D. Moretti, A. Garenne, F. Poulletier de Gannes, E. Haro, I. Lagroye, P. Levêque, B. Veyret & N. Lewis

Effets de signaux RF GSM-1800 sur l’activité électrique de réseaux de neurones en culture

Pourquoi cette étude ?

• SNC : cible des RF

2

• effets rapportés des RF sur

EEG/sommeil chez l’homme

• approche complémentaire

in vitro

Multi Electrode Array (MEA)

3

60 électrodes

Neurones corticaux

d’embryons de rat

4

20 ms

80 μV

Spikes & bursts

Système d’exposition

GSM-1800 DAS = 3.2 W/kg

Incubateur sec 37°C , 5% CO2

RF Charge 50 Ω

TEM

filtre - ampli

PC

MEA E

758 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 59, NO. 3, MARCH 2011

Fig. 5. Diagram of the setup used for temperature measurements. The Luxtronprobe positioning within the MEA resin holder is also shown in the figure inset.

respectively). The measurements were carried out both on an

empty TEM cell and on a MEA equipped one.

2) Temperature Measurements and SAR Evaluation: The

measurements were performed using a nonperturbing thermo-

metric method [28] with a Luxtron One version 2.82 probe

(uncertainty 0.1 C). The TEM cell equipped with MEA

and the sample holder filled with 3-mL RPMI solution was

exposed at an input power of 6.6 W delivered by a Rohde

& Schwarz signal generator (10 MHz-20 GHz) at 1.8 GHz

connected to an RFPA 8502000-10 power amplifier for 60 s,

as described in [27]. The measurement protocol included two

further dwell times before and after the MW exposure to guar-

antee temperature stability at the beginning of exposure, and

to verify temperature decay once the MW power was turned

off. A Labview code enabled a PC with an RS-232 interface to

automatically acquire temperature. Power monitoring during

exposure was performed using an HP33120A power meter

fitted with HPE4412A probes. A 5-mm-diameter hole was

made on the top and central TEM cell plates to enable probe

insertion without modifying system performance. The mea-

surement setup and Luxtron probe positioning within the MEA

resin holder are schematically depicted in Fig. 5.

Five consecutive temperature acquisitions were performed

on a single point on the center of the chip (see Fig. 5) where

the recording electrodes are positioned and neuron activity

takes place. As this region is very small (dimensions reported

in Section II-B) compared with the probe measurement volume

(about 1 mm ), temperature recordings on multiple points were

not feasible. The temperature measurements during exposure

(60 s) were interpolated using a linear regression method. The

slope of the straight line represents the term at , as

reported in the well-known equation (1), necessary to calculate

the SAR

(1)

The term is the sample specific heat, equal to

0.95 cal kg C and 4186 is the specific heat scale

factor.

To determine the thermal behaviors of the target in actual bi-

ological experiments, a set of temperature measurements within

an incubator at 37 C, 95% humidity (Binder APT line C 150)

was carried out using the same measurements setup (Fig. 5).

Since the open TEM cell presents slight EM fringe effects [26],

Fig. 6. SAR distribution in the whole biological sample volume obtained withelectrodes (80 and 40 m) and without them (SAR visualized in the half ofstructure). (a) In the case of isolated electrodes. (b) In the case of the metal-connected electrodes.

foam absorbers were used to cover the metallic walls of the incu-

bator to avoid any interference during MW exposure. A different

protocol was used, which includes 4-h MW exposure, with two

2-h periods of sham exposure both before and after. An input

power of 2.3 W was used during MW exposure.

III. RESULTS

A. Numerical Dosimetry

SAR distribution at 1.8 GHz is illustrated in Fig. 6 for both

modeling strategies (i.e., isolated electrodes and metal-con-

nected electrodes, see Fig. 4). This was calculated within the

whole biological sample using 80- and 40- m recording elec-

trodes and also without electrodes (Fig. 4). The SAR statistical

analyses are reported in Table III. In all cases, the mean SAR

values (with a maximum variability of only 15%) did not

depend on the presence and size of electrodes (Table III). This

result is understandable, as the contact volume of the electrodes

with the biological solution is very small.

Furthermore, in simulations using isolated electrodes, SAR

inhomogeneity, quantified as the coefficient of variation ( ,

i.e., the percentage ratio between the standard deviation and the

mean value, , reached a max-

imum of 67.5% when the recording electrodes were completely

5 [Merla et al., 2010]

Système d’exposition : Réseaux de neurones

Acquisition et Traitement du Signal

Matériels et Méthodes

Résultats

Réduction de 30% du nombre de spikes sur

3 min pendant l’exposition (p expo= 0,0003)

• Amplitude de l’effet en fonction des paramètres de

l’exposition RF (puissance, durée, modulation,

etc.)

• Effets sur la plasticité (pharmacologique, électrique)

• Mécanismes d’interaction

Sig

na

l

t

4,6 ms

1/8 7/8

0,57 ms

Signal utilisé pour

l’exposition:

GSM 1800 MHz

MEA

amplificateur

50!

charge

RF

TEM

signaux

neuronaux

Perspectives

No

mb

re d

e s

pik

es s

ur

3 m

in.

