Spinal sensory and motor tract activation after epidural electrical stimulation in the cat

Somatosensory evoked potentials (SEPs) after peripheral nerve stimulation and motor evoked potentials (MEPs) after transcranial stimulation have been routinely used as monitors of the viability of pathways in the posterior and anterior spinal cord, respectively, in patients undergoing spinal cord surgery. To assess total spinal cord function, both of these procedures must be performed simultaneously, which is both cumbersome and technically difficult. The objectives of this study were to demonstrate both sensory and motor spinal tract activation after epidural electrical stimulation of the cat spinal cord. Thirty-seven adult mongrel cats were anesthetized with ketamine, intubated, and maintained with Ethrane and nitrous oxide. Stimulating electrodes were placed over the right dorsolateral spinal cord epidurally at T7 after a laminectomy. Recording electrodes were placed over the right L3 spinal cord epidurally, on the right L7 dorsal and ventral nerve roots, on the right and left sciatic nerves in the popliteal fossa, and in the right gastrocnemius muscle. After epidural stimulation of the spinal cord at T11, distinct reproducible potentials were recorded at each site. Activity in the L7 dorsal root implicated activation of spinal sensory tracts. Activity in the L7 ventral root and in the gastrocnemius implicated activation of spinal motoneurons.(ABSTRACT TRUNCATED AT 250 WORDS)     

Successful monitoring of neurogenic mixed evoked potentials elicited by anterior spinal cord stimulation through thoracoscopy during spine surgery.

STUDY DESIGN: Neurogenic mixed evoked potentials were recorded after thoracoscopic spinal cord stimulation in patients undergoing video-assisted spine surgery. OBJECTIVE: To demonstrate the feasibility and value of thoracoscopic spinal cord monitoring. SUMMARY OF BACKGROUND DATA: Video-assisted thoracic surgery recently has been proposed as a new technique for thoracic spine surgery. It can be used for anterior spinal release of patients with severe spinal deformities and for thoracic hernia removal. METHODS: Five patients undergoing video-assisted thoracic surgery for spinal fusion were studied. Neurogenic mixed evoked potentials were elicited by electrodes seated into intervertebral discs through thoracoscopy and recorded from peripheral nerves of the lower limbs. Moreover, the study included the case of a patient with a thoracic hernia who underwent video-assisted thoracic surgery with combined thoracoscopic neurogenic mixed evoked potential and standard somatosensory evoked potential monitoring. RESULTS: Neurogenic mixed evoked potentials were recorded consistently after spinal cord stimulation in all patients. For the patient with a thoracic hernia, neurogenic mixed evoked potentials suddenly disappeared, whereas somatosensory evoked potentials were not significantly modified, leading to surgery interruption. Neurogenic mixed evoked potentials progressively reappeared after a 30-minute delay. Postoperation examination revealed a Brown-Sequard's syndrome with incomplete right motor deficit. CONCLUSIONS: Neurogenic mixed evoked potentials evoked by anterior stimulation through thoracoscopy are of interest for spinal cord monitoring when posterior electrical stimulation is impossible, and they provide reliable information regarding spinal motor pathways.     

"Backfiring" in spinal cord monitoring. High thoracic spinal cord stimulation evokes sciatic response by antidromic sensory pathway conduction, not motor tract conduction.

Spinal cord stimulation has been advocated as an alternative to motor cortex stimulation for motor tract activation. To test this theory, evoked responses were recorded from lumbar spinal cord (L2; n = 14), spinal roots (L4-L7; n = 112), peripheral nerves (sciatics; n = 28), and hind limb muscles (n = 28) after epidural stimulation of the T1-T2 segment of the spinal cord in dogs (n = 12), cats (n = 2), and monkeys (n = 2). The spinal response evoked by spinal cord stimulation was resistant to a dorsal hemisectioning (depth, 7-8 mm) of the midthoracic spinal cord. A minimal attenuation of latency and amplitude occurred with dorsal hemisectioning, suggesting signal transmission through descending or ascending pathways in the ventrolateral and ventral quadrants of the spinal cord. The sciatic nerve response was abolished by a dorsal column transection (depth, 3-4 mm) or ipsilateral lumbar dorsal rhizotomy (four dorsal roots). This shows that the evoked response recorded from the sciatic nerve in our animals was not travelling, as we expected, through the ventral roots, but rather was conducted antidromically through sensory fibers in dorsal roots.     

