Effect of high SARs produced by cell phone like radiofrequency fields on mollusk single neuron

Abstract

During exposure to the cell phone electromagnetic field (EMF), some neurons in the brain at areas of peak specific absorption rate (SAR) absorb more electromagnetic energy than is permitted by existing guidelines. The goal of the present work was to investigate the influence of cell phone-like EMF signal on excitability and memory processes in single neurons. A Transverse Electromagnetic Cell (TEM Cell) was used to expose single neurons of mollusk to the EMF. Finite-Difference Time-Domain (FDTD) method was used for modeling the TEM Cell and the EMF interactions with living nerve ganglion and neurons. Neuron electrophysiology was investigated using standard microelectrode technique. SAR deposited into the single neuron was calculated to be 8.2 W/kg with a temperature increment of 1.21°C. After acute exposure, the threshold of firing of action potentials (AP) was significantly decreased (p ≈ 0.001). Time of habituation to stimulation with the intracellular current injection was increased (p ≈ 0.003). These results indicate that acute exposure to EMF at high SARs impairs the ability of neurons to store information.

Full text

Electromagnetic
Biology
and
Medicine
,
Early Online: 1–11,
2012
Copyright Q Informa
Healthcare USA,
Inc.
ISSN:
1536-8378
print /
1536-8386
online
DOI:
10.3109/15368378.2012.701190
Effect of high
SARs
produced
by cell
phone like
radiofrequency
fields on mollusk single
neuron
B. Partsvania, T. Sulaberidze
&
L.
S
hoshiashvil
i
Department of
Biocybernetics,
Institute of
Cybernetics
of the Georgian Technical
University, Tbilisi,
Georgia
During
exposure
to the cell phone
electromagnetic
field
(EMF), some neurons
in the brain at
areas of peak specific absorption rate
(SAR)
absorb more electromagnetic energy than is
permitted by
existing guidelines. The goal
of the
present
work
was
to
investigate
the
influence
of
cell phone-like
EMF signal
on
excitability
and
memory processes
in
single neurons.
A
T
ransverse
Electromagnetic Cell
(TEM Cell) was used
to
expose single neurons
of
mollusk
to the
EMF.
Finite-Difference Time-Domain
(FDTD)
method
was used
for
modeling
the
TEM Cell and
the
EMF
interactions
with living
nerve ganglion and neurons. Neuron electrophysiology was
investigated
using standard microelectrode technique.
SAR
deposited
into the
single neuron was
calculated
to be
8.2
W/kg with a
temperature increment
of
1.21
8
C.
After
acute exposure,
the
threshold
of
firing of action potentials
(AP) was significantly decreased
(
p
<
0.001). Time
of habituation to
stimulation with the intracellular current injection was
increased
(
p
<
0.003). These
results
indicate that acute exposure to
EMF
at high
SARs
impairs the ability of neurons to store
information.
Keywords
Electromagnetic
field,
Mollusk neuron,
Action potential, Habituation
INTRODUCTION
The
widespread use
of mobile phone over the past
decades
has
generated
interest
about possible health effects of EMF. Therefore, various public organizations
throughout the world
have established safety guidelines
for
EMF absorption.
Specific
absorption rate (SAR) limit, specified in IEEE
C95.1:
2005, has been updated to
2
W/kg
over any 10 g
of
tissue. This new
SAR
limit
i
s
comparable
to the limit specified
in the International Commission on Non-Ionizing Radiation Protection (ICNIRP,
1998) guidelines of
1.6
W/kg for 1g tissue averaging and
2
W/kg for
10
g tissue
averaging.
The spatial distribution of the SAR in the head depends on many
diff
erent
parameters,
including the
frequency
band, phone model, and position of the phone
in relation to the head (Kainz et al.,
2005;
Wiart et al.,
2008).
Earlier, Bernardi et al.
(2000) performed electromagnetic and thermal analysis of the head exposed to
various types of cellular phones available on the market. The obtained results
showed,
for a radiated power of
600
mW, the maximum
SAR values averaged
over
Address correspondence
to
B. Partsvania,
Department of
Biocybernetics,
Institute of
Cybernetics
of
the
Georgian
Technical University, 5,
Sandro
Euli str., Tbilisi,
Georgia,
0186.
E-mail: besari2@yahoo.com
1

2
B. Partsvania
et al.
Electromagnetic Biology
and
Medicine
1g
of
tissue
were from
2.2–
3.7
W/kg depending on the phone
considered.
Kuo and
Chuang (2003) performed FDTD computation of fat layer effects on the SAR
distribution in a head model proximate to a dipole antenna at
900/1800
MHz. They
showed that
peak-SAR
was about
18
W/kg at
900
MHz, and
24
W/kg at
1800
MHz.
De
Salles
et al.
(2006)
studied
SAR
produced by mobile
phones
in the
head
of adults
and children. The frequencies used to feed the antennas were
1850
MHz and
850
MHz. It
was shown
that under similar conditions
1g SAR
calculated
for adult was
5.97
W/kg, while for children the
SAR was
higher than that for the adults at
6.2
W/kg.
