In 15 patients who underwent open exploration of the brachial plexus, the somatosensory evoked potentials and nerve action potentials recorded at the time of operation were useful as guides to the most appropriate surgical procedure, and also in predicting the outcome in certain lesions. In three patients the apparent normality of the upper trunk of the plexus was concealing a more proximal lesion which was irrecoverable. The presence of a somatosensory evoked potential showed functional continuity in three patients in whom the C7 root was clinically involved and who recovered after operation. In five patients proximal stumps of ruptured C5 roots showed functional central continuity; this indicated their suitability for grafting. These patients recovered except one who suffered from co-existing disease. The electrophysiological studies also confirmed the clinical diagnosis of avulsion of the C8 and T1 roots and therefore prevented unnecessary dissection

Somatosensory evoked potentials (SSEPs) measure the conduction pathways from the periphery to the brain and can demonstrate the site of neurological impairment in a variety of locomotor conditions. SSEPs were studied in 44 children (64 feet) with surgically corrected club feet. Four children had unreproducible responses, 18 showed abnormal recordings and 22 showed normal responses. In a further 31 feet (21 children) subjected to motor electrophysiological tests, 16 (52%) were abnormal. Overall, 44 of 95 feet (46%) showed abnormal SSEPs or motor electrophysiological tests. Neurological abnormality was related both to the severity of the deformity and the surgical outcome. It was seen in 38% of feet with grade-2 and in 53% of feet with grade-3 deformity. A fair surgical result was obtained in 36% of feet with a conduction deficit and in only 6% with no abnormality. These results suggest an association between neurological abnormality as demonstrated by SSEPs or motor electrophysiological studies and the severity of deformity in club foot and its response to surgical treatment

We have determined whether somatosensory evoked potentials (SEPs) were detectable after direct mechanical stimulation of normal, injured and reconstructed anterior cruciate ligaments (ACLs) during arthroscopy. We investigated the position sense of the knee before and after reconstruction, and correlated the SEP with instability. Reproducible SEPs were detected in all 19 normal ACLs and in 36 of 38 ACLs reconstructed during a period of 13 months. Of the 45 injured ACLs, reproducible SEPs were detected in 26. The mean difference in anterior displacement in the SEP-positive group of the injured ACL group was significantly lower than that in the SEP-negative group. In the reconstructed group, the postoperative position sense was significantly better than the preoperative position sense. Our results indicate not only that sensory reinnervation occurs in the reconstructed ACL, but also that the response to mechanical loads can be restored, and is strongly related to improvement in position sense

We undertook a retrospective analysis of 306 procedures on 233 patients, with a mean age of 12 years (1 to 21), in order to evaluate the use of somatosensory evoked potential (SSEP) monitoring for the early detection of nerve compromise during external fixation procedures for limb lengthening and correction of deformity. Significant SSEP changes were identified during 58 procedures (19%). In 32 instances (10.5%) the changes were transient, and resolved once the surgical cause had been removed. The remaining 26 (8.5%) were analysed in two groups, depending on whether or not corrective action had been performed in response to critical changes in the SSEP recordings. In 16 cases in which no corrective action was taken, 13 (81.2%, 4.2% overall) developed a post-operative neurological deficit, six of which were permanent and seven temporary, persisting for five to 18 months. In the ten procedures in which corrective action was taken, four patients (40%, 1.3% overall) had a temporary (one to eight months) post-operative neuropathy and six had no deficit. After appropriate intervention in response to SSEP changes, the incidence and severity of neurological deficits were significantly reduced, with no cases of permanent neuropathy. SSEP monitoring showed 100% sensitivity and 91% specificity for the detection of nerve injury during external fixation. It is an excellent diagnostic technique for identifying nerve lesions when they are still highly reversible.

