The pain relieving properties of imipramine (100 mg orally), tramadol (150 mg orally), and anpirtoline (60 mg orally) were compared in 16 healthy subjects in a cross-over, double-blind, randomized, and placebo-controlled study. Anpirtoline exhibits analgesia which is possibly mediated via serotoninergic pathways, whereas tramadol exerts its effects at opioid receptors. The pain-relieving effect of the tricyclic antidepressant imipramine may involve both serotoninergic and opioid mechanisms. Chemo-somatosensory event-related potentials (CSSERP) were recorded after painful stimulation of the nasal mucosa with carbon dioxide. Subjects rated the perceived intensity of the stimuli by means of a visual analogue scale. In addition, acoustically evoked responses were recorded, the spontaneous EEG was analyzed in the frequency domain, the subjects' vigilance was assessed in a tracking task, and side effects of the drugs were monitored. Anpirtoline and tramadol produced a decrease of both CSSERP amplitudes and subjective estimates of pain, the effects of the former compound being greater. In contrast, after administration of imipramine no change of CSSERP amplitudes could be detected, whereas the subjective estimate of pain intensity decreased significantly. This was accompanied by a significant decrease of arousal indicating that pain relief produced by acute administration of imipramine was primarily related to its sedation action. The analgesic properties of anpirtoline were demonstrated in man. Tramadol was characterized as a week opioid analgesic. In contrast, imipramine appeared to produce its pain-relieving effects predominantly by non-specific actions. It is hypothesized that different analgesics may change ERP sources in a drug-specific manner.