potentials and
*Author for correspondence; e-mail: [email protected]
†Biomedical Engineering Centre, Tallinn Technical
University, Ehitajate tee 5, 19086 Tallinn, Estonia.
‡Estonian Institute of Experimental and Clinical Medicine,
Hiiu 42, 11619 Tallinn, Estonia.
reaction time under speciŽ c circumstances of sustained attention (Crasson et al. 1999). On the other
hand, the data are controversial and other research
teams have reported that EMF do not aVect the
EEG for an awake human (Röschke and Mann 1997)
or that of an asleep human (Wagner et al. 1998).
Some recent studies suggest that exposure to EMF
may modulate the response of EEG oscillatory activity at 8 Hz, speciŽ cally during cognitive processes
(Krause et al. 2000). It is also reported that RF Ž elds
may have a measurable eVect on human cognitive
performance (Koivisto et al. 2000).
Our previous studies (Lass et al. 1999, Hinrikus
et al. 2001) showed that a 7 Hz modulation frequency
had substantial window eVects on EEG, causing
alterations in brain electrical activity (mostly depression in a-wave levels). Usually these alterations
occurred together with variations in human performance and cognitive function. Other studies (Krause
et al. 2000, Koivisto et al. 2000) have showed that the
eVects of RF on cognitive tasks occur during exposure
from standard GSM phones. The radiation generated
by GSM mobile phones has a complicated pulse
modulation pattern.
This study covers the possible eVects of 7 Hzmodulated EMF radiation on attention and shortterm memory in visually presented tasks.
2. Materials and methods
2.1. Subjects
The group of volunteers consisted of 100 students
of Tallinn Technical University; because of their
similar age, educational background and computer
experience, the subjects can be considered as homogeneous. All subjects were healthy, without any
known medical or psychiatric disorders. The 37
female and 63 male subjects had an average age of
21.4 years. A computer program randomly assigned
the 100 subjects to the EMF-exposed (31 males,
19 females, average age 20.7Ô 2.1 years) or shamexposed group (32 males, 18 females, average age
21.7Ô 3.3 years). This was a single blind study: the
International Journal of Radiation Biology ISSN 0955-3002 print/ISSN 1362-3095 online © 2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080/0955300021015393 4
938
J. Lass et al.
subjects were blind to their experimental conditions
but the experimenters were not. However, the subjects were aware of the possibility of being exposed.
2.2. Battery of tasks
Three computerized psychological tasks of diVerent levels of complexity for measuring attention and
short-term memory were developed.
Task 1, the most complex test, was intended to
measure divided attention and short-term memory
simultaneously. The computerized task was an
example of a modiŽ ed trail-making test (Reitan 1955,
part B), involving attention and short-term memory.
A similar test was used for evaluating the eVects of
50 Hz magnetic Ž elds on man, where signiŽ cant
results demonstrated a decrease in performance
(Keetley et al. 2001). Our test consisted of 25 black
and 24 white numbers displayed on the PC screen
in randomized order as an array of seven columns
and seven rows. The task of the subject was sequentially to ‘click’ on the numbers in a predetermined
order, beginning from 1 black and 24 white, so that
the Ž rst one was 1 black, the second was 24 white,
the third was 2 black and the fourth was 23 white,
etc., until all the numbers were clicked. All subjects
had the same task with the same combination of
numbers. Each time a subject clicked a wrong
number, the error was registered and the subject was
asked by the software to Ž nd the correct number
instead. The subjects were asked to complete the task
as fast as possible. The number of errors and duration
to complete the task were recorded. The subject had
a maximum of 8 min to complete the task, but was
allowed to interrupt the task before the predetermined time elapsed in case he/she felt that it was
impossible to Ž nish the task. All subjects who did not
complete the task or made more than 50 errors were
considered as failures. The upper part of Ž gure 1
illustrates the task.
Task 2 aimed to measure visual short-term
memory. Twelve diVerent items (book, butter y, lock,
etc.) were demonstrated simultaneously on the display during 3 s. After a 2-s delay, a list of 24 words
was shown on the display, and the subjects had 1 min
to select all the words referring to the items in the
previous picture. The number of correct and false
selections were recorded. The middle part of Ž gure 1
demonstrates the task.