Origins and conducting tracts of evoked spinal cord potentials in cats.

The origins and the conducting tracts of the evoked spinal cord potential elicited by sciatic nerve stimulation (spinal somatosensory evoked potential: spinal SEP) were studied in cats. The potentials were recorded at various levels of the intact spinal cord and partially resected spinal cord and analyzed with special reference to four negative potentials (N1, N2, N3, and N4 waves). The conduction velocities of N1, N2, N3, and N4 waves were 101.8, 52.7, 88.7, and 42.7 m/s, respectively. N1 and N3 waves are considered to originate from Group I fibers and to be conducted by second-order neurons ascending the ipsilateral-dorsolateral funiculus (dorsal spinocerebellar tract) after synaptic delays. N3 wave may have a longer synaptic delay than N1 wave. N2 and N4 waves are considered to originate from nerve fibers other than Group I and to be conducted primarily in the posterior funiculus. N1 and N3 waves have synaptic delays at the L3 level, but N2 and N4 appear to be conducted without synaptic delay. Therefore, N2 and N4 waves precede N1 and N3 at the upper lumbar and thoracolumbar junction levels, respectively. After that, N1 and N3 waves are conducted with greater velocity than N2 and N4 waves. Because of different conduction velocities, N1 overtook N2 and N3 overtook N4 at the midthoracic region. N2 and N1 waves had the shortest latency in the upper lumbar to thoracolumbar junction levels and the upper thoracic region, respectively.     

Origins and conducting pathways of motor evoked potentials elicited by transcranial (vertex-hard palate) stimulation in cats

Spinal cord evoked potentials were elicited in cats by transcranial electrical stimulation with electrodes on the vertex and hard palate. Vertex motor evoked potentials (V-MEP) were also recorded. An extracellular microelectrode recording technique was then used to analyze the results by isopotential mapping. The relationship between the distribution of field potentials and the stimulation polarity was studied using the field potential distribution of the V-MEP in the lower thoracic spinal cord that had been represented on the isopotential maps. The first negative wave of the V-MEP showed maximal amplitude distribution in the anterior funiculus, which corresponds to the extrapyramidal tracts. This pattern was seen with both stimulation polarity arrangements: 1) stimulation with the cathode at the vertex and the anode at the hard palate, and 2) stimulation with the anode at the vertex and the cathode at the hard palate. When the cathode was at the vertex, the stimulation threshold was lower, and the response had higher amplitude than when the anode was at the vertex. Recording V-MEPs elicited by vertex cathode stimulation could provide an excellent method of monitoring the extrapyramidal tracts in cats.     

Spinal evoked response in the cat.

Summated evoked potentials to sciatic nerve stimulation were recorded from surface electrodes placed over the spine of cats. The response progressively increased in latency rostrally. It was largest and most complex in configuration in leads placed over the caudal spinal cord where sciatic nerve roots enter and begin to ascend the cord. The conduction velocity of the response was about 90 m/sec from rostral sacral to cervical regions. A comparison of surface-recorded evoked responses to stimulation of the sural nerve, the nerve to the medial head of the gastrocnemius muscle and the sciatic nerve suggest that the peripheral nerve fibers that mediate the response to sciatic nerve stimulation are primarily muscle nerve afferents. In surface, lamina, and dural recordings made over similar segmental levels, the response to sciatic nerve stimulation progressively increased in amplitude, duration, and wave form complexity from surface to depth. Failure of transmission across complete cord transections was demonstrated. Results in preparations with partial cord sections suggest that the surface-recorded response is mediated by multiple spinal cord afferent pathways which are situated primarily ipsilateral to the stimulated peripheral nerve. The data indicate that summated evoked responses arising in spinal cord afferent pathways can be recorded from surface-recording electrodes in cats. They suggest that this animal model may prove useful in the study of certain aspects of spinal cord pathology.     

