Somatosensory · tion timein somatosensorypathwayswemea-suredthepeakinglatencies ofErb'spointN9, cervical N13, scalp far-field P9 and P14, and contralateral parietal N20 potentials.

3'ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:301-308

Somatosensory evoked potentials aftermultisegmental upper limb stimulation indiagnosis of cervical spondylotic myelopathyDomenico Restuccia, Massimiliano Valeriani, Vincenzo Di Lazzaro, Pietro Tonali,Frangois Mauguiere

AbstractRadial, median, and ulnar nervesomatosensory evoked potentials (SEPs)were recorded, with non-cephalic refer-ence montage, in 38 patients with clinicalsigns of cervical myelopathy and MRIevidence of spondylotic compression ofthe cervical cord. Upper limb SEPs areuseful in spondylotic myelopathy becauseSEPs were abnormal in all patients for atleast one of the stimulated nerves andSEP abnormalities were bilateral in allpatients but one. Reduction of the ampli-tude of the N13 potential indicating asegmental dysfunction of the cervicalcord was the most frequent abnormality;it occurred in 93.4%, 84-2%, and 64-5% ofradial, median, and ulnar nerve SEPsrespectively. A second finding was thatthe P14 far-field potential was more sen-sitive than the cortical N20 potential toslowing of conduction in the dorsal col-umn fibres. The high percentage of N13abnormalities in the radial and medianrather than in the ulnar nerve SEPs cor-related well with the radiological com-pression level, mainly involving theC5-C6 vertebral segments. Therefore therecording of the N13 response is areliable diagnostic tool in patients withcervical spondylotic myelopathy and P14abnormalities, though less frequent, canbe useful in assessing subclinical dorsalcolumn dysfunction.

(7 Neurol Neurosurg Psychiatry 1994;57:301-308)

Department ofNeurology, CatholicUniversity, Rome,ItalyD RestucciaM ValerianiV Di LazzaroP TonaliDepartment ofFunctional Neurology,Hopital Neurologique,Lyon, FranceF Maugui6reCorrespondence to:Dr D Restuccia,Department of Neurology,Policlinico A Gemelli,Lgo A Gemelli 8, 00168Rome, Italy.Received 18 December 1992and in revised form on30 March 1993 and26 May 1993.Accepted 28 May 1993

Spondylotic changes of the cervical spine arethe most common cause of cervical myel-opathy or radiculopathy.' The diagnosis ofcervical spondylotic myelopathy is based onthe combination of signs suggesting involve-ment of long pathways (spastic paraparesisassociated with a variable degree of lowerlimb ataxia), and dysfunction of motor andsensory neurones in the cervical grey matter.2Nevertheless, sensory motor and reflexchanges in the upper limbs can be missing3and, in the absence of sensory deficits, cervi-cal spondylotic myelopathy can be confusedwith other degenerative diseases such as amy-otrophic lateral sclerosis.

Magnetic resonance imaging (MRI) of thecord can show several types of signal abnor-malities at the level of cord compression,45but gives no information on cervical cord dys-finction in cervical spondylotic myelopathy.

Therefore, it is clinically relevant to developcomplementary investigations for assessingcord dysfunction at the cervical level.

Somatosensory evoked potentials (SEPs)have been used to disclose abnormalities ofascending sensory pathways in cervicalspondylotic myelopathy. Previous studieswith cephalic reference montages showed thatdorsal column dysfunction can be demon-strated in 43% to 100% of patients by lowerlimb SEPs, in 57% to 74% of patients byulnar nerve SEPs, and in 24% to 59% ofpatients by median nerve SEPs.6-'0 Abnormallower limb SEPs are, however, of no value forlocalising the dysfunction at the cervical level;moreover, in previous studies upper limbSEPs rarely showed abnormalities in patientswithout sensory deficits.

Non-cephalic reference recordings of SEPsallow a separate analysis of the dorsal hornN13 response and of the P14 potential, thelatency of which reflects the transit time ofthe ascending volley up to the lower brain-stem level. Abnormalities of the N13 poten-tial have been found in diseases affecting thecentral grey matter"-'5 and in a selectedpopulation of patients with cervical spondy-lotic myelopathy but with normal sensation.'6Moreover, prolonged P14 latencies in relationto a conduction slowing in dorsal columnswere found in focal cervical cord lesions aswell as in multiple sclerosis.'7-20 Our studyconsiders the question whether the diagnosticyield of SEPs in cervical spondyloticmyelopathy can be improved by assessingseparately dorsal column and dorsalhorn responses to stimulation of median,radial, and ulnar nerves in a large populationof patients with cervical spondyloticmyelopathy.

