vemp

The LaryngoscopeLippincott Williams & Wilkins, Inc.© 2006 The American Laryngological,Rhinological and Otological Society, Inc.

Improving Vestibular Evoked MyogenicPotential Reliability by using a BloodPressure Manometer

Robby Vanspauwen, MSc; Floris L. Wuyts, PhD; Paul H. Van de Heyning, MD, PhD

Objective/Hypothesis: To improve the reliabilityof vestibular evoked myogenic potentials (VEMP), wepropose a feedback method making use of a readilyavailable blood pressure manometer with inflatablecuff to control the sternocleidomastoid muscle (SCM)contraction. Study Design: Prospective study. Materi-als and Methods: The feedback method for SCM con-traction consisted of subjects pushing with their jawagainst the hand-held inflated cuff to generate a spec-ified cuff pressure. This pressure level was monitoredby subject and investigator. First, we tested in agroup of healthy subjects whether there was a learn-ing or fatigue effect during four successive SCM con-traction runs when making use of the feedbackmethod. Then, we investigated the mean rectifiedvoltage (MRV) values of the SCM. Next, we examinedthe hypothesis that the MRV values of the SCM con-traction measured before and simultaneously withthe VEMP test were equal when using this feedbackmethod. Finally, we compared the VEMP amplitudevariability in two circumstances: with and withoutthe feedback method. Results: There was no learningor fatigue effect, and the MRV values measured be-fore the VEMP were not significantly different fromthose measured during the VEMP test. The VEMPamplitude variability was significantly lower whenapplying the feedback method than when no feed-back method was used. Conclusion: This feedbackmethod improves significantly the VEMP amplitudereliability. Key Words: Feedback method, vestibularevoked myogenic potential, saccule, vestibular sys-tem, methodology.

Laryngoscope, 116:131–135, 2006

INTRODUCTIONVestibular evoked myogenic potentials (VEMPs)

serve as a diagnostic tool for saccular function. VEMPs,measured from the contracted sternocleidomastoidmuscle (SCM), are inhibitory myogenic potentials thatresult from ipsilateral saccular stimulation by means ofloud clicks or tone bursts. Currently, the outcome isconfined to be binary (i.e., present or absent for a givenstimulus). However, the peak latencies and the peak-to-peak amplitude of the p13 to n23 waveform complex canbe important parameters when determining left-rightdifferences in the VEMP results. Because the VEMPamplitude is highly dependent on the strength of theSCM contraction, it is essential to take this tonic con-traction into account to obtain a correct interpretationof left-right amplitude differences.1,2,3 A large amountof variability is caused by the fact that the SCM con-traction levels are different between and even withinsubjects. Simultaneous monitoring of the mean rectifiedvoltages (MRV) of the SCM contraction allows adequatemaintenance of this contraction, which is required for acorrect interpretation of the VEMP waveform. However,this method requires a special experimental set-up,which is usually not readily available in standard clin-ical settings because most commercial devices forevoked potentials do not allow simultaneous electro-myography (EMG) recording when the VEMP test isperformed. The authors propose a feedback method tokeep the SCM contraction at a specified level based onthe use of a blood pressure manometer, which is indeedavailable in all clinics.

Study Design and ObjectivesThis study protocol consists of two parts. The pur-

pose of part 1 is to verify the hypothesis that the MRVvalues of the contracted SCM, measured just before theVEMP test, can be considered as indicators for the tonicSCM contraction measured during the VEMP test. Wefirst tested whether there was a significant learning orfatigue effect by comparing the MRV values of fourconsecutive runs (2 runs just before the VEMPs and 2runs simultaneously with the VEMPs) when making

From the Vestibular Function Laboratory, University Department ofOtorhinolaryngology, Head and Neck Surgery, Antwerp University Hospi-tal (R.V., F.L.W., P.H.V.D.H.), and the Department of Biomedical Physics,University of Antwerp (F.L.W.), Antwerp, Belgium.

Editor’s Note: This Manuscript was accepted for publication Septem-ber 14, 2005.

Study funded by the University Department of Otorhinolaryngology,Head and Neck Surgery of Antwerp University Hospital and the Univer-sity of Antwerp.

