MEDIA ADVISORY: Pitt Researchers Raise Significant Questions in Letter to JAMA About Report on Cell Phones’ Effect on Brain Chemistry

PITTSBURGH—University of Pittsburgh researchers raise several questions about the design and conclusions of a study published Feb. 23 in the Journal of the American Medical Association (JAMA) that suggests a correlation between cell phone use and increased glucose metabolism in the brain, according to a letter published in the May 25 edition of JAMA. A team of researchers led by the National Institute on Drug Abuse (NIDA) reported that 50 minutes of cell phone exposure resulted in increased glucose metabolism in the brain region closest to the phone’s antenna, but that this burst of activity was of unknown clinical significance.

Arthur Kosowsky and Eric Swanson, professors in Pitt’s Department of Physics and Astronomy, and Edward Gerjuoy, Pitt Professor Emeritus of physics and astronomy, write that the researchers’ analyses were inconsistent with their data; that their analysis method was prone to statistical biases; and that they did not use a model that realistically represents cell phone radiation and its propagation into the brain.

The Pitt authors first point out that the Feb. 23 report used oversimplified electromagnetic radiation. The NIDA authors assumed that cellular phone antennas are simple dipoles in which electrons oscillate the length of a linear antenna, such as on a radio. Modern cellular antennas are more often fractal, consisting of complex, repeating shapes that can change the antenna’s electric field pattern. In addition, the study authors assumed that the resulting radiation field acted as it would in a vacuum, which ignores the complex interaction of electromagnetic fields and living tissue. These factors could produce substantially different conclusions from those reported in JAMA, Kosowsky explained.

Kosowsky and his colleagues also fault the report’s data analysis for being inconsistent and incomplete. Graphs in the paper that illustrate brain glucose metabolism do not include error bars to indicate the influence of uncertainty and noise—factors unrelated to the experiment—on the researchers’ measurements. As a result, Kosowsky said, the effect of cellular phones on glucose metabolism appears to be highly significant in the plotted data, yet the NIDA team stresses in their paper that phone use resulted in only a modest increase in metabolism. In short, it’s difficult to determine whether the study’s results are significant or marginal, Kosowsky said.

The NIDA team members also did not conceal from themselves which brain scans were from active and inactive phones, making biases in data analysis more likely, according to the letter. At the same time, the researchers did not randomize the side of the head on which study participants placed the active phone, nor did they control for the possible effect of heat from an active phone.

“The brain is not a symmetrical material that operates in a vacuum. It’s a complicated organ with intricate electrical properties that we think were not adequately considered in the analysis of this experiment, as it was described in JAMA,” Kosowsky said.

“We want to illuminate some potential shortcomings in a study that has been presented as making a significant contribution to the question of whether cellular phones impact brain function and health.”

The Pitt letter is available on JAMA’s Web site.

###

5/25/11/tmw/lks/jdh