No Influence on Selected Parameters of HumanVisual Perception of 1970 MHz UMTS-Like Exposure

Transcription

1 Bioelectromagnetics 26:243^250 (2005) No Influence on Selected Parameters of HumanVisual Perception of 1970 MHz UMTS-Like Exposure Gernot Schmid, 1 * Cornelia Sauter, 2 Robert Stepansky, 2 Ingrid S. Lobentanz, 2 and Josef Zeitlhofer 2 1 ARC Seibersdorf research, Seibersdorf, Austria 2 Department of Neurology, University Clinic of Vienna,Vienna, Austria In recent years several studies regarding possible effects of radio frequency (RF) electromagnetic fields (EMFs) on cognitive brain function were reported. In many of these studies on awake humans the working tasks were presented visually to the test subjects, e.g., on a computer screen. Therefore, the question of where in the chain of visual perception, brain processing and response a possible effect could be induced seems to be of interest. In this study, possible effects of exposure to a generic 1.97 GHz UMTS-like signal on human visual perception were investigated in a double blinded, crossover study including 58 healthy volunteer subjects (29 male, 29 female), aged years (mean SD). Each test subject underwent a battery of four different clinical tests three times (two different exposure levels and sham exposure) to assess selected parameters of visual perception. The generic signals applied to the subjects head represented the RF emissions of an UMTS mobile phone under constant receiving conditions and the under condition of strongly varying transmit power, i.e., the signal envelope contained low frequency components. In the high exposure condition the resulting average exposure of the test subjects in the cortex of the left temporal lobe of the brain was 0.63 W/kg (1 g averaged SAR) and 0.37 W/kg (10 g averaged SAR). Low exposure condition was one tenth of high exposure and sham was at least 50 db (corresponding to a factor of ) below low exposure. Statistical evaluation of the obtained test results revealed no statistically significant differences in the investigated parameters of visual perception between the exposure conditions and sham exposure. Bioelectromagnetics 26: , ß 2005 Wiley-Liss, Inc. Key words: visual performance; cognition; radio frequencies; visual processing; human response INTRODUCTION In view of the extraordinary increasing number of mobile phones in combination with concerns regarding health effects caused by the electromagnetic fields (EMFs) emitted by these devices, several studies have been conducted in recent years in order to investigate possible effects of mobile phone radiation on brain function. However, no clear picture can be drawn from these findings so far. Several studies on awake humans have suggested that mobile phone radiation can affect evoked related potentials and task related EEG [Eulitz et al., 1998; Freude et al., 1998, 2000; Krause et al., 2000a,b]. Regarding cognitive functions and perceptual tasks Preece et al. [1999] reported a significant decrease of choice reaction time, but no differences in memory tasks during low frequency (LF) modulated 915 MHz exposure. Similar findings were reported by Koivisto et al. [2000a,b], who exposed subjects with a GSM phone (902 MHz, TX mode only, no DTX mode) either switched on or off in a crossover study. They also reported a speed up of response in simple reaction time and vigilance tasks [Koivisto et al., 2000a] and also a speed-up in response time at high load memory tasks [Koivisto et al., 2000b]. Lee et al. [2001] compared a group of users of GSM phones with a group of non-users with respect to three measures of attention and found a facilitated performance in the group of GSM phone users in one out of the three measures. Further indication for a facilitating effect of GSM radiation was Grant sponsor: Forschungsgemeinschaft Funk e.v., Rathausgasse 11a, Bonn, Germany. *Correspondence to: Gernot Schmid, ARC Seibersdorf research, A-2444 Seibersdorf, Austria. gernot.schmid@arcs.ac.at Received for review 31 March 2004; Final revision received 11 August 2004 DOI /bem Published online in Wiley InterScience ( ß 2005 Wiley-Liss, Inc.