Gosselin
et al. (2008) investigated influences of
age-dependent
tissue parameters
and anatomical
structures
on
SAR
and temperature
increase
in the head of cellul
ar
phone users. They showed
that the maximal
value
of
SAR reached 5
W/kg.
Ragha
and
Bhatia
(2009) calculated
SAR variation
s
in the
head
model and their
dependence
on
the distance between the head and the dipole antenna. They showed that the
maximal
SAR value was 17.45
W/kg. Christ et al.
(2010)
studied influence of the head
anatomy
and
tissue properties
on the
energy absorption
and
temperature
i
ncr
ease
of
mobile phone
users.
Anatomically correct head models from different age groups
(children and adults) were
used, as
well as
age-dependent tissue
dielectric values.
They
showed
that that maximum
10
g
SAR
in the child’s head
exposed
to a generic
phone with monopole
antenna
at
900
MHz
was
5
W/kg. Virtanen et al.
(2007)
studied
the effect of authentic metallic implants on the
SAR
distribution of the head. The
results indicate that some of the implants
caused
a notable enhancement in peak
mass-averaged
SAR
(8
–
10
W/kg). Ebrahimi-Ganjeh and Attari (2007) computed
radiation
efficiency
of the
PIFA handset antennas
in the
presence
of
head
and hand.
The calculated peak
SAR was 7.27
W/kg.
Khalatbari et al.
(2006) showed
that, for
a
dipole
antenna,
the local maximum
SAR
deposited to the head was
54.14
W/kg. It was also shown that the maximum SAR
values decreased
dramatically when the
distance between
the head model and the
radiated antenna
increased.
Hence, investigation of the
single
neuron functioning
under high radiation
doses
comparable to those of mobile phone
exposure
is an
issue
of interest.
Influence of
radiofrequency EMFs
with high
SAR
on the
single
neuron biophysics
has been
s
tudied
.
At high radiation
doses (SAR 6.8
–
100
W/kg), isolated neurons
responded
to both continuous and pulsed
microwaves (1.5
GHz) with a
decrease
in
spontaneous
activity, an
increase
in membrane
conductance,
a prolongation of the
refractory period following depolarization, and a
decrease
of the survival time
(Wachtel
et al.,
1975; Seaman
and
Wachtel, 1978).
Arber and Lin
(1985) showed
that
exposure
of snail neuron
s
to
continuous-wave microwaves
for
60
min at
12.9
W/kg
inhibited
spontaneous
activity and
reduced
input
resistance
at certain temperatures.
Exposure of
ne
uron
s
to noise-modulated microwaves at 6.8 and
14.4
W/
kg
predominately
caused excitatory responses characterized
by
augmented
membrane
resistance
and the appearance of greater activity. Arber et al. (1986) found that
exposure
to a
2.45
GHz microwave field at
21
8
C
(SAR 12.9
W/kg for
60
min) caused
minor changes in Golgi complexes and a
s
light swelling of the endoplasmic
reticulum. Ginsburg et al. (1992) sought
effects
of
microwaves
(continuous-wave,
2.45-GHz, SAR 12.5,
or
125
W/kg) on input
resistances
and
AP interval
s
of n
euron
s
i
n
ganglia of the snail Helix
Aspersa,
at
20
8
C.
At
12.5
W/Kg, input
resistance
did not
change during irradiation, but
increased
afterward. At
125
W/kg, input resistance
during irradiation
was
lower than in non irradiated
controls. Later,
Field et al. (1993)
investigated effects
of
pulsed microwaves (2.45-GHz, 10 microseconds, 100 pps,
SAR:
81.5
kW/kg
peak, 81.5
W/kg
average)
on
membrane
input
resistance
and
AP
interval
statistics in spontaneously active ganglion neurons of Helix Aspersa. Statistical
comparison with sham-irradiated n
euron
s
revealed a
s
ignifican
t
increase in the

High SAR effects
on single neuron 3
Copyright
Q
Informa Healthcare
USA,
Inc.
mean input
resistance
of neuron
s
exposed
to pulsed
microwaves.
Bolshakov and
Alekseev (1992) investigated
bursting
responses
of Lymnea neuron
s
to microwave
radiation. The firing rate of
neurons increased
at
SARs
of a few W/kg. The
effect
was
dependent
on modulation, but not on modulation
frequency
and it had a threshold
SAR
near
0.5
W/kg.
These studies investigated
the
effects
of
EMF
on the
biophysical characteristics
of
neurons.
However,
there has been no
research
into
EMF’s
influences on neuronal
electrophysiology,
neuron excitability, and memory processes.
In the
present
work we
investigated
the influence of high
dose
cell phone signal
like electromagnetic fields on the single neuron excitability and ability to store
information.
METHO
DS
Exposure to
EMF
A
specially designed generator was used as
the
source
of the
RF EMF.
The generator
produces RF EMF
in the
range
1728
–
1802
MHz. The emitted power is tunable up to
10
W. Low-frequency modulation of the
generator
produces cell phone signal like
modulation. Signals consisted of an 1800-MHz carrier frequency shaped by a
rectangular pulse train. The width of
each
pulse
was
tunable from
500
–
577
m
s.
