There is a high incidence of intraoperative awareness during cardiac surgery. Mid-latency auditory evoked potentials (MLAEP) reflect the primary cortical processing of auditory stimuli. In the present study, we investigated MLAEP and explicit and implicit memory for information presented during cardiac anaesthesia. PATIENTS AND METHODS. Institutional approval and informed consent was obtained in 30 patients scheduled for elective cardiac surgery. Anaesthesia was induced in group I (n = 10) with flunitrazepam/fentanyl (0.01 mg/kg) and maintained with flunitrazepam/fentanyl (1.2 mg/h). The patients in group II (n = 10) received etomidate (0.25 mg/kg) and fentanyl (0.005 mg/kg) for induction and isoflurane (0.6-1.2 vol%)/fentanyl (1.2 mg/h) for maintenance of general anaesthesia. Group III (n = 10) served as a control and patients were anaesthetized as in I or II. After sternotomy an audiotape that included an implicit memory task was presented to the patients in groups I and II. The story of Robinson Crusoe was told, and it was suggested to the patients that they remember Robinson Crusoe when asked what they associated with the word Friday 3-5 days postoperatively. Auditory evoked potentials were recorded awake and during general anaesthesia before and after the audiotape presentation on vertex (positive) and mastoids on both sides (negative). Auditory clicks were presented binaurally at 70 dBnHL at a rate of 9.3 Hz. Using the electrodiagnostic system Pathfinder I (Nicolet), 1000 successive stimulus responses were averaged over a 100 ms poststimulus interval and analyzed off-line. Latencies of the peak V, Na, Pa were measured. V belongs to the brainstem-generated potentials, which demonstrates that auditory stimuli were correctly transduced. Na, Pa are generated in the primary auditory cortex of the temporal lobe and are the electrophysiological correlate of the primary cortical processing of the auditory stimuli. RESULTS. None of the patients had an explicit memory of intraoperative events. Five patients in group I, one patient in group II, and no patients in group III showed implicit memory of the intraoperative tape message. They remembered Robinson Crusoe spontaneously when they were asked their associations with Friday. In the awake state AEP peak latencies were in the normal range. During general anaesthesia in group I, the peaks Na, Pa did not increase in latency or decrease in amplitude before and after the audiotape presentation. The primary cortical complex Na/Pa could be identified as in the awake state. In contrast, in group II Na, Pa showed a marked increase in latency and a decrease in amplitude or were completely suppressed. CONCLUSIONS. During general anaesthesia auditory information can be processed and remembered postoperatively by an implicit memory function, when the electrophysiological conditions of primary cortical stimuli processing is preserved. Implicit memory can be observed more often when high-dose opioid analgesia is combined with receptor-binding agents like the benzodiazepines than under non-specific anaesthetics like isoflurane. Non-specific anaesthetics seem to provide a more effective suppression of auditory stimuli processing than receptor-specific agents.