2 244 Schmid et al. provided by the work of Edelstyn and Oldershaw [2002] who reported better performance of subjects with respect to capacity and processing speed after exposure by a 900 MHz GSM phone. In a recent study, a sex dependent effect was reported by Smynthe and Costall [2003]. They found an improved performance in memory recall tasks in male, but not in female subjects when they were exposed to a 1800 MHz GSM phone during the acquisition (learning) and the distraction phase (between acquisition and first recall phase). Subjects were not exposed during the recall phases. Finally, the Finnish group of Koivisto added a further piece to the puzzle which underlines the difficulties of getting a clear picture from the current body of study results. In their recent work Haarala et al. [2003] attempted to replicate their earlier findings reported in Koivisto et al. [2000a]. They used the same GSM mobile phone (902 MHz, TX mode only) and exactly the same cognitive tests as in the original study; however, they substantially improved the methodology, e.g., by introducing a double blind design, using a larger sample size, and performing multi center testing in which exactly the same study was carried out in a Finnish and a Swedish laboratory. The outcome of this study showed that the results of the original study could not be replicated, i.e., no significant effects of 902 MHz GSM mobile phone radiation on the investigated reaction time tasks could be found. All the above mentioned studies focused on the EMFs of mobile phones of the GSM system, which belongs to the second generation of mobile communication systems. Regarding EMFs emitted by the third generation systems, for example the UMTS system which is currently introduced in many countries in Europe, to our knowledge the only existing study was published by Zwamborn et al. [2003], which became known as the TNO-study. In this work two groups of test subjects, one with subjective complaints and one without, were exposed to base station-like signals of the GSM900, GSM1800, and UMTS system. Interestingly, beside some other effects, the authors reported a significant effect of the UMTS fields on visual selective attention in both groups of subjects. We report results from a study investigating possible influences of the EMFs of the UMTS system on selected parameters of human visual perception. MATERIALS AND METHODS Test Subjects In total 58 healthy volunteers (29 female, 29 male), age between 20 and 40 years [ years, mean standard deviation (SD)], participated in the study. All subjects were screened for medical illness and underwent ophthalmologic examination. Only ophthalmologically normal subjects (no abnormality detected in visual acuity, binocular vision, color perception, visual field examination, slit lamp examination, Amsler grid, fixation disparity) were selected; and all of them were regularly users of mobile phones with an average use per day between 2 and 90 min (mean 23.8 min), according to their own estimates as assessed by a questionnaire prior to the experiments. The education level of the selected test subjects was at least the level of qualification for university entrance. The whole study protocol was approved by the local ethics committee for research. All subjects gave their written informed consent to the test procedure and the applied exposure levels. Visual Perception Tests For assessing selected parameters of visual perception a battery of four different computer-based tests, well established in clinical diagnostics, were used [Schuhfried, 2004]. The Critical Flicker and Fusion Frequency Test, the Visual Pursuit Test, and the Tachistoscopic Traffic Test Mannheim belong to the Vienna Test System (Dr. Gernot Schuhfried GmbH, Mödling, Austria), which is well established in psychological diagnostics. The CompuVist system (Rodenstock Instruments GmbH, Ottobrunn-Riemerling, Germany), assessing the Contrast Sensitivity Threshold of the subjects, was used as the fourth test. During the Critical Flicker and Fusion Frequency Test, which measure the visual discrimination ability and the activation level of the central nervous system, the subject has to look into a tube in which a blinking light is presented. In ascending mode the blinking frequency of the light is increased stepwise starting at 25 Hz up to a maximum of 70 Hz, and the subject has to respond by a button as soon as the blinking light appears to him or her as a continuous light. The blinking frequency corresponding to the moment of the subject s response is automatically recorded as the critical fusion frequency. In descending mode the blinking frequency is decreased stepwise, starting at a blinking frequency of 70 Hz, which certainly appears as a continuous light, and the test subject has to respond at the moment of perceiving the light as not continuous anymore, the critical flicker frequency. In order to eliminate artifacts, the procedure is repeated four times and mean values of the critical fusion and flicker frequency are automatically calculated and recorded by the test system. The Visual Pursuit Test is used for the registration of concentrated targeted perception and selective attention in the visual area. It is designed as a line