Patients and methodsPATENTSWe studied 38 patients (mean age 56 (range37-77) years; 29 men) with cervical spondy-losis confirmed by MRI. All patients showedspastic weakness of the lower limbs, brisklower limb tendon jerks, and a unilateral orbilateral Babinski sign. Mild weakness andwasting in upper limbs or reduction orabsence of at least one of the upper limb ten-don reflexes was found in 22 patients. Nopatient complained of pain or paraesthesia inthe upper limbs. Joint and touch sensation inthe upper limbs was impaired in 17 patients;there was segmental pain and temperaturesensation impairment in the upper limbs of

301 on January 24, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.3.301 on 1 M

arch 1994. Dow

nloaded from

http://jnnp.bmj.com/

Restuccia, Valeriani, Di Lazzaro, Tonali, Mauguiere

Table 1 Normative data

Limit ofnormal valuesMean(SD) Range (mean (3 SD))

Radial nerve SEPs (16 control subjects):Interpeak intervals (ms) P9-N13 4 4(0 3) 3-8-5 5-3

P9-P14 5.4(0.3) 4-7-5-8 6-3P14-N20 4 7(0 5) 3-8-5-7 6-2P9-N20 10-1(0-5) 9-10 9 11-6

Ni 3/P9 amplitude ratio (logarithmic values) 0-291(0-125) 0-079-0-602 -0-0840-82*

Median nerve SEPs (20 control subjects):Interpeak intervals (ms) P9-N13 4 3(0 4) 3-6-5-4 5-5

P9-P14 5 4(05) 4-2-6-6 6-9P14-N20 4-9(0 5) 3-8-5-9 6-4P9-N20 10-3(0 5) 9-2-11-2 11-8

N13/P9 amplitude ratio (logarithmic values) 0-274(0 09) 0-125-0-505 0-0041*

Ulnar nerve SEPs (16 control subjects):Interpeak intervals (ms) P9-N13 4-4(0-4) 3-8-5-2 5-6

P9-P14 57(05) 4-9-6-7 72P14-N20 4-7(0 5) 3-6-5-4 6-2P9-N20 10-4(0-5) 9-4-11-3 11.9

N13/P9 amplitude ratio (logarithmic values) 0-252(0-116) 0-102-0-602 -0 0090-81*

*Corresponding absolute value.

13 patients. In nine patients there was an

increased signal on T2-weighted MRI scans

of the cord at the cervical level; in one patientMRI showed a segmental atrophy of the cer-

vical cord.Brain MRI, and CSF and blood tests were

also performed to exclude other pathologicalconditions such as multiple sclerosis or vita-min B-12 deficiency. All patients had nerve

conduction and concentric needle EMGexaminations. Motor and sensory nerve con-

duction velocity studies were performed inthe upper limbs with standard techniques.2'EMG activity in upper limb muscles was con-

sidered abnormal when there were fibrilla-tions and positive sharp waves in two or more

areas of the muscle under study. Upper limbnerve conduction velocities were within nor-

mal limits in all patients. Concentric needleexamination showed abnormalities confinedto upper limb muscles in 22 patients.

SEP RECORDING PROCEDUREFor SEP recording, patients lay on a couch ina warm and semidarkened room. Stimuli (0-3ms square pulses) were delivered at the rateof 5 Hz with skin electrodes (cathode proxi-mal) at motor threshold intensity for medianand ulnar nerve SEPs and at three times thesensory threshold intensity for radial nerve

SEPs. Stimulations were delivered at thewrist for median and radial nerves and abovethe ulnar groove at the elbow for the ulnarnerve. The filter bandpass was 10-3000 Hz(-3 dB at cut Qff point, 6 dB per octave); theanalysis time was 50 ms with a bin width of98,us. Samples with excess interference were

automatically edited out of the average. Twoaverages of 2048 or 4096 trials each were

obtained and drawn out by the computer on

an X-Y plotter. The recording electrodes(impedance below 5 kohm) were placed inthe supraclavicular fossa (Erb's point), over

the spinous process of the 6th cervical verte-bra (Cv6) and in the parietal scalp regionscontralateral and ipsilateral to stimulation.The Erb's point electrode was referred to Fzand the parietal scalp electrodes to the shoul-der contralateral to the stimulated side. For

