Modulation of neuronal activity and plasma membrane properties with low-power millimeter waves in organotypic cortical slices

Abstract

As millimeter waves (MMWs) are being increasingly used in communications and military applications, their potential effects on biological tissue has become an important issue for scientific inquiry. Specifically, several MMW effects on the whole-nerve activity were reported, but the underlying neuronal changes remain unexplored. This study used slices of cortical tissue to evaluate the MMW effects on individual pyramidal neurons under conditions mimicking their in vivo environment. The applied levels of MMW power are three orders of magnitude below the existing safe limit for human exposure of 1 mW cm(-2). Surprisingly, even at these low power levels, MMWs were able to produce considerable changes in neuronal firing rate and plasma membrane properties. At the power density approaching 1 microW cm(-2), 1 min of MMW exposure reduced the firing rate to one third of the pre-exposure level in four out of eight examined neurons. The width of the action potentials was narrowed by MMW exposure to 17% of the baseline value and the membrane input resistance decreased to 54% of the baseline value across all neurons. These effects were short lasting (2 min or less) and were accompanied by MMW-induced heating of the bath solution at 3 degrees C. Comparison of these results with previously published data on the effects of general bath heating of 10 degrees C indicated that MMW-induced effects cannot be fully attributed to heating and may involve specific MMW absorption by the tissue. Blocking of the intracellular Ca(2+)-mediated signaling did not significantly alter the MMW-induced neuronal responses suggesting that MMWs interacted directly with the neuronal plasma membrane. The presented results constitute the first demonstration of direct real-time monitoring of the impact of MMWs on nervous tissue at a microscopic scale. Implication of these findings for the therapeutic modulation of neuronal excitability is discussed.

Full text

Modulation of neuronal activity and plasma membrane properties with low-power millimeter
waves in organotypic cortical slices
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2010 J. Neural Eng. 7 045003
(http://iopscience.iop.org/1741-2552/7/4/045003)
Download details:
IP Address: 137.78.179.52
The article was downloaded on 29/07/2010 at 17:14
Please note that terms and conditions apply.
View the table of contents for this issue, or go to the journal homepage for more
Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING JOURNAL OF NEURAL ENGINEERING
J. Neural Eng. 7(2010) 045003 (9pp) doi:10.1088/1741-2560/7/4/045003
Modulation of neuronal activity and
plasma membrane properties with
low-power millimeter waves in
organotypic cortical slices
Victor Pikov1,5, Xianghong Arakaki2, Michael Harrington2,
Scott E Fraser3and Peter H Siegel3,4
1Neural Engineering Program, Huntington Medial Research Institutes, Pasadena, CA, USA
2Molecular Neurology Program, Huntington Medial Research Institutes, Pasadena, CA, USA
3Division of Biology, California Institute of Technology, Pasadena, CA, USA
4Division of Engineering and Applied Sciences, California Institute of Technology, Pasadena, CA, USA
E-mail: pikov@hotmail.com
Received 25 December 2009
Accepted for publication 30 April 2010
Published 19 July 2010
Online at stacks.iop.org/JNE/7/045003
Abstract
As millimeter waves (MMWs) are being increasingly used in communications and military
applications, their potential effects on biological tissue has become an important issue for
scientific inquiry. Specifically, several MMW effects on the whole-nerve activity were
reported, but the underlying neuronal changes remain unexplored. This study used slices of
cortical tissue to evaluate the MMW effects on individual pyramidal neurons under conditions
mimicking their in vivo environment. The applied levels of MMW power are three orders of
magnitude below the existing safe limit for human exposure of 1 mW cm−2. Surprisingly,
even at these low power levels, MMWs were able to produce considerable changes in neuronal
firing rate and plasma membrane properties. At the power density approaching 1 μWcm
−2,
1 min of MMW exposure reduced the firing rate to one third of the pre-exposure level in four
out of eight examined neurons. The width of the action potentials was narrowed by MMW
exposure to 17% of the baseline value and the membrane input resistance decreased to 54% of
the baseline value across all neurons. These effects were short lasting (2 min or less) and were
accompanied by MMW-induced heating of the bath solution at 3 ◦C. Comparison of these
results with previously published data on the effects of general bath heating of 10 ◦C indicated
that MMW-induced effects cannot be fully attributed to heating and may involve specific
MMW absorption by the tissue. Blocking of the intracellular Ca2+-mediated signaling did not
significantly alter the MMW-induced neuronal responses suggesting that MMWs interacted
directly with the neuronal plasma membrane. The presented results constitute the first
demonstration of direct real-time monitoring of the impact of MMWs on nervous tissue at a
microscopic scale. Implication of these findings for the therapeutic modulation of neuronal
excitability is discussed.
