Schoeni A, Roser K, Röösli M. Memory performance, wireless communication and exposure to radiofrequency electromagnetic fields: A prospective cohort study in adolescents. Environ Int. 2015 Oct 13;85:343-351. doi: 10.1016/j.envint.2015.09.025.
Highlights
• This is a prospective cohort study with approx. one year of follow-up.
• Self-reported and operator recorded mobile phone use data were collected.
• The cumulative RF-EMF dose for the brain and for the whole body was calculated.
• Associations were stronger for RF-EMF dose than for use of wireless devices.
• RF-EMF exposure might impair memory performance in adolescents.
Abstract
BACKGROUND: The aim of this study is to investigate whether memory performance in adolescents is affected by radiofrequency electromagnetic fields (RF-EMF) from wireless device use or by the wireless device use itself due to non-radiation related factors in that context.
METHODS: We conducted a prospective cohort study with 439 adolescents. Verbal and figural memory tasks at baseline and after one year were completed using a standardized, computerized cognitive test battery. Use of wireless devices was inquired by questionnaire and operator recorded mobile phone use data was obtained for a subgroup of 234 adolescents. RF-EMF dose measures considering various factors affecting RF-EMF exposure were computed for the brain and the whole body. Data were analysed using a longitudinal approach, to investigate whether cumulative exposure over one year was related to changes in memory performance. All analyses were adjusted for relevant confounders.
RESULTS: The kappa coefficients between cumulative mobile phone call duration and RF-EMF brain and whole body dose were 0.62 and 0.67, respectively for the whole sample and 0.48 and 0.28, respectively for the sample with operator data. In linear exposure-response models an interquartile increase in cumulative operator recorded mobile phone call duration was associated with a decrease in figural memory performance score by -0.15 (95% CI: -0.33, 0.03) units. For cumulative RF-EMF brain and whole body dose corresponding decreases in figural memory scores were -0.26 (95% CI: -0.42, -0.10) and -0.40 (95% CI: -0.79, -0.01), respectively. No exposure-response associations were observed for sending text messages and duration of gaming, which produces tiny RF-EMF emissions.
CONCLUSIONS: A change in memory performance over one year was negatively associated with cumulative duration of wireless phone use and more strongly with RF-EMF dose. This may indicate that RF-EMF exposure affects memory performance.
Excerpts
… From a public health point of view potential effects of chronic exposure are more relevant, which needs to be investigated with epidemiological studies. So far there has only been one community-based epidemiological study investigating effects of mobile phone use on adolescents’ memory. Abramson et al. (2009) showed in a cross-sectional analysis of 317 seventh grade students from Australia that mobile phone use was associated with faster and less accurate response on a number of tasks involving the memory. Since similar associations were found in relation to the number of SMS (short text messages), which produces negligible RF-EMF exposure, they speculated that these behaviours may have been learned through the frequent use of a mobile phone and may not be the consequence of mobile phone radiation. In a follow-up investigation one year later, in 236 of these students, an increase in mobile phone use was associated with a reduction in response time in one out of three tests involving the memory (Thomas et al., 2010). This study relied on self-reported mobile phone use only, which has been shown to be inaccurate. Adolescents tend to substantially overestimate their amount of mobile phone use (Aydin et al., 2011; Inyang et al., 2009).
Regular mobile phone use may affect adolescents in various ways. Thus, the main challenge for research consists in differentiating between RF-EMF radiation effects and other non-RF-EMF related effects from mobile phone use. For instance, frequent texting or gaming on a mobile phone may facilitate cognitive processes (Abramson et al., 2009). It was also observed, that calling and sending texts during night was associated with poor perceived health symptoms such as tiredness, rapid exhaustibility, headache and physical ill-being (Schoeni et al., 2015; Van den Bulck, 2007). Other studies showed that frequent mobile phone use was associated with anxiety (Jenaro et al., 2007), unhealthy lifestyle (Ezoe et al., 2009), depression (Yen et al., 2009) and psychological distress (Beranuy et al., 2009). Thus, to address RF-EMF effects of wireless communication devices, the development of a RF-EMF dose measure, which incorporates all exposure relevant factors, is inevitable.
Memory performance was assessed with a standardized, computerized cognitive test battery (IST, Intelligenz-Struktur-Test 2000R (Liepmann et al., 2006)). Verbal and figural memory was measured with the subtest of the IST. In the verbal memory task, word groups have to be memorized in one minute time. After 1 min the study participants give an account of the word groups that have been memorized. In total 10 points can be achieved by remembering the correct word groups. In the figural memory task, pairwise symbols have to be memorized in one minute time. After 1 min one part of the pairwise symbols is shown and the matching part has to be found. A total of 13 points can be achieved. For both the verbal and figural tests, 2 min is given to complete the test. Memory performance is considered as the right number of remembered word groups or symbols, respectively. For the statistical analyses of verbal and figural memory the continuous test score values
were used as outcome. Every test was conducted once at baseline and once at follow-up investigation.