The
width of
each
batch
was 4.6
ms.
Each
batch might contain
1–
7 pulses.
Nervous ganglions were exposed
to the EMF in a
transverse electromagnetic
cell
(TEM Cell). A modified TEM Cell as described by Merla et al. (2009) was used.
However,
our
TEM
Cell is
symmetrical. Biological tissue
i
s
positioned on the septum.
Besides,
there is no hole in our TEM Cell
as
opposed to their one. The TEM cell
consists
of a
section
of
rectangular coaxial transmission
line
tapered
at
each
end to
adapt to standard coaxial
connectors.
The Inner conductor (the septum) is a long,
flat plate with a
shaped ends
at both
sides. The generator was connected
to the TEM
Cell input. Output of the TEM cell was loaded with a matched 50V load. The RF
energy
flows longitudinally along the line.
The standing wave
ratio
SWR
for the TEM
cell was determined using a HP 8341B
Synthesized Sweeper
and an Automatic
Network Analyzer 8757A (Global Test
S
upply, Wilmington, NC, USA). It was
established
that the maximal value of
SWR observed
at
1800
MHz
was
equal to 1.3.
Consequently,
the ratio of output power to input power (Pout/Pin)
was 97%.
At this
frequency,
the
TEM
cell
generates reasonably
uniform
electric
and
magnetic fields
in
any
transverse cross section
of the cell at points
away
from the
edges
of the septum.
Biological tissue (nervous ganglion with dimensions approximately
2
mm3) was
placed in foam-plastic small dish
(1
cm3 with the Ringer solution) and was
positioned at the
center
of the
septum. The nervous ganglion was
thus
exposed
to the
propagating fields.
FDTD model included the
RF source,
the TEM cell with input and output tapered
transition sections, foam-plastic dish and its contents, and the matched load.
Electrical properties of the foam plastic are approximately equal to those of air.
A proprietary computer program
package
from Zaridze et al.
(2005), based
on the
FDTD method of Taflove and
Hagness (2000), was used
for modeling the TEM cell
and the EMF interactions with the living nerve ganglion and neurons. The
input
power was
selected
to be
10
W. Calculation of the
SAR
deposited into the single
neuron
gave
a value of
8.2
W/kg. The
calculated temperature
increment
was
1.21
8
C.
In the calculation, we
used
the
12
point
SAR
(Caputa et al.,
1999).
The temperature
increased
over a period of
50
min after which dynamic equilibrium
was
established.
The
final
temperature
of the neuron preparation
was increased
by
1,21
8
C
(above
the
room temperature).

4
B. Partsvania
et al.
Electromagnetic Biology
and
Medicine
Electrophysiology
The
i
so
lated
nervous system of the
mollusk Helix
pomatia was used
in
the
experiments.
Each snail was anesthetized by an injection of isotonic MgCl2 as
was
described
elsewhere (Partsvania
et al.,
2008).
The
nervous system was
then
separated
from the
body.
Ganglia
were treated with
0.5% Pronaze
solution
(Protease
from
S
tr
ep
to
my
ce
s
griseus
–
“
Si
gma-
Aldrich
”)
for
30
min at room temperature. After proteolytic
treatment, the conjunctive
tissue was
carefully removed using fine micro
-s
ci
ss
or
s.
Then the
ganglia were washed
se
ve
ra
l
times
with a
Ringer
solution
consisting
of: NaCl
80
mM, KCl
4
mM, CaCl2
35
mM, MgCl2
z
6
H2O
5
mM, and Tris
7
mM, at pH 7.5.
Thirty-five
sham-exposed
and 36 EMF-exposed
ne
uron
s
were
s
tu
di
ed
.
Immediately after exposure to the EMF in the TEM Cell for
60
min, the nervous
system was
placed in a Petri dish with
Ringer
solution and positioned in a Faraday
cage to filter out any environmental electromagnetic noise. Two identified giant
neuron
s
(neuron
#3
of the Right Pallial
(RPG#3)
ganglion and neuron
#3
of the Left
Pallial ganglion (LPG#3)) were selected for investigations. We investigated the
threshold for AP firing and the time of habituation. The time of habituation is
determined as time interval between the leading edge of deposited intracellul
ar
current and of the last
AP
fired by the neuron. For recording and
analyzing
of data,
the data acquisition
system
ML866 PowerLab/4/30 with the
software
“Chart 5 with
peak
parameters extension
”
(AD instruments Co, Australia)
was used.
The neuron
was
impaled with two
glass microelectrodes
(ME) filled with 2.5Mole KCl. For this
purpose the “Piezo Mikromanipulators
–
PM 20” (Ma¨rzha¨user,
Wetzlar,
Germany)
was used.
Microelectrodes were fabricated using a microelectrode
puller
P-30
(Sutter
Instrument Co, Novato,
USA).
Capillaries were of borosilicate
glass
BF
150
75 10
–
with filaments (Sutter Instruments Co.) The
size
of the microelectrode tip
was
less
than
1
mm. The
resistance
of each microelectrode did not
exceed
15 megaohm.