the recording of the cervical N13 potential weconnected grid 1 of the amplifier to the Cv6electrode and grid 2 to an electrode locatedimmediately above the thyroid cartilage. Thiselectrode site is referred to in the text and fig-ures as anterior cervical (AC). The rationalefor this Cv6 to AC montage has been dis-cussed in detail in previous studies'3 22; firstlyit records the activity generated by the trans-verse dipolar source of the Ni 3 potential witha maximal amplitude. Secondly it permits theselective assessment of the amplitude of thedorsal horn response as it does not recordpotentials generated above the foramenmagnum and tends to cancel the N 1I poten-tial, which reflects the ascending volley in thedorsal columns23-26 and is picked up by bothCv6 and AC electrodes.2224 A stationary P9potential, reflecting the positive front of theafferent volley in cervical roots,27 is alsopicked up by both Cv6 and AC electrodesand the waveform resulting from the algebraicsubtraction of the larger AC P9 from thesmaller Cv6 P9 is made of a small negative-positive diphasic deflection preceding thecervical N13."322

NORMATIVE DATATo match our control subjects for age withpatients with cervical spondylotic myelopathywe selected from the laboratory normativedata'6 those collected in normal subjects over40 years of age (20 subjects (age range 40-82,mean 52-7, seven men) for median nerveSEPs, 16 subjects (age range 40-82, mean53-2, seven men) for radial and ulnar nerveSEPs), see table 1. For assessing the conduc-tion time in somatosensory pathways we mea-sured the peaking latencies of Erb's point N9,cervical N13, scalp far-field P9 and P14, andcontralateral parietal N20 potentials. Toeliminate interindividual variations related toarm length, P9-N13, P9-P14, P14-N20, andP9-N20 interpeak intervals were also calcu-lated.The amplitude of the N13 potential was

assessed by calculating the N13/P9 amplituderatio using the Cv6-AC traces'3-'6 as shown infig 1.



j.com/

J Neurol N


arch 1994. Dow

nloaded from


Evoked potentials after upper limb stimulation in diagnosis ofcerical spondylotic mydopathy

Figure 1 SEPs evoked bystimulation of (A) nghtradial, (B) median, and(C) ulnar nerves in a46-year-old control subject.Two traces obtained on twosuccessive runs of2024(for the medianand ulnar nerves) or 4096(for the radid nerve) aresuperimposed. Erb-Fz =Erb's point referred to aforehead electrode (Fz);the P14 potential pickedupfrom the frontalelectrode is injected as an"N14" negatit*y; Cv6-AC = spinous process ofthe sixth cervical vertebra,referred to an anteriorcerical electrode (AC);Parc-Sh = partetalelectrode contralateral tothe stimulus, referred to theshoulder contralateral tothe stimulus (noncephalicreference electrode); Parc-Par, = parietal electrodecontralateral to thestimulus, referred to theparietal electrode ipsilateralto the stimulus; using thisderivation the subcorticalfar-feld potentialspreceding the N20 responseare canceled out. (D)Method usedfor measuringthe amplitude ofP9 (1)and N13 (2) as well as theN13/P9 amplitude ratio inthe Cv6-AC trace.

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Pari

A

0

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Pari

C

-11 gv

Erb-Fz

Cv6-AC

Parc-Sh

Par,-Pari

Time (ms)30

R

4AI-N9<< N14>>

N13N9P4

N20

0Time (ms)

.1 V

N9<< N14>

D

Cv6-AC: N13N9 P14

I N

N20

0Time (ms)

30

2 RV

0Time (ms)

ResultsAll patients but one showed abnormal SEPson both sides, at least for one of the stimu-lated nerves. In only one patient SEP abnor-malities were limited to the right side. Thelatency of the N9 and P9 responses were

always within normal limits, as well as theP9-N13 interpeak interval and the N13latency, when the N13 response was identifi-able. Abnormalities in N20 potential were

always associated with an abnormal P14potential.

N13 ABNORMALITSReduced or absent Ni 3 potential was found

after stimulation of at least one nerve in allpatients but two. N13 potential was absent orreduced in 71/76 of radial, 64/76 of median,and 49/76 of.ulnar nerve SEPs.

