The effects of midazolam on the EEG were related to plasma midazolam concentrations in 8 healthy male volunteers in order to develop a pharmacokinetic-pharmacodynamic model. The EEG parameters were derived by aperiodic analysis. The EEG was recorded between Fp1-M1 and Fp2-M2. Following a 15-minute baseline EEG registration, midazolam 15 mg was given intravenously over 5 minutes. Venous blood samples were taken until 8 hours after the start of the infusion. Within 2 to 4 minutes of starting the infusion all subjects became asleep, with loss of eyelid reflex. The most obvious EEG changes, in the beta frequency range (12 to 30 Hz), were observed within 2 minutes of the start of drug administration. Seven subjects awoke 60 to 70 minutes after the start of the infusion and 1 awoke after 45 minutes. The EEG parameter that best characterised the effect of midazolam was the total number of waves per second in the frequency range 12 to 30 Hz (TNW12-30). This was used as the effect parameter in the pharmacokinetic-pharmacodynamic modelling. The plasma concentration-time data were characterised by a triexponential function for all subjects. To allow for a possible delay between plasma midazolam concentration and EEG effect, a hypothetical effect compartment was included in the pharmacokinetic-pharmacodynamic model. A sigmoid maximum effect (Emax) model was used to characterise the effect compartment midazolam concentration-TNW12-30 data. The plasma drug concentration corresponding to half the maximum increase in TNW12-30 (EC50) was 290 +/- 98 micrograms/L.(ABSTRACT TRUNCATED AT 250 WORDS)