Microelectrodes were connected (through an Ag/AgCl wire) to the intracellul
ar
electrometer IE-251A (Warner Instruments, Hamden, USA). One microelectrode
served for registration and the other for intracellular stimulation. Intracellul
ar
current injection
was selected
for the neuron stimulation. Let
us denote
intracellul
ar
current injected to the neuron with the term “stimulus.”
Selection
of the stimulus
duration depended on neuron reaction. If the stimulu
s
value was below than
threshold, its span was usually of several
seconds.
If the stimulu
s
value was of
threshold, it
was
terminated
several seconds
later after firing of the
single AP.
If the
stimulu
s
value was
higher than
threshold,
it would
cause
firing of a
series
of
AP.
The
stimulu
s
would be terminated
several seconds
after firing of the last
AP
in the series.
For intracellular stimulation,
a
“
Pic
oamper
source
K
261”
(Keithley
Instruments
Inc.,
Cleveland,
OH,
USA) was
used.
At the beginning of each recording the threshold value of the stimulus
was
determined
(i.e.,
the
value
that
depolarized
the neuron up to the
threshold
and neuron
fireda single AP). For
t
hi
s
purpose, the
amplitude
of the stimulus was
first
set at
0.01
nA.
Experiments revealed
that
t
hi
s
value
always caused
neuron depolarization below the
threshold and did not
cause AP
firing. Then the stimulus
was
i
ncre
as
ed
to threshold
value, so that neuron fired only one AP.
This
value
depends
on neuron excitability.
To
force a
neuron to
generate
se
ve
ra
l
APs, a
s
li
gh
tl
y
higher
s
ti
mu
lu
s
would
need
to be
deposited
to the neuron. At
such a
kind of stimulation, the neuron would habituate to
the
s
ti
mu
lu
s.
Habituation
i
s
expressed as a
decline of
AP
firing by the neuron.
RESULTS
Neuron, like other biological
systems,
i
s
variable. The threshold of the stimulu
s
i
s
randomly variable. Its mathematical
expectation
(mean value) is unknown. By this

High SAR effects
on single neuron 5
Copyright
Q
Informa Healthcare
USA,
Inc.
reason we calculated 95% confidence interval for the unknown mean value of
threshold of the stimulu
s
(see
the Appendix). This is the interval within which with
95% probability that the mean value of the threshold stimulus would fall.
Calculations were
performed for
sham
and
as
well for
exposed neurons.
Calculations
were
based
on experimental
measurement
data. We obtained that these intervals
correspondingly are
0.21
–
0.38
nA for sham and
0.11
–
0.19
nA for
exposed
neurons.
The interval
s
do not overlap each other. This means that the mean value of the
threshold stimulus for exposed neurons is lower than that for
sha
m-exposed
neurons. Corresponding statistical analysis show that the threshold value of the
stimulu
s
for
exposed
neurons is significantly lower than for
sham-exposed
neurons
FIGURE 1 (a)
Reactions
of the
sham-exposed
neuron
(RPG#3)
to stimulus with different amplitudes.
The first stimulus
was
below the threshold value and
equaled
to
0.2
nA. This stimulus
causes
only
depolarization shift of the membrane potential. Value of the second stimulus was
0.3
nA. This
stimulus
was
of the threshold value and
caused
firing of a
single AP.
This value of the stimulus is
within the 95% confidence interval
(0.21
–0.38
nA). (b) Reactions of an
EMF-exposed
neuron
(RPG#3)
to stimulus at different amplitudes. The first stimulus was below the threshold value
(0.05
nA) and
does
not
cause AP
firing.
The second stimulus was
of the
threshold value
(0.15
nA) and
causes firing of a
single
AP. Stimulus of this value was within the 95% confidence interval
(0.11
–0.19
nA). Time intervals of stimulation are shown as solid lines under the recordings.
Calibration
25
mV, 5
s
.

6
B. Partsvania
et al.
Electromagnetic Biology
and
Medicine
(with p
<
0.001 -see
the appendix).
Fig.
1
shows samples
of
AP
by sham- and EMF-
exposed neuron
s
as responses
to stimulus. The stimulation
scheme
is
as
follows: at
first, the stimulu
s
value was set to be below the threshold. This caused neuron
depolarization without firing of AP. Then, the stimulu
s
value was set to be of the
threshold value.
As
a result, the neuron fired a
single
AP.
The results show that
EMF-exposed neuron
s
are more excitable, have lower
threshold, and accordingly need lower stimulus for firing of AP.
It is obvious that habituation might be
observed
after a neuron
has
fired at least
several APs.
For this
purpose,
s
tim
ulus
with higher value than threshold has to be
applied to the neuron to
cause AP. Usually,
neuron
reactions
are not deterministic.
The minimal value of a
s
timulu
s
that
causes
firing of AP series also represents
random variable.