Task 3, a less complicated task, was a ‘corrective’
test intended to measure attention and is similar to
the well-known Symbol Digit Modalities Test (Smith
1968). The task consisted of 10 rows of letters in
random order, 60 letters in each row. The task
of the subject was to Ž nd all predetermined
combinations of two letters from the text during
1 min. The number of letters looked through, marked
correctly and incorrectly, and the combinations
missed were recorded. From those parameters, the
error rate was calculated as the sum of missed and
incorrectly marked letters. Processing speed was calculated by dividing the number of letters looked
through by the time. The lower part of Ž gure 1
displays the task.
2.3. EMF exposure
The EMF exposure conditions were the same for
all subjects in the exposed group. During each test
session, the experimenter knew the exposure condition for each subject. The exposure conditions were
selected identical to those in our previous work (Lass
et al. 1999).
The 450 MHz EMF was generated by the Rhode
& Swartz (Munich, Germany) signal generator model
SML02. The RF signal was 100% amplitude modulated by the pulse modulator SML-B3 at the frequency of 7 Hz (duty cycle 50%). The 7 Hz frequency
was chosen, based on Ž ndings (Lass et al. 1999),
which indicates that this modulation frequency had
the greatest eVect on human EEG a-waves. The
generator signal was ampliŽ ed with the Dage
Corporation (Stanford, CT, USA) power ampliŽ er
model MSD-2597601. The 1 W EMF output power
was guided by coaxial to the 13 cm quarter-wave
antenna NMT450 RA3206 by Allgon Mobile
Communications AB (Taby, Sweden), located at
10 cm from the skin of the right side of the head.
The speciŽ c absorption rate (SAR) calculation
inside the brain was based on the measured Ž eld
power density on the skin. The EMF was measured
by a Ž eld power density meter P 3-20 (Russia) and
monitored during the experiments by IC Engineering
(Thousand Oaks, CA, USA) Digi Field C Ž eld
strength metre. The measured
Ž eld power density at
Õ
the cortex was 0.158 mW cm 2 . The SAR calculated
using SAR5 sE 2 /2r for brain
conductivity at
Õ 1
450 MHz Õ where s 5 1.18 S m
and
density r 5
Õ
1000 kg m 3 was 0.351 W kg 1 . This formula considers neither the real pattern of Ž eld power density
distribution inside an inhomogeneous body nor the
re ection from the body surface (Hinrikus et al. 1995).
The results of numerical calculations based on the
digital anatomical model took into account the frequency dependence and possible variations of tissue
dielectric value and conductivity (Hurt et al. 2000,
GajsÏ ek et al. 2001). This allowed us to obtain a more
reliable estimation of the whole-body and localized
SAR. In our case (brain tissue,
450 MHz frequency),
Õ
the normalized SAR (W kg 1 ) to EMF power density
EÚ ects of modulated radiation on visually presented tasks
Table 1.
Exposed group (n 5 31)
Times
Errors
939
Results of statistical analysis of task 1.
Sham-exposed group (n 5 36)
F-test for variances
t-test
Mean
SD
Mean
SD
p (two-tailed)
p (two-tailed)
243.6
8.6
67.2
9.2
222.3
5.2
64.6
4.5
0.812
0.030
0.191
0.18
Õ
(mW cm 2 ) ratio factor was 0.06 (GajsÏ ek et al. 2001).
Based on these
conditions, the calculated SAR was
Õ
0.0095 W kg 1 . The levels of power density as well
as SAR calculated by both methods were so low that
thermal eVects are extremely unlikely (Michaelson
and Elson 1995).
2.4. Experimental procedure
The experimental procedure was explained to the
subject who was seated in a chair at the table. He/she
performed the task on the PC monitor by clicking a
mouse. After Ž nishing the Ž rst task, the subject moved
to the next one, thus completing the three tasks
during 10–20 min. The order of the tasks was the
same for all subjects. Each subject performed all
tasks with the EMF on or oV.