Introduction. Somatosensory evoked potential (SSEP) monitoring allows for assessment of the spinal cord and susceptible structures during complex spinal surgery. It is well validated for the detection of potential neurological injury but little is known of surgeon's responses to an abnormal trace and its effect on neurological outcome. We aimed to investigate this in spinal deformity patients who are particularly vulnerable during their corrective surgery. Methods. Our institutional neurophysiology database was analysed between 1. st. October 2005 and 31. st. March 2010. Monitoring was performed by a team of trained neurophysiology technicians who were separate from the surgical team. A significant trace was defined as a 50% reduction in trace amplitude or a 10% increase in signal latency. Patients suffering a significant trace event were examined post-operatively by a Consultant Neurologist who was separate from the surgical team. Results. 2386 consecutive operations (F:1719, M:667 median age 16 yrs) were performed in the time period and 72 operations reported a significant trace event (‘red alert’). From these cases 47 (65%) had a clearly documented intervention by the surgeon and 7 patients overall suffered a lasting neurological deficit (0.3%). The most common timing events were during instrumentation (50%) and during correction/distraction (16%). Most common responses were optimisation of patient/monitoring set-up (23%) and adjustment of metalwork (22%). There were 18 wake-up tests performed. We found SSEP monitoring to have a sensitivity of 100%, specificity 97.4%, positive predictive value 14% and negative predictive value 100%. A Chi-square test (p=0.016) was significant suggesting intervention had a beneficial effect on neurological outcome. Conclusion. We would advocate the use of SSEP monitoring in all patients undergoing spinal deformity surgery. These patients tend to be young, neurologically intact pre-operatively and are particularly vulnerable to the large corrective forces their surgery requires

Introduction. Evidence suggests that intra-operative spinal cord monitoring is sensitive and specific for detecting potential neurological injury. However, little is known about surgeons' responses to trace changes and the resultant neurological outcome. Objective. To examine the role of intra-operative somatosensory evoked potential (SSEP) monitoring in the prevention of neurological injury, specifically sensitivity and specificity, and whether the abnormalities were reversible. Methods. 2953 consecutive complex spine operations (male 36% female 64%, median age 25yrs) prospectively performed using spinal cord monitoring at a single institution (2005–2009). All traces and neurophysiological events were prospectively recorded by the neurophysiology technician. All patients with a significant neurophysiology event were examined clinically by a neurologist, separate from the spinal surgery team. Significant trace abnormality was defined as a decrease in signal amplitude of 50% or a 10% increase in latency. Timing of trace abnormality, surgeon's response and prospective neurological outcome were recorded. Sensitivity, specificity, positive/negative predictive value were calculated. A Chi-squared test was performed to assess the impact of intervention on neurological outcome (p < 0.05). Results. 2953 operations involving SSEP monitoring were performed and 106 recorded a significant trace abnormality. This most often occurred during instrumentation and the most common reaction was adjustment of metalwork. SSEP monitoring had a sensitivity of 100%, specificity 97.3%, PPV 24%, NPV 100%. There were 79 false positives and no false negatives in this series. Chi-squared test was not significant (p=0.18) suggesting that intervention might not affect neurological outcome in this cohort. Conclusions. Triggering events are uncommon and the development of a persistent neurological deficit is rare with an incidence of 0.85% in this series of 2953 operations. In the majority of cases detection of a monitoring abnormality prompts a corrective reaction by the surgeon. Of those with an abnormal trace 76% were neurologically normal at follow up

Aims

Spinal deformity surgery carries the risk of neurological injury. Neurophysiological monitoring allows early identification of intraoperative cord injury which enables early intervention resulting in a better prognosis. Although multimodal monitoring is the ideal, resource constraints make surgeon-directed intraoperative transcranial motor evoked potential (TcMEP) monitoring a useful compromise. Our experience using surgeon-directed TcMEP is presented in terms of viability, safety, and efficacy.

An electrophysiological system for monitoring the spinal cord during operations for scoliosis is described. During the development of the technique the recording of cortical somatosensory evoked potentials from the scalp and spinal somatosensory evoked potentials from the laminae or spines was superseded by the positioning of recording electrodes in the epidural space cephalad to the area to be fused. All recordings were made in response to stimulation of the posterior tibial nerve at the knee. Results in 138 patients are presented and the findings in three patients who exhibited neurological deficits after operation are described. It is concluded that spinal somatosensory evoked potentials are sensitive to minor spinal cord impairment, possible due to ischaemia, and that these changes may be reversed when the cause is quickly remedied. The monitoring system interferes minimally with anaesthetic and surgical procedures and is now performed as a routine

