The clinical reliability of vestibular evoked myogenic potentials

Vestibular evoked myogenic potential (VEMP) testing has gained popularity as a diagnostic modality in otolaryngology and audiology. To maximize the utility of this test, examiners need the availability of ideal test settings and reliable norms. We conducted a prospective study of 8 subjects with no history of neurotologic symptoms to examine the test-retest consistency of VEMP testing and to analyze the impact of stimulus type and muscle tension monitoring. All subjects underwent VEMP testing with two stimuli: a 500-Hz tone and a click. With each stimulus, testing was completed with and without monitoring of sternocleidomastoid muscle tension. All subjects participated in an initial testing session and then returned for a repeat testing session 2 to 4 weeks later. We measured the amplitude of primary waveforms P13 (first positive peak) and N23 (first negative peak) and analyzed the reliability and reproducibility of the mean amplitude asymmetry of these VEMP peaks. The P13 component of the VEMP (specificity: 86.25%) demonstrated a more stable amplitude than did the N23 component (specificity: 70.50%). Therefore, our statistical analysis of the effect of stimulus type and muscle tension monitoring on test-retest reliability was limited to the P13 waveform. We found that neither the type of stimulus nor the presence or absence of muscle tension monitoring had any statistically significant effect on amplitude asymmetry. We concluded that in VEMP testing, the P13 component was more specific than the N23 component in identifying normal subjects and that the P13 component provided consistent results across test sessions, regardless of the type of stimulus or the presence or absence of muscle tension monitoring.

Latest articles from "Ear, Nose & Throat Journal":

Ear mold impression material as an aural foreign body(September 1, 2014)

Keratocystic odontogenic tumor(September 1, 2014)

Actinomycosis of the pharynx(September 1, 2014)

Role of the otolaryngologist in the management of asthma(September 1, 2014)

Stylohyoid syndrome, also known as Eagle syndrome: An uncommon cause of facial pain(September 1, 2014)

Solitary myofibroma of the oropharynx causing airway obstruction in an adult(September 1, 2014)

Bilateral nontuberculous mycobacterial middle ear infection: A rare case(September 1, 2014)

Other interesting articles:

Study of New Wall Materials Design Based on TRIZ Integrated Innovation Method
Management Science and Engineering (December 1, 2012)

Hearing Aids and Cell Phones
Hearing Health Foundation (U.S.). Hearing Health (January 1, 2010)

Sentiments of Disdain and Practices of Distinction: Boundary-Work, Subjectivity, and Value in the Hindi Film Industry
Anthropological Quarterly (January 1, 2012)

Social Spying
The New American (February 20, 2012)

The burden of hearing loss in Kaduna, Nigeria: A 4-year study at the National Ear Care Centre
Ear, Nose & Throat Journal (April 1, 2012)

Modern Drummer : MD (March 1, 2012)

A Balanced Blend
Minnesota Monthly (November 1, 2011)

Publication: Ear, Nose & Throat Journal
Author: Bush, Matthew L
Date published: April 1, 2010


Vestibular evoked myogenic potential (VEMP) testing is a readily available and easily performed audiologic test used to assess saccular function and the integrity of the inferior vestibular nerve. Its use is becoming more widespread in both otolaryngology and audiology diagnostic practices. VEMP testing has been used in patients with cerebrovascular accidents, ' acoustic neuromas,2 vestibular neuritis,3 herpes zoster oticus,3 Meniere disease,3 and superior semicircular canal dehiscence.4

In order to obtain the maximum benefit from VEMP testing, examiners need to know the ideal techniques for conducting testing and established ranges of normal results. Few studies have been conducted to evaluate the reproducibility of VEMP results in normal subjects.5 It is essential in the recording of VEMPs to maximize the reliability of the amplitude of the primary waveforms P13 (first positive peak) and N23 (first negative peak). Also, the type of stimulus that provides the most reliable VEMP waveform has not been determined. Moreover, there is some evidence that monitoring muscle tension can have an impact on the variability of these waveforms.6 Finally, some standard VEMP recording units are equipped with a built-in muscle tension monitor, but some authors prefer to use a conventional blood pressure cuff.7

In this article, we describe our investigation to evaluate the specińcity of VEMP testing in normal subjects and to determine the impact that the type of stimulus and the presence or absence of sternocleidomastoid muscle control have on test reliability.

Subjects and methods

After we obtained Institutional Review Board approval for this investigation, we recruited study subjects from among patients who were being evaluated for nonotologic problems at our tertiary care otolaryngology clinical practice. Patients with a history of vertigo, neurologic deficits, and hearing loss were excluded from candidacy, as were those who were unable to participate in the testing. Eight adults volunteered for this prospective investigation.

