In order to monitor descending pathways during neurosurgical operations on the spinal cord, motor evoked potentials (MEP) were recorded from the epidural space of the spinal cord and the cauda equina following transcranial electrical cortex stimulation in a total of 40 patients. It was the aim of our study to test this invasive recording technique with regard to recordability of potentials as well as correlation to post-operative neurological conditions. In 15 of 23 patients (65.2%) intraoperative potentials were obtained from the spinal cord and in 33 of 40 patients (82.5%) they were obtained from the cauda equina. On the basis of acceptable changes in amplitudes of up to 50% at the end of the operation, 13 of 15 (86.7%) recordings from the spinal cord and 28 of 33 (84.8%) from the cauda equina correlated correctly with the post-operative motor status. At the same time, there were false positive results in 2 (13.3%) and 5 (15.2%) recordings, respectively. False negative results were not observed. Post-operative deterioration of the motor status as observed in 3 patients coincided in 2 patients with a permanent reduction in amplitudes of more than 50% of the baselines and in 1 patient with an intra-operative loss of potentials. Thus both spinal cord and cauda equina recording of MEP elicited by transcranial electrical stimulation were proven to be sensitive methods for detection of impending neurological complications. Cauda equina recording, however, seems to allow recording in more cases.