3 Visual Perception During UMTS Exposure 245 tracking test and consists of 80 different items. In each item a tangle of nine different lines leading to nine different endpoints is presented to the subjects on a computer screen. The starting point of one out of the nine lines is marked and the subjects have to follow this line with their eyes to find the corresponding endpoint and to respond accordingly on a special keyboard (a component of the Vienna Test System). The performance of the subject is scored automatically, considering the number of correct answers, median time for correct answers and total working time. Test duration for each subject was approximately 5 10 min per run. The Tachistoscopic Traffic Test Mannheim is generally used to assess the optical perception performance and perceptive speed in traffic psychology. In each of the 20 different items, consecutively processed by the subjects, a picture of a typical traffic scene is presented to the subjects on a computer screen for only 1 s. Immediately after the presentation of the picture the subject has to specify whether he or she had seen pedestrians, motor vehicles, cyclists, motorcyclists, traffic signs or traffic lights. Subjects again respond via buttons on a keyboard and the number of correct answers is recorded automatically by the test system. During the Contrast Sensitivity Threshold of the test subjects visual system, the subjects repeatedly underwent the task to detect one out of five available symbols, randomly presented, appearing with stepwise increasing contrast from a dark grey background on the CompuVist screen. Twenty different contrast levels are available. Upon detecting the presented symbol the subject has to respond by pressing the corresponding button on the CompuVist keyboard. The contrast level at which the test subject gives a correct response was rated. Exposure System The developed exposure system allows double blind, computer controlled exposure of the test subjects head at three different treatment levels, a High exposure level, a Low exposure level, and Sham exposure. Figure 1 shows a schematic overview of the exposure system. During all experiments the test subjects were sitting inside a laterally shielded exposure cabin with the inner surface covered by radio frequency (RF) absorbing material (pyramidal absorbers, type Ecosorb VHP-12-NRL, Emerson & Cuming, Westerlo, Belgium). The RF power was applied to the test subject s head by a special head set consisting of a modified stereo head phone (Model K44, AKG acoustics, Munich, Germany) from which all metallic parts were removed and small matched helical antennas were built in (see Fig. 2). The antennas were mounted in the Fig. 1. Schematic overview of the exposure system (frontal wall removed). headset in a way to achieve an exposure distribution in the head similar to the case of mobile phone usage. Switching of the different exposure levels was accomplished by the control software, which served also for continuously recording the delivered RF power every 30 s (forward as well as reverse power measurements). As signal source a generic UMTS Signal generator (Model GUS6960S, University of Wuppertal, Fig. 2. Headset wornbya subject.theantennasareprotectedbya cap oflowlossmaterial (white).

4 246 Schmid et al. Wuppertal, Germany) was used, providing a generic UMTS test signal developed for use in bio-electromagnetic experiments [Ndoumbè Mbonjo Mbonjo et al., 2004]. This test signal represents a wideband code division multiple access (WCDMA) signal at a carrier frequency of 1.97 GHz, which is close to the frequency band allocated to the uplink of UMTS in frequency division duplex (FDD) mode. The signal composition is based on the assumption of transmitting six data channels (at 960 kbps) and one control channel (at 15 kbps). Coding for scrambling and spreading, modulation and pulse shaping are implemented according to the UMTS specification leading to the typical 5 MHz wide UMTS-FDD signal. In order to simulate low frequency variations in the envelope of the RF signal as expected in practical systems due to fading phenomena in combination with the extremely efficient power control algorithm implemented in UMTS, the applied test signal provides intervals of constant output power as well as intervals of strong variations of