The exposure chamber design was principally the same as described before [Pakhomov et al., 2000]. The chamber, provided with a continuous but laminar flow of fresh artificial cerebrospinal fluid (ACSF) of about 20 mm depth, was mounted on top of a waveguide opening sealed with a sapphire matching plate covered with 0.5 mm Plexiglas, which served as a chamber bottom.
A brain slice was centered in a small frame (also made of 0.5 mm Plexiglas), which was fixed to the chamber bottom and aligned with the waveguide axis. MWexposures were performed essentially the same way as described before [Pakhomov et al., 2002].
The peak E field in the waveguide reached 1.57 MV/m.
Measurement and calculation details
Incident and reflected powers in the waveguide were measured via directional couplers by a power meter with powersensors. Reflection from the exposure chamber into the waveguide was less than 3%. Local SAR values were measured in the ACSF using a microthermocouple technique [Pakhomov et al., 2000, 2002]. For field mapping, the MTC was driven in three orthogonal directions by means of a micromanipulator. An analytical approach proposed by Alekseev and Ziskin  was employed to account for differences in dielectric properties of the braintissue and ACSF. The presence of the plastic bottom cover and the brain slice holding frame in the exposure chamber decreased local SAR by 50-60%, and at the same time, reduced variations at the brain slice location to ±20%.