Radial, median, and ulnar spinal SEPabnormalities were combined as follows. N13was abnormal for the three tested nerves in49 upper limbs (64-5%). These abnormalitieswere bilateral in 23 patients and unilateral inthree. When N13 was abnormal for only twoof the three tested nerves, the pattern con-sisted in all cases of abnormal N13 in radialand median nerve SEPs with normal ulnarnerve spinal SEPs. This was found in 15upper limbs (19-7%). These abnormalities

30

30

303

L;

on January 24, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.3.301 on 1 March 1994. D

ownloaded from



Table 2 Correlation between clinical signs and distribution ofSEP abnormalities

Abnormal Normal

Clinical signs Radial, median, ulnar Radial, median Radial In all three nerves

N13 responseLower motor neuron signs 31/49 upper limbs 8/15 upper limbs 3/7 upper limbs 0/5 upper limbsPain and temperature 16/49 upper limbs 4/15 upper limbs 3/7 upper limbs 0/5 upper limbs

segmental hypesthesiaJoint and touch hypesthesia 22/49 upper limbs 4/15 upper limbs 1/7 upper limbs 4/5 upper limbs

P 14 responseLower motor neuron signs 19/30 upper limbs 3/6 upper limbs 5/11 upper limbs 15/29 upper limbsPain and temperature 8/30 upper limbs 6/6 upper limbs 4/11 upper limbs 5/29 upper limbs


N 20 responseLower motor neuron signs 18/29 upper limbs 3/4 upper limbs 3/8 upper limbs 18/35 upper limbsPain and temperature 7/29 upper limbs 4/4 upper limbs 4/8 upper limbs 7/35 upper limbs


were bilateral in four patients and unilateralin seven. When N13 was abnormal for onlyone nerve, only radial nerve SEPs wereaffected. This was the situation in sevenupper limbs (9.2%). These abnormalitieswere bilateral in two patients and unilateral inthree.A normal N13 was found whatever the

stimulated nerve in five upper limbs; thesenormal spinal responses were bilateral in two

N14~~II%

t~~!r~~\.

!Y~--I~xi-~~ ~ ~ ~ ~ ~ ~ ~

patients and found after stimulation of theleft side in one patient. Table 2 shows corre-lations betweeq N13 findings and clinicalsigns in the corresponding upper limb.

P14 ABNORMALITIESThe P14 potential was absent or delayed afterstimulation of at least one nerve in 24patients (63-1%). Abnormal P14 was foundin 30, 36, and 47 of the 76 radial, median,

TrTtxX. .e r~~~~~~~'T

,, i,. .: . . . ;. ...... i.~~4~

1ieY:]E''''_ a.fZi-

Figure 2 (A) Delayed or absent P14 with normal latency N20 after stimulation of the median nerve. Right and left median nerve SEPs are illustratedwith the same abbreviations as in fig 1. The peak latencies ofN9 and P9 responses were within normal limits. The N13 potential was absent afterstimulation of the right median, and reduced after stimulation of the left median nerve. Whereas the N20 latencies, as weUl as the P9-N20 intervals, were

stiU within normal limits, the P9-P14 interval was slighdy increased after stimulation of the right median nerve and the P14 was unrecognisable afterstimulation of the left median nerve. (B, C) The cervical spinal cord MRI (Tl- and T2-weighted) showed a stenosis of the cervical cord with spondyloticcord compression at the C3-C4, C4-C5, and C5-C6 levels.



j.com/

J Neurol N


arch 1994. Dow

nloaded from


Evoked potentials after upper limb stimulation in diagnosis of cervical spondylotic myelopathy

A B C

+1 1 AV

N9:<<Nl4>>

N 13 ?

pg P14_A

Erb-Fz

Cv6-AC

Parc-Sh

~N0: Parc-ParN20

lime (ms)30 0

-1J2 [tV

N9<<N14>>

N13 ?

P14P9

N20

Time (ms)

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Pari

30 0

<<Nl4>> +I2VN9 + 2 RV

_fN13 ?

P9P14

N20

Time (ms)30

Figure 3 Association of normal scalp SEPs and abnormal spinal SEPs after stimulation of all three nerves. (A) Right radial, (B) median, and (C)ulnar nerve SEPs. Same presentation as infigs 1 and 2. The peaking latencies ofN9, P9, P14, and N20 responses, as well as P9-P14 and P14-N20interpeak intervals were within normal limits. N13 potential was absent or reduced after stimulation of all three nerves.

D E

Figure3contd (DE) Cervical spinal cordMR(lad2wihe)hwd

compression at CS-C6 level

and ulnar nerve SEPs respectively; P14 was

abnormal for the three tested nerves in 30upper limbs (39 5%). These abnormalitieswere bilateral in 13 patients and unilateral infour. In all cases where the P14 was abnormal

for only two of the three tested nerves thepattern consisted of abnormal P14 in medianand ulnar nerve SEPs with normal radialnerve scalp SEPs. This was found in sixupper limbs (7.9%). This type of SEP abnor-mality was bilateral in one patient and unilat-eral in four. When P14 was abnormal for onlyone nerve, only ulnar nerve SEPs were

affected. This was found in 11 upper limbs(145%), bilaterally in five patients, and uni-laterally in one.Normal radial, median, and ulnar nerve

P14 potentials were found in 29 upper limbs(38- 1%). This finding was bilateral in 14patients and unilateral in one. Table 2 gives

the correlations between P14 findings andclinical signs in the corresponding upperlimb.