Introduction: Somatosensory evoked potentials are monitored during the surgical treatment of spinal disorders to reduce the risk of cord injury. Whilst studies have examined its role in patients undergoing correction of idiopathic and neuromuscular scoliotic curves, its effectiveness in patients undergoing operative treatment for spinal injury is less certain. Methods and Results: We reviewed the medical records of patients who underwent surgery for spinal trauma. between 1995 and 2000. There were 82 patients with adequate data for analysis who underwent 83 spinal reconstructive procedures. We recorded the age at injury, diagnosis, time of operation, levels instrumented, systolic and diastolic blood pressures and surgical approach. The intraoperative somatosensory evoked potential (SSEP) traces were examined. The SSEP at insertion of electrode was taken as the control level. The highest and lowest intraoperative somatosensory evoked potentials and SSEP at closure were noted and expressed as a percentage of the control value. Forty patients (48%) had a pre-operative neurological deficit. Neurological deterioration occurred postoperatively in three patients. Eighty-three traces from 82 patients were available for analysis. Fifty-seven patients had a fall in trace amplitude by more than 25% of the control, 25 by more than 50% and eight by more than 75%. With an SSEP amplitude loss of 60%, both sensitivity and specificity for the prediction of post-operative neurological injury were optimised at 67 and 81% respectively, with one false negative result. SSEP rise at completion of spinal reconstruction and highest intraoperative SSEP rise was compared with neurological outcome in the 40 patients with abnormal pre-operative neurology. Neurology improved in all patients in this group who had a trace amplitude more than 60% above the control value at end of operation. None had neurological deterioration. There was no correlation between intraoperative SSEP rise and neurological outcome. Conclusion: Loss of trace amplitude more than 50% is common during spinal reconstructive surgery after trauma, however a 60% threshold for SSEP fall improves specificity by reducing the rate of false positive results. A trace amplitude 60% above the control value at completion of operation is specific but not sensitive for postoperative neurological improvement

Segmental vessel ligation during anterior spinal surgery has been associated with paraplegia. However, the incidence and risk factors for this devastating complication are debated. We reviewed 346 consecutive paediatric and adolescent patients ranging in age from three to 18 years who underwent surgery for anterior spinal deformity through a thoracic or thoracoabdominal approach, during which 2651 segmental vessels were ligated. There were 173 patients with idiopathic scoliosis, 80 with congenital scoliosis or kyphosis, 43 with neuromuscular and 31 with syndromic scoliosis, 12 with a scoliosis associated with intraspinal abnormalities, and seven with a kyphosis. There was only one neurological complication, which occurred in a patient with a 127° congenital thoracic scoliosis due to a unilateral unsegmented bar with contralateral hemivertebrae at the same level associated with a thoracic diastematomyelia and tethered cord. This patient was operated upon early in the series, when intra-operative spinal cord monitoring was not available. Intra-operative spinal cord monitoring with the use of somatosensory evoked potentials alone or with motor evoked potentials was performed in 331 patients. This showed no evidence of signal change after ligation of the segmental vessels. In our experience, unilateral segmental vessel ligation carries no risk of neurological damage to the spinal cord unless performed in patients with complex congenital spinal deformities occurring primarily in the thoracic spine and associated with intraspinal anomalies at the same level, where the vascular supply to the cord may be abnormal

We have investigated the changes in anterior laxity of the knee in response to direct electrical stimulation of eight normal and 45 reconstructed anterior cruciate ligaments (ACLs). In the latter, the mean time from reconstruction was 26.7 months (24 to 32). The ACL was stimulated electrically using a bipolar electrode probe during arthroscopy. Anterior laxity was examined with the knee flexed at 20° under a force of 134 N applied anteriorly to the tibia using the KT-2000 knee arthrometer before, during and after electrical stimulation. Anterior tibial translation in eight normal and 17 ACL-reconstructed knees was significantly decreased during stimulation, compared with that before stimulation. In 28 knees with reconstruction of the ACL, in 22 of which the grafts were found to have detectable somatosensory evoked potentials during stimulation, anterior tibial translation was not decreased. These findings suggest that the ACL-hamstring reflex arc in normal knees may contribute to the functional stability and that this may not be fully restored after some reconstructions of the ACL

We have assessed the efficacy of free nerve grafts in 90 cases of brachial plexus injury. Relatively good recovery of the elbow flexor and extensor muscles and of those of the shoulder girdle was found but recovery of the flexors and extensors of the forearm and of the intrinsic muscles of the hand was extremely poor. Poor results were found when spinal nerve roots seemed normal to the touch and appeared intact but had abnormal somatosensory evoked potentials or myelography. Recovery of the deltoid and infraspinatus muscles was better when injury had occurred to the circumflex and suprascapular nerves rather than to the plexus itself, perhaps because these nerves were explored in their entirety to determine the presence of multiple lesions. It is important to visualise the entire nerve thoroughly to assess the overall condition. Thorough exploration of the plexus and the use of intraoperative recording of somatosensory evoked potentials are essential