Send Correspondence to Dr. Floris Wuyts, Vestibular Function Lab,Department of ENT, Antwerp University Hospital, University of Antwerp,Wilrijkstraat 10, 2650 Edegem, Belgium. E-mail: [email protected]

DOI: 10.1097/01.mlg.0000187405.57567.ae

Laryngoscope 116: January 2006 Vanspauwen et al.: Feedback Method to Improve VEMP Reliability

131

use of the feedback method. Furthermore, we comparedthe MRV values of the SCM contraction measured be-fore the VEMP test with the MRV values of the SCMcontraction measured simultaneously with the VEMPs.

The purpose of part 2 was to show that the VEMPamplitude variability is significantly lower when mak-ing use of our proposed feedback method in comparisonwith measurements obtained without feedback method.We compared the VEMP amplitude variability of mea-surements obtained with and without feedback method.

MATERIALS AND METHODS

Part 1Subjects. All subjects in this study (parts 1 and 2) had

normal hearing sensitivity (20 dB hearing level, American Na-tional Standards Institute, 1996) and no history of vestibular,neurologic, or inner ear pathology. All subjects gave their in-formed consent, approved by the institutional review board com-mittee of the Antwerp University Hospital. All tests were per-formed in a sound-proof booth. In part 1, we tested 10 healthysubjects (5 men, 5 women) ranging in age from 21 to 34 (average26) years.

Equipment and electrode location. The experimentswere performed with two identical and connected auditory evokedpotential systems (Nicolet Viking, Viasys Healthcare, Consho-hocken, PA), equipped with EMG software. Before placing theAg/AgCl surface electrodes (Blue Sensor, Ambu, Denmark), theskin was cleansed and scrubbed with an impedance lowering gel.The active electrodes were placed on the medial portion of thecontracted SCM muscle belly, the reference electrode on the up-per part of the sternum but below the attachment points of theSCM muscles, and the ground electrode on the forehead. In everysubject, the distances between the active electrodes and the ref-erence electrode was kept constant for each side. Skin imped-ances were less than 5 k�.

MRV measurement. Per side, MRV values were measuredduring two runs of 30 seconds just before the VEMP test as wellas during two runs of 30 seconds simultaneously with the VEMPtest. For each run, one average MRV value was calculated as anindicator of the SCM contraction. These averaged MRV valueswere used in the statistical analysis.

VEMP recording and acoustic stimulation. The re-sponses were obtained from each side separately using tonebursts of 500 Hz (95 dB nHL, rise/fall time � 2 ms, plateau time� 2 ms, repetition rate � 5.1 Hz), which were delivered unilat-erally with insert earphones (Nicolet, Viasys Healthcare, Consho-hocken, PA, model TIP-300). The acoustically evoked VEMP re-sponses were amplified, band-pass filtered (10 Hz–1.5 kHz), andaveraged. Each side was subjected to two runs of 150 tone bursts,which were averaged. Analysis sweep time was 100 ms. TheVEMP response in this study (part 1 and 2) was only consideredas being reliable if the following conditions were fulfilled: 1) thep13 and n23 peaks were reproducible in each run, and 2) theVEMP amplitude was 1.5 times greater than the residual noise inthe trace preceding the p13. All results fulfilled these conditions.

SCM contraction. To contract the SCM during the MRVmeasurement and VEMP test, a feedback method was appliedbased on the use of a blood pressure manometer with inflatablecuff (Welch-Allyn, Skaneateles Falls, NY). The subjects wereseated in a comfortable chair with their back against the backsupport of the chair (the head was not positioned against a sup-port). The cuff was inflated to a standard preset level of 20 mm Hgso that a cushion was formed. When testing the left side (i.e.,measuring the MRV values or performing the VEMP test on thatside), the subject had to flex the head approximately 30 degrees

forward and rotate it approximately 30 degrees to the right side,taking care not to bend the head to the right shoulder. Whileholding the cuff between right hand and jaw (Fig. 1), the subjectpushed with his or her head against the hand-held cuff to gener-ate a cuff pressure of 40 mm Hg. In this way, a controlled con-traction in the left SCM was achieved. The subject was clearlyand continuously instructed to maintain the cuff pressure at aconstant level (small variations of 2 mm Hg above and below 40mm Hg were accepted). All MRV and VEMP measurements inpart 1 were performed at this cuff pressure of 40 mm Hg.

