In 11 consecutive patients, intraoperative electromyographic (EMG) recordings were made from the facial muscles during microvascular decompression for hemifacial spasm. In one patient, recordings could not be obtained for technical reasons, and two patients had no abnormality. In the remaining eight patients, the abnormal response resolved before decompression in two, resolved immediately at the time of decompression in five, and failed to resolve in one. All patients were relieved of their hemifacial spasm. In the five patients whose abnormalities resolved at the time of decompression, there was a precise intraoperative correlation between decompression of the nerve and disappearance of the abnormal EMG response. In three cases, this was a useful guide to the need to decompress more than one vessel. These results confirm the findings of Møller and Jannetta, support the use of this technique for intraoperative monitoring of facial nerve decompression procedures, and provide strong circumstantial evidence that vascular cross-compression is an important etiological factor in hemifacial spasm.