This is a prospective study of 107 repairs of obstetric brachial plexus palsy carried out between January 1990 and December 1999. The results in 100 children are presented. In partial lesions operation was advised when paralysis of abduction of the shoulder and of flexion of the elbow persisted after the age of three months and neurophysiological investigations predicted a poor prognosis. Operation was carried out earlier at about two months in complete lesions showing no sign of clinical recovery and with unfavourable neurophysiological investigations. Twelve children presented at the age of 12 months or more; in three more repair was undertaken after earlier unsuccessful neurolysis. The median age at operation was four months, the mean seven months and a total of 237 spinal nerves were repaired. The mean duration of follow-up after operation was 85 months (30 to 152). Good results were obtained in 33% of repairs of C5, in 55% of C6, in 24% of C7 and in 57% of operations on C8 and T1. No statistical difference was seen between a repair of C5 by graft or nerve transfer. Posterior dislocation of the shoulder was observed in 30 cases. All were successfully relocated after the age of one year. In these children the results of repairs of C5 were reduced by a mean of 0.8 on the Gilbert score and 1.6 on the Mallett score. Pre-operative electrodiagnosis is a reliable indicator of the depth of the lesion and of the outcome after repair. Intra-operative somatosensory evoked potentials were helpful in the detection of occult intradural (pre-ganglionic) injury

Purpose. To compare the incidence and nature of ‘neurophysiological events’ identified, post hoc, by a consultant neurophysiologist with those identified intra-operatively by clinical physiologists, before and after intervention(s). Methods. The IOM wave-recordings, event-logs and reports of all spinal deformity cases conducted by a team of clinical physiologists from April to June 2009 (Group 1) were reviewed retrospectively by the same, experienced clinical neurophysiologist, (MG). Interventions were then agreed. The first was to alter the IOM report document to drop down menus. The second was to arrange a series of teaching sessions for the clinical physiologists on a variety of aspects of IOM. Finally during these teaching sessions recent cases were brought to review in an informal setting to discuss. Following implementation of the interventions a further review from April to June 2010 (Group 2) was carried out in the same manner. The clinical physiologists did not know the time periods over which the review would be taking place. Results. From April to June 2009 (Group 1) thirty two patients were studied and from April to June 2010 (Group 2) thirty four patients were studies. Group 1. Twenty seven of these had been monitored using ‘multimodal’ IOM consisting of cortical (CSEP) and spinal (SSEP) somatosensory evoked potentials and motor (MEP/CMAP) evoked potentials. Two patients were inappropriate for MEP recording and two were monitored using epidural SSEP recording. During 10 operations (31%) the surgeons were notified of an ‘intra-operative neurological event’ judged by the clinical neurophysiologist as potentially requiring a surgical response. When the results were audited, only 2 (6%) of these ‘events’ were considered by the consultant clinical neurophysiologist to represent ‘true positive’ intra-operative neurophysiological findings. Group 2. Twenty six of these had been monitored using ‘multimodal’ IOM consisting of cortical (CSEP) and spinal (SSEP) somatosensory evoked potentials and motor (MEP/CMAP) evoked potentials. Four patients were inappropriate for MEP recording and had a combination of SSEP and CSEP. The remainder had IOM with unimodal. No epidural IOM was used during this period. During 4 operations (12%) the surgeons were notified of an ‘intra-operative neurological event’ judged by the clinical neurophysiologist as potentially requiring a surgical response. Post-operative examination of all the patients in both groups revealed that no ‘false negative’ conclusions had been reached. Conclusion. In this series clinical physiologists were found to alert the surgeons 5 times more frequently than was likely to have been the case with an experienced consultant clinical neurophysiologist (31% and 6% respectively). However the increased reporting of intra-operative events did not result in any alteration of the ultimate surgical strategy in any operation although tactical changes were sometimes necessary during the operation in order to test the reversibility of the flagged event. The implementation of two simple interventions resulted in the clinical physiologists alerting the surgeons only 1/3 of that previously (12%) of cases. The log indicated that on all occasions appropriate surgical action had been taken with no residual neurological deficit. This study, owing to its size, cannot answer the key question of safety. Further work to estimate the statistical power required of such a study is being sought. In the interim proving a track record of successful cases provides evidence of efficacy. Ethics Approval: None. Interest Statement: None

