Spinal cord potentials in traumatic paraplegia and ... · JournalofNeurology, Neurosurgery, andPsychiatry, 1980, 43, 823-830 Spinal cordpotentials in traumatic paraplegia andquadriplegia

Journal of Neurology, Neurosurgery, and Psychiatry, 1980, 43, 823-830

Spinal cord potentials in traumatic paraplegiaand quadriplegiaE M SEDGWICK, E EL-NEGAMY, AND H FRANKEL

From the Wessex Neurological Centre, Southampton, The Higher Institute oJ Physical Therapy,University of Cairo, Egypt, and The National Spinal Injuries Centre, Stoke Mandeville Hospital,Stoke Mandeville

S U M MA R Y Cortical, cervical and lumbar somatosensory evoked potentials were recordedfollowing median and tibial nerve stimulation in patients with traumatic paraplegia and quadriplegia.The isolated cord was able to produce normal potentials even during spinal shock if the verticalextent of the lesion did not involve the generator mechanisms. The cervical potentials showedsubtle changes in paraplegia at Th5 levels and below. In high cervical lesions the early cervicalpotentials may still be present but the later potentials were absent or, in partial lesions, delayed.

The first attempt to record spinal cord poten-tials in humans was by Pool,' who notedepileptic-like waveforms from the distal cord ina paraplegic patient. Sawa2 inserted bipolar elec-trodes into the cords of psychotic patients andnoted spikes and waves which were probablyinjury potentials. Magladery et al were able toidentify root and spinal potentials from intra-thecal electrodes in the lumbar region followingtibial nerve stimulation. Recently it has becomepossible to record potentials from the skin sur-face which originate in the spinal cord followingperipheral nerve stimulation.4-7 Potentials fromthe lumbar region have been identified as origin-ating in the roots and cord by Dimitrijevic et al,'7Delbeke et al,8 and El-Negamy and Sedgwick6while those recorded from over the cervicalarea originate from the brachial plexus,9 thedorsal horn of the cervical spinal cord,6 andfrom more rostral structures.

This investigation was carried out with twoaims. Firstly, to determine whether the spinalcord could produce potentials when isolated fromrostral neuroaxis and secondly, to see if spinalcord potentials could be of any clinical use in thediagnosis and assessment of the spinal injurypatient.

Electrophysiological methods have been usedpreviously to judge the extent of spinal cordAddress for reprint requests: Dr EM Sedgwick. Wessex NeurologicalCentre, Southampton General Hospital, Tremona Road, South-ampton S09 4XY

Accepted 20 March 1980

injury and its prognosis in experimental ani-mals'0-'5 and in man,'6-18 but all these workersstudied the cortical somatosensory evoked poten-tial. The animal studies showed that, after spinalcord injury, the cortical somatosensory evokedpotential may be preserved for a few minutes ifthe injury was mild, but then disappeared. Earlyrecovery of the potential preceded good func-tional recovery but Singer et all' claim goodfunctional recovery in monkeys even when thecortical potential had been absent for as long as19 days.Experience with humans shows that cerebral

evoked potentials may be absent or attenuatedearly after injury but a gradual return and nor-malisation of potentials from stimuli originatingbelow the site of injury heralds good functionalrecovery before any significant clinical improve-ment occurs.16-'8There have been no previous studies of the

spinal cord potentials after spinal injury in manalthough they have been used in the diagnosisof multiple sclerosis'9 and they are of abnormalform in cervical spondylosis.20

Methods

The methods have been described in detail byEl-Negamy & Sedgwick.6 Briefly, the mediannerve at the wrist or the tibial nerve at thepopliteal fossa was stimulated at three timessensory threshold or sufficient to produce amoderate m-ascle twitch in the absence of sensa-

823

F

Protected by copyright.

on July 12, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.9.823 on 1 S

eptember 1980. D

ownloaded from

http://jnnp.bmj.com/

824

Number of patientsand site of spint lesionPartial Completelesions lesions

.1

.... ....

elx: 2v 3I 4

3

lx

8 3C

: 4

a 5

Cauda equlna

Fig 1 The number of patients and the level of thespinal cord lesion is shown diagrammatically. Thepatient marked X had two lesions. Those withclinically complete lesions are shown on the right andthose with partial or incomplete lesions on the left.