2.5. Statistical analyses
Four diVerent tests to estimate statistically significant diVerences between the groups were used.
Conformity to the normal distribution was analysed
by the Kolmogorov–Smirnov test, and whenever
the normality of distribution was not conŽ rmed, the
transformation (square root) was applied. The F-test
for variance tested the homogeneity of variances
between the groups. The two-tailed Student’s t-test
compared the means; when variances were diVerent,
the Welch approximation of the t-test was applied.
The x2 t-test compared statistically the diVerences
between the failed and successful subjects for task 1.
p < 0.05 were considered as signiŽ cant.
3. Results
Histograms of the measured variables distribution
for tasks 1 (errors, time), 2 (correct answers, errors)
and 3 (errors, speed) are shown in Ž gures 2–4,
respectively. The parallel bars illustrate the impact
on the exposed and the sham-exposed subjects.
The results of the statistical analyses for task 1 are
shown in table 1. The only data included in statistical
analysis were from the subjects who actually completed the task. The subgroup consisted of 67 subjects, 36 sham-exposed (21 male, 15 female) and 31
exposed (22 male, nine female). It can be seen from
table 1 that a slight increase in time occurred in the
exposed group, but it was not statistically signiŽ cant.
The error rates have diVerent variances between the
groups and the raw data were not normally distributed. Although the diVerences between the groups
for the number of errors were not signiŽ cant, the
diVerence in the means exceeded by 50% those in
the exposed group. Also, the number of failed subjects
was greater in the exposed group (19, seven male
and 10 female) than in the sham-exposed group (14,
six male and three female). But the x2 -test did not
reveal any signiŽ cant diVerences between the number
of the failed and Ž nished subjects in groups ( p < 0.2).
The results of the statistical analyses for task 2 are
presented in table 2 and Ž gure 3. They show that no
statistically signiŽ cant diVerences existed between the
groups in view of correct answers but there was a
signiŽ cant diVerence in errors: the sham-exposed
group identiŽ ed more incorrect words than the
exposed group.
The results of the statistical analyses for task 3 are
shown in table 3 and Ž gure 4. As with task 1, no
signiŽ cant changes occurred in the means between
the groups but the variance of errors diVers signiŽ cantly between the groups: the exposed group had
greater variance in errors.
4. Discussion
A clear increasing tendency in the number of
errors in the exposed group for task 1 (table 1) was
apparent. The means for the number of errors in the
task aimed at divided attention exceeded that of the
sham-exposed group by 50%. Despite diVerent tendencies in errors between the exposed and shamexposed groups, the diVerence in the number of
errors was not statistically signiŽ cant.
It is interesting to note that more subjects failed
task 1 in the exposed group (19) than in the shamexposed group (14). This Ž nding is in good agreement
with the tendency of an increase in errors. The x2 test applied for statistical evaluation did not reveal
any signiŽ cant diVerences between the number of
failed and those of completed subjects in a group
( p < 0.2).
J. Lass et al.
940
Cancel
test
Figure 1.
Illustrations of psychological task battery. From top to bottom: task 1, modiŽ ed trail-making test; task 2, short-term
visual memory test; task 3, corrective test.
EÚ ects of modulated radiation on visually presented tasks
941
Histogram of errors made during task 1
Number of subjects
12
10
8
Exposed
6
Sham
4
2
0
0-1
2-3
4-5
6-12
13-33
Range of errors
Histogram of time spent completing task 1
Number of subjects
10
8
6
Exposed
Sham
4
2
0
130-170 171-200
201-230
231-275
276-300 301-480
Time periods (s)
Figure 2.
Histograms of errors made and time spent for task 1 (trail-making test) with and without (sham) EMF exposure.
Table 2.
Exposed group (n 5 50)
Correct answers
Errors
Results of statistical analysis of task 2.