Statistics. Statistical analysis in this study was performedusing SPSS 12.0 (SPSS Software, SPSS, Chicago, IL). For eachanalysis, a significance level of 5% was adopted. Given the paireddesign (i.e., subjects were compared with themselves) of part 1,each statistical analysis took this repeated measures design intoaccount by using an univariate analysis of variance with subjectas a “random factor.”

Part 2Subjects. In this part, we tested two groups of normal

subjects. The first group (further denoted as group 1) was com-posed of 15 healthy subjects (7 men, 8 women) ranging in agefrom 21 to 37 (average 23 ) years. The second group (group 2)consisted of 12 healthy subjects (6 men, 6 women) ranging in agefrom 21 to 32 (average 24) years.

MRV measurement. Only in group 1 was the feedbackmethod applied. The same procedure with respect to the SCMcontraction (using the feedback method) as already described inpart 1 was used. Cuff pressures of 30, 40, and 50 mm Hg wereapplied. At each of these cuff pressures, the MRV values wererecorded before the VEMP test during 1 run, separately for bothsides.

VEMP recording and acoustic stimulation.GROUP 1. At each cuff pressure (30, 40, and 50 mm Hg), 500

Hz tone bursts (100 dB nHL, rise/fall time � 2 ms, plateau time� 2 ms, repetition rate � 5.1 Hz) were presented. For each side,two runs of 100 tone bursts were performed. The results of bothruns were averaged, providing the final response, from which thepeak-to-peak amplitude (p13–n23) and absolute latencies (p13and n23) were calculated.

GROUP 2. In this group, we performed the VEMP protocolthat we applied previously (i.e., without making use of a feedback

Fig. 1. Feedback method to control the contraction of the rightsternocleidomastoid muscle (SCM). In this picture, the read out ofthe manometer is turned to the reader to clarify the testing method.When performing the test, this read out is turned to the subject.


132

method). To test the left side, the subjects were asked to bend thehead 30 degrees forward and rotate it 30 degrees to the right sideas they had to push against the hand of the investigator, whichwas placed against the right jaw of the subject. This resulted inthe contraction of the left SCM. The subjects were instructed tomaintain a constant contraction. Tone bursts of 500 Hz were used(100 dB nHL, rise/fall time � 2 ms, plateau time � 2 ms, repeti-tion rate � 5.1 Hz). For each side, three runs of 100 tone burstswere presented and averaged.

RESULTS

Part 1There was no significant evolution in MRV levels

during the four subsequent recordings, neither for the leftSCM (F[3,27] � 0.55, P � .66) nor for the right SCM(F[3,27] � 2.34, P � .10). This indicated that there was nolearning or fatigue effect.

For the next analysis, we pooled for each subject theMRV values of both runs performed before the VEMP testand both runs performed simultaneously with the VEMPtest (these averages are further denoted as “pre” and “perMRV data,” respectively). There was no significant differ-ence between the pre MRV data and per MRV data, nei-ther for the left SCM (F[1,9] � 0.11, P � .75) nor for theright SCM (F[1,9] � 2.42, P � .15).

The mean difference between the pre and per MRVvalues per subject was �3.3 �V (SD � 14.9 �V) for thepooled data of left and right sides. These differences werenot significantly different from 0 (one sample t test: t[19]� �1.001; P � .33). The 95% prediction interval (mean �2SD; mean �2 SD) of the differences between pre and perMRV values was (�29; 29) �V.

Part 2In this part, we compared the VEMP amplitude vari-

ability in both groups. The average absolute VEMP am-plitude difference between the left and right side was usedas the measure for repeatability. To calculate the VEMPamplitude difference of the measurements obtained withthe feedback method, the data from the three applied cuffpressures were pooled. In the group tested without feed-back, this absolute difference yielded 104 (SD � 72, se �21) �V with a range from 6 to 282 �V. In the group testedwith the feedback method, this value was 34 (SD � 25, se� 6) �V, ranging from 0 to 106 �V. These values weresignificantly different (2 samples t test: t[11.8] � 3.3; P �.007). The difference in VEMP amplitude variability inboth groups is depicted in Figure 2.