(Some figures in this article are in colour only in the electronic version)
5Author to whom any correspondence should be addressed.
1741-2560/10/045003+09$30.00 1© 2010 IOP Publishing Ltd Printed in the UK

J. Neural Eng. 7(2010) 045003 VPikovet al
1. Introduction
Pervasive use of electromagnetic waves in modern society
has strong potential implications for human health, and
particularly for the activity of the nervous system. Wide-
spread use of infrared waves and radiofrequencies (RF) in
wireless communication has already stimulated an interest into
their possible effects on the nervous tissue and led scientists
to the discovery of considerable effects at 2 μm (150 THz)
(Wells et al 2005a,2005b) and 333 mm (900 MHz)
(Hillert et al 2008, Wiholm et al 2009). Most recently,
novel wireless video communication devices (Lawton 2008)
and non-lethal military weapons (LeVine 2009) have been
developed within the extremely high-frequency band of RF.
This band spans 30–300 GHz, or wavelengths from 10 to
1 mm, and is formally designated as the millimeter wave
(MMW) band. MMWs have low energy levels (0.12–1.2 meV)
that are associated with molecular motion and intermolecular
interactions, and are strongly absorbed by dipole-oriented
molecules, such as water and phospholipids (Enders and Nimtz
1984, Liburdy and Magin 1985, Cametti et al 1988, Beneduci
2008, Ramundo-Orlando et al 2009). At 60 GHz, a commonly
utilized MMW frequency, the absorption coefficient for water,
tissue media and skin tissue ranges from 50 to 55 cm−1, with
almost all energy being absorbed within 0.4 mm (Zhadobov
et al 2008). This strong absorption has been exploited for more
than a half century for simple heating of biological tissue, but
as the authors show in this paper, MMWs may also produce
more subtle effects on the neuronal tissue at power levels well
below the existing safe exposure limit of 1 mW cm−2(Chou
and D’Andrea 2005). Assuming that they show no long-term
health impact, these effects might be exploited for regulating
neuronal firing and, perhaps, other cellular activities.
The effects of MMWs on neural tissue have been
previously examined only in whole-nerve preparations. In
the first detailed study (Pakhomov et al 1997), MMWs (40–
52 GHz) were applied to an isolated sciatic nerve in the frog.
Using 20 min long exposures at the incident power density
(IPD) of 2–3 mW cm−2, a small increase (1–3%) in nerve
firing was seen in several trials, while in others no changes
were observed. In a more recent study by another group
(Alekseev et al 2009), MMWs (42 GHz) were applied to the
hind paw skin of an anesthetized mouse. During 1 min of
MMW exposure to the skin area, firing of the sensory sural
nerve was evaluated. At the IPD of 45–220 mW cm−2,skin
temperature was increased by 1.7–4.5 ◦C and strong (down to
44% of control) suppression of the nerve firing was observed.
A radiant IR source, which resulted in equivalent skin heating
of 4.5 ◦C, produced a similarly strong (40% of control)
suppression of the nerve firing. In addition to the suppression
of firing, the authors saw a transient (20–40 s) facilitation
of the nerve firing, which was evident immediately after the
MMW exposure, but not after the IR exposure. This capability
of MMWs to produce opposite—excitatory and inhibitory—
effects on the nerve activity is puzzling and warrants a more
detailed evaluation of the MMW-induced neuronal effects with
single-cell resolution. This study attempts to address this
important question by using patch-clamp recording in a slice
preparation of cortical tissue.
2. Methods
2.1. Animals and electrophysiological setup
Animal procedures were approved by the HMRI IACUC.
Three neonatal P13–P16 Sprague-Dawley rats were deeply
anesthetized by ketamine (87 mg kg−1, IP) and xylazine
(13 mg kg−1, IP) and decapitated. The brains were removed
and placed in a beaker for perfusion with ice-cold (0 ◦C)
artificial cerebrospinal fluid (aCSF): a mixture of (in mM)
126 NaCl, 2.5 KCl, 2.4 CaCl2,1.3MgCl
2,1.2NaH
2PO4,
26 NaHCO3and 10 mM glucose (pH 7.4). The osmolarity
of aCSF was 310–330 mOsm, as measured by depression
of freezing point (Micro-Osmette, Precision Systems, Inc.,
Natick, MA). Slices (300 μm) containing the cerebral cortex
were cut in the coronal plane with a motorized vibratome
(World Precision Instruments, Sarasota, FL) and placed
on a custom-made slice chamber for incubation at room
temperature (20 ◦C) in an artificial cerebrospinal fluid solution
(aCSF) (bubbled with 95% O2,5%CO
2, pH 7.4) for at least 1 h
before patch clamp recording. All slices were used on the same
day. For electrophysiological recording, the slice was placed
in a tissue chamber (BT-1-18, Cell MicroControls, Norfolk,
VA) and was continuously perfused with room-temperature
aCSF using an inlet and outlet designed for turbulence-
free laminar flow. The aCSF flow (1–2 ml min−1)was
controlled by gravity. Prior to MMW exposure, the amount
of aCSF was minimized to allow better MMW penetration.
