ID Number		1145
Study Type		Human / Provocation
Model		900 MHz (GSM) exposure (including isolated 2, 8, 217 Hz, harmonic components) to human volunteers and analysis of EEG
Details		Human subjects (n=10) exposed to 900 MHz (GSM) RF using a phone set a full power (250 mW max avg) and measured for awake EEG recordings (with eyes closed). A phone used in standby mode was used as a control. Statistically significant changes (p < 0.05) between sham and exposed were observed for both the alpha and beta bands in the central region and in the beta band in the occipital region. No effects in other bands or with the phone in standby mode. The study also reported a general decrease in EEG in EEG power over most of the occipital EEG rhythms with exposure, with the decrease statistically significant in five of the twenty-one rhythms examined. The authors conclude mobile phone exposure may influence human brain activity, although they emphasize that this should be considered only a pilot study. In a subsequent study (attempting to replicate the results of Acherman and Huber, 2003), participants were exposed to a 900 MHz RF signal either with GSM modulation in DTx mode, unmodulated (CW), or with the ELF component from battery switching using a custom designed handset for exposure (designed to replicate the modulations of real GSM exposures in frequency content, SAR level and distribution, and current density level and distribution). Exposures were for 3 periods (for each signal type) with each period consisting of 4 x 7.5 minute exposures / sham exposures separated by a 1 minute pause. The authors report artifacts due to coupling of RF on the EEG leads during exposure, so only pre- and post-exposure data were analyzed. No changes in alpha band power due to RF exposure were detected, and thus the findings of Huber et al were not replicated. In a study to optimize their exposure, the authors characterized the ELF emissions from several mobile phones operating in both talk and DTx mode (generated by the battery, oscillators, and other internal circuitry) using two different single axis shielded loop probes. The first had a 34 mm diameter with good spatial resolution to detect the main 217 Hz spectral component. The second loop probe had a 127 mm diameter and a corresponding flat (+/- 1dB) broadband response over a wider range from 20 Hz to 70 kHz. These fast response probes allowed the relative amplitude, phase, and time domain characteristics of the pulsed magnetic field to be captured. A detailed calibration process and theory for the two loops was provided correlating the captured spectral components with the output voltage in the probe instrumentation chain. They observed a maximum magnetic field at the front and back surfaces of the phones they evaluated of 22.4 and 44.4 microTesla, respectively. Spectral components at 2.1, 8.3, 217 Hz and other harmonic products were detected. These findings were in general agreement with a previous characterization by Tuou and Kuster et al given at the 2005 joint BEMS / EBEA meeting in Dublin Ireland. They then replicated these ELF components by using a small loop antenna placed 1 cm underneath the front fact of their exposure handset with current sufficient to generate a peak field of 25 microTesla at the phone surface. The system allows the ability to turn on and off the ELF and RF components independently to evaluate their respective contribution to any observed effects. In an initial study using this modified mock handset exposure system, EEG was recorded in human volunteers (n = 72) over four 30-minute test periods. During the first and last 5 minutes of each EEG recording period, the subjects were not exposed. During the middle 20 minute of the EEG recording period, subjects were either sham exposed or exposed to 900 MHz (CW), 900 MHz (GSM), or ELF signals. Each of the four 30-minute testing periods was separated by a 4 minute break. The exposure handset was positioned at the right side of the head while a non-radiating handset was placed on the left side to avoid lateralization of participants attention. The order of exposures was randomly assigned using a counter balanced, double blind cross-over design. In the current paper the authors only report data for the pure ELF from a simulated DTX frame structure including 2, 8, 217 Hz components and their harmonics. Using numerical simulations the authors demonstrated peak currents generated within participants heads did not exceed 100 ¼A/m2, or 1% of the ICNIRP exposure limit. RESULTS: The authors report that ELF exposure did not change overall EEG alpha band activity as compared to pre-exposure levels. However, ELF exposure did appear to & suppress the tendency of alpha activity to increase throughout the half hour test interval. Specifically, alpha band recordings on the ipsilateral side of exposure were lower (i.e., did not increase to the same extent) than during the sham exposure (p = 0.022). There was no ipsilateral difference observed after ELF exposure, and no contralateral differences. The authors indicate that their observations are supported by previous reports using ELF exposures (Cook et al., Bioelectromagnetics 2005, 26:367-76) and conclude A decrease in the alpha frequency band was observed during ELF exposure in the exposed hemisphere only. This result suggests that ELF fields as emitted from GSM handsets during DTX mode may have an affect on the resting alpha band of the human EEG.
Findings		Effects
Status		Completed With Publication
Principal Investigator		RMIT University, Melbourne Australia - irena.cosic@rmit.edu.au
Funding Agency		ACEBR, Australia
Country		AUSTRALIA
References		, , () :- Perentos, N et al. 30th Annual International IEEE EMBS (Eng Med Biol Soc) Conference, Vancouver British Columbia, (2008) August 20-24:5680-5683 Perentos, N et al. Australasian Phys Eng Sci Med. , (2008) 31:235-242 Perentos, N et al. Australas Phys Eng Sci Med, (2007) 30:274-280 D'Costa, H et al. Australian Physical and Engineering Sciences in Medicine, (2003) 26:162-167
Comments		Cosic et al (2007) did NOT replicate Huber et al (2003) - using a different exposure system there were no alpha band changes due to RF exposure. The authors suggest a) different SAR distributions, b) total exposure duration was 15 minutes vs Huber et al 30 minutes, and c) this exposure differed from COsic et al earlier exposures (effects in alpha bands) by also including the 2 Hz and 8 Hz components of DTx (prior work only included 217 Hz and 1736 Hz components)