tion (about 2x motor threshold). Recordings fromnthe skin overlying the vertebral column wereamplified with a bandwidth of 10Hz to 10kHZand recorded on magnetic tape using a FM sys-

Normal Subject

FzC3Cv2 I

Median NerveStiunlation I

Fz -C3Cv2

Scapufla-- ----

I

Stimulation

10 20 30m stc

E M Sedgwick, E El-Negamy, and H Fralnkeltem with a bandwidth of 0 to 10 kHz. Off-lineaveraging by a PDP 12 computer produced thecervical or lumbar somato-sensory evoked poten-tials which were plotted and analysed.Twenty-nine subjects aged between 19 and

66 years had all suffered traumatic complete orpartial spinal cord section at the levels shownin fig 1. Two additional female patients aged 19and 21 years had hysterical paraplegia and twoothers had cauda equina lesions, which weretraumatic in one and a sequel to laminectomyperformed at age three months for spina bifidain the other. Recordings were technically unsatis-factory from one other subject and he is excludedfrom this study. All but one were receiving in-patient treatment and all but one had recoveredfrom spinal shock. Nineteen had received theirinjuries in the preceding eight months and theremainder two of 24 years earlier. They weretaking their usual medication and were in goodgeneral health without pressure sores or urinaryor other infection.The recordist had no knowledge of the level

or extent of the lesion and this information wasnot available until after the initial analysis ofthe results.

Results

Potentials from a normal subject are shown onthe left side of fig 2. The upper trace shows thecortical potential recorded from C3-Fz (10-20system) after stimulation of the median nerve.The second trace, from electrodes placed over

Patint with CompleteLI SpInal Lesion(Sugical Removal ofSpinal Segnmnt)

Fig 2 On the left are

potentials from a normalsubject following mediannerve (upper 2 traces) andtibial nerve stimulation.The traces on the right arefrom a subject who hadsurgical removal of a

damaged cord segment atLI. He had a normal NIOpotential from L4 but nospinal or cortical potentials.Median nerve stimulationgave normal potentials.

10 20m sec

30



j.com/

J Neurol N


eptember 1980. D

ownloaded from


Spiinal cord potentials in traumatic paraplegia and quadriplegia

the 2nd cervical vertebra (Cv2) and at Fz, showsthe response to median nerve stimulation; fournegative peaks, Nil, N13, N14, and N20, namedaccording to their polarity and approximatelatency are evident. We follow the conventionof denoting negativity at the "active" electrodeby an upward reflection in the figures. The N20potential is a cortical event and is best recordedby an electrode placed near the cortical sensoryarea as shown in the top trace. Tibial nervestimulation at the popliteal fossa produced adouble negative wave with peaks N1O and N14from over the lumbar vertebrae, while only N14is seen at LI or lower thoracic levels. The Cv2-Fz electrodes recorded a low amplitude, longduration potential of uncertain origin which hasbeen studied by Jones and Small21 but not in detailby us. The scalp electrodes signal the arrival ofthe afferent volley at the cortex and the time delayfrom NIO is a measure of transmission time alongthe cord and through the sub-cortical structures.

The latencies and amplitudes of all thesepotentials were measured in all subjects andcompared with recordings taken from 31 normalsubjects in the case of the cervical potentials,and from 18 normal subjects in the case of thelumbar potentials. The amplitudes were muchmore variable than the latencies, but even thelatencies showed considerable fluctuation due todifferent lengths of arms and legs. To overcomethe problem of limb length, the latencies of thesecond and subsequent peaks were measuredfrom the initial peak (N9) which indicates thearrival of the afferent volley in the brachialplexus.9 The N9 potential was recorded by elec-trodes on the 7th cervical vertebra.

Spinal potentials caudal to the lesionThe tibial nerve contains afferents for L4 and5, SI and 2 and it is these spinal segments whichgive rise to the N14 lumbar potential. Afterspinal injury above this level the function ofthe caudal segments of the cord is severely im-paired. Is there a similar impairment of theability of the cord to produce the N14 potential?

