Objective. Our group has undertaken a confirmation study of the work performed by the Swedish group of Leif Salford, which reported the occurrence of blood-brain barrier (BBB) permeation and brain damage (presence of dark neurons), 50 days after a single 2-hour exposure of rats to a GSM-900 signal. Preliminary results were presented at the BioEM05 meeting (Haro et al, P-A-133). We improved the Swedish analysis procedure by adding the detection of apoptosis and a more specific staining method for dark neurons. The detection of dark neurons and BBB permeability was assessed 14 and 50 days after GSM-900 exposure. Apoptosis was detected 14 days after exposure. Methods. The exposure setup was the loop antenna that allows for head-only exposure. Five groups of 16 Fisher 344 rats (14-week old) were exposed to the GSM-900 signal during 2 hours at various brain averaged SAR levels (0, 0.14 and 2.0 W/kg), or were used as cage control or positive controls. Positive controls were treated with kainic acid (10 mg/kg) or by cold injury (dry ice during 5 minutes). After exposure, rats were kept alive during 14 or 50 days to study brain damages. Then, they were anesthetized with urethane (i.p. 1.5 mg/kg), perfused with PBS and fixed with paraformaldehyde 4 % (PAF 4 %). Brains were extracted and put in cold PAF 4 % during the following night, then placed in cold 20 % sucrose during 2-3 days, frozen in isopentane and placed at -80 C. Coding was done on brains. Frozen brains were cut in 3 different areas (frontal, median and posterior). Slices were stained with cresyl violet and FluoroJade B to detect the presence of dark neurons. Apoptosis was investigated using the TUNEL method and endogenous albumin visualized using immuno-histochemical labeling as a marker of leakage of the blood-brain barrier. The numbers of dark neurons labeled with cresyl violet, apoptosis and albumin leakage were evaluated by counting. Fluorojade B staining was measured by image analysis using the Aphelion software. Analyses in brain sections were performed in particular in the motor cortex and the hippocampus (CA1, CA2, CA3 and DG). Kruskal-Wallis statistical tests were used. Results. Using the two methods to detect dark neurons, we found no significant difference between sham and exposed animals at both 14 and 50 days after exposure. These results are in agreement with the absence of neurons dying via apoptosis 14 days after exposure. For all tested SARs, GSM-900 did not induce significant albumin leakage in any brain area. Discussion and Conclusion. Our study did not confirm that GSM-900 exposure induce brain damages as previously described. Our data, together with those of the collaborative groups (Brooks Air Force Base, USA and the National Institute of Public Health, Japan) , should help to reach a conclusion about the effect of GSM-900 signals on the brain