Perentos N, Croft RJ, McKenzie RJ, Cvetkovic D, Cosic I. (2007). Comparison of the effects of continuous and pulsed mobile phone like RF exposure on the human EEG. Australas Phys Eng Sci Med. 30(4):274-280.
This study attempted to replicate results from an earlier study of Global System for Mobiles (GSM) phone radiation and how it affects resting brain function (Huber et al., 2002), and to extend those results to include more realistic and generalizable exposures. Earlier studies and replications often conflicted, possibly because of variable experimental conditions.
With no currently accepted theories or physical mechanisms to explain how such radiofrequency (RF) exposure would result in bioeffects, the authors contend that further empirical tests are required, especially for the particular modulation characteristics of GSM exposure. The present study explores one of the most consistent findings, that of increased resting electroencephalogram (EEG) alpha-band activity.
Twelve volunteers were included (six males/six females), from 19 to 32 years of age. Over a two-hour period, subjects were exposed to 15 minutes of (un-modulated) continuous wave RF exposure, 15 minutes of pulse-modulated RF exposure, and 15 minutes of sham-exposure. EEG recordings were by a standard fitted cap. Exposure was with a crude representation of a GSM monopole antenna device, which lacks thermal and auditory queues. It was placed in a standard “touch” ear to mouth position at the left hemisphere. Two signals were used: an un-modulated and pulse-modulated 900 MHz RF signal. Standard Fourier analyses were complemented with a non-linear Approximate Entropy method.
From the traditional Fourier power spectra analyses, statistically significant differences were observed in the delta band from looking at differences in the source of the exposure (sham, modulated, un-modulated). No interactions were significant. And the statistically significant result was not further supported upon exploration with post hoc tests.
From the Approximate Entropy method analyses, there were no statistically significant results. However, the new method indicated that 80% of cases contained statistically significant non-linear elements.
Discussion and Interpretation
After using a more realistic GSM exposure profile, the study failed to replicate earlier results for alpha-band activity in a resting wakeful state. Results were unclear concerning the advantages of investigating non-linear features of EEG data using the Approximate Entropy analysis. The authors identify three issues that may have prevented replication of the earlier Huber et al. 2003 results. One, the more realistic exposure in the present study would have resulted in lower deep tissue exposure. Two, the exposure duration was half that of the earlier study (which was 30 minutes), though the authors do note that positive findings have been observed with even shorter exposures. Three, there were differences in the aspects of the exposure modulation among earlier studies.