95%
confidence
i
nterval
s
for mean value of this random variable
occur
between
0.42
–
0.56
nA for
sham-exposed
and
0.19
–
0.35
nA for EMF-exposed
neurons. The result was obtained based on measurements data. There is no
overlap here also. Corresponding statistical analysis (see
the appendix)
showed
that
EMF-exposed
neuron
requires
significantly smaller stimuli for firing of the
AP
series
FIGURE 2
Case study
of the habituation to stimulus of
sham-
and
EMF-exposed neurons.
(a) Sham-
exposed
neuron
(LPG#3)
is stimulated with a
s
timulus of
0.5
nA value. The stimulus is within the
95%
confidence
interval
(0.42
–0.56
nA). Neuron fired
5 APs.
Time of habituation
was 8 s.
(b) Neuron
(LPG#3)
from
another
preparation
was
stimulated with a
s
timulus of
0.2
nA
value. The
stimulus was
within the confidence interval
0.19
–0.35
nA. Neuron fires 7
APs.
Time of habituation is 11 s. This
neuron
was exposed
to the EMF in the TEM cell. (c)The
same
neuron is stimulated with stimulus
0.5
nA.
Stimulus
of this
value corresponds
to the
95%
confidence
interval for
sham-exposed
neurons
(0.42
–0.56
nA). The neuron fired 23
APs.
Time of habituation was prolonged up to 60 s. Time
intervals of stimulation are shown with solid lines under the
recordings.
Calibration
25
mV. 10 s.

High SAR effects
on single neuron 7
Copyright
Q
Informa Healthcare
USA,
Inc.
than
sham-exposed neuron
s
( p
<
0.003; see
Appendix). Usually, neurons react to
the higher stimulu
s
longer and fires more
APs (see
the appendix). Thereby, if an
EMF-exposed
neuron is stimulated with stimuli from the interval
0.42
–
0.56
nA (i.e.,
95% confidence
interval for
sham-exposed neurons),
it would fire
APs
longer than
if
it was stimulated with stimuli from the
0.19
–
0.35
nA interval (i.e.,
95%
confidence
interval for
EMF-exposed
neurons). Fig. 2 illustrates this statement. Thereby, it
i
s
obviou
s
that
exposed
neuron
requires
more time for habituation than
sham
exposed
neuron if stimulus is one and the same.
DISSCUSION
FTDT modeling
showed
that the temperature increment of nervous ganglion
as
a
result of
exposure
to the EMF
was
1.21
8
C.
However,
after finishing of the exposure,
the ganglion
was
moved from the foam plastic chamber with
1
cm3 volume to the
Petri dish with
10
cm3 volume. Therefore, it is obviou
s
that at the beginning of
recordings
the
temperature
of ganglion returns to the room
temperature. Thereby,
i
t
might
be stated
that
effects observed
in
experiments arise
during
exposure.
However,
despite
cooling
back
the
effects are retained.
Mollusk Helix
Pomatia is
could blooded
animal and its neurons fire
APs
at ambient
temperatures
approximately from
5
8
C
up to
32
8
C.
The
mollusk
whole organism
and
nervous system
is
adapted
to
such
wide
temperature
interval.
Thereby,
it
i
s
not probable that if
temperature
will be risen on
1.21
8
C
over the room temperature (for example,
22
8
C)
this will change animal’s
ability to store information.
Therefore, observed effects
might be
regarded as
non
themal effects.
Experiments revealed
that absorption of high EMF
energy causes increase
of the
neuron excitability in
comparison
with
sham-exposed neurons
( p
<
0.001).
Statistical
analysis shows
that
reactions
of
EMF-exposed
neuron
are prolonged
with comparison
to
sham-exposed ones
( p
<
0.003).
This
means
that
neurons
fire more
APs
and time
necessary
for habituation is longer for
EMF-exposed
neurons in comparison with
sham-exposed ones.
Habituation
i
s
regarded
as a form of non
associative
learning
(Kandel, 1975). Thereby,
habituation at a
single
neuron
scale may be considered as
a
process of information
store
by the neuron. Consequently, a neuron’s ability of
information
storage
is impaired after
exposure
to
EMF
at high SARs.
These results suggest
that absorption of high
EMF energy
at a point of
peak
SARs
could alter a neuron’s normal functioning. In our opinion, it is worthwhile that
effects
of high-peak
SARs
on the neuron
have
to be taken into consideration in the
setting of guidelines of
EMF-exposure
in the future.
ACKNOWLEDGEMENTS
The
work
was supported
by a grant #GNSF/ST09_80_6_274 from the
Rustaveli
National
Science
Foundation.
The authors are grateful
to Dr. Henry Lai for his help
in the writing of this article and his valuable comments and advice.
Declaration
of
interest
The
authors
report no conflicts of
interest.
The
authors
alone
are responsible
for the
content and writing of the paper.
REFER
ENCES
Arber,
S.
L., Lin,
J.
C. (1985). Microwave-induced
changes
in nerve cells: effects of modulation and
temperature. Bioelectromagnetics
6:257
–270.

8
B. Partsvania
et al.
Electromagnetic Biology
and
Medicine
Arber,
S.
L., Neilly,
J.
P., Lin,
J.
C., Kriho, V.
(1986).
The effect of
2450
MHz microwave radiation on the
ultrastructure of snail
neurons.