In this study we considered objectively recorded data on mobile phone use collected from the Swiss mobile phone operators as well as self-reported data on wireless communication devices usage obtained from a written questionnaire referring to the 6 months period prior to each examination. In terms of RF-EMF related exposure measures we inquired about call duration with own or any other mobile phone (referred to as duration mobile phone calls), call duration with cordless (fixed line) phone and duration of data traffic on the mobile phone, e. g. for surfing and streaming. The duration of gaming on computers and TV and number of all kind of text messages (SMS, WhatsApp etc.) are not, or only marginally relevant for RF-EMF exposure and were thus inquired to be used as negative exposure control variables in the analyses.
Informed consent to obtain objectively recorded mobile phone use data from the mobile phone operators was given by 234 out of 439 study participants and their parents. This included duration of each call and on which network (GSM or UMTS) it started, number of SMS (text messages) sent per day and amount of volume of data traffic (MB/day). Data were obtained for up to 18 months, 6 months before baseline until follow-up investigation.
A particular strength of this study is the longitudinal design. To the best of our knowledge this is the first longitudinal study on memory performance in adolescents using not only mobile phone call duration as an exposure proxy, but calculating RF-EMF dose measures derived from objectively recorded operator data and propagation modelling. Compared to a cross-sectional design where changes over time cannot be assessed andwhere reverse causality is of concern, longitudinal studies allow for more robust conclusions.
.. Most relevant contributors for the brain dose are calls on the GSM network (on average 93.3% for the whole sample based on self-reported data and 58.7% for the sample with operator data using operator recorded information) followed by calls with the cordless phones (4.2% and 21.0%, respectively). For the whole body dose, calls on the GSM network (on average 66.9% for the whole sample and 19.5% for the sample with operator data), the use of computer/laptop/tablet connected to WLAN (12.0% and 29.1%, respectively) and data traffic on mobile phones over WLAN (8.1% and 22.3%, respectively) counted for the most part. Less important for the dose measures were exposure from radio and TV broadcast transmitters (brain dose: 0.1% and 0.4%, respectively; whole body dose: 0.3% and 0.9%, respectively) and mobile phone base stations (brain dose: 0.6% and 3.5%, respectively; whole body dose: 2.0% and 4.8%, respectively).
… media usage measures which are not, or only marginally associated with RF-EMF were not associated with figural memory performance (e.g. sending text messages, playing games, and duration/volume of data traffic on the mobile phone). On the other hand, mobile and cordless phone use,which involves RF-EMF exposure, tended to be negatively correlated, although not statistically significant, whereas the dose measures were significantly correlated in many models. The relative high correlation between dose measures and self-reported and objectively recorded mobile phone call duration respectively, limits the possibility to disentangle effects due to RF-EMF exposure or due to other factors associated with mobile phone use.
Since we found stronger associations between RF-EMF doses and figural memory but not verbal memory, one could speculate that this might be due to different brain areas involved in the verbal and figural memory tasks. The type of information being processed determines the brain activity during encoding and retrieval and as a consequence brain activity patterns during figural memory tasks differ fromthose observed during verbal memory tasks. During figural memory processes, encoding elicits bilateral prefrontal activity and retrieval increases the activity in bilateral or right-sided temporal regions and in bilateral prefrontal regions (Beason-Held et al., 2005; Roland and Gulyas, 1995; Wagner et al., 1998). During verbal encoding increases in prefrontal and temporal brain activity in the left hemisphere can be seen (Heun et al., 2000; Iidaka et al., 2000; Reber et al., 2002; Strandberg et al 2011) and during verbal retrieval the activity in bilateral or rightsided prefrontal regions, bilateral or left-sided temporal regions and the anterior cingulate are increased (Beason-Held et al., 2005; Buckner et al., 1998; Cabeza et al., 1997). Stronger overall effects observed for figural memory processes predominantly involving the right hemisphere compared to the verbal memory tasks mostly involving the left hemisphere is compatible with the fact that 81.2% of the study participants reported at follow-up to mainly use mobile phones on the right side but only 18.8% on the left side or with no laterality preference. Strikingly, our laterality analyses indicated indeed stronger associations for right side users for the figural memory task whereas the reverse pattern was seen for the verbal task. However, the sample size of the laterality analysis was small for the subgroup with left side or no side preference for mobile phone use (n= 80).
We considered a number of potential confounders and adjusted model estimates were relatively similar to the crude model estimates, which indicates that confounding seems not to have a substantial impact on the results. Nevertheless, we cannot exclude that we have missed a relevant confounder ….
ConclusionThe observed striking pattern with more consistent associations for RF-EMF dose measures compared to usage measures and no indications of associations for negative control exposure variables may indicate that RF-EMF exposure affects the figural memory of adolescents. However, given the complex correlation structure for various exposure measures and the uncertainty in the RF-EMF dose calculation, the observed associations need to be interpreted with caution.