Cortical layer 2/3 pyramidal neurons were visually identified
under differential interference contrast (DIC) optics and IR
illumination on an upright Nikon FN1 microscope with a 40×
(0.8 N.A.) lens and a high-speed CCD camera (Qimaging).
Glass recording electrodes (OD 1.5 mm, ID 1.12 mm, World
Precision Instruments, Sarasota, FL) had a resistance of 4–
7Mwhen filled with a pipette solution containing (in
mM) 140 K gluconate, 5 KCl, 4 NaCl, 10 HEPES, 0.3
Na3GTP, 2 MgATP (pH 7.2–7.3, osmolarity 280–300 mOsm).
For evaluation of the neuronal activity in the absence of
intracellular calcium signaling, the intracellular calcium stores
were buffered by adding 10 mM EGTA to the pipette solution.
Whole-cell current-clamp recordings were obtained using a
MultiClamp 700B patch amplifier, Digidata 1440 digitizer and
a computer running pCLAMP software (all from Molecular
Devices, Sunnyvale, CA). Data records were digitized at
10 kHz. The junction potential was around 10 mV, and
all voltage recordings shown were not corrected for these
potentials. Intracellularly injected current was held at 0 or
negative values for 75% (15 s) of each 20 s cycle. Positive
current (40–200 pA) was injected at 25% duty (5 s) during each
cycle and the intracellular potential was recorded. Intracellular
voltage measurements were analyzed in the beginning at
0.5 s after the initiation of current injection to avoid the initial
variability in the AP firing rate. The action potential amplitude
was defined as the difference between the pCLAMP-detected
AP threshold and the positive peak. Time constants for rise
and decay (τrise and τdecay) of the action potential (AP) were
calculated from the single-exponential functions, describing
the full rising and decaying halves of the AP. These functions
were derived from the data points by applying the Chebyshev
2

J. Neural Eng. 7(2010) 045003 VPikovet al
Figure 1. Left: electrophysiological setup with integrated MMW source. Right, top: close-up view of the tissue chamber and the hollow
waveguide with an open-ended curved tip, allowing MMWs to be irradiated toward the tissue chamber. Right, bottom: schematic
illustration of the MMW propagation from the tip of the waveguide. See the text for more details.
fitting algorithm. The apparent input resistance was calculated
by dividing the change in the resting membrane potential
by the amplitude of the injected depolarizing current: Rn=
V/I.
2.2. MMW exposure
The MMW exposure system (figure 1) consisted of a
custom-assembled MMW source (injection-locked IMPATT
oscillator) operating at 60.125 GHz and producing up to
185 mW of continuous wave power at the exit of the open-
ended WR-15 waveguide. Power was continually monitored
through a directional coupler and calibrated RF power meter
(HP 436A, Agilent Technologies, Santa Clara, CA). The power
emitting on the tissue chamber was controlled by a rotary vane
attenuator that could reduce the power to below detectable level
(<1μW) without throwing an electrical switching discharge.
The rectangular waveguide with 3.8 ×1.9 mm2aperture was
carefully positioned above the tissue chamber, blocking the
path of the microscope. The MMW power was directed
perpendicular to the plane of the chamber for more uniform
irradiation of the area occupied by the tissue slice. The tip of
the waveguide was placed just beyond the far field distance of
the tissue with an air gap of 4.8 mm to the surface of the aCSF
solution and approximately 7 mm to the top of the slice. The
linearly polarized RF fields exiting the rectangular waveguide
expand at a half angle of ∼27◦in the H plane (parallel to
the wide wall) and 56◦in the E plane (along the short wall)
creating a half-power ellipse of 5.5 mm2(major and minor
axes of 14.2 and 4.9 mm, respectively) at the aCSF surface.
The RF beam is refracted upon entering the solution and forms
a half-power ellipse of 6.5 mm2(major and minor axes of 15.2
and 5.4 mm, respectively) at the top of the slice, 2.2 mm below
the aCSF surface. A large portion of the incident RF signal is
reflected upon contact with the fluid surface and is assumed to
be radiated into the surrounding space without absorption by
the fluid. Using published data for the complex permittivity
of water and water-based media at 60 GHz (Zhadobov et al
2008), we calculated the reflected power to be ∼45%, leaving
55% of the available power for absorption by aCSF. Using
the published data for the same frequency and similar ionic
media (Zhadobov et al 2008) and Beer’s law (I=I0e−ax )
for the intensity drop with penetration depth x(in cm), we
estimated the loss tangent for MMW absorption by aCSF to
be 1.48. This translates into the absorption coefficient of
52 cm−1, indicating the power drop of 99.5% by 1 mm of aCSF.
