BACKGROUND Recent evidence suggests that anesthetic action within the spinal cord is important in suppressing somatic responses to painful stimuli. Whether the brain influences this response is not clear. This study was designed to test the hypothesis that the brain affects anesthetic requirements. METHODS Six goats were anesthetized with isoflurane. After tracheal intubation and femoral arterial cannulation, bilateral neck dissections were performed to isolate the external carotid arteries and external jugular veins. The occipital arteries were ligated bilaterally. Control isoflurane requirements as defined by the minimum alveolar concentration (MAC) were determined by using a dew-claw clamp as a painful stimulus. Cranial venous blood was drained into a bubble oxygenator in which an isoflurane vaporizer was placed in line with the gas flow, and arterial blood was infused into a carotid artery with a roller pump. This arrangement permitted selective control of the delivery of anesthetic to the head and to the systemic circulation. Isoflurane concentration in the arterial blood delivered to the head was estimated from the isoflurane concentration in the oxygenator exhaust. While isoflurane concentration in the head was maintained at approximately 0.2-0.3%, MAC for the body was determined. After return to the native circulation, MAC was determined again. RESULTS During bypass with cranial isoflurane concentration at 0.2-0.3%, all animals showed varying, intermittent degrees of light anesthesia, including spontaneous head movement, chewing, swallowing, and eye opening. Isoflurane MAC was 1.4 +/- 0.2% (mean +/- SD) at baseline, decreased to 0.8 +/- 0.1% during bypass (P < 0.05), and increased to 1.2 +/- 0.2% after bypass (P not significant compared with baseline). CONCLUSIONS These results verify the importance of volatile anesthetic action at an extracranial site vis à vis purposeful movement in response to a noxious stimulus. Furthermore, the results confirm that the brain affects anesthetic requirements.