menu Contact Us Dizzy Patients Health Care Providers Research Acknowledgements Disclaimer Quoting

Ocular Vestibular Evoked Myogenic Potential (oVEMP) Testing

Timothy C. Hain, MD Page last modified: June 5, 2019 button Return to testing index


Recent workers have claimed that bone-conducted tone bursts activate the vestibular system, producing potentials in the extraocular muscles (e.g. Jombic and Bahyl, 2005).  There is also literature stating that sound can do the same.

These are called "oVEMP" responses, where the "o" stands for ocular. The term is a little ambiguus as there are many ocular muscles, that might reasonably have input from different vestibular structures. Nevertheless, it is short, and we will just use it as is conventional.

We will use the terminology "oVEMP" to denote vestibular evoked myogenic potentials elicited from the ocular muscles. When we use the terms "cVEMP" or tVEMP or whatever (e.g. gVEMP), the small letter indicates that a muscle other than the eye muscle is being monitored - - such as cervical or triceps or other skeletal muscles.


How are oVEMPs generated ?

In general, one needs to consider the input, central processing, and output for physiological responses. Basically, things are very messy.

Ovemp Anatomy
Stimulation of the otolith organ is transmitted via the vestibular nerve and nuclei and the MLF resulting in activation of the oculomotor nuclei and the extraocular muscles. (Kantner, Gurkov, 2012. )


Input for the oVEMP:

The vibration stimuli as well as sound stimuli that are used to elicit oVEMPs excite many structures -- the cochlea, the otoliths, the semicircular canals, proprioceptive input, so there are many choices for input.

There is some evidence that these are utricular in origin. (Manzari et al, 2012a). Curthoys et al (2016) reported in guinea pigs that both bone and air conducted inputs causes "phase locking" of afferents in both the utricle and saccule, and furthermore that this produces a limit on neural firing. Phase locking occurs up to 1500 for bone and 3000 for air, implying that guinea pig afferents can fire as high as 3000 spikes/second (thats very high !). On the other hand, phase locking at low frequencies (i.e. 100 hz) limits responses, presumably because afferents cannot fire less than about 100 spikes/sec. Fife et al (2017) stated that "it is not known whether oVEMP/cVEMP responses accurately identify vestibular function specifically related to the saccule/utricle". We think this is fair, as of 2017. We have unproven claims.

Sound can excite the semicircular canals as well as the utricle, and does this especially so in SCD (superior canal dehiscence). The corallary is that sound (in disease states), could excite the superior canal, and cause a VOR in the vertical eye muscles -- including the inferior rectus and inferior oblique. This means that the input AND output might be different in disease than in normal subjects.

If oVEMPS are indeed utricular, they should enter the brain through the superior vestibular nerve. If oVEMPS also have some saccule component, there could also be some input through the inferior vestibular nerve. Of course, if one considers the multiple other structures that can be activated by sound (i.e. the cochlea, the semicircular canals), or vibration (i.e. all of the above plus proprioceptors), there are a large number of potential generators for oVEMPS, that might vary according to whether one is working with a normal individual or a patient with pathology.

Chihara et al (2009) reported that oVEMPs do not require either a cochlear or facial nerve. This does not mean that oVEMPs have no input at all from these nerves, but merely that it is not required. In other words, cochlear and facial nerves are not necessary. They may still be sufficient. Similarly, oVEMPs do not require an eyeball.

According to Naranjo, VEMP amplitudes are increased by threat and fear (Naranjo et al, 2016) showing that there are also other inputs to consider. The implication is that cVEMPs are not simple saccule reflexes, but have multiple inputs. This could also be the cases for oVEMPs.


According to Oh et al (2016), the ocular VEMPs (oVEMPs) reflect the function of the vestibular nuclei and the crossed vestibulo-ocular reflex (VOR) pathways, mostly contained in the medial longitudinal fasciculus (MLF). Therefore, lesions involving the vestibular nuclei can present abnormalities of both cVEMPs and oVEMPs. The medullary lesions involving the descending MLF or the spinal accessory nucleus impair cVEMPs. In contrast, the lesions involving the MLF, the crossed ventral tegmental tract, oculomotor nuclei and the interstitial nucleus of Cajal can impair oVEMPs.


The output system for oVEMPs is the surface EMG under the eyes. It is generally assumed that the output is the extraocular muscles. For the central signal to activate the extraocular muscles, it must traverse one or more of the oculomotor nerves (3,4,6), and the neuromuscular junction of the eye muscle (s). Logic would suggest that the inferior oblique muscle is the one that is being picked up by the usual electrode placement.