DISCUSSIONIt has already been shown that the VEMP amplitude

highly depends on the strength of the SCM contrac-tion.1,2,3 There are also several other factors, such as thestimulus intensity, muscle anatomy, subcutaneous fatlayer, skin impedance, and location of active surface elec-trodes, that jeopardize the interpretation of the VEMPoutcome.4,5 Anatomic factors causing variability cannot beruled out, but by applying appropriate methodologies, theinfluence of specific factors can be reduced as much aspossible. Several of these methods have already been de-

scribed (e.g., concerning the placement of the electrodesand the stimulus settings).5,6,7,8

For a correct interpretation of left-right differences inVEMP amplitudes, the amount of SCM contraction has tobe taken into account. This issue is usually only dealt within research papers rather than in clinical papers. Themain reason for this fact is that standard auditory evokedpotential systems, which are used in most clinical prac-tices to measure the VEMP response, do not usually offerthe possibility to measure the MRV generated by thecontracted SCM, and if they do, the equipment seldomallows simultaneous measurements of MRV and VEMP.

In this study, we present an alternative methodbased on the use of a blood pressure manometer withinflatable cuff and an auditory evoked potential system.The advantage of this method is that the equipment (i.e.,a blood pressure manometer), is standard and readilyavailable in all clinics. The cuff method is based on the factthat a contraction of the SCM is needed to squeeze the cuffbetween hand and jaw. However, this is only possiblewhen the cuff is inflated with a basal pressure of 20 mmHg. Squeezing the cuff will generate an increase in pres-sure, which can be monitored on the manometer by subjectand investigator. All reported cuff pressures in this paperare expressed as additional pressures with respect to thisbasal cuff pressure. Depending on the cuff geometry, thisbasal cuff pressure can be different (in our laboratory, weused 20 mm Hg).

In part 1 of this study, we showed first that there wasno significant training or fatigue effect, either for left orright sides, by comparing the MRV values of four subse-quent measurement runs (2 runs before the VEMP testand 2 runs simultaneously with the VEMP test performedat identical cuff pressures). Based on these findings, weconclude that making use of the manometer to contractthe SCM is a good method to keep the SCM contraction ata constant level throughout the test. Next, we demon-strated that there was no significant systematic differencebetween the average MRV values of both runs performedbefore the VEMP test (pre MRV value) and the averageMRV values of both runs performed simultaneously with

Fig. 2. Boxplot of difference in VEMP amplitude variability betweenthe test protocol with (group 1) and without (group 2) feedbackmethod for the sternocleidomastoid muscle (SCM) contraction.


133

the VEMP test (per MRV value). Therefore, when it ispossible to measure the MRV values but impossible tomeasure these simultaneously with the VEMPs, the MRVvalues measured before the VEMP test can be applied asindicators for the SCM contraction performed during theVEMP test when the feedback method is applied at anidentical cuff pressure. The 95% prediction interval of thedifferences between pre and per MRV values was (�29;29) �V, which is relatively small. The small differencesbetween pre and per MRV values were possible becausewe tolerated little variations of 2 mm Hg above and belowthe specified cuff pressure. We showed that under similarconditions (i.e., within the same person and for the sameelectrode placement) a relatively good correlation existsbetween the SCM contraction levels (represented by theMRV values) and the cuff pressures.9

In part 2, we showed that, when making use of thefeedback method to control the SCM contraction, theVEMP amplitude variability is significantly lower whenapplying the feedback method (34 �V, SD � 25 �V, se � 6�V) than when no feedback method is used (104 �V, SD �72 �V, se � 6 �V). Therefore, it is possible to determinemore reliable left-right amplitude differences in VEMPresults when the feedback method is applied.

CONCLUSIONTo improve the reliability of the VEMP test, which is

inherently dependent on the SCM tension, we propose thefollowing recommendations.

1. When it is possible to measure EMG values but impos-sible to record them simultaneously with the VEMPmeasurements, the blood pressure manometer with in-

flated cuff allows for a control of the SCM contraction bymeans of the MRV measures. First, record, just prior tothe VEMP test, the average MRV value of the SCMcontraction in one side for a certain cuff pressure. Next,measure the MRV value for the SCM contracted on theopposite side. Consequently, the investigator deter-mines which cuff pressures result in the same averageMRV values for both sides. These cuff pressures, whichcan be different for both sides but result in identicalaverage MRV values, are then applied during the sub-sequent VEMP test. This is the first goal of the feedbackmethod. The second goal consists of the maintenance ofa constant SCM contraction level throughout the VEMPrecording. Both methods largely improve the VEMPreliability. The procedures of the feedback method are,together with the instructions for the subjects, summa-rized in Table I.