Sham-exposed group (n 5 50)
F-test for variances
t-test
Mean
SD
Mean
SD
p (two-tailed)
p (two-tailed)
7.0
2.76
1.95
1.73
7.0
3.56
2.02
1.93
0.801
0.456
0.919
0.032
The results of task 1 showed clear tendencies
produced by the eVect of modulated microwave
radiation, but the statistical values of diVerences
between the exposed and sham-exposed groups were
not signiŽ cant: therefore, further investigations are
necessary.
The most signiŽ cant changes were observed in the
accuracy of a short-term memory task (task 2). In
that task, the number of wrong answers was higher
in the sham-exposed group.
In contrast, the results of the divided attention task
1 showed that the number of errors was higher in
the exposed group. In spite of diVerent statistical
values of the opposite tendencies, none of these
prevail. The reason probably lies in the diVerent
levels of cognitive processes involved during the two
tasks rather than in the opposite character of the
radiation eVect. Authors and the subjects regarded
task 1 as the most complicated. This required ‘parallel
processing’ of two random series of numbers. This
was the only task not feasible to complete for all
subjects from the exposed and sham-exposed groups.
In less complex tasks, the brain’s ability to adapt can
compensate for the eVect of a low-level stressor. The
compensatory mechanisms can even improve the
quality of cognitive processing, which was shown in
task 2. The tendency of an increase in the number
of errors in the exposed group in the conditions of
the most diYcult task should not be neglected. It
seems a possible neuropsychological phenomenon
that diVerent cognitive tasks have diVerences in error
tendencies such that a weak stressor can, at Ž rst,
aVect and trigger changes in more complex cognitive
tasks (Pelosi et al. 2000). This may be important and
should be studied further.
In task 3, changes in the number of errors had the
same tendency as in task 1: they appeared higher in
the exposed group than in the sham-exposed group,
J. Lass et al.
942
Histogram of errors made during task 2
Number of subjects
12
10
8
Exposed
6
Sham
4
2
0
0
1
2
3
4
5
6
7
Number of errors
Number of subjects
Histogram of correct answers task 2
25
20
15
Exposed
Sham
10
5
0
0-1
2-3
4-5
6-7
8-9
10-11
Number of correct answers
Figure 3.
Histograms of errors and correct answers for task 2 (visual memory test) with and without (sham) EMF exposure.
Table 3.
Results of statistical analysis of task 3.
Exposed group (n 5 50)
Mental speed (letters/s)
Errors
Sham-exposed group (n 5 50)
F-test for variances
t-test
p (two-tailed)
Mean
SD
Mean
SD
p (two-tailed)
5.914
7.1
2.553
5.1
5.741
6.5
2.183
2.7
< 0.0001
but not signiŽ cantly. Relative changes were lower than
in task 1 (table 3). This task, less complicated than task
1, was aimed at attention. The character of changes
caused by modulated radiation in task 3 is in good
agreement with the hypotheses used above. The compensatory mechanisms of the brain were more capable
of reducing the eVect of low-level stressor in the less
complicated task than in the more complicated task 1.
However, the compensatory mechanisms do not
appear to enhance the cognitive processing as was the
case with the less complicated memory task 2.
Statistically, the most signiŽ cant diVerences were
associated with variances of errors between the
exposed and the sham-exposed groups in tasks 1 and
3. A statistically signiŽ cant increase in variances of
errors in the exposed group in comparison with the
sham-exposed group occurred in both tasks. A higher
increase in variances was observed in task 1. There
0.275
0.477
0.501
is no reason to associate this fact with personal
diVerences in the vigilance and cognitive abilities of
the subjects. All 100 subjects, randomly divided into
two experimental groups, were of similar age, educational background and computer experience. If there
were any individual diVerences in personal vigilance
and cognitive abilities of the subjects, the relatively
large sample size in each group is likely to compensate
for these diVerences. The fact that the variance of
errors diVered between the exposed and the shamexposed groups in tasks 1 and 3 indicates that the
eVect of modulated radiation varies for diVerent
subjects. An external stressor increased the variability
of the real personal cognitive abilities. Personal sensitivity to radiation as well as the brain compensatory
mechanisms vary from person to person. This is the
most likely explanation why variances in errors
signiŽ cantly increased in the exposed group.
