Aim:The cerebral state index (CSI) was recently introduced as an electroencephalographic monitor for measuring the depth of anesthesia. We compared the performance of CSI to the bispectral index (BIS) as electroencephalographic measures of sevoflurane effect using two combined sigmoidal E max models. Methods: Twenty adult patients scheduled for laparotomy were studied. After induction of general anesthesia, sevoflurane concentrations were progressively increased and then decreased over 70 min. An analysis of the BIS and CSI with the sevoflurane effect-site concentration was conducted using two combined sigmoidal E max models. Results: The BIS and CSI decreased over the initial concentration range of sevoflurane and then reached a plateau in most patients. A further increase in sevoflurane concentration produced a secondary plateau in the pharmacodynamic response. The CSI was more strongly correlated with effect-site sevoflurane concentration (R 2 =0.95±0.04) than the BIS was (R 2 =0.87±0.07) (P<0.05). The individual E max and C eff50 (effect-site concentration associated with 50% decrease from baseline to plateau) values for the upper and lower plateaus were significantly greater for BIS (12.7±7.3, 1.6±0.4, and 4.2±0.5, respectively) than for CSI (3.4±2.2, 1.2±0.4, and 3.8±0.5, respectively) (P<0.05). The remaining pharmacodynamic parameters for the BIS and CSI were similar. Conclusion: The overall performance of the BIS and CSI during sevoflurane anesthesia was similar despite major differences in their algorithms. However, the CSI was more consistent and more sensitive to changes in sevoflurane concentration, whereas the measured BIS seemed to respond faster. The newly developed combined E max model adequately described the clinical data, including the pharmacodynamic plateau.