0

2

4

6

8

10

12

15 15 15 15 15 18 18 18 18 18 19 21

EXPOSUREBEFORE

EXPO

AFTER

DIV

MF

R [

Hz]

0

2

4

6

8

10

12

14 14 14 14 14 17 17 17 17 17 18 20

SHAM EXPOSURE

DIV

MF

R [

Hz]

Exposition réelle

Exposition fictive

Jours de cultures

Jours de cultures

Cellule Transversale Electromagnétique (TEM)

Signal composé : activité spontanée + interférence GSM

Protocole

Signal composé après filtre fréquentiel

* * * * *

* * * * *

AVANT

EXPO

APRES

Interférence GSM

Système d’exposition : Réseaux de neurones

Acquisition et Traitement du Signal

Matériels et Méthodes

Résultats

Réduction de 30% du nombre de spikes sur

3 min pendant l’exposition (p expo= 0,0003)

• Amplitude de l’effet en fonction des paramètres de

l’exposition RF (puissance, durée, modulation,

etc.)

• Effets sur la plasticité (pharmacologique, électrique)

• Mécanismes d’interaction

Sig

na

l

t

4,6 ms

1/8 7/8

0,57 ms

Signal utilisé pour

l’exposition:

GSM 1800 MHz

MEA

amplificateur

50!

charge

RF

TEM

signaux

neuronaux

Perspectives

No

mb

re d

e s

pik

es

sur

3 m

in.

0

2

4

6

8

10

12

15 15 15 15 15 18 18 18 18 18 19 21

EXPOSUREBEFORE

EXPO

AFTER

DIV

MF

R [

Hz]

0

2

4

6

8

10

12

14 14 14 14 14 17 17 17 17 17 18 20

SHAM EXPOSURE

DIV

MF

R [

Hz]

Exposition réelle

Exposition fictive

Jours de cultures

Jours de cultures

Cellule Transversale Electromagnétique (TEM)

Signal composé : activité spontanée + interférence GSM

Protocole

Signal composé après filtre fréquentiel

* * * * *

* * * * *

AVANT

EXPO

APRES

Filtre

6

t

Signal GSM 1/8

Protocole d’exposition

PRE SHAM EXPO

POST

3 min 3 min t 3 min

PRE GSM EXPO

POST

t

1er jour = sham

3 min 3 min 3 min

2ème jour = expo

7

0

100

200

300

400

500

14 14 14 14 14 14 17 17 17 17 17 17 17 17 18 20

BR

[b

urs

ts/m

in]

Age [DIV]

SHAM PRE EXPO POST

8

Résultats : Bursting rate [BR]

0

100

200

300

400

500

15 15 15 15 15 15 18 18 18 18 18 18 18 18 19 21

BR

[b

urs

ts/m

in]

Age [DIV]

EXPO

0

100

200

300

400

500

14 14 14 14 14 14 17 17 17 17 17 17 17 17 18 20

BR

[b

urs

ts/m

in]

Age [DIV]

SHAM PRE EXPO POST

9

Résultats : Bursting rate [BR]

10

Analyse statistique “avant-après”

n=16 cultures

11

Analyse statistique “pendant l’exposition”

n=16 cultures

Analyse de sensibilité : effet = f (âge)

12

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

14 15 16 17 18 19 20 21 22

Age (DIV)

Réduction de BR (%)

n=16 cultures

ΔT (3 min) = 0,06 °C

Mécanisme de l’effet : échauffement ?

13

Constante de temps = 13 min

exposition

[Merla et al., 2010]

• Étude de faisabilité

Publiée : Moretti et al Bioelectromagnetics 2013

Conclusion

• Effets du GSM sur l’activité de neurones “jeunes”

• Effet réversible

15

Expérimentation en cours

– effet de la modulation des RF

(TDMA vs CW)

– niveau de DAS

– durée d’exposition

Perspectives

16

- Approche intégrative sur modèle animal

D!DE+($F(, ; $G$C#, (2, &$C&9#2#; #+$" :$7+3. &" ("8) $HCC7I$G$J +K, 0#* +&#$" :$6+; 0"932+&3+$, &5$F0, 2&$7+3. &" ("82+9$H6F7I$G$L&, M(2&8$7+3. &" ("82+9$:"0$6+; 0"932+&3+$G$NNNO22#O2#$G$, E+(OM(, ; P 22#O2#$

L(+3#0"5+9$"&$#. +$93, (K$0+3"05$3. , &8+9$2&$#. +$+(+3#023, ($:2+(5$5; +$#"$M0, 2&4N25+$, 3#2%2#) $"932((, #2"&9O$

_` RY$

a* 9b$

L(+3#0"+&3+K. , ("80, K. ) $HLL/ I$D&, () 929$" :$("3, ($+(+3#023, ($:2+(5$K"#+&#2, (9$

- Effets des RF sur la plasticité neuronale

Michela Chiappalone

17

Financement : Université Bordeaux 1, CNRS et Bouygues Telecom

Noëlle Lewis

Philippe Lévêque Emmanuelle Haro

Florence Poulletier de Gannes

Isabelle Lagroye

André Garenne

Daniela Moretti

bernard.veyret@ims-bordeaux.fr

18