Physiol. Chem. Phys. Med. Nucl. Magn. Reson.
18:243
–249.
Bernardi,
P., Cavagnaro,
M.,
Pisa, S., Piuzzi, E. (2000). Specific
absorption rate and
temperature
increases
in the head of a cellular-phone
user.
IEEE Trans. Microwave Theor. Techniq.
48:1118
–1126.
Bolshakov,
M. A.,
Alekseev, S.
I.
(1992).
Bursting
responses
of Lymnea neurons to microwave radiation.
Bioelectromagnetics
13:119
–129.
Bolshev,
L. N., Smirnov, N. B.
(1983).
Tables of
Mathematical
Statistics
Moscow:
Nauka (in Russian).
Brownlee,
K. A.
(1965).
Statistical Theory and Methodology in Science and Engineering, 2nd
ed.
New York:
John
Wiley &
Sons,
Inc.
Caputa,
K.,
Okoniewsky,
M.,
Stuchly,
M.
(1999).
An algorithm for computation of the power deposition
in
human
tissue.
Ant. Prop. Mag.
41:102
–107.
Christ, A.,
Gosselin,
M., Christopoulou, M., et al.
(2010). Age-dependent tissue-specific exposure
of cell
phone
users.
Phys. Med. Biol.
55:1767
–1783.
de
Salles,
A. A., Bulla, G.,
Rodriguez,
C. E.
(2006). Electromagnetic
absorption in the head of adults and
children due to mobile phone operation
close
to the
head.
Electromagn. Biol. Med.
25:349
–360.
Ebrahimi-Ganjeh,
M.
A.,
Attari,
A. R. (2007).
Interaction of dual band helical and pifa
handset antenas
with
human head and hand. PIER
77:225
–242.
Field, A.
S.,
Ginsburg, K.,
Lin,
J. C. (1993). The effect
of
pulsed microwaves
on
passive electrical
properties
and interspike intervals of snail
neurons.
Bioelectromagnetics
14:503
–520.
Ginsburg,
K.
S.,
Lin,
J. C.,
O’Neill,
W.
D.
(1992). Microwave effects
on input
resistance
and action potential
firing of snail
neurons.
IEEE Trans. Biomed. Eng.
39:1011
–1021.
Gosselin,
M.
C.,
Christ,
A.,
Murbach, M., et al.
(2008).
Influences of
age dependent tissue parameters
and
anatomical
structures
on
SAR
and temperature
increase
in the
heads
of cellular Phone
Users.
XXIX
Gen. Assem. Int. Union Radio Sci. (URSI).
Chicago,
Illinois,
August
7–
16.
International Commission on Non-Ionizing Radiation Protection
(ICNIRP)(1998). Guidelines
for limiting
exposure
to time-varying electric, magnetic and
electromagnetic
fields (up to
300
GHz). Health Phys.
74:494
–522.
IEEE C95.1: 2005 (2005). Safety levels
with
respect
to human
exposure
to radio
frequency
electromagnetic
fields,
3
kHz to
300 GHz.
IEEE
Standard
,
New York,
1–
237.
Kainz, W., Christ, A., Kellom, T., et al. (2005). Dosimetric comparison of the specic anthropomorphic
mannequin (SAM) to 14 anatomical head models using a novel denition for the mobile phone
positioning. Phys. Med. Biol.
50:3423
–3445.
Kandel, E.
R. (1976).
Cellular Basis of Behaviour.
San Francisco:
W.H.
Freeman
and Company.
Khalatbari,
S.,
Sardari, D., Mirzaee, A. A., Sadafi,
H.
A. (2006). Calculating
SAR
in two
models
of the human
head
exposed
to mobile phones radiations at 900 and
1800
MHz. Progr. Electromagn. Res. Symp.
Cambridge,
MA, March
26
–29 104
–109.
Kuo, L-C, Huey-Ru Chuang, H-R (2003). FDTD
computation of fat
layer effects
on the
SAR
distribution in a
multilayered
s
uperquadric-ellipsoidal head - model irradiated by a dipole antenna at
900/1800
MHz.
Electromagn Compat., EMC ’03.2003 IEEE Int. Symp. Istanbul,
Turkey.
May
11
–16: 323
–326.
Merla,
C.,
S
a
¨
ıghi,
S.,
Arnaud-Cormos,
D., et al.
(2009).
A
Real
time
exposure system
for neuronal networks
in the MW band. Joint Mtg. Bioelectromagn. Soc. Eur. BioElectromagn. Assoc. Davos, Switzerland
(pp.
48
–50).
Partsvania, B., Shoshiashvili,
L.,
Sulaberidze, T., Modebadze,
Z.
(2008). Extremely
low
frequency
magnetic
fields
effects
on the snail
single neurons.
Electromagn. Biol. Med.
27:409
–418.
Ragha,
L.
K., Bhatia,
M.
S. (2009).
Numerical
evaluation
of
SAR
for
compliance testing
of
personal
wireless
devices.
Int. J. Recent Trends Eng.
6:69
–72.
Seaman,
R. L.,
Wachtel,
H.
