Effects of inhaling three levels of nitrous oxide in oxygen on event-related brain potentials (ERPs) and pain report were examined in 10 volunteers undergoing painful electrical stimulation of tooth pulp. Previous work by the authors demonstrated that inhalation of nitrous oxide 33 per cent in oxygen, iv injection of 0.1 mg fentanyl, oral administration of 975 mg aspirin, and electrical acupunctural stimulation all reduced ERP amplitudes obtained a vertex during painful tooth pulp stimulation. The authors report here the demonstration of a dose-response relationship between increasing concentrations of nitrous oxide in oxygen and measures of ERP amplitude and pain report. Subjects inhaled room air, nitrous oxide 25 per cent, 37 per cent, and 50 per cent in oxygen while ERPs were recorded and pain reports were given. The procedure was repeated on three separate days with each subject experiencing all levels of treatment on each day. Analyses of variance revealed that both ERP amplitude and pain report significantly decreased as dosage increased, and a significant linear trend was observed for the positive-going ERP wave-form deflection between 160 and 240 ms. Pain report scores decreased significantly (P less than 0.001) and proportionally as dosage increased, but there was not a significant linear trend. Inhalation of nitrous oxide in oxygen increased peak latency for the negative component at 50 ms and the positive component at 90 ms but not for later components. These outcomes demonstrate that amplitude measures of the vertex ERP obtained with dental dolorimetry correlate consistently with pain and analgesia. Simultaneous assessment of brain electrical activity and subjective report appears to be a useful approach for the assessment of analgesia in humans.