Using a further approximation that the energy in the RF beam
is uniformly distributed over the tissue within the half-power
ellipse, the maximal power of 185 mW exiting the waveguide
port produces a power density of approximately 90 mW cm−2
transmitted into the aCSF and less than 1 μWcm
−2at the tissue
level, 2.2 mm below the surface. Therefore, in this study, the
highest applied IPD of MMWs at the level of slice neurons
is 1000 times lower than the most conservative current safe
exposure level of 1 mW cm−2(Chou and D’Andrea 2005).
During MMW exposure, the aCSF temperature containing
the cortical slice was monitored using a thermocouple probe
positioned near the bottom of the well 3 cm away from the
center of the chamber. The probe was connected to a TC2BIP
temperature controller with 0.1 ◦C accuracy and digital readout
screen (Cell MicroControls, Norfolk, VA). To compensate for
the delay in horizontal heating transfer from the center to the
periphery of the chamber, thermal recordings were calibrated
with a probe positioned at the depth of 2.2 mm either directly
under the waveguide or 3 cm away from the waveguide.
The applied corrections for different IPDs were +0.1 ◦Cfor
280 nW cm−2,+0.3◦C for 560 nW cm−2and +0.5 ◦Cfor
840 nW cm−2.
3

J. Neural Eng. 7(2010) 045003 VPikovet al
Table 1 . Analysis of the AP-related and membrane parameters in two subsets of cortical neurons. Statistical significance (p) indicates
difference from the pre-exposure control, based on the Dunnett post hoc test, and is marked as follows: <0.05 (∗), <0.01 (∗∗).
Neuronal Pre-MMW During MMW exposure at following IDP ranges (nW cm−2)
group Parameter exposure 30–50 70–90 140–200 250–330 490–600 700–800
#1 (n=4) AP amp. (mV) 72 ±273±174±275±276±476±582±5∗
Firing rate (Hz) 3.2 ±0.2 2.8 ±0.3 3.2 ±0.3 3.1 ±0.4 2.0 ±0.6 0.9 ±0.2∗1.3 ±0.4∗
τrise (ms) 0.73 ±0.15 0.63 ±0.1 0.51 ±0.06 0.56 ±0.08 0.4 ±0.06 0.28 ±0.06∗0.25 ±0.06∗
τdecay (ms) 9.7 ±1.7 8.8 ±3.7 7.9 ±2.0 6.5 ±1.9 3.8 ±1.1 2.6 ±0.9∗1.0 ±1.1∗∗
Rn(M) 308 ±69 397 ±56 287 ±77 283 ±54 229 ±45 148 ±18∗173 ±34∗
#2 (n=4) AP amp. (mV) 42 ±441±543±645±552±650±845±13
Firing rate (Hz) 5.6 ±0.3 5.9 ±0.1 6.0 ±0.3 6.4 ±0.4 7.1 ±0.5 8.1 ±1.1∗6.5 ±0.5
τrise (ms) 1.1 ±0.15 1.07 ±0.18 0.98 ±0.14 0.92 ±0.15 0.55 ±0.09 0.42 ±0.05∗∗ 0.36 ±0.18∗
τdecay (ms) 5.1 ±0.5 5.5 ±0.9 5.1 ±0.7 4.5 ±0.9 3.2 ±0.7 2.3 ±0.7∗1.1 ±0.6∗∗
Rn(M) 423 ±42 448 ±56 394 ±38 372 ±41 274 ±29 243 ±34∗208 ±86∗
Table 2 . Analysis of the AP-related and membrane parameters of cortical neurons during MMW exposure and bath heating. Statistical
significance (p) indicates difference from the pre-exposure control, based on the Dunnett post hoc test, and is marked as follows: <0.05 (∗),
<0.01 (∗∗).
Pre-MMW During exposure During exposure Pre-bath During bath
Parameter exposure at 490–600 nW (cm−2) at 700–800 nW (cm−2) heatingaheatinga
Bath T(◦C) 20.0 22.1 23.2 23 33
AP amp. (mV) 57 ±661±757±12 90 ±11 76 ±8
Firing rate (Hz) 4.4 ±0.5 5.0 ±1.6 4.8 ±1.8 N.A. N.A.