the signal envelope (30 db variation of output power). The corresponding power control algorithm follows also the UMTS specifications, i.e., the output power level is updated every 0.67 ms (maximum power step size 3 db), giving the typical 1500 Hz component in the spectrum of the signal envelope (beside a 8 12 Hz component due to the simulated fading, see Fig. 3). More details about the applied signal can be found in Ndoumbè Mbonjo Mbonjo et al. [2004]. A detailed dosimetric analysis of the exposure system was based on the power recordings obtained during the experiments and extensive numerical simulations using the SEMCAD simulation platform (Version 1.6 b107, Schmid & Partner Engineering AG, Zurich, Switzerland). Using the anatomical European male head model (Schmid & Partner Engineering AG) at a spatial resolution of mm yielded a mean 1 g averaged SAR in the brain cortex of the temporal lobe of 0.63 W/kg (minimum 0.25 W/kg, maximum 1.49 W/kg) in the High exposure condition, corresponding to 0.37 W/kg (minimum 0.16 W/kg, maximum 0.84 W/kg) in terms of 10 g averaged SAR (Figs. 4 and 5). Low exposure condition was defined to be one tenth of High exposure and Sham was at least 50 db below Low exposure. The resulting value of SAR in the High exposure condition indicated above is in the same order of magnitude as can be expected from UMTS mobile phones. The technical specifications of the (so far still few) commercially available UMTS phones indicate 10 g averaged SAR values between <0.1 W/kg and 0.98 W/kg, based on specifications of models Siemens U10, Siemens U15, Nokia 7600, Samsung SGH-Z105, Sony Ericsson Z1010, and Motorola A835. However, it must be noted that the SAR values given in Fig. 3. Envelope ofthe applied1.97 GHz RF signal (see also Ndoumbe Mbonjo Mbonjo et al. [2004]).

5 Visual Perception During UMTS Exposure 247 Fig. 4. Qualitative spatial distribution of SAR (peak SAR) in the numerical anatomical head model in a plane through the antenna feed point. Calculation of the mean exposure of the subjects was calculated to be 0.63 W/kg (maximum value of1g averaged SAR in the brain cortex at the temporallobe). the technical specifications of the mobile phones are measured during compliance testing procedures in homogeneous phantoms at maximum output power level and therefore give a conservative estimate of the maximum possible SAR in the human head. The above minimum and maximum values of SAR to be expected in the test subjects temporal cortex were obtained by an uncertainty analysis, considering worst case combinations of the antenna s position with respect to the head, different head size (scaling the model by 15%) and 20% variations in the dielectric properties of head tissues. Values according to Gabriel et al. [1996] were taken as nominal values for conductivity and permittivity of head tissues. The differences in the antenna s position with respect to the head were taken into account in the uncertainty analysis based on four geometrical parameters: three angles defining the direction of the antenna axis and the distance between antenna feed point and the surface of the head (Fig. 6). Representative values of these parameters were obtained on a sample of 20 randomly selected subjects (10 male, 10 female) and are listed in Table 1. The main contribution (85%) to the total exposure uncertainty was due to the variations of the antenna s position with respect to the head, caused by different head sizes and shapes. Variations in dielectric properties and variations in antenna feed point power contribute comparatively less to the total exposure uncertainty. The validity of the numerical model of the small helical antennas used was verified prior to the dosimetric analysis by comparisons of experimental and numerical results. The horizontal antenna pattern was measured in an anechoic chamber (ARC Seibersdorf Research, Seibersdorf, Austria) and calculated using the SEMCAD simulation platform, showing deviations of less than 5%. Furthermore the SAR inside the homogeneous specific anthropomorphic mannequin (SAM) phantom, which is derived from the 90th percentile male head data reported by Gordon et al. [1988], was measured using an automated SAR test facility (DASY 3, Schmid & Partner Engineering AG). The measurement results were compared to corresponding SEMCAD calculations (using the numerical SAM phantom) and showed agreement within 10% with respect to the maximum 1 g averaged SAR. Fig. 5. Spatialdistributionof SAR (peak SAR) onthe surface ofthe left sided brain cortex. The lower portion of the left temporal lobe wasthe most exposed part ofthe brain. Fig. 6. Numerical head model with head set and definition of parameters for assessment of uncertainty due to variations of antenna positionwith respect to thehead.