N20 ABNORMALITIESN20 potential was absent or delayed afterstimulation of at least one nerve in 22patients (57-9%). The N20 potential was

abnormal in 29/76 of radial, 33/76 of median,and 41/76 of ulnar nerve SEPs. Radial,median, and ulnar scalp N20 abnormalitieswere combined as follows. N20 was abnormalfor the three tested nerves in 29 upper limbs(38'1%). These abnormalities were bilateralin 13 patients and unilateral in three. WhenN20 was abnormal for only two of the threetested nerves the pattern consisted in all cases

of abnormal N20 in median and ulnar nerve

SEPs with normal radial nerve N20. This was

found in four upper limbs (5'3%), bilaterallyin one patient, and unilaterally in two. WhenN20 was abnormal for only one nerve, onlyulnar nerve SEPs were affected. This was thesituation in eight upper limbs (10-5%). Theseabnormalities were bilateral in three patientsand unilateral in two.The N20 potential was normal for all

tested nerves in 35 upper limbs (46%). Thisfinding was bilateral in 16 patients and uni-

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Pari

.

305


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



A B

N9

N13 ?

P14P9

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Pari

N20

0 30Time (ms)

j1 1 RV

+12 AV

N9 <<N 14»>N13 ?

P14P9

Erb-Fz

Cv6-AC

Parc-Sh

Parc-PariN20

0 30Time (ms)

C+I 2 [V

,\<14>> ?

N9

N13P14?

I-

N20

30Time (ms)

0

Figure 4 Association ofabnormal scalp and spinal SEPs. (A) Right radial, (B) median and (C) ulnar nerve SEPs. Same presentation as infigs 1-3.The peak latencies ofN9 and P9 responses were within normal limits. The N13 potential was absent after stimulation of the radial and median nerves,

but still normal after stimulation of the ulnar nerve. The peak latencies of the P14 response, as well as the P9-P14 intervals, were delayed after stimulationof the radial and median nerves and not evaluable after stimulation of the ulnar nerve.

lateral in three.Thus a normal N20 coexisted with a

clearly abnormal P14 for at least one of the

stimulated nerves in five patients. Fig 2 illus-trates this pattern; it was found in six ulnar

and three median nerve SEPs and one radial

nerve SEP.Table 2 gives the correlations between N20

findings and clinical signs in the correspond-ing upper limb.

COMBINATIONS OF SCALP AND SPINAL SEP

U_ l l . _?= i . ., :.M~ANORAXESAbnormalities of scalp and spinal SEPs,defined respectively as abnormal N20 or P14

..,..onthe one hand and abnormal N13 on the

i_I,:,:'other,for at least one of the stimulated

nerves, were combined in three different pat-terns, found bilaterally in 36 of our 38

patients and distributed as follows: 13patients (34%) had normal scalp SEPs withabnormal spinal responses (26/76 upper

limbs; fig. 3); 21 patients (55%) showed

abnormalities of both scalp and spinal SEPs(42/76 upper limbs; fig. 4); in two patients

;.lonly (5%) normal segmental spinal responseswere associated with abnormal scalp SEPs onboth sides (4/76 upper limbs).Of the two remaining patients one had nor-

mal SEPs on one side with abnormal spinaland normal scalp SEPs on the other; in thesecond patient scalp and spinal SEPs were

both abnormal on one side whereas only

spinal SEPs were affected on the other side.

Thus more than one third of our patients

: 'with an isolated amplitude abnormality of thesegmental spinal N13 response would have

been considered as having normal upper limb

SEPs if recorded with a conventional frontalreference montage, which does not allow a

selective assessment of this variable.14

Figure 4 contd (D) Cervical spinal cord MRI (Tl-weighted) showed a stenosis of thecervical canal with compression at C4-C5. (E) T2-weighted scan showed an CLIICAL CORRELATIONS (TABLE 3)

intramedullary hyperintensity (arrow). Abnormal P14 and N20 were significantly

Erb-Fz

Cv6-AC

Parc-Sh

Parc-Par1

306

_


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Evoked potentials after upper limb stimulation in diagnosis of cervical spondylotic myelopathy

correlated with loss or reduced joint andtouch sensation (X2, p < 005), but neitherwith segmental pain and temperature hyper-aesthesia, nor with absent or reduced tendonreflexes in the corresponding upper limb.Abnormal N13 was correlated with absent

or reduced tendon reflexes in the correspond-ing upper limb (X2, p < 0-05), but neitherwith segmental pain and temperature hypes-thesia, nor with lost or reduced joint andtouch sensation.