(1978). Slow
and rapid
responses
to CW and pulsed
microwave
radiation by
individual
Aplysia pacemakers.
J. Micro. Power
13:77
–86.
Taflove,
A.,
Hagness, S. (2000).
Computational
electrodynamics: the finite difference time-domain method
New York: Artech House.
Virtanen, H.,
Keshvari, J.,
Lappalainen, R.
(2007).
The effect of authentic metallic implants on the SAR
distribution of the head
exposed
to
900, 1800
and
2450
MHz dipole near field. Phys. Med. Biol. 52:
1221
–1236.
Wachtel,
H.,
Seaman, R., Joines,
W.
(1975). Effect
of low-intensity
microwaves
on isolated
neurons.
Ann.
NY Acad. Sci.
247:46
–62.
Wiart,
J.,
Hadjem,
A., Wong,
M.
F., Bloch,
I.
(2008). Analysis
of
RF exposure
in the
head tissues
of children
and adults. Phys. Med. Biol.
53:3681
–3696.
Zaridze,
R. S.,
Gritsenko, N., Kajaia, G.,
et al.
(2005).
Electro-thermal
computational suit for
investigation
of
RF power
absorption and
associated temperature change
in human
body.
IEEE AP-S
Int. Symp. USNC/
URSI Nat. Radio Sci. Mtg.
Washington
D.C. (pp.
175
–179).

High SAR effects
on single neuron 9
Copyright
Q
Informa Healthcare
USA,
Inc.
1
2
s
2
s
2
2
S2
1
n;m ¼
q
ffiffiffiffiffiffi ffiffi
0
APPENDI
X
As described above,
two
issues were
of
interest:
(1)
What
i
s
the influence of exposure
to the EMF with high
SAR
on the neuron excitability? and (2) Does the exposure
change
neuronal habituation?
Corresponding experiments were
carried out to
give answer
to the first question,
i.e., to determine whether there is any difference in excitability
between
sham- and
EMF-exposed neurons. Neuron is regarded minimally excited if stimulus is of
threshold value and neuron
fires
only on AP.
The control
(sham-exposed)
group consisted of 35 neuron
s
and EMF-exposed
group consisted of 36 neurons. The threshold value of the stimulus is a
r
an
dom
variable with unknown mathematical expectation and unknown dispersion.
Calculation
showed
that the
95%
confidence interval for the unknown mean value
of the stimulus threshold is
0.21
–
0.38
nA for sham and
0.11
–
0.19
nA for EMF-
exposed neurons (Brownlee, 1965). These interval
s
do not overlap. The interval
0.21
–
0.38
nA is higher than
0.11
–
0.19
nA. Thus, we
see
that the mean value of the
stimulu
s
i
s
lower for
EMF-exposed
neurons than for
sham-exposed ones.
However,
this
issue needs
additional confirmation with corresponding
s
tatist
ical methods of
hypothesis testing.
This
testing
is
necessary because confidence
i
nterval
s
are close
to
each
other. The
hypothesis
about equality of stimulus threshold
value average
levels
must be
tested
for sham- and
EMF-exposed neurons. Dispersions
of the stimuli are
unknown.
By
this
reason,
firstly, the
hypothesis about
equality of
dispersions
must be
tested. Let us denote unknown dispersion of the stimulus threshold value for
exposed
neuron
as
s
2
and unknown dispersion of the stimulus threshold value for
sham-exposed
neuron
as
s
2
.
The following
hypothesis
must be tested:
s
2
H 0 : 1 ¼
1
ð
main
hypothes
is
Þ
2
s
2
H 1 : 1
,
1
ð
altern
ative
hyp
othesis
Þ
2
Such a
kind of
alternative hypothesis is
conditioned by calculation of the so-called
sampling dispersions. Correspondingly, S2 ¼ 0:017 and S2 ¼ 0:062. Let us take
1 2
a
¼ 0.05 as significance level (i.e., 5% significance level). Numerical value of 5%
upper critical point occur
F
34;35;0
;
05
<
1:8
(Bolshev
and Smirnov,
1983).
Numerical
value of the criteria statistics is
f
~
¼ S2
<
3:65: The null hypothesis H must be
discarded
and the
alternative hypothesis
H1 must
be accepted because
f
~
.
F
34;35;0
;
05
.
This
means
that the
dispersions
s
2
and
s
2
are
significantly different from
each
other.
1
2
Consequently,
t-test could not be used for comparison of equality of stimulus
threshold mean values for
sham-
and EMF-exposed neurons. Here ari
ses
the
so-called Behr
ens
–
Fisher
problem.
Therefore,
approximating method is relevant
in this case (Bolshev and Smirnov, 1983).
S
ta
ti
stics
of cri
te
ri
a
Z
~
X
n
2
Y
m ; n ¼ 36; m ¼
35, has asymptotic
normal distribution. Let us denote
S2 S2
1 2
n
þ
m
as
m1 unknown
average
value of stimulus threshold for
EMF-exposed neuron
s
a
nd
with m2 for
sham-exposed
neurons. The following
hypothesis
must be tested:
H 0 : m1
2
m2 ¼
0
main
hypothe
sis
;
H 1 : m1
2
m2
,
0
alternative hypothes
is
:
This form of the alternative hypothesis is conditioned by two
reasons:
(1) by
positions of the obtained confidence
i
nterval
s
relative to each other; (2) by the

10
B. Partsvania
et al.