EÚ ects of modulated radiation on visually presented tasks
943
Histogram of errors made during task 3
Number of subjects
35
30
25
20
Exposed
Sham
15
10
5
0
0-2
3-5
6-8
9-11
12-32
Number of errors
Number of subjects
Histogram of mental speed task 3
18
16
14
12
10
8
6
4
2
0
Exposed
Sham
1.5-3.0
3.1-6.0
6.1-7.5
7.5-9.0
9.1-10.0
Speed (letters/s)
Figure 4.
Histograms of errors and mental speed for task 3 (corrective test) with and without (sham) EMF exposure.
Tasks 1 and 3 aimed at the measurement of
attention gave similar results. The results showed a
signiŽ cant increase in the variances of errors and
increasing tendencies of the mean numbers of errors
in the exposed group. The results for task 2 aimed
at memory were diVerent: variances of numbers of
wrong and correct answers remained unaltered and
a signiŽ cant decrease in wrong answers for the
exposed group occurred. The results for the tasks
aimed mainly as attention or at memory probably
diVered due to the diVerent levels of task complexity
rather than to the diVerent type of the tasks only.
The present results give further support to the
assumption that modulated radio frequency Ž elds
have an eVect on human cognitive processing (Preece
et al. 1999, Koivisto et al. 2000, Krause et al. 2000).
Several aspects of the three tasks suggest group
diVerences associated with the presence or absence
of 7 Hz-modulated RF exposure. The results show
that the accuracy of tasks is most likely in uenced
by the modulated radiation. No signiŽ cant diVerences
were found in the speed or time alterations during
task performance. These results contrast with
Koivisto et al. (2000), where the increase of mental
speed was observed. However, alterations were not
found in the accuracy of performance in the working
memory task.
5.
Conclusions
The results suggest that low-level 7 Hz-modulated
microwave radiation has diVerent eVects on the
cognitive tasks of diVerent complexity levels. These
eVects might cause a decrease in the number of
errors for less complicated neuropsychological tasks
and an increase in the errors for more complicated
ones.
Based on the eVects of low-level 7 Hz-modulated
RF radiation on human performance of the cognitive
tasks, the following conclusions can be drawn:
The eVect on human performance is present
during cognitive tasks.
The eVects vary from subject to subject.
The physiological mechanisms underlying such
eVects are not yet fully understood. Studies should
be continued to replicate and extend the Ž ndings.
Acknowledgements
This study was supported by Grant No. 5143 of
the Estonian Science Foundation. The authors are
thankful to Kairi Vaher for valuable discussions in
neuropsychology.
944
EÚ ects of modulated radiation on visually presented tasks
References
Bawin, S. M., Gavalas-Medici, R. J. and Adey, W. R., 1973,
EVect of modulated very high frequency Ž elds on speciŽ c
brain rhythms in cats. Brain Research, 58, 365.
Borbe ly, A. A., H uber, R., Graf, T ., Fuchs, B., Gallmann,
F. and Achermann, P ., 1999, Pulsed high-frequency
electromagnetic Ž eld aVects human sleep and sleep electroencephalogram. Neuroscience Letters, 275, 207– 210.
Chizhenkova, R. A., 1998, Slow potentials and spike unit
activity of the cerebral cortex of rabbits exposed to
microwaves. Bioelectromagnetics , 9, 337.
Crasson, M., Legros, J. J. and Legros, P . W., 1999, 50 Hz
magnetic Ž eld exposure in uence on human performance
and psychological parameters: two double blind experimental studies. Bioelectromagnetics , 29, 474–486.
GajsÏek, P ., Hurt, W. D., Ziriax, J. M. and Mason, P . A.,
2001, Parametric dependence of SAR on permittivity
values in a man model. IEEE Transactions on Biomedical
Engineering , 48, 1169–1177.
Hinrikus, H., Krasavin, J., Beilenhoff, K. and Hartnagel,
H. L., 1995, Calculation of microwave radiometric signal
in multilayered biological tissue. In Proceedings of the IEEE
Engineering in Medicine & Biology 17t h Annual Conference,
pp. 1583–1584.