DiscussionThis study shows that the recording of bothspinal and scalp SEPs in patients with cervi-cal spondylotic myelopathy can reveal twomain types of cord dysfunction at the cervicallevel.The most frequent SEP abnormality was

the absence or reduction of the spinal N13potential, which was evident in all but twopatients, at least for one of the tested nerves.

This abnormality was often associated withnormal scalp SEPs, the loss of N13 with nor-

mal P14 and N20 potentials having beenfound in 28 of 76 upper limbs. A loss of N13with normal P14 and N20 potentials isknown to occur in lesions of the cervical grey

matter leaving dorsal columns unaffected,such as cervical syrinxes11-14 and intramedul-lary tumours.'5 Therefore, the isolated abnor-mality of the N13 potential in patients withcervical spondylotic myelopathy is thought toreflect an anatomical damage limited to thecervical grey matter. In a previous study, we

proposed that this finding could reflect a

decreased blood supply to cervical cord fromanterior spinal artery.16 The 11 patients ofthis previous study have been selected fromour cervical spondylotic myelopathy seriesbecause they presented normal sensations.The present study, which includes both theseprevious patients and other unselected cases,confirms that the finding of the isolated lossof the spinal N13 is not unusual in patientswith cervical spondylotic myelopathy. A firstreason for our high rate of N13 abnormali-ties, compared to the lower detection rates ofupper limb SEPs in earlier studies,-10 is thatwe did not use a neck to forehead montagefor the recording of the N13 potential. The

cervical response recorded with this montageis an amalgam of the cervical N13 and scalpP14 potentials1728 and is not significantly

affected by selective lesions of the cervicalgrey matter.'3'4 In fact, earlier SEP studies inpatients with cervical spondylotic myelopathyfailed to reveal any selective loss of the N13potential in association with normal scalppotentials. In one of these reports,8 some

upper limb abnormalities in SEPs were

reported as reflecting a "proximal root/cordlesion" when showing abnormal cervicalresponses. The N20 potential was also abnor-mal, however, in all of these patients. Anotherreason for the rate of N13 abnormalities inour patients pertains to the use of multiplenerve stimulation, including the radial nerve,which corresponds to the more frequentlyinvolved level.Even though correlations between clinical

signs and N13 abnormalities are not simple,abnormalities of the N13 potential are

strongly related to the compression level as

revealed by radiography and MRI investiga-tions. Thus the higher rate of reduced N 13potentials in radial and median nerve SEPsthan in ulnar nerve SEPs (93-5% and 84-6%v 65-3%) correlates well with the involvementof the C5 level as shown by MRI in all ofour patients with this SEP abnormality.Moreover, 31 of 38 patients had no MRI evi-dence of cord compression below the sixthcervical vertebra. Conversely, in the twopatients showing a normal N13 response afterstimulation of all three nerves on both sides,the cord compression was located at theC2-C3 levels.

Electromyographic signs of denervation inupper limb muscles were present in many ofour patients. Although we cannot exclude a

radicular damage, it seems more likely thatthis finding was due to the involvement of theanterior hom cells; our patients in fact alwayspresented a clear involvement of the spinallong tracts without radicular pain or paresthe-siae.

Another frequent SEP feature in our

patients -was the abnormality of the P14potential, which was found in 37-6% ofradial, in 41% of median, and in 62-8% ofulnar scalp SEPs. It is generally agreed thatthe P14 scalp far-field potential is generatedin the brainstem tracts of the ascending lem-niscal pathways close to the cervicomedullaryjunction."7-20 Thus both increased P9-P14interval or absent P14 with a normal P9, as

found in all patients with abnormal P14 inthis series, directly reflect time dispersion of

Table 3 Sensation, reflexes, and N13, P14 and N20 SEPs

Clinicalfeatures in corresponding upper limb

Pain and temperature Joint and touchTendon reflexes sensation sensation

Normal Reduced Normal Reduced Normal Reduced

N 13:Normal (5 upper limbs) 5* 0* 5 0 1 4Abnormal (71 upper limbs) 29* 42* 48 23 44 27

P 14:Normal (29 upper limbs) 14 15 24 5 29* 0*Abnormal (47 upper limbs) 20 27 29 18 16* 31*

