Brain stem auditory (BAERs) and median verve somatosensory evoked responses (MnSSERs) were recorded from normal neural pathways during the induction of pentobarbital coma in six patients undergoing elective excision of complex ateriovenous malformations. Each patient received a 33-minute infusion of pentobarabital at a rate of 0.6 mg/kg (total dose, 19.8 mg/kg). This regimen resulted in burst suppression or isoelectricity of the electroencephalogram in all patients. Although statistically significant changes in latency and amplitude occurred, both BAERs and MnSSERs were readily recordable in all patients throughout the infusion. For the BAER, there were significant increases in the latencies of Waves III and V. However, these increases are sufficiently small that misinterpretation of these changes as an evolving neurological injury is unlikely, e.g., Wave V latency increased from 6.25 ± 0.25 (SD) to 6.58 ± 0.16 ms (P < 0.006). For the MnSSER, changes of greater potential clinical relevance were observed. There were substantial increases in the latencies of the early components of the primary cortical response and in the central conduction time (e.g., CCT of 6.1 ± 0.6 ms preinduction vs. 7.7 ± 1.1 ms at t = 33 min, P < 0.003), and the amplitude of the early cortical response (N20-P25) decreased by a mean of 45% (P<0.02). By contrast, subcortical components of the MnSSER (brachial plexus, upper cervical spine) were only minimally affected. We conclude that, in patients who are initially neurologically intact, BAERs and MnSSERs can be monitored effectively during pentobarbital coma and that the loss oof these responses should not be ascribed to the effects of this drug alone. However, the interpretation of evoked response changes, particularly MnSSER changes, that occur during pentobarbital administration should take into account the dose-related changes in latency and amplitude that we have observed.