The so-called carousel setup has been widely utilized for testing the hypotheses of adverse health effects on the central nervous system (CNS) due to mobile phone exposures in the frequency bands 800-900 MHz. The objectives of this article were to analyze the suitability of the setup for the upper mobile frequency range, i.e., 1.4-2 GHz, and to conduct a detailed experimental and numerical dosimetry for the setup at the IRIDIUM frequency band of 1.62 GHz. The setup consists of a plastic base on which ten rats, restrained in radially positioned tubes, are exposed to the electromagnetic field emanating from a sleeved dipole antenna at the center. Latest generation miniaturized dosimetric E field and temperature probes were used to measure the specific absorption rate (SAR) inside the brain of three rat cadavers of the Lewis strain and two rat cadavers of the Fisher 344 strain. A numerical analysis was conducted on the basis of three numerical rat phantoms with voxel sizes between 1.5 and 0.125 mm3 that are based on high resolution MRI scans of a 300 g male Wistar rat and a 370 g male Sprague-Dawley rat. The average of the assessed SAR values in the brain was 2.8 mW/g per W antenna input power for adult rats with masses between 220 and 350 g and 5.3 mW/g per W antenna input power for a juvenile rat with a mass of 95 g. The strong increase of the SAR in the brain with decreasing animal size was verified by simulations of the absorption in numerical phantoms scaled to sizes between 100 and 500 g with three different scaling methods. The study also demonstrated that current rat phantom models do not provide sufficient spatial resolution to perform absolute SAR assessment for the brain tissue. The variation of the SAR(brain)(av) due to changes in position was assessed to be in the range from +15% to -30%. A study on the dependence of the performance of the carousel setup on the frequency revealed that efficiency, defined as SAR(brain)(av) per W antenna input power, and the ratio between SAR(brain)(av) and SAR(body)(av) are optimal in the mobile communications frequency range, i.e., 0.8-3 GHz.
Copyright 2003 Wiley-Liss, Inc.