τrise (ms) 0.92 ±0.12 0.36 ±0.04∗∗ 0.33 ±0.11∗0.90 ±0.27 0.51 ±0.13∗
τdecay (ms) 7.4 ±1.2 2.4 ±0.5∗∗ 1.1 ±0.3∗∗ 2.8 ±1.1 1.7 ±0.4∗
Rn(M) 366 ±43 202 ±27∗∗ 196 ±51∗389 ±151 254 ±96∗
aThe neuronal parameters from a published report (Lee et al 2005), where bath heating was produced by an in-line
water heater.
We used general linear model (GLM) followed by the
Dunnett post hoc test on predefined comparisons against the
pre-exposure control using the Minitab software (Minitab Inc.,
State College, PA). Significance was set at p<0.05 (∗) and
p<0.01 (∗∗). The values in tables 1and 2, and figure 3are
expressed as mean ±standard deviation.
3. Results
Whole-cell recordings were made in freshly prepared slices
of the cerebral cortex of neonatal rats. Neurons of pyramidal
phenotype located in the layer 2/3 of the cortex were selected
using infrared videomicroscopy with DIC enhancement. A
total of eight neurons were patched in eight slices from three
rats. After the neurons were patched and their microscopic
visualization was no longer necessary, the MMW exposure
system was moved into place and used to illuminate the tissue
chamber with a series of IPDs in random order. Prior to the
MMW exposure, the fluid flow through the chamber was halted
and aCSF was partially removed from the well containing the
tissue slice until visual confirmation of a concave meniscus
formed at the fluid surface. The remaining amount of aCSF
in the chamber was 0.9–1.1 ml and the depth of solution
covering the slice was between 2.1 and 2.3 mm. This depth was
sufficient to fully stabilize the temperature in the fluid above
the slice and allow sufficient nutrition to the neurons during
the measurement sequence lasting 25–30 min. A thermal
sensor was located close to the slice, assuring that the slice
was exposed to an equivalent amount of aCSF-damped heating
from the MMW source. The actual power, generated by the
MMW source, was measured throughout the experiment using
an inline directional coupler and the calibrated power meter
(figure 1).
In order to improve neuronal viability and stability of the
neuronal firing in the whole-cell recordings, the depolarizing
current was injected intracellularly at a 25% duty cycle
(figure 2). Several MMW power levels were applied to the
slice in randomized order to remove any possible effects of
cumulative exposure. Each exposure lasted for 60 s (or
three 20 s cycles); thus, three 5 s neuronal activity records
were collected during the exposure. In a presented example
(figure 2), exposure of the neuron n14 with a low MMW
IPD of 71 nW cm−2produced no significant effect on the
neuronal firing rate during the exposure and a small increase
in firing rate after the exposure. Higher IPD levels (from
284 to 737 nW cm−2) produced considerable suppression of
neuronal firing during the exposure and strong facilitation of
firing immediately after the exposure.
Next, we examined the latency of neuronal responses to
MMWs. We evaluated the changes in the membrane input
resistance Rn, estimated as the change of the resting membrane
potential during the injection of depolarizing current (V/I).
This parameter was not dependent on the presence of the APs
and thus, was not susceptible to confounding variability in
the AP firing rate. An averaged Rnvalue was calculated for
each 5 s period of current injection in 20 s intervals. The IPD
4

J. Neural Eng. 7(2010) 045003 VPikovet al
Figure 2. Sample recordings of neuronal activity in neuron n14 selected to show four IPD levels of MMW exposure, applied in random
order during the experiment. For each record, two traces are shown: the top trace shows the intracellular voltage and the bottom shows the
injected current. Red bars indicate the duration of the MMW exposure with the value above the bar indicating the IPD level.
Figure 3. Latency of the MMW-induced suppression in Rnat
different IPD levels. The red bar indicates the duration of the MMW
exposure and values in the graph legend indicate the IPD levels.
Error bars indicate the standard deviations.
exposures for different neurons were grouped into six levels,
and the Rnvalues for these levels were averaged across the
neurons for each time point (figure 3). Maximal Rndrop
was evident at 50 s into the MMW exposure. Thus, the
detailed evaluation of MMW effects on the neuronal activity
and plasma membrane parameters in subsequent figures 4–8
and tables 1and 2was done using the values at 50 s into
the exposure. After the exposure, the Rnvalues exhibited a
gradual recovery, which was complete (or nearly complete for
the highest IPD) after about 2 min. Following each exposure
trial, we have allowed at least 3 min before commencing the
next exposure to provide sufficient time to assure the recovery
of baseline neuronal activity.
