ID Number		150
Study Type		Human / Provocation
Model		900 MHz (GSM) exposure to humans and analysis of EEG, sleep, and hormone levels (while sleeping)
Details		Human subjects (n = 27) were exposed to 900 MHz (GSM) RF from a mobile phones transmitting at 8 watts at a distance of 40 cm away for 2 nights while they slept (11 pm to 7 am). Initial studies (Neuropsychobiology (1996) 33:41-7) using field strengths of ~0.5 mW/cm2 were reported to shorten sleep onset latency and reduce duration and percentage of REM sleep. In initial studies, subjects were observed to fall asleep about 3 minutes faster than sham exposed subjects (p<0.005) and had about 15 minutes less REM sleep (p<0.05). REM sleep latency was non-significantly increased. Total sleep times, sleep efficiency index, and the other conventional sleep parameters were not affected. The authors also initially reported qualitative alterations of the EEG alpha band signal (8-13 Hz) during REM sleep, with an increased spectral power density in all frequency bands (delta, theta, alpha1, alpha2, and beta), the most pronounced effects occurring in the alpha1 and alpha2 bands (p<0.05). Decreasing trends in these frequency bands were observed during the other sleep stages. Significant effects on duration and percentage of REM sleep, as well as the trend toward an increase in REM latency, were also reported in a subsequent study (Neuropsychobiology (1998) 38:251-256). In other follow up night-time sleep EEG studies (Neuropsychobiology (2000) 42:207-212; Bioelectromagnetics (1998) 19:199-202; Bioelectromagnetics (1997) 18: 2, 172-6), however, sleep latency and duration and percentage of REM sleep were not statistically affected, even at levels 100-250 times higher than used in previous studies. The authors suggest that the different characteristics of the RF exposure may have caused the seemingly discrepant findings. In awake EEG studies, volunteers (n=34) were exposed (or sham exposed) to a 900 MHz GSM signal again positioned 40 cm away beneath a bed the subjects were lying on and having an average power density of 0.05 mW/cm2. EEGs were recorded in 2 x 10 minute sessions while the subjects were awake, lying down with their eyes closed. No statistically significant effects on EEG spectral power densities were observed. Eventual sleep latency and length of REM sleep were not significantly affected (in contrast to initial reports), although there was a trend towards a decrease. Related studies (Neuroendocrinology (1998) 67: 2, 139-4) found no changes in the levels of growth hormone, luteinizing hormone, or melatonin in exposed subjects, although there was a transient increase in cortisol levels. Additional studies reported no change in heart rate or rhythm with RF exposure during sleep. In a review article on the subject (Sleep Medicine Reviews (2004) 8:95-107), the authors suggest that despite anecdotal reports from individuals, the epidemiologic literature to date provides no consistent evidence that sleep disturbances are a major concern with RF or mobile phone exposure, although lab studies show "some consistency regarding a slight sleep-promoting effect and an increase of the alpha power of the sleep EEG" based on a "very limited" number of experiments.
Findings		Effects
Status		Completed With Publication
Principal Investigator		University of Mainz, FRG
Funding Agency		Deutsche Telekom, Germany
Country		GERMANY
References		Mann, K et al. Sleep Medicine Reviews, (2004) 8:95-107 Wagner, P et al. Neuropsychobiology, (2000) 42:207-212 Mann, K et al. Neuropsychobiology, (1998) 38:251-256 Wagner, P et al. Bioelectromagnetics, (1998) 19:199-202 Mann, K et al. Neuroendocrinology, (1998) 67:139-144 Roschke, J et al. Bioelectromagnetics, (1997) 18:172-176 Mann, K et al. Neuropsychobiology, (1996) 33:41-47
Comments		Krause (NeuroReport 2000) also reported increases in alpha band from exposure) Exposures in the follow up studies were less than half of that used in the original studies (20 mW/cm2 as opposed to 50 mW/cm2) and utilized a different antenna design. These differences were offered as a possible explanation for the failure to replicate the statistically significant effect on sleep latency and length of REM sleep observed previously. There were also other inconsistencies between the original and follow up study designs such as the length of time required by exposed subjects to enter REM sleep. The review by Mann et al summarizes key sleep and EEG related studies by the authors plus that of Akerstedt et al, Graham and Cook, Achermann, Borbély et al., Wittersheim, Lebedeva, Reite et al., Ruhenstroth-Bauer et al., Pasche et al, etc. on many sleep-related parameter including heart rate, EEG, sleep quality, sleep architecture, latency, daytime performance and well-being, hormone levels, etc. The authors did make the point in the review that individuals suffering from these endpoints often try to attribute the cause to easily identifiable factors (i.e., mobile phones) and further that epidemiological evidence does not show strong evidence for any direct correlation. Other studies that have also found increased alpha signals: " Mann & Röschke (1996) 0.5 W/m² average power for 8-hrs = increased alpha 7.5  15 Hz (REM) " Mann et al., (1998) & Wagner et al., (1998) 0.2 W/m² average power for 8-hrs = no significant effects " Wagner et al., (2000) 50 W/m ² submaximal power for 8-hrs = No significant effects " Borbély et al., (1999) 1 W/Kg Peak SAR at 15 mins on/off for 8-hrs = increased alpha 11.5 - 12.25 Hz & 13.5 - 14 Hz " Huber et al., (2000) 1 W/Kg Peak SAR for 30 mins prior to daytime sleep = increased alpha 9.75  11.25 Hz and 12.25  13.25 Hz " Huber et al., (2002) 1 W/Kg Peak SAR for 30 mins prior sleep = increased alpha 12.25  13.5 Hz " Loughran et al., (2005) 2W, Peak SAR 0.6 W/Kg for 30 mins prior sleep = increased alpha 11.5  12.25 Hz