6 248 Schmid et al. TABLE 1. Variations of Uncertainty Parameters Obtained on 20 Adults (Compare Fig. 6) a b g d Median mm Range þ þ mm a, measured against a vertical axis through the center of the earpiece. b, measured against a plane through the center of the earpiece and parallel to the saggital plane of the body (negative sign means inclined towards the head). g, measured against a plane through the center of the earpiece and parallel to the saggital plane of the body (negative sign means inclined towards the head). d, distance between head surface and surface of protecting cap of antenna. Background ambient field strength at the test subjects workplace was measured in the frequency range from DC up to 18 GHz and was found to be negligible, i.e., only the typical electromagnetic environment caused by computer screens was found (maximum values, e.g., 50 Hz: 3 V/m, 35 nt; 72 Hz: 1 V/m, 240 nt; 92.5 khz: 100 nt). Experimental Procedure Each of the 58 subjects was tested on a single week day between 3:30 and 7:00 p.m. Within this time period each subject underwent the four tests used for assessing visual perception (see above) under three different exposure conditions, High exposure, Low exposure, and Sham exposure, i.e., each subject serves as his or her own control. Each subject underwent the different clinical tests in the same order; application of exposure conditions, however, was controlled randomly and double blinded by the control software. In order to minimize learning effects due to running the same test twice successively, the three different test runs of the four different clinical test were nested (i.e., first run of test 1- first run of test 2- first run of test 3- first of test 4- second run of test 1- second run of test 2-...). By definition of the study design subjects were exposed only at the left side of the head, but they were not informed about this fact (they were informed only that exposure can occur at the left, the right, or on both sides). Additionally, all tests except the Critical Flicker and Fusion Frequency Test were presented to the subjects under a positive control condition (flicker light) at the end of the subject s session. The flicker light was applied in the exposure cabin from the back in order to avoid blinding effects to the subjects. Statistical Evaluation For all tested parameters of visual performance, statistical analyses with respect to the different exposure conditions were performed. Furthermore the obtained test results were analyzed with respect to possible learning effects due to repetition of the tests by the subjects. In case of homogeneity of variances and normal distribution, analysis of variances were calculated. Friedman tests were used for all other data. Type I error level was globally set to a ¼ Because of multiple testing (k ¼ 12) Bonferroni adjustment was applied yielding a i ¼ RESULTS Table 1 shows mean, standard deviation, and range of all parameters obtained from the subjects under the three different exposure conditions and the positive control condition, except for the Critical Flicker and Fusion Frequency Test, which did not allow application of the positive control. For data showing non-homogeneity of variances and normal distributions only the median and range is given. Results of statistical analyses are indicated by corresponding F and P values in cases where analysis of variances could be calculated and by w 2 and P values in cases where the Friedman test was used for statistical analysis. As indicated by Table 2 none of the investigated parameters of visual perception showed a statistically significant difference between sham and the two different levels of RF exposure. Including also the positive control condition (application of flicker light into the exposure cabin) in the analysis, a statistically significant difference became apparent in the number of correct answers in case of the Tachistoscopic Traffic Test Mannheim and the Contrast Sensitivity Threshold levels, showing these parameters to be sensitive to the positive control stimulus. Results of the Visual Pursuit Test and the number of incorrect answers in the Tachistoscopic Traffic Test Mannheim did not show this kind of sensitivity. Analysis of the dataset with respect to learning effects showed a statistically significant effect as a consequence of repeated measurements in case of the Tachistoscopic Traffic Test. This effect, reflected by an increase in correct answers and a decrease of incorrect answers, might have some effect on the results regarding this test. All other tests did not show significant effects of learning. DISCUSSION This double blinded crossover study is one of the first studies investigating a possible influence of the EMFs radiated by third generation mobile phones on human cognitive functions. The focus within this work was put on visual perception, which was assessed by a test battery of four well established clinical tests, processed by each subject under two different levels of