The firing rate and the AP amplitude were somewhat
variable across the 25–30 min of recording, as can be seen
in figure 2in the trial with IDP of 492 nW cm−2. In order
to reduce the confounding influence of this variability on the
effects of MMWs, the overall baseline value of the firing rate
and AP amplitude was calculated by averaging the values from
the baseline periods prior to each MMW exposure trial. The
data for individual exposure trials were then normalized using
a ratio of the pre-exposure baseline value in that particular trial
to the overall baseline value.
No normalization was used for the pre-exposure
variability in firing rate and AP amplitude across the neurons.
Instead, we decided to group the neurons based on their
5

J. Neural Eng. 7(2010) 045003 VPikovet al
(A) (B)
Figure 4. The effect of the IPD on the firing rate (A) and AP amplitude (B) in eight neurons.
Figure 5. Sample AP profiles in neuron n14 before MMW exposure and at 50 s after the beginning of exposure at different IPD levels
(indicated in red). In the experiment, these IPDs were applied in random order.
(A) (B)
Figure 6. TheeffectoftheIPDonτrise (A) and τdecay (B) in eight neurons.
baseline firing rate. Two groups were identified: first (n=4)
with low (∼3 Hz) and second (n=4) with high (5–6 Hz) firing
rate (figure 4and table 1). The low-firing group exhibited
strong suppression of neuronal firing by the MMWs, while
the neurons in the second group exhibited either facilitation
(n=2) or remained relatively unchanged (figure 4(A)). The
neurons in the first group also exhibited higher baseline AP
amplitudes (∼70 mV) as compared to the second group (30–
50 mV). Low baseline AP amplitude in the neuron n22 can
be potentially attributed to uncompensated capacitance in the
electrode. In both groups, there appeared to be a small (if any)
effect of the MMWs on the AP amplitude even at the highest
IDP levels.
To examine the possible changes in the plasma membrane
characteristics, we examined the shape of the individual APs
and the input resistance of the membrane. The profiles of
APs were compared prior to MMW exposure and at 50 s
during the MMW exposure at several IPD levels (figure 5).
Noticeable narrowing of the peaks can be readily observed at
higher IPD levels. This narrowing was completely reversed
6

J. Neural Eng. 7(2010) 045003 VPikovet al
Figure 7. TheeffectoftheIPDonRnin eight neurons.
to the baseline level within 2 min after the exposure (data not
shown).
Quantification of the time constants for the AP rise and
decay phases (τrise and τdecay) confirmed the remarkable
decrease in both parameters proportional to the increasing
strength of IPD and this effect was clearly dose dependent
on the level of IDP (figure 6).
Rnwas markedly affected by the MMW exposure. The
Rndrop, similar to the peak narrowing, was strongly IPD dose
dependent, and in most neurons at high IPDs, the Rnvalue
dropped below 200 M(figure 7).
Out of the eight neurons studied, two neurons (n13 and
n14) were examined in the absence of intracellular calcium
signaling, accomplished by buffering the intracellular Ca2+
stores with 10 mM EGTA, added to the pipette solution.
As evident from figures 4,6, and 7, there was no apparent
difference in the responses of these neurons (n13 and n14)
as compared to other neurons in this study. Their baseline
values and MMW-induced changes in the τrise ,τdecay and Rn
values were indistinguishable from the others, and their firing
rates and AP amplitudes were close to those of the non-Ca2+-
buffered neurons in group 1, n11 and n12.
Statistical evaluation of AP-related and membrane
characteristics of two identified neuronal groups is presented
in table 1. In group 1, MMW exposure reduced the firing rate
to one third of the pre-exposure value, while in group 2 it was
slightly increased. The AP amplitude in group 1 is increased
by 14%, while in group 2 it remained unchanged. Both groups
of neurons exhibited similar shortening of the rise and decay
times of the AP peak and similar reduction of Rn.
Both groups of neurons exhibited similar reduction in their
plasma membrane properties (Rn,τrise, and τdecay ): therefore,
the data from all neurons were combined to evaluate their
relative reduction as the function of MMW IPD (figure 8).
These membrane properties, which are inversely correlated
with the membrane conductivity, were dose-dependently
reduced. At the highest IPD level, Rnwas suppressed by
more than 40%, and τrise and τdecay were suppressed by more
than 60% from their pre-exposure values.
Figure 8. Relative reduction in Rnτrise and τdecay as a function of
the MMW IPD, averaged from eight neurons.
The neuronal changes during exposure at the two strongest
IPD levels were compared with changes, observed during
heating of the aCSF bath with an in-line fluid heater, as reported
previously by others (Lee et al 2005). While the temperature
rise of aCSF during MMW exposure at the two highest IPD
levels was only 2.1 and 3.2 ◦C versus 10 ◦C during bath heating,
the observed MMW-induced decreases in τrise ,τdecay and Rn
were considerably more pronounced (table 2). Specifically, the
MMW exposure caused an 83% decrease in the AP width and
46% decrease in Rn, while the 10 ◦C bath heating produced
only a 41% decrease in the AP width and a 35% decrease
in Rn.