If sound were exciting other structures than the utricle, such as the superior semicircular canal (which happens), then one would expect that there would be vestibular ocular reflexes in all of the muscles measured by the oVEMP (i.e. inferior oblique, inferior rectus), and laterality would be gone as well. In other words, one could not depend on the left eye reflecting the output of the right utricle.

Other factors. The oVEMP is a tiny response and the person interpreting the test has a strong influence on oVEMP latency and amplitude (Ertl et al, 2016). This is not the case for cVEMPs. This suggests that oVEMP studies could easily be affected by bias, and that ideally they should use blinded interpreters.

Methodology of oVEMPs

Fife et al (2017) published a review of VEMP usage including methodology of oVEMP.
Ovemp method
Ovemp methodology from Katner and Gurkov, 2012.

The input signal for an oVEMP is either sound or vibration, and the output is the response triggered average of EMG from the ocular muscles -- generally those just under the eye. While oVEMPs were developed using bone vibration, as of 2017, the most common method of eliciting oVEMPs is with sound, in a way similar to cVEMPs. The best frequency of sound is reportedly similar to cVEMPs, about 500 Hz, but Piker et al (2013) suggested that 750 or 1000 might be attempted in older adults. In Piker et al (2013), 750 was the best frequency overall, and one would wonder why not just use 750.

For the bone stimulus, subjects are positioned lying supine and looking upward, and undergo repeated applications of the stimulus through a source of vibration, commonly a "minishaker". Some studies are done using a B71 bone vibrator. A "minishaker" such as the Bruel and Kjer 4810 is a common method of providing a vibration input. The much cheaper bone oscillator, such as the Radioear B-71, is not sufficiently loud to produce a reliable output (Iwasaki et al, 2008). Obviously, for bone vibration, both ears are being stimulated at the same time, and one must hope that the "wiring" is unilateral and not partially crossed as nearly everything else in the brainstem seems to be. So in other words, for clinical use, the air input seems more logical.

For either method, the electrodes are placed under the eye, with the goal of recording EMG activity from the inferior oblique muscle. Becuse the inferior oblique is located under the eyeball, it is activated by looking upward. The inferior rectus is relaxed. Although the superior rectus is also activated, it is on the other side of the eye-socket. Thus the idea is that one is mainly measuring the inferior oblique. These are small muscles and likely rather sensitive to electrode placement.

oVEMPS are not due to the corneoretinal potential used for ENG testing. This is a DC potential that is filtered out by the oVEMP amplifier settings.

oVEMP obtained in a patient using sound. The latency here is 11-12 msec, and the amplitude is only 2.5-7.5 uV

This produces a rather small (compared to the cervical VEMP) potential, at about 10 msec (i.e. a shorter latency than the cervical or triceps VEMPs).

oVEMPS can be elicited as early as the age of 3 years old, making them potentially a good test to use for vesibular assessment in children (Young, 2015). One does have the problem here of getting the child to look upward.

Venhovens et al (2015) reported that oVEMP amplitudes have fair-good test-retest reliability, but latencies had a poor reliability (when repeated 1 week later). This would suggest that latencies are not a good parameter for diagnosis.

Mattingly et al (2015) reported a single case of bilateral hearing loss immediately after oVEMP and cVEMP testing. They suggested limits for sound exposure for VEMPS. Practically, we do not think that the risk is high in as much as persons who operate loud construction equipment do not generally have significant risk for sudden hearing loss. Noise exposure is very well studied and we see this case as an unfortunate event alone.

Iwasaki et al (2008) reported that oVEMPS are asymmetrical when stimuli are presented in the midline, and simultaneous recordings are made from each eye. From other studies of oVEMPS, one would think that this protocol would be less sensitive than AC oVEMPS.

Other methods of eliciting oVEMPs:

Obvious technical pitfalls in oVEMPS.

There are a HUGE number of issues with oVEMP studies -- both technical as well as methodological.

A common logical lapse in procedures used to measure oVEMPs is lack of a method of calibrating the surface EMG. As far as we tell, the present attititude in the literature is just to live with the logical lapse, presumably thinking that it is better to do the test than not test at all. Because the eye muscles are much smaller than the neck muscles, there is less "wiggle-room" to make a mistake with electrode positioning with oVEMPs, compared to cVEMPs.

Another logical issue is that there is generally no method of being sure that the patient is cooperating with the oVEMP and keeping their eyes positioned upward. Again, researchers seem to take the general approach that it is better to do this fairly simple test and see how it works without controlling for obvious technical problems. Patients with nystagmus may not be able to do this, and for this reason, have no oVEMP (Yang et al, 2017).