Spinal cord mapping as an adjunct for resection of intramedullary tumors: surgical technique with case illustrations

OBJECTIVE: Resection of intramedullary spinal cord tumors may result in transient or permanent neurological deficits. Intraoperative somatosensory evoked potentials (SSEPs) and motor evoked potentials are commonly used to limit complications. We used both antidromically elicited SSEPs for planning the myelotomy site and direct mapping of spinal cord tracts during tumor resection to reduce the risk of neurological deficits and increase the extent of tumor resection. METHODS: In two patients, 3 and 12 years of age, with tumors of the thoracic and cervical spinal cord, respectively, antidromically elicited SSEPs were evoked by stimulation of the dorsal columns and were recorded with subdermal electrodes placed at the medial malleoli bilaterally. Intramedullary spinal cord mapping was performed by stimulating the resection cavity with a handheld Ojemann stimulator (Radionics, Burlington, MA). In addition to visual observation, subdermal needle electrodes inserted into the abductor pollicis brevis-flexor digiti minimi manus, tibialis anterior-gastrocnemius, and abductor halluces-abductor digiti minimi pedis muscles bilaterally recorded responses that identified motor pathways. RESULTS: The midline of the spinal cord was anatomically identified by visualizing branches of the dorsal medullary vein penetrating the median sulcus. Antidromic responses were obtained by stimulation at 1-mm intervals on either side of the midline, and the region where no response was elicited was selected for the myelotomy. The anatomic and electrical midlines did not precisely overlap. Stimulation of abnormal tissue within the tumor did not elicit electromyographic activity. Approaching the periphery of the tumor, stimulation at 1 mA elicited an electromyographic response before normal spinal cord was visualized. Restimulation at lower currents by use of 0.25-mA increments identified the descending motor tracts adjacent to the tumor. After tumor resection, the tracts were restimulated to confirm functional integrity. Both patients were discharged within 2 weeks of surgery with minimal neurological deficits. CONCLUSION: Antidromically elicited SSEPs were important in determining the midline of a distorted cord for placement of the myelotomy incision. Mapping spinal cord motor tracts with direct spinal cord stimulation and electromyographic recording facilitated the extent of surgical resection.     

Origins and conducting pathways of motor evoked potentials elicited by transcranial magnetic stimulation in cats.

Spinal cord motor evoked potentials (MEPs-S) were elicited in cats by transcranial magnetic stimulation. The MEPs-S recorded from the epidural electrode at the first lumbar (L1) level consisted of four negative peaks (N1, 2.56; N2, 3.19; N3, 4.06; N4, 4.99 ms) followed by small, multiphasic waves. The mean conduction velocities of N1-N3 of the MEPs-S were calculated to be 90 to 115 m/s, which is consistent with extrapyramidal tract activation. A direct brain stem electrical stimulation and sequential transection of the spinal cord studies showed that N1-N3 of the MEPs-S originated primarily from the brain stem and conducted in the ventral funiculus of the spinal cord, which corresponded to the extrapyramidal tracts (e.g., the reticulospinal and vestibulospinal tracts). We also showed that N1 and N2 of the MEPs-S originated mainly from the lower medulla or upper cervical and that N3 originated from the lower pons or upper medulla (vestibular nucleus). On the other hand, N4 of the MEPs-S had a conduction velocity of 70 m/s, which disappeared by ablation of the sensorimotor cortex and pyramidotomy. A dorsal hemisection of the spinal cord, resulting in the disappearance of N4 of the MEPs-S, indicated that it conducted in the dorsolateral funiculus of the spinal cord. These results suggest that N4 of the MEPs-S originates from the sensorimotor cortex and conducts in the lateral corticospinal tract. MEPs-S by transcranial magnetic stimulation in cats constituted complex responses in both the pyramidal and extrapyramidal tracts.     

Effects of spinal cord lesioning on somatosensory and neurogenic-motor evoked potentials

Somatosensory (SEPs) and neurogenic-motor evoked potentials (NMEPs) were elicited from 16 hogs and two humans before, during, and after spinal cordotomy, dorsal, or ventral root rhizotomy. Results indicated that SEPs appear to be insensitive to the effects of motor tract lesioning in hogs and humans. In every case of motor paraplegia, SEPs remained unchanged in the presence of abnormal ischiatic/sciatic NMEPs. These results suggest that SEPs are not adequately sensitive to the functional status of the motor system in hogs and humans. Ischiatic/sciatic NMEPs remained unchanged after sensory tract lesioning, suggesting that these NMEPs are insensitive to the functional status of the sensory system. These results suggest that SEPs and NMEPs should be used in combination when monitoring spinal cord function during surgeries that place that structure at risk.     