Fig 2 shows the potentials recorded from apatient with a complete lesion of the cord at LI.(The level given is the lowest segment with nor-mal function.) The cervical and cortical responsesto median nerve stimulation were normal but, asexpected, no cortical or cervical potentials fol-lowed tibial nerve stimulation. A normal caudaequina or NIO potential was present at the L4electrode, but no spinal or N14 potential at LI.This suggests that the isolated cord cannot pro-duce potentials, but this patient had evidence of

35.

30*

z1V 2-0

10

5 10 15 20ms

Fig 3 Amplitude and latency of NJO and N14potentials following tibial nerve stimulation havebeen plotted from patients with lesions at ThJO andabove. The ellipses represent the 90% limits fornormal subjects. The one high amplitude N14potential was recorded from a subject in spinalshock.

damage to the lower segments. The patientsuffered root pains at the level of the transec-tion and, in an effort to control them, the scartissue in the spinal canal had been removedallowing an opportunity to confirm the complete-ness of his lesion. After operation, however,he lost a number of reflexes including his reflexbladder; it was concluded that further damagehad been produced in the caudal segment of cordperhaps by -interference with the blood supply.The absence of N14 in this patient was consi-dered as due to this damage.Other patients showed potentials of normal

amplitude and latency from the caudal segment.In fig 3 the amplitude and latency of NIO andN14 for normal subjects have been plotted asellipses so that the amplitude and latency pointsof 90% of normal subjects would be expectedto fall within them. Fig 3 also shows the valuesfrom 19 patients who had two identifiable poten-tials and who had lesions at ThiO or above. Itis of interest,that the one point which lies beyondthe normal range was obtained from a patientin spinal shock five days after a complete lesionat Th6.Fig 4 shows the potentials from patients with

lesions at ThlO and below. We have recorded apotential in a subject with a lesion as low as L3,but in other cases no potentials could beobtained.We conclude that the isolated lumbar cord can

produce normal potentials even though it lacksall descending control. Smaller isolated segments,

825



j.com/

J Neurol N


eptember 1980. D

ownloaded from


E M Sedgwick, E El-Negamy and H Frankel

,Uv

5 10 15 20ms

Fig 4 Amplitude and latency of the lumbarpotentials in patients with lesions at ThJO and below.One patient had a lesion at L3 and gave normalpotentials. Six other patients, listed on the graph,had no recognisable potentials. CE=Cauda equinalesion.

due to lesions below ThlO, were less able toproduce recordable potentials. This could bebecause the generating segments were themselvesinvolved in the lesion which usually extends overa few segments. The patients with cauda equinalesions showed no responses at all which was tobe expected.

Spinal potentials rostral to the lesionIn fig 5 the ellipses represent the 90% limits ofthe amplitude and latency of the cervical N11,N13 and N14 potentials of normal subjects. Thelatencies were computed from the peak of theN9 brachial plexus potential, which overcomesvariability due to different arm lengths. Theplotted points are from 17 patients with lesionsat Th5 or below. The cervical potentials ap-peared normal (fig 2), but the points do not liewithin the expected limits so a more detailedstatiscal analysis was undertaken and the results

0 2 4 6 8 10 12

ma

Fig 5 The ellipses represent the 90%0/. limits of theamplitude and latency of the Nil, N13 and N14cervical somatosensory evoked potentials in normalsubjects. Latencies were measured from the peak ofthe N9 potential. The plotted points are fromsubjects with lesions at Th5 or below (including onewith a cauda equina lesion). The final point, notencompassed by an ellipse, is of the N20 response.

are given in the table.The table shows that the N9 potential was of

the same amplitude in the paraplegic patients asin the normal subjects, but the latency wassignificantly shorter, 7-96 ms compared to 9 0 ms(p=0-001). We believe this to be due totemperature differences in the laboratories wherethe studies were performed. The patients withspinal injury who cannot thermoregulateefficiently, were in a warm environment at about24°C while the normals were studied inuniversity laboratories which were not so wellheated.