4. Discussion
This study provides the first real-time examination of single
neuronal activity during MMW exposure in a slice preparation.
The key findings in this study are the MMW-induced fully
reversible effects on neuronal firing, the shape of the APs and
Rn. These short-lasting effects were observed at extremely
low IPD levels (<1μWcm
−2) of MMW and are stronger than
those induced by 10 ◦C bath heating.
The observed MMW effects on the neuronal firing
might be neuron subtype specific. A subset of layer 2/3
cortical neurons with low firing rate (group 1) has been most
susceptible to suppression of neuronal firing by the applied
MMWs, with a reduction to one third of the pre-exposure level.
This dichotomy of responses might indicate the presence of
different subtypes of pyramidal cells. Further studies with
blockage of synaptic activity might be able to clarify the
existence of such pyramidal neuron subpopulations and the
mechanisms of their MMW-induced suppression of neuronal
firing.
While the AP amplitude was not significantly affected by
the MMW exposure, other AP parameters, τrise and τdecay
were strongly affected by the MMWs, with the total AP
duration reduced from 8.3 to 1.4 ms (83% decrease) for the
highest IPD level. A key characteristic of the membrane
conductivity, Rn, was similarly dose-dependently reduced by
7

J. Neural Eng. 7(2010) 045003 VPikovet al
increasing IPD levels, indicating the channel opening. The
coincident decrease in Rn,τrise and τdecay is not surprising,
as the narrowing of AP spikes is strongly correlated with
decreased Rn(Trevelyan and Jack 2002). The 10 ◦C bath
heating of cortical neurons has a negligible effect on the AP
amplitude and significantly reduces Rn,τrise and τdecay (Lee
et al 2005).
If MMW action on the neurons is largely thermal, then
some of the same mechanisms implicated in the effects of
heating might be involved. One such proposed mechanism
for the heating-dependent changes in the neuronal membrane
properties involves a temperature-dependent increase in K+
channel function (presumably due to faster channel activation),
while Na2+ channel function remains relatively constant
(Volgushev et al 2000b, Cao and Oertel 2005). A modeling
study of Rnin the cortical pyramidal neurons also indicated
an important role of the increased membrane conductance in
the thermally induced effects (Trevelyan and Jack 2002). The
MMWs effects in this study were considerably stronger than
those evoked by 10 ◦C bath heating, while the temperature
rise at two highest IPD levels reached only 2.1 and 3.2 ◦C.
It may be possible that some of the MMW-induced
effects in this study were mediated through non-thermal
mechanisms, involving specific absorption of the MMWs
by neural tissue. Specific absorption of MMWs by plasma
membrane-bound water, phospholipids and other dipole-
oriented molecules has previously been demonstrated (Enders
and Nimtz 1984, Liburdy and Magin 1985, Cametti et al
1988, Beneduci 2008, Ramundo-Orlando et al 2009). The
involvement of specific plasma membrane ion channels in
the non-thermal mechanism of MMW action remains to be
explored.
In an attempt to differentiate the direct effects of MMWs
on the voltage-sensitive channels in the plasma membrane
from possible indirect activation of the membrane-bound ion
channels through intracellular second messenger-mediated
signaling, we blocked the intracellular Ca2+-dependent second
messenger signaling in two neurons. This was accomplished
using a pipette solution with a calcium chelator (10 mM
EGTA), which buffers the intracellular calcium stores and
thus eliminates the possible effect of intracellular calcium
fluctuations on the membrane potential (Maravall et al 2000,
Harks et al 2003). The obtained results suggest that Ca2+ -
dependent second messenger signaling does not significantly
affect the induced changes in the AP generation and plasma
membrane properties, implying the importance of direct
voltage-sensitive channel opening for mediating the MMW
effects.
In summary, the key finding of this study is the
demonstration that low levels of MMW power can induce
changes in neuronal firing and a profound reduction of Rn,
which in turn may lead to increased neuronal excitability
(Volgushev et al 2000a). In this study, neurons were effectively
inhibited at a depth of 2 mm under the aCSF, but it remains
to be seen whether the MMWs can produce similar changes
in neuronal activity when applied in the human skin and/or
central nervous system. More studies are needed to evaluate
the possibility of inducing long-term changes in the neuronal
excitability by MMWs in order to establish the feasibility of
using MMW for therapeutic neuromodulation.
Acknowledgments
The authors would like to acknowledge administrative
and technical support from Professor David B Rutledge,
California Institute of Technology, and financial support from
the Huntington Medical Research Institutes and the Chief
Scientist’s Office of the Jet Propulsion Laboratory. Helpful
suggestions from the anonymous reviewers are also gratefully
acknowledged.