Modulating mechanism of the dorsal column stimulation for pain relief (author's transl)]

The neural mechanism of the modulating effects of spinal cord stimulation upon intractable pain was studied in cats under local anaesthesia. The electrical activity at the centre median nucleus of the thalamus (CEM), which responded to noxious stimuli and was inhibited by nitrous oxide inhalation, was recorded as an indicator of the degree of pain sensation. The suppressed effect upon the evoked potential responding to sciatic stimulation (1 Hz) was recorded during and after train stimulation of various parts of the spinal cord by means of silver ball bipolar electrode. Whenever inhibitory modulating effect upon CEM response was observed by spinal cord stimulation, the bilateral dorsal tractotomy cranial, caudal or both to the stimulating point in order to examine the influence upon the CEM potential. The results are as follows: 1) The CEM evoked potential responding to sciatic stimulation is inhibited by about 30-45% of its amplitude by dorsal column stimulation. This effect corresponds to the inhibitory effect by 75% N2O gas inhalation. Upon dorsal column stimulation, the intraspinal electrical activity is found in the central gray matter the antero-lateral quandrants of the spinal cord as well as the dorsal column. 2) Following lateral or anterior column stimulation as well as dorsal column stimulation, the CEM evoked potential responding to sciatic stimulation is inhibited by about 50% of its amplitude. 3) Following bilateral dorsal tractotomy cranial, caudal or both to the stimulating point, the CEM evoked potentials responding to sciatic stimulation are inhibited by dorsal column stimulation. According to the experimental results, it may be concluded that the inhibitory modulating mechanism by dorsal column stimulation for pain relief is not only mediated through the dorsal column, but also through other ascending spinal pathways.     

The effects of spinal gray activation by strychnine on the motor evoked potential in the rat

Spinal motor evoked potentials were elicited by electrical stimulation of the motor cortex in 14 rats before and after the application of strychnine to the surface of the spinal cord. Strychnine applied to the high cervical cord resulted in the emergence of additional peaks at the site of application and in electrodes positioned distally on the middle and lower thoracic cord. The strychnine-induced peaks occurred earlier and were larger in amplitude (P less than 0.01) in the distal spinal cord. Strychnine applied to the lower thoracic cord resulted in similar peak generation in the lower thoracic cord and in the spinal cord proximal to the application of strychnine. These findings demonstrate that strychnine-induced motor evoked potential changes arise from spinal gray activity induced by efferent pathways activated during transcortical stimulation. The role of the propriospinal tract in mediating the distal effects of strychnine is discussed. In conclusion, the strychnine-induced peaks of the motor evoked potential may be used as a measure of spinal gray integrity in experimental spinal cord injury models.     

Delayed detection of motor pathway dysfunction after selective reduction of thoracic spinal cord blood flow in pigs

OBJECTIVE: Clinical monitoring of myogenic motor evoked potentials to transcranial stimulation provides rapid evaluation of motor-pathway function during surgical procedures in which spinal cord ischemia can occur. However, a severe reduction of spinal cord blood flow that remains confined to the thoracic spinal cord might render ischemic only the descending axons of the corticospinal pathway. In this situation lower-limb motor evoked potentials could respond relatively late compared with a similar spinal cord blood flow reduction of the lumbar spinal cord that renders predominantly motoneurons ischemic. METHODS: Selective thoracic and lumbar spinal cord ischemia was induced by sequential clamping of segmental arteries during continuous assessment of laser-Doppler spinal cord blood flow at the thoracic and lumbar spinal cord. Myogenic motor evoked potentials were recorded from the upper and lower limbs. The time to loss of motor evoked potentials was compared (n = 11) during reduction of laser-Doppler spinal cord blood flow below 25% of baseline (ischemic segment), and flow was maintained at greater than 75% of baseline in the nonischemic segment, both during thoracic and lumbar spinal cord ischemia. RESULTS: Average laser-Doppler spinal cord blood flow in the ischemic segment was similar during thoracic (26% +/- 15% [+/- SD]) and lumbar (26% +/- 16%) ischemia, whereas normal flow was maintained in the nonischemic segment. The time to motor evoked potentials loss was considerably longer after thoracic spinal cord ischemia (15 +/- 11 minutes) than after lumbar spinal cord ischemia (3 +/- 2 minutes, P <.005). CONCLUSION: In this experimental model of selective spinal cord ischemia, a severe reduction of lumbar spinal cord blood flow results in rapid loss of myogenic motor evoked potentials, whereas a similar blood flow reduction in the thoracic spinal cord results in relatively slow loss of motor evoked potentials. The effectiveness of motor evoked potentials to rapidly assess spinal cord integrity might be limited when spinal cord ischemia is confined to the thoracic segments.     