Table The cervical somatosensory potential from 31 normal subjects and 17 with spinal trauma below theThl segment. The amplitudes of each component are given in ,V and the latentcy to peak from thepreceding peak or from stimulus in the case of N9. Three normal subjects had no N9 component and onehad no N14

N9 N9-NlI Nll-NI3 N13-Nl4 N14-N20 N9-N20

N9 NIlI N13 N14 N20Amp Lat Anmp Lat Amip Lat Amp Lat Lat Amp Lat

Mean 0-80 9 0 1 8 2-44 2-2 1 35 1 30 1 22 5-68 1 30 10-71Normal subjects SD 0 40 0-80 0-7 0 50 0-80 0 44 0 90 0-34 0-89 0 40 0.91

n 28 28 31 28 31 31 31 31 30 31 28Mean 0-63 7-96 1 28 2 12 2 17 2-39 0 95 1-36 4-35 1 22 10 16Spinal trauma SD 0-20 0 34 0 33 0 54 0-54 0-63 0 50 0-32 0-67 0 43 0-76n 17 17 17 17 17 17 17 17 17 17 17t 1*15 5-06 2-89 2 02 0 14 6-72 1*48 1*38 5-36 0-64 2-08df 43 43 46 43 46 46 46 46 45 46 43p NS 0 031 0 01 0-05 NS 0 091 N3 NS 0 001 NS NS

826



j.com/

J Neurol N


eptember 1980. D

ownloaded from


Spinal cord potentials in traumatic paraplegia and quadriplegia

The amplitude of NIl was slightly less in thepatients with spinal injury (1P28AV comparedwith 1-8gV), and the latency of the peak wasalso less (2-12 ms compared with 2-44 ms). Theplotted points of NIl lie within or just to theleft of the NIl ellipse.The delay between NI1 and N13 was pro-

longed by about 1 ms (2-39 compared with 1-35 msin the normal group). The amplitudes of the N13potential were the same in both groups. Theother difference between the groups was ashorter delay in the spinal injury group betweenthe N14 and N20 potentials (4-35 ms comparedwith 5-68 ms). However, in spite of these changesthe total conduction time from the N9 potentialof the brachial plexus to the N20, the firstcortical event, was the same in both groups.The lengthening of the Ni 1-N13 delay in

spinal injury was an unexpected finding. NI1was thought to be generated in the dorsal hornof the cord,6 but our more recent work suggeststhat NIl may be presynaptic and that N13 is apostsynaptic dorsal horn neurone potential.Whatever the true origin of these potentialsthere seems to be some slight alteration in theirgenerators after spinal injury some distancecaudal to them. One explanation was that thedegenerative and regenerative processes whichfollow trauma could produce a temporarychange lasting perhaps seven months from thetime of injury. This idea was tested by comparingthe Ni I-N13 delay in 12 subjects who hadsuffered trauma in the previous seven months,and in whom the degeneration process may nothave been totally completed, with five subjectswhose trauma occurred at least 18 months be-fore recording, allowing ample time for com-pletion of the degeneration process. The N ll-N13delay was 2-40 ms for the recent trauma groupand 2-16 ms for old trauma. Both groups wereprolonged compared to normal and althoughthe recently injured group showed a longer delaythan the older group, the difference was notstatistically significant.

Patients with cervical lesionsThere were 13 patients with lesions at Thi orabove. Lesions in this region disturbed thecervical potentials as can be seen from fig 6.Definite potentials could not be identified inseven patients. This may have been because thepotentials were absent, but there were technicaldifficulties with recording from patients withhigh lesions; muscle relaxation was difficult toobtain (probably because they needed to use theiraccessory muscles of respiration even when lying

"v

5 10 15 20ms

Fig 6 Plots of amplitude and latency of the cervicalsomatosensory evoked potential of patients withlesions at and above Thl. Latencies were measuredfrom the point of stimulation (wrist) so that theN9 ellipse is included. This makes the 90% limits forNil, N13 and NJ4 broader to accommodatedifferences in arm length. Note the absence of laterpotentials in three cases and of the delayed potentialsin two others.

still), and the EMG signal from these musclesmay have swamped the potentials.Of the five patients from whom potentials

were recorded all had a normal N9. In threepatients with complete lesions at Cv5 and Cv6there was a recognisable potential of a latencyapproporiate for Ni1, but no subsequentpotentials. One subject with a partial lesion atThl, and another with a partial lesion at Cv7,gave small but recognisable Nll potentials butsubsequent waveforms were delayed and the N20appeared at 22 5 ms, whereas it is normally seenat 19-7 ms.

