Use of sensory-evoked potentials recorded from the human occiput for intraoperative physiologic monitoring of the spinal cord

Abstract

Study design: This is a report of a prospective case series.

Objective: The purpose of this study was to determine whether evoked potential activity recorded from occipital scalp electrodes in humans is similar to that recorded in animals, and to evaluate the independence of this activity from the classical somatosensory-evoked potential.

Summary of background data: Intraoperative somatosensory-evoked potentials can be of limited usefulness in predicting spinal cord injury because they are transmitted primarily through the dorsal columns, and therefore do not reflect integrity of important ventral pathways. It recently has been shown in animal studies that a sensory-evoked potential recorded from the cerebellum is mediated via ventral tracts and is useful as an adjunct to the somatosensory-evoked potential in spinal cord monitoring.

Methods: Twenty-five patients undergoing spinal or posterior fossa surgery were consecutively entered into the study. Evoked responses were recorded transcranially from over the cerebellar hemispheres and from the vertex. Recordings were made directly from the surface of the cerebellar hemispheres in seven of these patients.

Results: Waveforms could be recorded and reproduced in all but one of the patients. The overall appearance of the occipital recordings was similar to the appearance of responses obtained in animals. The mean latency of the first negative peak recorded from the ipsilateral occiput was 33.0 msec (standard deviation, 3.7 msec) compared with 52.4 msec (standard deviation, 6.1 msec) for the somatosensory-evoked potential. In addition, the amplitude of this response (0.35 microV; standard deviation, 0.20 microV) was independent of the amplitude of the somatosensory-evoked potential (0.76 microV; standard deviation, 0.69 microV). In five cases, one evoked potential could be recorded in the absence of the other. Recordings from the surface of the cerebellum were of the same morphology, but of greater amplitude than the transcranial recordings.

Conclusion: Evoked responses can be reliably recorded from over the occiput and show characteristics independent of the classical somatosensory-evoked potential. These responses are very similar to the cerebellar-evoked potential recently characterized in animals and may provide a method for assessing the physiologic integrity of the ventral tracts of the spinal cord in humans.