After providing informed consent, the 8 volunteers were screened to confirm normal hearing. They were taken to a soundproof testing room, where thresholds for the octave frequencies from 500 to 4,000 Hz were established with a calibrated GSI 6 1 audiometer (GrasonStadler; Eden Prairie, Minn.) to confirm normal symmetrical hearing. Tympanometry was also performed. Exclusion criteria included evidence of middle ear disease, an asymmetry of 5 dB or greater, or an abnormal hearing sensitivity. All 8 subjects exhibited normal middle ear spaces and normal symmetrical hearing.

We performed VEMP testing in all subjects and identified the primary waveforms (P 13 and N23) and measured the amplitudes. The waveform amplitudes of the left and the right were used to calculate a mean amplitude asymmetry ratio (AAR). All subjects underwent testing ( 1 ) with two types of stimuli and (2) with and without sternocleidomastoid muscle contraction monitoring. Subjects were retested 2 to 4 weeks after their initial evaluation. The application of the testing conditions was performed in a randomized fashion at both the initial testing session and the retesting session. Subjects were asked to lie in a supine position on the testing table. VEMP testing was performed with the SmartEP system (Intelligent Hearing Systems; Miami). At the onset of testing, subjects were asked to elevate their head from the table and turn it at a 45° angle in the direction of the stimulus. Two separate recordings were made at a 10-minute interval, one with a 500-Hz tone and one with a click stimulus, both at 100 dB nHL. Each test was captured by traditional electrode montage and recording techniques with the recording electrodes placed on the sternocleidomastoid muscle ipsilateral to the ear being tested and at the high forehead with a ground electrode on the midforehead. Recordings were made of the latency and amplitude of Pl 3 and N23; no effort was made to record the N34 and P44 peaks.

The use of a muscle tension monitor was another variable in this study. Sternocleidomastoid muscle contraction was encouraged in all subjects. Constant contraction was initiated by instructing subjects on proper contraction and then maintained by either coaching and observation by the test administrator or by allowing the subjects to monitor their own muscle contraction with a hand-held monitor.

Subjects were asked to return to the clinic 2 to 4 weeks following the initial test session for a repeat test to determine the reliability of the initial results. Administration of the repeat test was identical to that of the initial test except that the screening pure-tone audiometry was not performed.

Results were quantified in terms of AAR,7 which is calculated according to this equation:

AAR = 100 × (amplitude^sub left^ - amplitude^sub right^) / (amplitude^sub left^ + amplitude^sub right^)

There is no absolute standard that determines the cliriical significance of any particular AAR, but a range from 30 to 47% is commonly used.7 Individual institutions have been advised to determine their own normative values. Our institution has designated an AAR of 34% or higher as abnormal, and that is the threshold we used in this study.

Our statistical analysis took into consideration the specific waveform, the type of stimulus, and the presence or absence of muscle monitoring at each testing session. Group means for AAR and for VEMP specificity at each testing session were obtained.


VEMP test results were easily obtained in all subjects.

Amplitude asymmetry ratios. Analysis of variance (ANOVA) revealed that the mean AAR for P13 was significantly lower (p < 0.001 ) than the mean AAR for N23 during seven of the eight comparisons of each (figure 1 ), indicating that P13 is a more reliable waveform. In view of this finding, further statistical analyses were performed only on Pl 3 values.

Some variability among P 1 3 AARs was seen across all tests (range: 15 to 25), but a paired comparison analysis with the Student itest revealed that the differences were not statistically significant (p= 0. 14). Likewise, there were no significant differences in P13 values across all tests vis ą vis the type of stimulus (p = 0.32) and the presence or absence of sternocleidomastoid muscle contraction monitoring (p = 0.60).

Specificity. The overall specificity of the P 13 mean AAR was 86.25%, and that of the N23 mean AAR was 70.50% (figure 2).


The usefulness of a diagnostic test is dependent on its specificity and sensitivity. By manipulating a test setting, one can develop a highly sensitive test in which all patients, regardless of their true status, will yield an abnormal result. Conversely, with stringent settings, tests can be made highly specific so that results in all patients will be normal, regardless of the presence or absence of pathology. Ideally, any new test should be first evaluated in a normal patient population that has no known pathology. This allows examiners to identify conditions and variables that are consistent with normal findings, to establish quantifiable ranges of normative values, and to develop appropriate testing conditions. Obviously, we would expect that any reliable diagnostic test would have a high specificity in a group of normal subjects.


The purpose of our study was to determine the reliability and specificity of VEMP vis ą vis 4 variables: primary waveform (P13 and N23), type of stimulus (500-Hz tone and click), sternocleidomastoid muscle contraction monitoring (with and without), and the timing of the testing session (initial test and retest). We found that the mean AAR of the P 13 waveform was statistically superior to that of the N23 component and that this superiority was consistent regardless of the type of stimulus used, the status of muscle monitoring, and the timing of the test.