Radiological diagnosis is not the only tool in detection, monitoring of progress and making easy to undertake a decision about the surgical scoliosis correction. The below presented algorithm of scoliosis monitoring with complex and repetitive (comparative) neurophysiological examinations facilitates the doctor’s decision about method of the conservative treatment or just the moment of surgical intervention [3, 14]. Neurogenic changes in muscles can be found in early stages of the spine deformation – usually when the Cobb’s angle is over 100 [1]. Vertebral rotation and curvature progression follow simultaneously leading to deformation of the spinal cord together with the local ventral roots compression and sometimes inflammation of them. The structure of the grey matter especially in the ventral horn changes its form more on the convex side of scoliosis. Cell bodies together with the axonal hillocks in the motoneuronal pools show deformations comparing to the analogical area of the concave side. This produce discrete unilateral axonopathy in both efferent fibers of peroneal and tibial nerves in scoliotic patients at the age of about 10. This can be found in electroneurographical (ENG) recordings of M and F potentials even at the angle of scoliosis of 100 [10, 14]. Both parameters of the amplitudes and conduction velocities in M-wave studies are decreased and the frequency of F wave recording is diminished what suggests pathological asymmetrical changes just at the level of the ventral root. That is why electromyographical (EMG) recordings show asymmetrical, according to the ventral root somatotopical innervation, selective (found only in some muscles) deficits in frequency and amplitude of motor units action potentials, predominantly in girls. These girls have scoliosis accelerating the most with angle changes of 50 per year [2] that rapidly deepens the neurogenic changes. Other significant evaluation of the scoliosis acceleration is using the somatosensory evoked potentials (SEPs) for recording progression of pathology in the afferent transmission within the long ascending spinal cord pathways running in dorsal, dorsolateral and lateral funiculi [4, 5]. Changes in parameters more amplitude than conduction velocity from SEPs studies recorded at the cervical level are more visible on the concave than convex side of scoliosis. These changes are correlated with increasing the Cobb’s angle at the apical thoracic vertebrae (Th7–8) while peripheral sensory transmission remains only slightly disturbed [6, 7]. These changes were found to be twice greater when recording of SEPs was performed over cranially on the contralateral side of the scalp to the stimulation site at the ankle (tibial nerve than peroneal nerve fibers excitation) both in mothers and their daughters [4]. This points at the strong inhibition of the afferent transmission at the level of the brain stem (probably thalamus or medial lemniscus). During the comparative SEPs recordings at the cervical level, when parameters of waves change dramatically (or even they disappear), this may suggest that the lateral angle of scoliosis exceeded 450 with great acceleration of the torsion [9]. Somatosensory evoked potential recordings during the surgical correction of scoliosis showed only rarely the immediate improvement of the afferent transmission [7, 8, 11]. However, they make sure a surgeon about lack of blockade within the spinal pathways which comes from derotation and distraction procedures performed on the spine during implantation of the corrective instrumentation. First visible results of improvement in the SEPs parameters recorded postoperatively are usually seen a week after the surgery [14]. The above analogical phenomena but referring to the efferent transmission were shown in motor evoked potentials studies which were induced with the magnetic field (MEP) in areas of motor cortex and recorded from centres of cervical and lumbosacral spinal cord as well as from nerves and muscles of upper and lower extremities [12,13, 15]. Usually when AIS reaches the Cobb’s angle of 200 at the age of 25 and does not progress more it can be assumed, that its development is finished. In these patients the signs of neurogenic changes found in EMG examinations performed in lower extremities, paravertebral and gluteal muscles do not progress, too [14]

To examine the effect of lateral spine curvature on somatosensory evoked potentials (SSEP) in patients with adolescent idiopathic scoliosis (AIS) compared to normal controls. We hypothesise that patients with AIS will show increased latency in their SSEPs when bending into their curve suggesting that their spinal cord is more sensitive to this increased lateral curvature. Patients were recruited from the paediatric scoliosis clinic in a single centre. Inclusion criteria were: diagnosis of AIS, age 10–18 years, major thoracic curve measuring greater than 10 degrees on Cobb measurement, and undergoing nonoperative management. Exclusion criteria were: any detectable neurologic deficit, and previous surgery on the brain or spine. SSEP recordings were obtained via stimulation of the posterior tibial nerve with surface electrode and measurement of the cortical response over the scalp. All recordings were performed three times: with the patient in neutral standing and maximum right and left side bending. SSEP recordings show that when AIS subjects bend into their curve, latency slows by an average of 0.5ms. However there was a bimodal distribution with most subjects showing minimal change (3ms). This subset was statistically different from both a control group, and the larger AIS group. There appears to be a subset of patients with AIS who have subclinical spinal cord dysfunction demonstrated by abnormal SSEPs. This may place these patients at slightly higher risk of neurologic injury at the time of surgery