Spinal shockOnly one subject with spinal shock was studied,but the observations seem worth recording.Fig 7 shows the lumbar potentials recorded fivedays after an injury which produced a completelesion at Th6. At the time of recording he wascompletely areflexic and atonic with no bladderfunction or anal sphincter tone. Twelve dayslater the second recording was done at a timewhen reflex function was beginning to return.The N10 was of equal size on both occasionsbut the N14 was large when first tested butwithin normal limits on the second occasion.

Discussion

The N14 Lumbar potential is a segmental

827



j.com/

J Neurol N


eptember 1980. D

ownloaded from


828

Tibial nervestimulation

A

1_,

2

2

3

5ms

Fig 7 Lumbar somatosensory evo

following tibial nerve stimulation ispinal shock. The potentials in A wthe period of spinal shock, five dajsuffered a complete Th5 lesion. Tobtained 12 days later at a time wiwas beginning to return.

response6 7; Ertekin22 has reamplitude but similar potentialelectrodes. The available evidethe potential is postsynaptic anis generated in the dorsal hornwhere low threshold afferent filpostsynaptic neurones in theknown to be under suppraspin,alter their responsiveness in anicord is cooled or cut.23 So it I

find a normal, Lumbar N14 powith spinal cord section (parnseveral segments rostral tosegments. This implies thatneurones are able to respondincoming volley. This is in contmechanisms whose function isafter injury.Lack of N14 from patients u

ThiO, and in some cases at Thto the lesion extending severaldown the cord from the mainThis is a common finding at autno anatomical knowledge of t]lesions in our patients.

E M Sedgwick, E El-Negamy, and H Frankel

The presence of a potential corresponding toNil recorded from Cv7 after median nervc

2 V stimulation has been interpreted as evidence that+ NllI is generated in the cord at segmental level,';

probably by dorsal horn synapses. Jones,9 how-ever, suggested that Nll is a dorsal root or dorsalcolumn potential and that N13 might begenerated by segmental neurones. Further care-ful study of potentials in patients with highcervical lesions will help to resolve this point butgood recordings from these subjects aretechnically difficult to obtain.

Finding minor abnormalities in the cervicalB potentials in patients with lesions at Th5 and

below was unexpected. The rostral part of thecord is normally subject to some influences fromcaudal regions via the propriospinal system. Thechanges in potentials could be due to a re-organisation of synaptic inputs to neurones whichlost some of their normal synapses due to de-generation. Such changes in experimentalanimals have been shown anatomically by

15 25 Illis,24 25 and functionally by Devor.26 Anoked potentials alternative explanation is that the degenerativein a patient in process in the cord alters the micro-environmentvere obtained during and produces slowed conduction in the remainingys after the subject dorsal column fibres. Patients with old lesions'hose in B were (over 18 months), however, still showed an in-hen reflex function creased delay between NIl and N13. The Nll-

Ni3 delay cannot be explained by the relativelyearly appearance of Ni 1 in the spinal injury

corded a higher patients, for the N9-Nl3 time also is prolonged.using intrathecal N14 is thought to be a potential generated innce suggests that the thalamus and one would not expect ad that it probably shortened delay between it and N20 which is theof the spinal cord first cortical potential. The N9-NI4 time isbres synapse. The prolonged in the spinal injury patients but thedorsal horn are shortened delay from N14-N20 allows some

al control and to " catching up " so that N20 is at approximatelyimals if the spinal the same latency in both groups when measuredwas interesting to from N9.tential in patients Whatever the explanation of these changes, ittial or complete) is clear that disruption of the sensory pathway

the generator from the legs results in changes in the pathwaythe dorsal horn from the arms. Perot"' noted attenuation of thenormally to an median nerve cortical evoked potential for one

'rast to the motor week after spinal section, but he made no com-severely disturbed ment on latency changes (which appear quite

marked in his figure 28-1 ) nor offered anyvith lesions below explanation.O, could be due It is known from animal work that the caudalsegments up and spinal cord normally influences the activity ofpoint of impact. rostral segments, an action often called the-opsy but we have Schiff-Sherrington phenomenon.27 The alteredhe extent of the cervical potentials may be a reflection of a similar

lprocess. This would be an interesting area for



j.com/

J Neurol N


eptember 1980. D

ownloaded from


Spinal cord potentials in traumatic paraplegia and quadriplegia

further investigation of the adaptation of theremaining CNS to a well-defined lesion.

