A research paper published in the latest issue of Proceedings of the National Academy of Sciences describes an experiment to measure heating of brain tissue exposed to 1.9GHz RF energy. The researchers believe data from this experiment may be useful in the search for possible health risks of cell phones.
The experiment setup the researchers used was for me the coolest thing about it. They used an MRI to measure temperatures throughout a sample of brain tissue with a 1 degree C resolution. All materials inside the MRI cage had to be nonferromagnetic. They used a non-ferromagnetic coax, balun, and dipole antenna. They used a VNA to measure S11 (i.e. power reflected from the antenna because of poor matching) to tune the antenna.
They ran a dead 1.9GHz carrier (CW) into the sample for 12 minutes at various output levels from 125mW to 2W. The tests under 500mW did not detect a temperature rise because it was less than 1C [see update below], but the data from the other output levels agreed with mathematical models, making it possible to extrapolate heating at lower power levels.
They would like to repeat the experiment on a living animal’s brain. They expect heating to be much less than that observed in the dead brain tissue because blood flow moves the heat away.
What does all this tell us about mobile phones?
A Spectrum magazine article on the paper says “Years of studies to determine whether cellphones can cause brain tumors have yielded one popular consensus: More studies are needed.” This is possibly the popular consensus, but the scientific consensus is that there is no reason to believe there are health risks.
Dr. Kenneth Foster, who was quoted in the Spectrum article, explained that this research is a form RF exposure assessment, since heating is a function of exposure. There are good models for RF exposure, so MRI-based research will be useful only if it discovers a problem with those models.
Dr. Foster said people have speculated about RF energy causing microheating of small areas well beyond the average temperature rise and about health effects unrelated to heating. Research has not been able to detected these. 
I asked him why we have specific absorption rate (SAR) rules limiting exposure from mobile devices if we cannot find evidence of health effects. Many wireless devices contain a warning in their owner's manual that you must maintain a certain distance between the device and the user’s head to comply with SAR limits. Dr. Foster explained that scientists determined how much RF energy would affect lab animals due to heating. They divided that by a fudge factor of 10 to 50 to provide work out safe maximums for humans.
Another researcher quoted in the Spectrum article, Dr. William Happer, responded to my questions about this research by saying no “scientific question is ever fully settled.” Things can be settled, he said, in “religion, but not in science.” This got me thinking. What’s the difference between scientific claims and religious claims? Scientific claims are experimentally falsifiable and always open to new evidence. It seems like the hypothesis of health effects of low-level RF signals has indeed been falsified, but in science we’re always open to new evidence.
One of the authors of the original study, Dr. David Gultekin, sent me copy of the paper but did reply to my questions.
Conclusion
I agree with Dr. Happer that we should never rule out re-testing a hypothesis in a new way. Something may come just from the new test methodology. It’s unfortunate, though, to see the progression from a scholarly paper to a science article with a title claiming this “could solve cellphone radiation problems” to the non-scientific world where the perception is scientists don’t know anything about this topic.
Update on Jan 8, 2013
Dr. Gultekin responded to my questions and this article. Contrary to what I reported above, temperature resolution is better than 1C and heating is measurable even at 0.125W.
Dr. Gultekin explains the reason for this research is that it is a non-invasive non-contact way of measuring radiation absorption with a high spacial and temporal resolution. The methodology could be used for other research involving radiation absorption, including research in living tissue.
