BACKGROUND: We tested the hypothesis that escalating drug concentrations of sevoflurane are associated with a significant decline of cerebral blood flow in regions subserving conscious brain activity, including specifically the thalamus. METHODS: Nine healthy human volunteers received three escalating doses using 0.4%, 0.7% and 2.0% end-tidal sevoflurane inhalation. During baseline and each of the three levels of anaesthesia one PET scan was performed after injection of . Cardiovascular and respiratory parameters were monitored and electroencephalography and bispectral index (BIS) were registered. RESULTS: Sevoflurane decreased the BIS values dose-dependently. No significant change in global cerebral blood flow (CBF) was observed. Increased regional CBF (rCBF) in the anterior cingulate (17-21%) and decreased rCBF in the cerebellum (18-35%) were identified at all three levels of sedation compared to baseline. Comparison between adjacent levels sevoflurane initially (0 vs. 0.2 MAC) decreased rCBF significantly in the inferior temporal cortex and the lingual gyrus. At the next level (0.2 MAC vs. 0.4 MAC) rCBF was increased in the middle temporal cortex and in the lingual gyrus, and decreased in the thalamus. At the last level (0.4 MAC vs. 1 MAC) the rCBF was increased in the insula and decreased in the posterior cingulate, the lingual gyrus, precuneus and in the frontal cortex. CONCLUSION: At sevoflurane concentrations at 0.7% and 2.0% a significant decrease in relative rCBF was detected in the thalamus. Interestingly, some of the most profound changes in rCBF were observed in structures related to pain processing (anterior cingulate and insula).