7 Visual Perception During UMTS Exposure 249 TABLE 2. Parameters of Visual Perception and Corresponding Statistical Evaluation With Respect to the Different Exposure Conditions and the Positive Control Condition (n ¼ 58) Condition Mean SD Range F-value P * Flicker and Fusion Fusion frequency (Hz) High n.s. Frequency Test Low Sham Flicker frequency (Hz) High n.s. Low Sham Visual Pursuit Test No. of correct answers High n.s. Low Sham Pos ctrl n.s. Mean reaction time High n.s. of correct answers (s) Low Sham Pos ctrl n.s. Score High n.s Low Sham Pos ctrl n.s. Tachistoscopic Traffic No. of correct answers High n.s. Test Mannheim Low Sham Pos ctrl Median w <0.001 sig No. of incorrect answers High n.s. Low Sham Pos ctrl Median w n.s. Contrast Sensitivity Contrast sensitivity High Median w n.s. Test threshold level Low Median Sham Median Pos ctrl Median w <0.001 sig. *n.s., not significant; sig., significant. exposure and sham exposure. We did not find any statistically significant differences in Critical Flicker and Fusion Frequency, Threshold of Contrast Sensitivity, and the number of correct answers in the Tachistoscopic Traffic Test (as a measure of optical perception performance and speed) between the two exposure conditions and sham exposure. Furthermore, the number of correct answers and the response time in the Visual Pursuit Test (aiming at visual orientation performance and selective attention in the visual area) was not altered by the exposure to the applied EMF. Comparing our results with other studies on cognitive effects possibly caused by exposure from mobile phones is difficult, because most of the studies carried out so far had their focus on the GSM system, which differs substantially from the UMTS. As a consequence this study s signal envelope in the time domain, widely considered as important, differs completely from the GSM signals applied in other studies. Furthermore, the resulting exposure of the brain, e.g., in terms of SAR, is mostly, if completely, defined relatively poorly in earlier studies. Another, probably much more important reason for difficulties in comparing earlier studies with our work is the difference regarding the investigated endpoints and applied test systems. However, in a global comparison, not taking into account the different test sets used and the completely different characteristic of the applied EMFs, our findings can be seen in contrast to earlier studies which reported decreased response times [Preece et al., 1999; Koivisto et al., 2000a,b] or facilitated performance of several cognitive measures [Lee et al., 2001; Edelstyn and Oldershaw, 2002; Smynthe and Costall, 2003] during GSM exposure. On the other hand, our findings are well in line with the replication study published by Haarala et al. [2003] which, in contrast to the original study [Koivisto et al., 2000a], also did not find any statistically significant changes of the investigated cognitive parameters due to exposure from a GSM mobile phone. The only study allowing a more detailed comparison is the work published by Zwamborn et al. [2003]. Beside investigations using GSM900 and GSM1800

8 250 Schmid et al. signals for the exposure of the test subjects they also used an UMTS-like signal, which however differs in frequency and shape of the time domain envelope from the signal we applied in our work. Interestingly, similar to our study, Zwamborn et al. [2003] also considered selective visual attention as one of their endpoints, and they reported statistically significant effects on this parameter during UMTS exposure, in contrast to our findings. This contrast is enhanced by the fact that the considered parameter of selective visual attention was the only parameter which shows up to be affected in both groups of their subjects, those with subjective complaints and those without them. Furthermore the mentioned contrast becomes even more interesting when considering