Intra-operative spinal cord monitoring during surgery for scoliosis using somatosensory evoked potentials

The method of intra-operative monitoring of spinal cord function by spinal somatosensory evoked potentials (SEP), as used at the Royal Alexandra Hospital for Children during surgery for scoliosis is described. Using a non-polarisable platinum spinal epidural electrode, SEP elicited by tibial and peroneal nerve stimulation in the popliteal fossa are recorded proximal to the level of spinal correction. The large amplitude and discrete waveform of the SEP enable rapid signal acquisition and easy interpretation. The spinal SEP is stable under both therapeutic hypotension and general anaesthesia. Electrophysiological monitoring has now superceded the 'wake-up' test as an index of spinal cord function during corrective surgery for scoliosis.     

Influence of moderate hypothermia on posterior tibial nerve somatosensory evoked potentials

Posterior tibial nerve somatosensory evoked potentials (PTN-SSEP) were recorded in eight patients during cardiac surgery with cardiopulmonary bypass and moderate hypothermia (25-28 degrees C). There was no correlation between changes in amplitude and temperature; however, latencies of potentials recorded over the tibial nerve in the popliteal fossa, the lumbar spinal cord, and the cortex increased linearly as temperature decreased. Latency changes correlated well with nasopharyngeal temperature, but only poorly with rectal and lower limb muscle temperatures. During perioperative monitoring of spinal cord function by means of PTN-SSEP, an increase of the first positive cortical peak (P1) greater than 3 msec is considered an indication for intervention. In this study P1 prolonged 1.15 msec/degree C (r = 0.89, P less than 0.001). This implies that a temperature decrease of 2-3 degrees C may prolong P1 latency by more than 3 msec.     

An experimental study on conductive spinal cord action potentials evoked by direct stimulation of the spinal cord--pathway and source of conductive action potentials in the spinal cord

The following experiments were carried out in adult cats to clarify the pathway and origin of the conductive spinal cord action potential evoked by dorsal epidural stimulation. (1) Comparison with the potential by surface stimulation of the spinal cord: Since waveforms, conduction velocity, and the relationships between the stimulation site and the change in threshold level were equal to those of surface stimulation, epidural stimulation and direct surface stimulation apparently induced the potential deriving from the same origin. (2) Recording of single fiber action potential: The mean conduction velocity of the dorsal column fiber was approximately 50 m/sec and that of the dorsolateral funiculus fiber was about 80 m/sec. 15-20% of potential N1 and 80-85% of potentials N2 and N3 were composed by the dorsal column fibers, whereas 80-85% of N1 and 15-20% of N2 and N3 were composed by the dorsolateral funiculus fibers.     

Evaluation of motor- and sensory-evoked potentials for spinal cord monitoring during thoracoabdominal aortic aneurysm surgery.

OBJECTIVE: To assess the utility of spinal cord monitorings for prediction of spinal cord ischemia, we investigated the role of both motor evoked potentials and sensory evoked potentials during thoracoabdominal aortic aneurysm surgeries. METHODS: We monitored two kinds of sensory evoked potentials; descending evoked spinal cord potentials from the lumbar enlargement after cervical spinal cord stimulation and segmental evoked spinal cord potentials at the lumbar enlargement elicited by peroneal nerve stimulation, and motor evoked potentials from the lumbar enlargement elicited by direct subcranial stimulation in 9 thoracoabdomonal aortic aneurysm surgeries. RESULTS: Postoperative paraplegia occurred in one case in which the patients died during the perioperative period. One case showed transient paraparesis, but recovered following rehabilitatation. These cases showed a decrease in the amplitude of descending evoked spinal cord potentials and motor evoked potentials. CONCLUSION: The recovery of the amplitude of the motor evoked potentials and the descending evoked spinal cord potentials after declamping correlated with the neurologic outcome.     