Type of stimulus. A trend has emerged in the literature in favor of a tone burst rather than a click stimulus in eliciting VEMP responses. Wu et al reported that the tone burst had an advantage over the click in terms of the latency and amplitude of each waveform, although they found no difference in AAR.8 Welgampola and Colebatch reported that a tone burst stimulus of 500 or 1,000 Hz elicited the greatest amplitude of both waveforms.9

Our study found that the type of stimulus had no significant effect on the AAR for P13.

Muscle monitoring. The consensus in the literature is that sternocleidomastoid muscle stimulation has a positive impact on VEMP amplitude stability. For example, Colebatch et al demonstrated a linear relationship between sternocleidomastoid electromyographic activity and VEMP waveform amplitude. ] ° However, the issue of whether monitoring of muscle stimulation makes any difference in VEMP testing results has yet to be settled. Vanspauwen et al found that waveform amplitude variability decreased when they used a blood pressure cuff to monitor muscle contraction.6

During our study, we used a handheld light box that provided immediate feedback to subjects regarding the quality of their muscle contraction. We found that muscle tension monitoring provided no statistically significant advantage; this finding should be confirmed in a larger study group. Omitting muscle tension monitoring from future testing protocols might simplify testing somewhat and decrease its cost, although our monitoring equipment is easy to use. The key to satisfactory muscle monitoring is to instruct patients on proper head elevation and to provide them with verbal feedback to discourage fatigue during testing. Testing patients with cervical problems, those who are uncooperative, and those who may not be physically able to maintain muscle contraction can be a challenge. Wang and Young suggested the use of effortless head rotation as an alternative to head elevation, but this technique may result in a lower waveform amplitude. ' '


In conclusion, VEMP testing is a readily available and easily performed audiologic test. Based on our results, VEMP responses are consistently reproducible, and P 1 3 amplitude asymmetry is significantly lower than N23 asymmetry among normal subjects. VEMP testing based on a Pl 3 analysis is more specific than that based on N23 in terms of correctly identifying normal subjects, and it provides consistent results across test sessions regardless of the type of stimulus and the presence or absence of muscle tension monitoring. VEMP testing is in wide clinical application, and our study provides additional information on its precision.


1. Chen CH, Young YH. Vestibular evoked myogenic potentials in brainstem stroke. Laryngoscope 2003;1 13(6):990-3.

2. Murofushi T, Matsuzaki M, Mizuno M. Vestibular evoked myogenic potentials in patients with acoustic neuromas. Arch Otolaryngol Head Neck Surg 1998;124(5):509-12.

3. Murofushi T, Takai Y, Iwasaki S, Matsuzaki M. VEMP recording by binaural simultaneous stimulation in subjects with vestibulo-cochlear disorders. Eur Arch Otorhinolaryngol 2005;262( 10):864-7.

4. Vanspauwen R, Salembier L1 Van den Hauwe L, et al. Posterior semicircular canal dehiscence: Value of VEMP and multidetector CT. B-ENT 2006;23):141-5.

5. Versino M, Colnaghi S, Callieco R, Cosi V. Vestibular evoked myogenic potentials: Test-retest reliability. Funct Neurol 2001;16(4):299309.

6. Vanspauwen R, Wuyts FL, Van de Heyning PH. Improving vestibular evoked myogenic potential reliability by using a blood pressure manometer. Laryngoscope 2006;1 16( 1 ):131-5.

7. Gans RE, Roberts RA. Understanding vestibular-evoked myogenic potentials (VEMPs). Audiology Today 2005; 17(1 ):23-5.

8. Wu HJ, Shiao AS, Yang YL, Lee GS. Comparison of short tone burstevoked and click-evoked vestibular myogenic potentials in healthy individuals. I Chin Med Assoc 2007;70(4): 159-63.

9. Welgampola MS, Colebatch JG. Characteristics of tone burst-evoked myogenic potentials in the sternocleidomastoid muscles. Otol Neuromi 2001;22(6):796-802.

10. Colebatch JG, Halmagyi GM, Skuse NE Myogenic potentials generated by a click-evoked vestibulocolic reflex. J Neurol Neurosurg Psychiatry l994;57(2):190-7.

11. Wang CT, Young YH. Comparison of the head elevation versus rotation methods in eliciting vestibular evoked myogenic potentials. Ear Hear 2006;27(4):376-81.

Author affiliation:

Matthew L. Bush, MD; Raleigh O. Jones, MD; Jennifer B. Shinn, PhD

Author affiliation:

From the Division of Otolaryngology, Department of Surgery, University of Kentucky College of Medicine, Lexington.

Corresponding author: Matthew L. Bush, MD, Division of Otolaryngology, UniversityofKentuckyCollegeofMedidne,800RoseSt.,SuiteC-236, Lexington, KY 40536-0293. E-mail:

Previous presentation: The information in this article has been updated from its original presentation at the annual meeting of the American Auditory Society; March 3-6, 2007; Scottsdale, Ariz.

The use of this website is subject to the following Terms of Use