References
Alekseev S I, Gordiienko O V, Radzievsky A A and Ziskin M C
2010 Millimeter wave effects on electrical responses of the
sural nerve in vivo Bioelectromagnetics 31 180–90
Beneduci A 2008 Review on the mechanisms of interaction between
millimeter waves and biological systems Bioelectrochemistry
Research Developments ed E M Bernstein (New York: Nova
Science Publishers) pp 35–80
Cametti C, De Luca F, Macri M A, Maraviglia B and Sorio P 1988
Audio to microwave frequency dielectric study of the
pretransition region in DPL-water systems Liq. Cryst. 3839–45
Cao X-J and Oertel D 2005 Temperature affects voltage-sensitive
conductances differentially in octopus cells of the mammalian
cochlear nucleus J. Neurophysiol. 94 821–32
Chou C-K and D’Andrea J (ed) 2006 IEEE Standard for Safety
Levels with Respect to Human Exposure to Radio Frequency
Electromagnetic Fields, 3 kHz to 300 GHz
Enders A and Nimtz G 1984 Dielectric relaxation study of dynamic
properties of hydrated phospholipid bilayers Ber. Bunsenges.
Phys. Chem. 88 512–7
Harks E G A, Torres J J, Cornelisse L N, Ypey D L and
Theuvenet A P R 2003 Ionic basis for excitability of normal rat
kidney (NRK) fibroblasts J. Cell. Physiol. 196 493–503
Hillert L, ˚
Akerstedt T, Lowden A, Wiholm C, Kuster N, Ebert S,
Boutry C, Moffat S D, Berg M and Arnetz B B 2008 The
effects of 884 MHz GSM wireless communication signals on
headache and other symptoms: an experimental provocation
study Bioelectromagnetics 29 185–96
Lawton G 2008 Wireless HD video heats up Computer 41 18–20
Lee J C F, Callaway J C and Foehring R C 2005 Effects of
temperature on calcium transients and Ca2+-dependent after
hyperpolarizations in neocortical pyramidal neurons
J. Neurophysiol. 93 2012–20
LeVine S 2009 The Active Denial System. A Revolutionary,
Non-lethal Weapon for Today’s Battlefield (Washington, DC:
National Defence University) pp 1–17
Liburdy R P and Magin R L 1985 Microwave-stimulated drug
release from liposomes Radiat. Res. 103 266–75
Maravall M, Mainen Z F, Sabatini B L and Svoboda K 2000
Estimating intracellular calcium concentrations and buffering
without wavelength ratioing Biophys. J. 78 2655–67
Pakhomov A G, Prol H K, Mathur S P, Akyel Y and Campbell C B
1997 Search for frequency-specific effects of millimeter-wave
radiation on isolated nerve function Bioelectromagnetics
18 324–34
Ramundo-Orlando A, Longo G, Cappelli M, Girasole M,
Tarricone L, Beneduci A and Massa R 2009 The response of
giant phospholipid vesicles to millimeter waves radiation
Biochem. Biophys. Acta 1788 1497–507
Trevelyan A J and Jack J 2002 Detailed passive cable models of
layer 2/3 pyramidal cells in rat visual cortex at different
temperatures J. Physiol. 539 623–36
8

J. Neural Eng. 7(2010) 045003 VPikovet al
Volgushev M, Vidyasagar T R, Chistiakova M and Eysel U T 2000a
Synaptic transmission in the neocortex during reversible
cooling Neuroscience 98 9–22
Volgushev M, Vidyasagar T R, Chistiakova M, Yousef T and
Eysel U T 2000b Membrane properties and spike generation in
rat visual cortical cells during reversible cooling J. Physiol.
522 Pt 1 59–76
Wells J, Kao C, Jansen E D, Konrad P and Mahadevan-Jansen A
2005a Application of infrared light for in vivo neural
stimulation J. Biomed. Opt. 10 64003
Wells J, Kao C, Mariappan K, Albea J, Jansen E D, Konrad P
and Mahadevan-Jansen A 2005b Optical stimulation of neural
tissue in vivo Opt. Lett. 30 504–6
Wiholm C, Lowden A, Kuster N, Hillert L, Arnetz B B, Akerstedt T
and Moffat S D 2009 Mobile phone exposure and spatial
memory Bioelectromagnetics 30 59–65
Zhadobov M, Sauleau R, Le Drean Y, Alekseev S I and Ziskin M C
2008 Numerical and experimental millimeter-wave dosimetry
for in vitro experiments IEEE Trans. Microwave. Theory Tech.
56 2998–3007
9