Electrophysiological, histological, and behavioral studies in a cat with acute compression of the spinal cord

We made an animal model of cervical spinal cord injury in a cat and performed electrophysiological, histological, and behavioral studies. The cervical spinal cord injury model was made by inserting a screw through the fourth cervical vertebral body of the cat. After the injury, muscle tonus of the limbs and trunk diminished. The cat was able to walk and to perform target reaching after the injury, although the trunk was unstable and the movement of the forelimb was ataxic. Atrophy of the back muscle remained. The cat was unable to stand bipedally. These behavioral disturbances suggested dysfunction of the ventral funiculus. After behavioral analysis for 3 months, an electrophysiological study was performed. Action potentials of the ventral funiculus evoked by stimulation of the lateral vestibular nucleus or the medial longitudinal fasciculus were recorded at several levels of the cervical cord. They were diminished at levels caudal to the compression site. After the recording, the cervical cord was studied histologically. There was demyelination and gliosis in the ventral funiculus and in part of the ventrolateral funiculus at or near the injured site. Electrophysiological and histological findings were in good agreement with the behavioral ones. Behavioral studies seem useful for evaluating the function of the spinal cord, especially for assessing the injured system in the spinal cord. Received: March 28, 2000 / Accepted: September 13, 2000     

Experimental investigation on the spinal cord evoked injury potential

Averaged somatosensory evoked potentials from the epidural space in response to sciatic nerve stimulation were recorded in bipolar and common reference mode in cats following various types of injury. An investigation was conducted on the development and properties of the spinal evoked response recorded from the center of the injury site, designated here as the "spinal cord evoked injury potential." Typically it is a two-peak monophasic positive potential, approximately 40 msec in duration, with a slight negative afterwave. With increasing distance from the site of injury, its amplitude rapidly decreases, whereas latency remains constant. The common reference recording technique resulted in an earlier and better demonstration of the evoked injury potential, especially when it was transitory or incomplete. When impairment of conduction developed gradually, the evoked injury potential developed gradually too. In serial recordings along the spinal cord axis, the transition from a normal triphasic to a monophasic evoked injury potential allowed a precise localization of the lesion. These data suggest that the diagnostic value of intraoperative spinal cord monitoring may be increased by adopting a technique that incorporates several epidural recordings with a common reference recording technique. The spinal cord evoked injury potential seems to be a more sensitive indicator of spinal cord injury than the cortical evoked potential. The findings are discussed in the light of the presently developing spinal cord monitoring techniques.     

Effects of epidural electrical stimulation modalities on spinal cord function and morphology in cats

The safety of direct electrical stimulation to the spinal cord as used in spinal cord monitoring is controversial. This investigation was undertaken to determine the critical condition of continuous electrical stimulation over a period of 30 min with varying intensities and pulse durations, and the level of safety of intermittent stimulation of 100 pulses with 10 mA current of 0.3 msec duration with varying resting intervals. The spinal cord was evaluated by spinal cord evoked potentials (SCEP) recording and histological examination. The intensity of continuous stimulation greater than 10 mA of 0.3 msec duration resulted in permanent changes in SCEP, focal colliquative necrosis of the cord, and acceleration of the permeability of the intrinsic vessels. In the intermittent stimulation with 1-min resting intervals, spinal cord lesions were not observed.     

Spinal cord stimulation evoked potentials during thoracoabdominal aortic aneurysm surgery.

Although monitoring of somatosensory evoked potentials elicited from stimulation of lower extremity peripheral nerves has been suggested as a method for assessing neural function during thoracoabdominal aortic aneurysm surgery, this technique has been reported to yield a large number of false positives. It was believed that direct stimulation of the spinal cord would eliminate some of the problems associated with peripheral evoked potentials. The present study compared in 18 patients the use of scalp recorded evoked potential following stimulation of either the posterior tibial nerve via percutaneous needles or the spinal cord via an epidural electrode previously placed fluoroscopically. In 10 patients in whom distal bypass or shunt was not used, peripheral evoked potentials totally disappeared within 5-30 min of aortic clamping. Spinal cord stimulation evoked potentials disappeared permanently in 2 patients shortly after aortic cross-clamping; 1 died shortly after the procedure, and the other awoke densely paraplegic and died the next day. When distal perfusion was maintained by shunt or bypass, the disappearance of both peripheral and spinal evoked potentials accurately predicted the neurologic outcome of 1 paralyzed patient. Loss of spinal cord stimulation evoked potentials was found to be correlated with adverse neurologic outcome. Over the period of aortic clamping a gradual decrease in mean amplitude (50% at 45 min [P less than 0.05]) and a 20% increase in mean latency time were observed. Maintenance of adequate distal perfusion may permit the use of peripheral evoked potentials in the assessment of spinal cord ischemia during aortic cross-clamping.(ABSTRACT TRUNCATED AT 250 WORDS)