Hinrikus, H., Lass, J., P arts, M. and Tuulik, V., 2001,
Modulated microwave eVects on EEG. In Proceedings
of 5t h International Congress of the European BioElectromagnetics Association (EBEA), pp. 289–291.
Hinrikus, H . and Riipulk, J., 1996, Modulated microwave
eVects on the nervous system. Medical and Biological
Engineering and Computing, 34(suppl. 1), 129– 130.
Hurt, D. H ., Ziriax, J. M. and Mason, P . A., 2000, Variability
in EMF permittivity values: implications for SAR calculations. IEEE Transactions on Biomedical Engineering, 47,
396–401.
Johnson, C. and Guy, A. W., 1972, Non-ionizing electromagnetic wave eVects in biological materials and systems.
Proceedings of IEEE, 60, 692– 701.
Keetley, V., Wood, A., Sadafi, H. and Stough, C., 2001,
Neuropsychological sequelae of 50 Hz magnetic Ž elds.
International Journal of Radiation Biology, 77, 735– 742.
Koivisto, M., Revonsuo, A., Laine, M. and HaÈmaÈlaÈinen, H .,
2000, The eVects of electromagnetic Ž eld emitted by
GSM phones on working memory. NeuroReport , 11,
1641–1643.
Krause, C. M., SillanmaÈki, L., Koivisto, M., HaÈggqvist, A.,
Saarela, C., Revonsuo, A., Laine, M. and H aÈmaÈlaÈinen,
H., 2000, EVects of electromagnetic Ž elds emitted by
cellular phones on the electroencephalogram during a
visual working memory task. International Journal of
Radiation Biology, 76, 1639–1667.
Lass, J., Tuulik, V. and H inrikus, H ., 1999, Modulated microwave eVects on EEG alpha-waves. Medical and Biological
Engineering and Computing, 37(suppl. 1), 105–108.
Lu, S.-T., Mathur , S. P ., Stuck, B., Zwick, H., Ziriax, J.,
Merrit, J., Lutty, G., McLeod, D. and Johnson, M.,
2000, EVects of high power microwaves on the retina of
the rhesus monkey. Bioelectromagnetics , 21, 439–454.
Michaelson, S. M. and Elson, E. C., 1995, Interaction of
nonmodulated and pulse modulated radiofrequency Ž elds
with living matter: experimental results. In C. Pol and
E. Postow (eds), Handbook of Biological EÚ ects of
Electromagnetic Fields (Boca Raton: CRC Press),
pp. 435–533.
P elosi, L., Slade, T ., Blumhardt, L. D. and Sharma, V. K.,
2000, Working memory dysfunction in major depression:
an event-related potential study. Clinical Neurophysiology,
111, 1531– 1543.
P reece, A. W., Iwi, G., Davies-Smith, A., Wesnes, K., Butler,
S., Lim, E. and Varey, A., 1999, EVect of a 915-MHz
simulated mobile phone signal on cognitive function in
man. International Journal of Radiation Biology, 75, 447–456.
Reitan, R. M., 1955, The relation of the trail-making test to
organic brain damage. Journal of Consulting Psychology,
19, 393–395.
RoÈschke, J. and Mann, K., 1997, No short-term eVects of digital
mobile radio telephone on the awake human electroencephalogram. Bioelectromagnetics , 18, 172– 176.
Smith, A., 1968, The symbol digit modalities test. Learning
Disorders, 3, 83–91.
Vander Vorst, A. and Duhamel, F., 1996, 1990–1995 advances
in investigating the interaction of microwave Ž elds with
the nervous system. IEEE Transactions on Microwave Theory
and Techniques, 44, 1898– 1909.
Wagner, P ., RoÈschke, J., Mann, K., H iller, W. and Frank,
C., 1998, Human sleep under the in uence of pulsed
radiofrequency electromagnetic Ž elds: a polysomnographic study using standardized controls. Bioelectromagnetics, 19, 199–202.