As part of a study to determine if decompressive surgery benefits patients who have incomplete lesions of the spinal cord, the somatosensory evoked potential (SEP) has been employed as an intraoperative spinal cord monitor. This procedure was used to see if decompression results in any rapid changes in spinal cord conductivity and to correlate these findings with the patient's subsequent clinical course. In addition to trauma, however, other factors also affect the SEP, and these must be recognized if intraoperative monitoring is to be successful. Frequency of stimulation, wakefulness, and anesthetic agents alter the wave-form and amplitude of the SEP. Using nitrous oxide, oxygen, meperidine (or morphine), and a muscle relaxant for anesthesia, and stimulating at frequencies of 1 or 2 per second, 11 patients with cervical or upper thoracic lesions have been monitored. There was no loss or diminution of an SEP and no patient was neurologically poorer afterwards. In 4 patients, the SEP "improved" soon after decompression. Clinical outcome, however, was not related to whether the SEP changed or remained the same. In total, 8 patients subsequently improved functionally and 3 did not.