We report our experience of the monitoring of spinal somatosensory evoked potentials in 60 patients with neuromuscular scoliosis. In 15 cases a significant change occurred in the trace when a sublaminar wire was tightened. There were no postoperative neurological deficits attributable to the surgery

Spinal nerve roots often sustain compression injuries. We used a Wistar rat model of the cauda equina syndrome to investigate such injuries. Rapid transient compression of the cauda equina was produced using a balloon catheter. The results were assessed by daily neurological examination and somatosensory evoked potential (SEP) recording before surgery and ten weeks after decompression. Compression of the spinal nerves induced changes in the SEP which persisted for up to ten weeks after decompression, but it had no effect on the final neurological outcome. Our study shows the importance of early surgical decompression for cauda equina syndrome

Introduction. The rate of total hip arthroplasty (THA) surgery continues to dramatically rise in the United States, with over 300,000 procedures performed in 2010. Although a relatively safe procedure, THA is not without complications. These complications include acetabular fracture, heterotopic ossification, implant failure, and nerve palsy to name a few. The rates of neurologic injury for a primary THA are reported as 0.7–3.5%. These rates increase to 7.6% for revision THA. The direct anterior total hip arthroplasty (DATHA) is gaining popularity amongst orthopedic surgeons. Many of these surgeons elect to use the Hana® table during this procedure for optimal positioning capability. Although intraoperative mobility and positioning of the hip joint during DATHA improves operative access, select positions of the limb put certain neurologic structures at risk. The most commonly reported neurologic injuries in this regard are to the sciatic and femoral nerves. To our knowledge, the use of neuromonitoring during DATHA, especially those using the Hana® table, has not been described in the literature. Methods. The patient was a 60-year-old male with long standing osteoarthritis of the right hip and prior left THA. Somatosensory evoked potential (SSEP) leads were placed bilaterally into the hand (ulnar nerve) as well as the popliteal fossae (posterior tibial nerve). Unilateral electromyography leads were placed into the vastus medialis obliquus, biceps femoris, gastrocnemius, tibialis anterior, and abductor hallucis of the operative limb (Fig. 1). Once the patient was sterilely draped, a direct anterior Smith-Peterson approach to the hip was used. Results. After the patient completed standard pre-operative protocol, neuromonitoring leads were placed as described above. There were no complications, neuromonitoring remained stable from baseline, and the patient tolerated the procedure well. Moreover, the senior author routinely uses a prophylactic cable around the calcar, particularly in patients with osteoporotic bone, as was the case with this patient. The patient's post-operative course has been without complications as well. Conclusion. There are a few studies that have examined the pressure changes around the femoral nerve during a DATHA and found that the nerve was at most danger with misplacement of a retractor near the anterior lip of the acetabulum. Furthermore, the popularity of DATHA and the Hana® table make neuromonitoring more amenable for use since the whole limb does not need to be sterilely prepped as with other approaches to the hip. The reported rates of neurologic injury during any THA along with those developed from passage of prophylactic cerclage cables and the goals of reducing surgical complications make this novel technique intriguing. It allows the surgeon yet another safe and effective tool to decrease the likelihood of neurologic injury during DATHA. For any figures or tables, please contact the authors directly

We used evoked spinal cord potentials (ESCP) for intraoperative diagnosis in 17 cases of traumatic brachial plexus palsy. Forty spinal nerves were directly stimulated during exploration of the brachial plexus and ESCP recorded from the cervical epidural space were compared with simultaneously observed somatosensory evoked potentials (SEP) and myelographic findings. Both SEP and ESCP could be evoked in 21 spinal nerves but ESCP were always more distinct and five to ten times greater in amplitude than SEP. In four nerves, ESCP but no SEP were produced, suggesting that there was continuity from the nerves to the spinal cord. ESCP were obtained from two spinal nerves which appeared to be abnormal on the myelogram. The results show that intraoperative electrodiagnosis by epidural ESCP recordings can provide useful information on the lesions of traumatic brachial plexus palsy