that Zwamborn et al. [2003] exposed their subjects at very low field levels, using base station antennas, corresponding to maximum local 10 g averaged SAR values in the range of mw/kg. So far we do not have an explanation for the differences in the comparable results from Zwamborn et al. [2003] and our study. However, beside the already mentioned differences regarding the exposure signal it should also be noted that Zwamborn et al. [2003] not only exposed their subjects, in the case of UMTS exposure, laterally, as we did, but also from the front. Unfortunately in the statistical analysis of Zwamborn et al. [2003], these two situations are not treated separately, so that we were not able to clarify if this could serve as a an explanation. CONCLUSION Our results indicate that EMF exposure as it can be expected from mobile phones of the third generation mobile communication systems in WCDMA-FDD mode has no significant effect on the investigated parameters of human visual perception. However, because the presented results are in contrast to findings of other authors, further research, including also other cognitive functions, is desirable. Furthermore it must be pointed out that this study only considered acute effects during exposure. Conclusions about possible effects after exposure and long term effects can not be drawn from this work. ACKNOWLEDGMENTS The authors thank Forschungsgemeinschaft Funk e.v., Bonn, Germany for funding and all test subjects for participating in this study. REFERENCES Edelstyn N, Oldershaw A The acute effects of exposure to the electromagnetic field emitted by mobile phones on human attention. Neuroreport 13: Eulitz C, Ullsperger P, Freude G, Elbert T Mobile phones modulate response pattern of human brain activity. Neuroreport 9: Freude G, Ullsperger P, Eggert S, Ruppe I Effects of microwaves emitted by cellular phones on human slow brain potentials. Bioelectromagnetics 19: Freude G, Ullsperger P, Eggert S, Ruppe I Microwaves emitted by cellular telephones affect human slow brain potentials. Eur J Appl Physiol 81: Gabriel S, Lau RW, Gabriel C The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues. Phys Med Biol 41: Gordon CC, Churchill T, Clauser CE, Bradtmiller B, McConville JT, Tebbets I, Walker RA Antropomorphic survey of US Army personnel: Methods and summary statistics. Technical Report NATICK/TR-89/044. US Army Natick Research, Development and Engineering center, Natick, Massachusetts, Sept Haarala C, Bjönberg L, Ek M, Laine M, Revonsuo A, Koivisto M, Hämäläinen H Effect of a 902 MHz electromagnetic field emitted by mobile phones on human cognitive function: A replication study. Bioelectromagnetics 24: Koivisto M, Revonsuo A, Krause C, Haarala C, Sillanmäki L, Laine M, Hämäläinen H. 2000a. Effects of a 902 MHz electromagnetic field emitted by cellular telephones on response time in humans. Neuroreport 11: Koivisto M, Krause CM, Revonsuo A, Laine M, Hämäläinen H. 2000b. The effects of electromagnetic field emitted by GSM phones on working memory. Neuroreport 11: Krause CM, Sillanmäki L, Koivisto M, Haggquist A, Saarela C, Revonsuo A, Laine M, Hämäläinen H. 2000a. Effects of electromagnetic field emitted by cellular phones on the EEG during a memory task. Neuroreport 11: Krause CM, Sillanmäki L, Koivisto M, Haggquist A, Saarela C, Revonsuo A, Laine M, Hämäläinen H. 2000b. Effects of electromagnetic field emitted by cellular phones on the electroencephalogram during a visual working memory task. Int J Radiat Biol 76: Lee TMC, Ho SMY, Tsang SYC, Yang SYC, Li LSW, Chan CCH Effect on human attention of exposure to the electromagnetic field emitted by mobile phones. Neuroreport 12: Ndoumbè Mbonjo Mbonjo H, Streckert J, Bitz A, Hansen V, Glasmachers A, Gencol S, Rozic D A generic UMTS test signal for RF bio-electromagnetic studies. Bioelectromagnetics 25: Preece AW, Iwi G, Davies-Smith A, Wesnes K, Butler S, Lim E, Varey A Effect of a 915-MHz simulated mobile phone signal on cognitive function in man. Int J Radiat Biol 75: Schuhfried G Mödling (A): Schuhfried GmbH. Accessible via Internet at: Smynthe JW, Costall B Mobile phone use facilitates memory in male, but not female subjects. Neuroreport 14: Zwamborn APM, Vossen SHJA, van Leersum B, Ouwens MA, Mäkel W Effects of global communication system radio-frequency fields on well being and cognitive functions of human subjects with and without subjective complaints. The Hague: TNO Physics and Electronics Laboratory; FEL- 03-C148.