Clinical applicationCortical potentials may be useful to determinethe completeness of a lesion. The re-appearanceof a cortical potential may herald clinical re-covery, but their continued absence or markedabnormality is a bad prognostic sign. Corticalpotentials were recorded in only three out of11 cases in this series with clinically in-complete lesions, so the technique can demon-strate only the incompleteness of a lesion ratherthan prove the totality of transection. We wouldstress, however, that the experimental protocolwas not the most appropriate to detect lowamplitude, long duration and delayed potentials.Also we only stimulated the nerves on one sidewith a modest strength. Some simple modifica-tion to tihe technique may increase the yield.Nevertheless, at present the evoked potentialtechnique is not as effective at demonstratingsurviving ascending axons as are the techniquesof Dimitrijevic et a128 for demonstrating descend-ing ones. Further studies are needed for it isbecoming clear that clinically complete lesionsare not necessarily neurophysiologicaHly complete.Demonstration of the survival of a few pathwaysmay have therapeutic implications in the future.

Cortical and spinal potentials could havelimriited value in checking hysterical andmalingering patients. Two hysterics, who had aparaplegia convincing enough to warrant theiradmission to a spinal injuries unit, both hadentirely normal potentials. While such a findingdoes not exclude some degree of spinal injury,the discrepancies in neurophysiological andclinical findings could be helpful. Thesetechniques could also be of use in assessingpatients early after injury especially those withother injuries which preclude their co-operationin clinical examinations.The potentials have no role in determining the

level of a lesion. They could have a role howeverin assessing the longitudinal extent, especially oflower thoracic and high lumbar lesions wherethe integrity of the badder reflex mechanisms isirn doubt. Also they could help in the assessmentof high cervical lesions, when the survival ofphrenic motoneurones is in doubt, and a decisionon respiration by phrenic nerve stimulation hasto be made.29

Finally, evoked potentials will play an im-portant role in assessing spinal cord functionduring any therapy which is aimed at restoring'spinal cord structure and function.

E El-Negamy was supported by the ArabRepublic of Egypt.

References

1 Pool JL. Electrospinogram (ESG), Spinal cordaction potentials recorded from a paraplegicpatient. J Neurosurg 1945; 3:192-8.

2 Sawa H. Spontaneous electrical activities obtainedfrom human spinal cord. Folia Psychiatr Neurollap 1947; 2:165-79.

3 Magladery JW, Porter WE, Park AM, TeasdaleRD. Electrophysiological studies of nerve andreflex activity in normal man. IV: The twoneurone reflex and identification of certain actionpotentials from spinal roots and cord. Bull JohnHopkins Hosp 1951; 88:499-519.

4 Matthews WB, Beauchamp M, Small DG.Cervical somato-sensory evoked response in man.Nature 1974; 252:230-2.

5 Cracco RQ. Spinal evoked response: peripheralnerve stimulation in man. ElectroencephalogrClin Neurophysiol 1973; 35:379-86.

6 El-Negamy E, Sedgwick EM. Properties of aspinal somatosensory evoked potential recordedin man. J Neurol Neurosurg Psychiatry 1978;41:762-8.

7 Dimitrijevic MR, Larsson LE, Lehmkuhl D,Sherwood A. Evoked spinal cord and nerve rootpotentials in humans using a non invasive record-ing technique. Electroencephalogr Clin Neuro-physicl 1978; 45:331-40.

8 Delbeke J, McComas AJ, Kopec SJ. Analysis ofevoked lumbosacral potentials in man. J NeurolNeurosurg Psychiatry 1978; 41:293-302.

