Figure 3. Sedation and memory scores plotted by time and normalized concentration, with baseline (B) values at the far left, the first three increasing target concentrations (TC1 to TC3) in the left panel, and the last three decreasing target levels (TC3 to TC5) in the right panel, with the highest target level, MaxTarget (TC3), repeated in both the left and right panels. Placebo and ondansetron groups are shown with open symbols, plotted against the TC axis, and data are presented in a time-related manner. Active drugs are represented by the black symbols, which from left to right represent the normalized concentrations at baseline, TC1, TC2, and so on. The x axis used for active drugs is the normalized serum concentration. For propofol and thiopental, which had less sedation than other drugs at TC3, the TC3 point occurs at an x-axis value (normalized concentration) similar to TC1-TC2 for other drugs. The value 5 on this axis is equivalent to the Cp50SEDfor each drug. For simplicity, data are presented only for the LR method of normalizing concentration. (A) Average sedation scores: Because the serum concentrations are normalized against sedation (x axis), the active drugs are superimposed. Note that propofol has the lowest maximum sedation, and thus its normalized concentration at MaxTarget is about half that of midazolam and fentanyl. Thiopental also reaches maximal sedative effect at a lower normalized concentration than midazolam or fentanyl. The placebo group demonstrates a significant placebo response with sedation scores in the placebo group being greatest at TC3. (B) Word recognition at the end of the study day: Values represent the percentage of words recognized from a list presented at a particular serum concentration earlier in the day. Note that propofol and midazolam have the same effect on word recognition, but these memory effects occur at different sedative effects, and thus different normalized serum concentrations. Thiopental has a maximal effect at TC3 that is only half of that obtained with midazolam or propofol. Up to that point, the memory impairment is similar to that of midazolam. Fentanyl and ondansetron have essentially no effect on word recognition. (C) Picture recall at the end of the study day: Values represent the cumulative percentage of recall of pictures presented earlier in the day at a particular serum concentration. Again, propofol produces similar impairment to midazolam, but at lesser sedative effect (lower normalized serum concentration). The maximal impairment achieved with thiopental is less than with midazolam or propofol. Picture recall with fentanyl and ondansetron is similar to placebo. Note the similarity with word recognition, shown in figure B. (D) Picture recognition at the end of the study day: Values represent the cumulative percentage of successful retrieval (including previous recall attempts) when participants were asked to recognize previously unrecalled pictures that had been presented earlier in the day at a particular serum concentration. Note that only propofol and midazolam produce any significant impairment with visual cueing for a memory. The recognition procedure increases memory retrieval compared with recall (panel C) to approximately 50% of pictures shown for midazolam and propofol, and 87% with thiopental. Data points for placebo, ondansetron, and fentanyl are superimposed at 100% recognition.