2. When no EMG equipment is available, we still proposeto apply the feedback method to keep the SCM contrac-tion constant and comparable for left and right sides. Assuch, the VEMP reliability will also improve.

AcknowledgmentsThe authors thank I. Naoemova, F. Scherf, K.

Vanderkelen, W. Moeyersons, and Dr. P. ten Broecke fortheir assistance and technical support.

BIBLIOGRAPHY1. Colebatch JG, Halmagyi GM, Skuse NF. Myogenic potentials

generated by click-evoked vestibulocollic reflex. J NeurolNeurosurg Psychiatry 1994;57:190–197.

2. Lim CL, Clouston P, Sheean G, Yianniks C. The influence ofvoluntary EMG activity and click intensity on the vestib-

TABLE I.Summary of Different Steps of Feedback Method with Instructions for Subjects.

Procedure of Cuff Feedback Method Instructions for Subjects

Step 1 Subject takes place in chair with back against back support andhead free to move. Electrodes are applied and connected by theinvestigator.

Step 2 Investigator shows how the subject has to bend his/her head(flexed 30 degrees forward and rotated 30 degrees to the rightside to contract the left SCM).

“Bend your head a little forward and turn itsideways”

Subject assumes the desired head position. “Don’t bend your head sideways to your shoulder”

Step 3 Investigator inspects that active electrode is located properly on themuscle belly of the left SCM while head of subject is rotated.

Step 4 Subject holds the inflated cuff between right hand and yaw. “Do not squeeze the cuff with your fingers. Holdthe cuff with a flat hand between hand and yaw”

Step 5 Subject pushes with head against hand-held cuff to generate aspecified cuff pressure.

“Push with your head so as to generate andmaintain a pressure of 40 mm Hg* on the readout”

Step 6 Investigator initiates the MRV measurement while inspecting thecuff pressure.

“Maintain the specified cuff pressure”

Step 7 Steps 2 to 6 are repeated for the right SCM.

Step 8 Investigator determines which cuff pressure measured from theright SCM results in comparable values as when contracting theleft SCM.

Step 9 Investigator initiates the VEMP recording with the previouslydetermined cuff pressures for both sides while inspecting the cuffpressure.

“Maintain the specified cuff pressure”

*This pressure has to result in a clearly visible sternocleidomastoid muscle (SCM) contraction and has to be comfortable for the subject. The magnitude ofthe pressure is depending on the geometry of the blood pressure manometer. VEMP � vestibular evoked myogenic potential; MRV � mean rectified voltage.


134

ular click evoked myogenic potential. Muscle Nerve 1995;18:1210–1213.

3. Robertson DD, Ireland DJ. Vestibular evoked myogenic poten-tials. J Otolaryngol 1995;24:3–7.

4. Ochi K, Ohashi T, Nishino H. Variance of vestibular-evokedmyogenic potentials. Laryngoscope 2001;111:522–527.

5. Sheykholeslami K, Murofushi T, Kaga K. The effect of sterno-cleidomastoid electrode location on vestibular evoked myo-genic potential. Auris Nasus Larynx 2001;28:41–43.

6. Cheng PW, Huang TW, Young YH. The influence of clicksversus short tone bursts on the vestibular evoked myogenicpotentials. Ear Hear 2003;24:195–197.

7. Ferber-Viart C, Duclaux R, Colleaux B, Dubreuil C. Myo-genic vestibular-evoked potentials in normal subjects: acomparison between responses obtained from sterno-mastoid and trapezius muscles. Acta Otolaryngol 1997;117:472–481.

8. Welgampola MS, Colebatch JG. Characteristics of tone burst-evoked myogenic potentials in the sternocleidomastoidmuscles. Otol Neurotol 2001;22:796–802.

9. Vanspauwen R, Naoemova I, Van der Kelen K, et al. Simple‘cuff pressure based’ feedback procedure for increasing thereliability of the VEMP response amplitude. Presented atthe XXIII Barany Society Conference, Paris, 2004.


135

vemp

Documents