Slashdot Mirror
Research Finds Effects of GSM Signals on Sleep
An anonymous reader writes "The effects of mobile phone radiation on sleep were studied in Sweden in a laboratory experiment where subjects were exposed either to 884 MHz GSM radiation or placebo.
The study finds that compared to placebo, in the radiation-exposed subjects there was a prolonged latency to reach the first cycle of deep sleep (stage 3). The amount of stage 4 sleep was also decreased. Moreover, participants that otherwise have no self-reported symptoms related to mobile phone use, appear to have more headaches during actual radiofrequency exposure as compared to sham exposure."
Well to an extent I did. I've been out in the middle of nowhere to the extent that you couldn't get a cell phone signal if your life depended on it (and sometimes it does!) and there is an odd sense of quiet.
I know it sounds nuts but on a windy night even with the trees moving it still seems more quiet but in an almost impossible to define way. Like there is something that you can't put your finger on NOT there.
I always thought it might be either radio singles or high pitch EM radiation from all the fun toys I have around it (yes, including a Wireless Router). So I'm not complaining, and I can sleep fine, but at the same time this study doesn't shock me at all.
i have occasionally slept with my phone under my pillow to use as an alarm. i noticed that when i did so, i did NOT sleep as well. I also tended to wake up with a headache. no phone = better sleep overall, phone = less restful sleep. it's pretty consistent, and i could duplicate these results any time by sleeping with my phone. take it for what you will, YMMV.
Uh, no. I don't see the words "double blind".
Then you need to look a bit more carefully.
I don't see any detailed description of how they did the placebo business. I don't see any description of how they tested for cheating. If there are two rooms, one for placebo and one for RF, or if the RF generator was in the same room, obviously the whole experiment is bogus.
None of those things need to be in the paper; the presumption in scientific papers is that the authors are familiar with the basic tools and methods of their research area. Unless you have a specific cause to doubt that, you have no justification for questioning their results because they did not include those details.
I think one room was where they strapped a black box to your head and either zapped you with RF or Placebo, and the other room had a bed and an EEG for the sleep testing.
It's pretty skinny on quantitative analysis. There's some numbers, and a mention of some preliminary results from a logistic regression. Quite why they've not got some final results from the logistic regression (it doesn't take long, it's not like there's masses of data) is interesting...
Well, there's also this little known effect, like that EM fields induce currents in conductors. The brain works based on electrical impulses. Can it cause induction?
I don't know whether it can or not, but I'd like to see that addressed just for once. You know, instead of the "it can't be anything but heating" handwaving. I'd like just once that someone addresses that point, even if to bury it finally, you know?
Second, exactly how do microwaves heat water. If you have one MW photon for each million mollecules of water, the way I remember quantum physics is that they _don't_ get a millionth of it each. One mollecule absorbs the whole photon, then bounces into the surrounding ones and spreads the energy around. I.e., for a really really tiny fraction of a second, you have a really high energy mollecule there, not just a bunch of slightly faster ones.
What if that one mollecule is a protein? What if it has a resonance on exactly that frequency or close enough?
What if it bounces into a protein? No, seriously, mis-folding for example is known to be a serious problem. (See mad cow disease or CVD for, admittedly, uncommonly extreme examples of what it can do.) Can it break other bonds or mollecules there? It only takes one protein matching something to fire a signal for example.
I'm curious, you know? Has anyone calculated the energies involved? Is everyone dead sure that it can't break some of the weaker bonds? We don't even really understand how all proteins are folded. (Or we'd give up on that whole branch and on Folding@Home and go do something else.)
No, I'm not one of the tinfoil hat gang, and I never attributed headaches to RF, but I like my science more exact nevertheless. If you're going to claim that it can be _nothing_ else, then I'll take that literally. I'd expect a thorough debunking of literally everything else conceivable there. Ionization is only one aspect of the problem.
I also recally one study where early adopters of cell phones did get slightly more often brain cancer. Ok, so those emitted a heck of a lot more power than cellphones nowadays, and it wasn't that horribly many people even then, so I'm not putting on the tinfoil hat any time soon. But that's one effect which, if true, can't be explained by the "but it's only a little warmth" hypothesis. _Something_ happened in there which we thought was only possible via ionizing radiation. What _is_ the explanation for that? I don't think anyone knows for sure yet.
A polar bear is a cartesian bear after a coordinate transform.
Each of the 71 subjects had one night of practice ('habituation') followed by either a night of real RF then a night of fake RF ('sham'), or vice-versa. Double-blind means that neither the subjects nor the scientists knew which one they were getting at the time.
According to the paper: 'Under the RF exposure condition, participants exhibited a longer latency to deep sleep (stage 3, meanRF=0.37, (SD=0.33), mean- Sham=0.27 hours (SD=0.12); F=9.34, p=0.0037)'. But I don't know how they did their statistics.
Because they had 71 subjects, you get the uncerainty of the mean of each measurement by dividing the SD (standard deviation) by sqrt(71), giving mean latencies and uncertainties thereof of: RF = 0.37 +/- 0.039; sham = 0.27 +/-0.014; delta = 0.10 +/- 0.041; yielding a significance of 2.4 sigma.
2.4 sigma should convince approximately no-one.
This simplistic statistical analysis ignores the fact that the distributions are non-Gaussian (which they definitely are). But as a working scientist, I have learned to never presume that authors did their statistics right. (Not that I have reason to doubt these particular scientists, but averaged over papers P(wrong statistics) is much much greater than the 0.0037 they calculate for their effect.)
On the ad hominen side, this paper was funded by the Mobile Manufacturer's Forum. Therefore, somehow, it must be an evil plot or something, although I don't see how.
That's great for the 2006 study. This abstract has a date of 2007 and doesn't claim to be double-blind, unfortunately.
Claiming in no uncertain terms that [double-blinding] was definitely NOT used in this experiment when you have no way of knowing, and simply because it wasn't mentioned, is rather silly.
Pot, kettle, black. Grandparent is correct, parent is a douche. The study did not claim it was double-blinded, and the grandparent simply pointed this out.
Also, if a study was double-blinded, the authors would almost certainly include that information in the abstract. They didn't, so it's fair to assume it wasn't.