A third logical problem is that studies are generally not blinded -- allowing for the possibility of bias. This is typical of an "early" adopter research base.

A fourth problem is that in disease (such as SCD or cholesteatoma), the semicircular canals may be sensitive to sound as well, which would make most of the assumptions about this being a contralateral utricular response go out the window.

A fifth problem is that in disease, the central wiring of VEMPs may be different. After all, what is going in is sound, not movement. Why couldn't sound be a sufficient stimulus by itself in persons with clear vestibular loss, and not require utricular stimulation ?

A 6th problem is that oVEMPs are usually done sitting upright, and the brain's weight on the dura may "autoplug" an open superior canal. This issue has not been dealt with in any great detail. It is not a problem with cVEMPs done supine (but some people do them upright -- a bad idea in our opinion).

One can also broadly criticize all studies of VEMPs in that the core dogma driving the enthusiasm in doing VEMP tests is implausible. The core dogma is that things are very simple, allowing one to make diagnostic inferences. While not emphasized much in the enthusiast literature, VEMP tests likely involve input from many senses, and are not confined to saccule or utricle inputs as has been suggested. It is also implausible that the wiring is completely "unilateral", and especially so in the very patients we are attempting to diagnose. Thus we have a substantial clinical database concerning oVEMPs that has not produced very impressive results, perhaps based on oversimplified ideas about how the ear is wired up to the nervous system.

Normative values for asymmetry ratio and latency.

Ocular VEMP Amplitude (monaural air-conduction, tone-burst)
  No. Subjects Age range Mean AR(%) N1-P1 SD Mean + 2SD (%)
Murnane et al, 2011 47 18-34 18.6   11.6 41.8
Piker et al, 2011 50 8-88 14   10.0 34
Chihara et al, 2007 10 23-59 19.3   8 35.3
Park et al, 2010 20 24-34 31   6 43
Versino et al, 2015













(first version of data table courtesy of Ploy Maroongrooge, Au.D., as of 2017)

This literature base is not impressive -- we have a little less than 130 normal subjects here in the world literature, and their ages are a bit variable. We think 35% is a reasonable pick for asymmetry ratio.

With respect to the latency, the mean latency (according to Kantnor and Gurkov, 2012) ranges from 13.3 to 17.9. Rosengren et al (2011) studied 61 subjects at 132 dB SPL (somewhat high), and suggested mean latency was 15.4+-1.3.

Frequency dependence of oVEMP amplitude.

Piker et al (2013) described the frequency and age dependence of oVEMPs.

ovemp frequency and amplitude

Subfigure C above from Piker et al.'s paper shows peak-peak amplitudes as a function of both age and frequency. Note that amplitudes at the best frequency, 750, range from about 5 to 15. Clearly 750 hZ is the best frequency for oVEMPS in normal subjects, and clearly oVEMP amplitude drops by roughly a factor of 3 between "young adult" and "old adult". It would seem imprudent to use an "absolute" criterion for amplitude, such as are recommended for SCD, given that they are so strongly dependent on age.

Curiously, Rosegren et al (2011), also reported rather low amplitudes for oVEMPs elicited by tone bursts in 100 subjects, with the mean being 3.78 uV p-p. They reported "There was a significant, moderate correlation between age and amplitude for the AC click oVEMP and contralateral BC tone burst oVEMP only, indicating that the responses became smaller with increasing age (clicks r = 0.33, P = 0.009; BC tone bursts r = 0.41, P = 0.001)". In as much as more recent papers (as well as our clinical data) find much higher amplitudes of oVEMPs in general, it would seem to us that technique must have advanced since 2011. However, it is also possible that the Piker et al paper (2013) is problematic instead. Rosengren et al observed that there are previous reports of declining amplitude with age.


Age dependence of oVEMP amplitude.

As cVEMPs clearly have a strong dependence on age, and the presumed reason for this is that the otoliths deteriorate with age, one would expect that oVEMP amplitude would also depend on age, and in particular, decline greatly with age. So it seems to us that oVEMPS must be reduced in older persons, just as are cVEMPs. Additionally, it seems to us that the utility of air conduction oVEMPs declines with age. If cVEMP amplitude deteriorated due to a different reason than otolith wear/tear, than the conclusion that this also applies to oVEMP would be wrong, but in our opinion, both deteriorate with age.