N 20:Normal (35 upper limbs) 17 18 27 8 33* 2*Abnormal (41 upper limbs) 17 24 26 15 12* 29*

*p<0.05 (2).



j.com/

J Neurol N


arch 1994. Dow

nloaded from



the ascending volley in the dorsal columns atthe cervical level. In earlier studies, whichused a cephalic reference montage, scalp SEPabnormalities in cervical spondylotic myel-opathy were constantly represented by theabolition or latency delay of the parietal N20response. The evaluation of abnormal dorsalcolumn function provided by a prolongedN20 latency or N13-N20 interval is indirect,and includes the transit times in medial lem-niscus and thalamocortical fibres; thereforeany intracranial conduction slowing in thesomatosensory pathways can cause an N20abnormality. Our findings show that the eval-uation of the scalp P14 far-field providesmore reliable information on dorsal columndysfunction than that of the parietal N20potential. Firstly, we found a greater percent-age of abnormalities for P14 than for N20. Infive patients, N20 was still recognisable andof normal latency, ,whereas the P14 was com-pletely abolished or delayed (fig 2). This isthe reason why the percentage of patientswith subclinical dorsal column dysfunctionsrevealed by upper limb SEPs (normal jointand touch sensation but abnormal scalpSEPs) was greater in our series (32%) than inthose reported earlier (Yu and Jones8 18%;Veilleux and Daube9 17%). The abolition orthe latency delay of the P14 potential with anormal N20 can be explained by the resyn-chronisation of the ascending volley in theintracranial somatosensory pathways.Secondly, by contrast with what has beenseen for the N20 potential,8 we never found anormal P14 after stimulation of an upperlimb presenting with lost or reduced joint andtouch sensation.

Concerning the correlation between scalpSEP abnormalities and compression level,our data confirm that ulnar scalp SEPs aremore often involved than median scalpSEPs.89 As the C6 and C7 myelomeres,receiving the sensory inputs from the mediannerve, are located at the level of the fifth andsixth cervical vertebrae, it might be expectedthat patients with normal median nerve andabnormal ulnar nerve scalp SEPs would havea compression level below the sixth cervicalvertebra. This assumption was not verified inour series, in which patients with isolatedabnormality of the ulnar nerve P14 showedeither a clear cut compression at the fifth andsixth cervical vertebra, or a multilevel com-pression involving myelomeres rostral toC8-TI. A greater sensitivity of the longerfibres to compression could account for thehigher rate of abnormal ulnar nerve SEPscompared with median nerve SEPs in cervicalspondylotic myelopathy. Due to the small dif-ference in length between the fibres incomingfrom median and ulnar nerves, however, thisexplanation seems unlikely, even though itmight hold for the high detection rate of tibialnerve SEPs in cervical spondylotic myel-opathy.8 9

1 Brain WR, Northfield D, Wilkinson M. The neurologicalmanifestations of cervical spondylosis. Brain 1952;75:187-225.

2 Adams C. Cervical spondylotic radiculopathy and myel-opathy. In: Vinken PJ, Bruyn GW, eds. Handbook ofclinical neurology, vol. 26: Injuries of the spine and spinalcord, part 2. Amsterdam: North Holland, 1976:97-112.

3 Ferguson RJ, Caplan LR. Cervical spondylotic myel-opathy. Neurol Clin North Am 1985;3:373-82.

4 Al-Mefty 0, Harkey LH, Middleton TH, Smith RR, FoxJL. Myelopathic cervical lesions and spondylotic lesionsdemonstrated by Magnetic Resonance Imaging. J7Neurosurg 1988;68:217-22.

5 Mehalic TF, Pezzuti RT, Applebaum BI. MagneticResonance Imaging and cervical spondylotic myel-opathy. Neurosurgery 1990;26:217-27.

6 El-Negamy E, Sedgwick EM. Delayed cervical somato-sensory potentials in cervical spondylosis. Y NeurolNeurosurg Psychiatry 1979;42:238-41.

7 Ganes T. Somatosensory conduction times and periph-eral, cervical and cortical evoked potentials in patientswith cervical spondylosis. J Neurol Neurosurg Psychiatry1980;43:683-9.

8 Yu YL, Jones SJ. Somatosensory evoked potentials incervical spondylosis. Correlation of median, ulnar andposterior tibial nerve responses with clinical and radio-logical findings. Brain 1985;108:273-300.

9 Veilleux M, Daube JR. The value of ulnar somatosensoryevoked potentials (SEPs) in cervical myelopathy.Electroencephalogr Clin Neurophysiol 1987;68:415-23.