9 Jones SJ. Short latency potentials recorded fromthe neck and scalp following median nervestimulation in man. Electroencephalogr ClinNeurophysiol 1977; 43:853-63.

10 Singer JM, Russell GV, Coe JE. Changes inevoked potentials after experimental cervicalspinal cord injury in the monkey. Exp Neurol1070; 29:449-61.

l l Croft TJ, Brodkey JS, Nulser FE. Reversiblespinal cord trauma. A model for electricalmonitoring of spinal cord function. J Neurosurg1972; 36:402-6.

12 D'Angelo CM, Van Gilder JC, Taub A. Evokedpotentials in experimental spinal cord trauma. JNeurosurg 1973; 38:332-6.

13 Deeke L, Tator CM. Neurophysiological assess-ment of afferent conduction in the injured spinalcord of monkeys. J Neurosurg 1973; 39:65-74.

14 Martin SH, Bloedel JR. Evaluation of experi-mental spinal cord injury. J Neurosurg 1973; 39:75-81.

15 Cusik JF, Mkylebust JB, Larson SJ, Sances A.Spinal cord evaluation of cortical evoked re-sponses. Arch Neurol 1979; 36:140-3.

16 Perot PL. Somatosensory evoked potentials in theevaluation of patients with spinal cord injury.

829



j.com/

J Neurol N


eptember 1980. D

ownloaded from


830

In: Morley TP, ed. Current Controversies inNeurosurgery. Philadelphia: WB Saunders, 1976:160-7.

17 Perot PL. The clinical use of somato-sensoryevoked potentials in spinal cord injury. ClinNeurosurg 1973; 20:367-81.

18 Rowed DW, McLean JA, Tator CM. Somato-sensory evoked potentials in acute spinal cordinjury. Prognostic value. Surg Neurol 1978; 9:203-10.

19 Small DG, Matthews WB, Small M. Subcorticalsomatosensory evoked potentials in multiplesclerosis. Electroencephalogr Clin Neurophysiol1977; 43:536-7.

20 El-Negamy E, Sedgwick EM. Delayed cervicalsomatosensory potentials in cervical spondylosis.J Neurol Neurophysiol Psychiatry 1979; 42:238-41.

21 Jones SJ, Small DG. Spinal and sub-corticalevoked potentials following stimulation of theposterior tibial nerve in man. ElectroencephalogrClin Neurophysiol 1978; 44:299-306.

22 Ertekin C. Comparison of the human evokedelectrospinogram recorded from the intrathecal,epidural and cutaneous levels. Electroencephalogr

EM Sedgwick, E El-Negamy, and H Frankel

Clin Neurophysiol 1978; 44:683-90.23 Wall PD. Dorsal Horn Electrophysiology. In:

Iggo A, ed. Handbook of Sensory Physiology.Berlin: Springer, 1973: 253-70.

24 Illis LS. Experimental model of regeneration inthe central nervous system I Synaptic changes.Brain 1973; 96:47-60.

25 Illis LS. Experimental model of regeneration inthe central nervous system. II: The reaction ofglia in the synaptic zone. Brain 1973; 96:61-8.

26 Devor M, Merrill EG, Wall PD. Dorsal horncells that respond to stimulation of distant dorsalroots. J Physiol (Lond) 1977; 270:519-32.

27 Granit R. The basis of motor control. London:Academic Press, 1970: 180-3.

28 Dimitrijevic MR, Faganel J, Lehmkuhl D,Sherwood A. Motor control in man with spinalcord injury. In: Desmedt JE, ed. Motor controlin man: suprasegmental and segmental mechan-isms. Progress in Clinical Neurophysiology, vol 8.Basel: Kargel; in press.

29 Glenn WWL, Holcomb WG, Shaw RK, HoganJF, Holschuh KR. Long term ventilatory supportby diaphragm pacing in quadriplegia. Ann Surg1976; 183:566-77. P

rotected by copyright. on July 12, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.43.9.823 on 1 Septem

ber 1980. Dow

nloaded from


Documents

Spinal cord potentials in traumatic paraplegia and ... · JournalofNeurology, Neurosurgery, andPsychiatry, 1980, 43, 823-830 Spinal cordpotentials in traumatic paraplegia andquadriplegia