Electromagnetic Biology
and
Medicine
difference between numerical values of the corresponding statistical estimations
(statistical averages)
X
36
<
0:1475 and
Y
35
<
0:2963 of unknown mathematical
expectations
m1 and m2 Numerical value of criteria
statistics
is:
Z
~
n
;
m
<
2
3
:
13
and
P
value is p
<
0.001.
Consequently, the main hypothesis must be discarded and the altrernative
hypothesis accepted
with
significance
level
a
¼
0.05.
This
means
that the stimulus
threshold mean value level for
EMF-exposed neuron
s
i
s
s
ig
nificantly lower than for
the
sham-exposed
neurons. The conclusion is that the
EMF-exposed
neurons are
more
excitable
and
sensitive
to the stimulus than the
sham-exposed
neurons.
Let us discuss another issue: Does exposure
to
EMF change
the habituation ability
of the neuron
?
What are
the minimal
levels
of the stimuli that
forces sham-
and EMF-
exposed neuron
s
to fire
several APs (series
of
AP)?
This question is fundamental for
giving answer to that question. Habituation might be observed if a neuron fires
several APs
(at
least generates
more than
1 AP)
at the initial stimulation. Habituation
occurs when
the neuron
stops
firing
APs despite
the
fact
that the stimulus
is going
on.
Let us denote
the minimal
level
of stimulus forcing neuron to
produce AP series
as
“stimulation level of
series (SLS).”
It is obvious that
SLS
is a random variable. Here
are two groups of neurons (as in the single AP case):
sham-
and
EMF-exposed
groups.
The number of
EMF-exposed
neurons is n ¼
36,
and the number of sham-
exposed neuron
s
i
s
m ¼
35.
A
95% confidence
interval of
SLS
unknown mean value
is
0.19
–
0.35
nA for EMF-exposed neurons and
0.42
–
0.56
nA for
sha
m-exposed
neurons.
FIGURE 3
Effect
of
increase
of the stimulus
value
on the time of neuron habituation. First stimulus
at
0.5
nA
caused
firing of 3
APs.
Time of habituation is
11 s. Second
stimulus at
1
nA
caused
firing of
10 APs. Time of habituation is 45 s. Third stimulus at
2
nA caused firing of 41 AP. Times of
habituation is
168 s.
Time intervals of stimulation are shown with solid lines under the recordings.
Calibration:
20
mV 10 s.

High SAR effects
on single neuron 11
Copyright
Q
Informa Healthcare
USA,
Inc.
s
2
s
2
1
f
~
¼
S
<
2
:
7
H
S
1
1
2
2 0
Here, as it was in the case of the single APs, these intervals do not overlap.
Moreover,
EMF-exposed
neurons need an
average
smaller
SLS
than
sha
m-exposed
ones. However, this conclusion
needs
enforcing by the
reason
of proximity of the
confidence intervals. Dispersions are unknown. By this reason, hypothesis of
equality of
dispersions has
to be
tested.
Namely:
s
2
H 0 : 1 ¼
1
main hypothesis;
2
s
2
H 1 : 1
,
1
alternative hypothesis:
2
With corresponding calculation, we obtain that S2 ¼ 0:060 for
EMF-exposed
neurons, and S2 ¼ 0:160 for
sham-exposed
neurons. For a
significance
level of
2 2
a
¼
0.005,
we obtained that
F
34;35;0
;
05
<
1
:
8
and criteria
statistics
is
f
~
¼ S2
<
2
:
7
:
For
2 S2
2
hypothesis
must
be discarded
and the
alternative hypothesis
H
1
1
must
be accepted. Consequently, the unknown
disperses
s
2
and
s
2
are
s
ignificantly
different. For testing the
average values,
the equality approximate method must be
used as
in the
case
of single APs.
Calculation of the
statistics
Z
~
n
;
m
gives
Z
~
n
;
m
<
2
2
:
81:
p
<
0.003. Consequently,
the
main hypothesis H0 must be discarded with significance level
a
¼
0.005
and the
alternative
hypothesis
H1 must be
accepted. Therefore,
stimulation level
SLS
for a
series
of
EMF-exposed neuron
s
i
s
s
ignifi
cantly
less
than for
sham-exposed
neurons.
As
is well known,
increase
of the stimulu
s
causes
a
s
ignifican
t
i
ncrea
se
of neuron
reaction. Consequently, there is
an
increase
in the number of
APs
and
a
prolongation
of the reaction time. For illustration of this idea we give one typical
case
in Fig. 3.
Consequently,
if
sham-
and
EMF-exposed neurons are
stimulated in
average
with
one and the
same
stimulus, the number of
APs
fired by the
EMF-exposed
neurons
would be significantly more. Habituation ability is impaired,
because
the time of
habituation
has
increased.