10 Perlik SJ, Fisher MA. Somatosensory evoked responseevaluation of cervical spondylitic myelopathy. MuscleNerve 1987;10:481-9.

11 Emerson RG, Pedley TA. Effect of cervical spinal cordlesions on early components of the median nervesomatosensory evoked potentials. Neurology 1986;36:20-6.

12 Urasaki E, Wada S, Kadoya C, Matsuzaki H, Yokota A,Matsuoka S. Absence of spinal N13-P13 and normalscalp far-field P14 in a patient with syringomyelia.Electroencephalogr Clin Neuiophysiol 1988;71:400-4.

13 Restuccia D, Mauguiere F. The contribution of mediannerve SEPs in the functional assessment of the cervicalspinal cord syringomyelia. Brain 1991;114:361-79.

14 Mauguiere F, Restuccia D. Inadequacy of the foreheadreference montage for- detecting abnormalities of thespinal N13 SEP in cervical cord lesions. Electro-encephalogr Clin Neurophysiol 199 1;79:448-56.

15 Ibafiez V, Fischer G, Mauguiere F. Dorsal horn and dor-sal column dysfunction in intramedullary cervical cordtumours. A somatosensory evoked potential study.Brain 1992;115:1209-34.

16 Restuccia D, Di Lazzaro V, Valeriani M, Tonali P,Mauguiere F. Segmental dysfunction of the cervicalcord revealed by abnormalities of the spinal N13 poten-tial in cervical spondylotic myelopathy. Neurology1992;42: 1054-63.

17 Mauguiere F, Ibafiez V. The dissociation of early SEPcomponents in lesions of the cervicomedullary junction:a cue for routine interpretation of abnormal cervicalresponses to median nerve stimulation. Electro-encephalogr Clin Neurophysiol 1985;62:406-20.

18 Yamada T, Ishida T, Kudo Y, Rodnitzky R, Kimura J.Clinical correlates of abnormal P14 in median SEPs.Neurology 1986;36:765-71.

19 Garcia-Larrea Lo, Mauguiere F. Latency and amplitudeabnormalities of the scalp far-field P14 to median nervestimulation in multiple sclerosis. A SEP study of 122patients recorded with a noncephalic reference montage.Electroencephalogr Clin Neurophysiol 1988;71:180-6.

20 Turano G, Jones SJ, Miller DH, de Bolay GH, Kakigi R,McDonald WI. Correlation of SEP abnormalities withbrain and cervical cord MRI in multiple sclerosis. Brain1991;114:663-81.

21 Daube JR. Nerve conduction studies. In: Aminoff MJ, ed.Electrodiagnosis in dinical neurology. New York: ChurchillLivingstone, 1986:265-306-.

22 Desmedt JE, Nguyen TH. Bit-mapped colour imaging ofthe potential fields of propagated and segmental sub-cortical components of somatosensory evoked potentialsin man. Electroencephalogr Clin Neurophysiol 1984;58:481-97.

23 Desmedt JE, Cheron G. Central somatosensory conduc-tion in man: neural generators and interpeak latencies ofthe far-field components recorded from neck and rightor left scalp and earlobes. Electroencephalogr ClinNeurophysiol 1980;50:382-403.

24 Desmedt JE, Cheron G. Prevertebral (oesophageal)recording to subcortical somatosensory evoked poten-tials in man: the spinal P13 component and the dualnature of the spinal generators. Electroencephalogr ClinNeurophysiol 1981;52:257-75.

25 Anziska BJ, Cracco RQ. Short latency SEPs to mediannerve stimulation: comparison of recording methodsand origin of components. Electroencephalogr ClinNeurophysiol 1981;52:531-9.

26 Lueders H, Lesser R, Hahn J, Little J, Klem G.Subcortical somatosensory evoked potentials to mediannerve stimulation. Brain 1983;106:341-72.

27 Yamada T, Kimura J, Nitz DM. Short latency somatosen-sory evoked potentials following median nerve stimula-tion in man. Electroencephalogr Clin Neurophysiol 1980;48:367-76.

28 Mauguiere F. Les potentiels evoques somesthesiques cer-vicaux chez le sujet normal: analyse des aspects obtenusselon le siege de l'electrode de' rference. Rev Electro-encephogr Neurophysiol Ctin 1983;13:259-72.



j.com/

J Neurol N


arch 1994. Dow

nloaded from


Somatosensory · tion timein somatosensorypathwayswemea-suredthepeakinglatencies ofErb'spointN9, cervical N13, scalp far-field P9 and P14, and contralateral parietal N20 potentials.

Documents