The purpose of the study was to investigate placebo and buprenorphine effects on event-related potentials (ERPs) in experimental pain and the potential benefit of population pharmacodynamic modelling in data analysis. Nineteen healthy volunteers received transdermal placebo and buprenorphine in a cross-over study. Drug plasma concentrations and ERPs after electrical stimulation at the median nerve with intensity adjusted to pain detection threshold were recorded until 144 hrs after administration. Placebo and concentration-effect models were fitted to data using non-linear mixed-effects modelling implemented in NONMEM (V7.2.0.). Pharmacodynamic models were developed to adequately describe both placebo and buprenorphine ERP data. Models predicted significant placebo effects, but did not predict significant effects related to buprenorphine concentration. Models revealed that ERPs varied both between subjects and between study occasions. ERPs were found to be reproducible within subjects and occasions as population variance was found to be eight times higher than the unexplained variances. Between-subject variance accounted for more than 75% of the population variance. In conclusion, pharmacodynamic modelling was successfully implemented to allow for placebo and variability correction in ERP of experimental pain. Improved outcome of ERP studies can be expected if variation between subjects and study occasions can be identified and described.