The previously referenced paper of Piker et al (2015) would support this idea. According to Piker et al (2015), "VEMPs in response to air conduction stimuli are bilaterally absent in a large percentage of older patients complaining of dizziness who otherwise have normal vestibular and auditory testing for their age. In combination with other abnormal vestibular findings, an absence of VEMP responses may be of value. However, the functional consequence of an isolated bilaterally absent VEMP is not known and may provide minimal information to an older patient's diagnostic picture. In cases where the response is bilaterally absent, a more intense AC stimulus should be used or bone conducted vibration should be considered."

Li, Layman, Carey and Agrawal (2015) recorded oVEMPs using head taps (rather than sound). This method is not used in the clinic at this time (2018), and thus this oVEMP data is mainly of historical interest. Potentials were very large (compared to those for other methods), and averaged about 25uV for persons less than 50 (there were only 12 of these). The average potential for persons 80 and above was about 11. While the numbers here are not very useful, as they are for an discarded methodology, this paper shows that oVEMP potentials are highly age dependent. It appears that "young adults" for Piker's study had about half the amplitudes as the <50 group in Li et al's study. Oddly, individuals of Black ethnicity had larger oVEMP amplitudes on average than whites (18.5 vs. 12.2). Looking at their graphics, this may be related to some outliers with very large potentials. The study design did not include temporal bone CT scans, which would be needed to exclude SCD. In our opinion, it is unlikely that there is this much of a racial difference with oVEMP amplitudes, and this question needs to addressed again.

Contrary to the assertion that oVEMP amplitudes decline with age, Versino et al (2015), found almost no dependence of oVEMP amplitude on age in roughly 50 normal subjects, even though the amplitudes were clearly smaller in older patients. This was likely just an insufficient sample as they comment about the immense variability of oVEMP amplitudes. This means that more subjects are needed to answer this question. This paper is also difficult to interpret as amplitudes were roughly an order of magnitude smaller (about 1 uV), compared to the much larger oVEMP amplitudes reported in the rest of the literature.

What data is there concerning particular disorders in the literature ?

The VEMP literature is rapidly increasing and it seems that there are considerable valuable diagnostic information to be obtained.

In the pages below, we review papers claiming a utility for VEMPs, convering both cVEMPs and oVEMPs. This is a difficult task as the literature is growing rapidly. One would think that as oVEMPs are much smaller than cVEMPs, and consequently more vulnerable to noise, they would perform worse than cVEMPS for diagnosis of nearly anything but disorders that are selective for either the utricle or superior vestibular division of the vestibular nerve. In other words, oVEMPs should reasonably have a narrower utility than cVEMPs. One would also think that AC oVEMPS would generally perform better for localization than BC oVEMPS, as the stimulus is confined to one ear. We think it is a very bad idea (for diagnosis) to use a BC stimulus, as one does not know what is being activated.

Probably the main reason to do oVEMPs is to diagnose SCD. These oVEMPs can be amazingly clear.

oVEMP obtained in a patient using sound. The right ear has clear SCD on temporal bone CT scan.


Below is a discussion of entities not covered in the pages above.

Behcet's syndrome.

Bayram et al (2015) found no difference between amplitude ratio and thresholds in patients with Behcet's. This has little significance given the rarity of Behcet's.

Enlarged vestibular aqueduct syndrome (EVAS)

Taylor et al (2012) reported that increased oVEMPs, compared to normal VEMPS, were found in a single patient with bilaterally enlarged vestibular aqueducts. Interesting I suppose, but this obviously needs to be confirmed with, lets say, 10 cases.

NPH (normal pressure hydrocephalus)

Bottcher et al (2016) reported that one third of 25 patients with suspected NPH had impaired otolith function. "Responders to STT only had a significant increase of oVEMP and thereby utricular input, probably due to a decrease of pressure. " We would wonder in this study if the age group being tested might not have VEMPs at all, as VEMPs are reduced with age.

Spinocerebellar degeneration (i.e. a variant of central vestibular disorder).

Riberio et al (2015) reported in 14 Machado-Joseph disease patients that 13/14 had various abnormalities in cVEMP or oVEMP. They included timing, amplitude and asymmetry among their "abnormalities". Given the rather broad nature of this report, somewhat of a fishing expedition, we think more work needs to be done.

Utricular disease

One would think that oVEMPs would be the perfect test for unilateral utricular disease, but we have a catch-22 here -- we have no practical way of diagnosing unilateral utricular disease aside from oVEMPs, and because of this, we cannot use one to calibrate the other. In other words, another example of the "circular reasoning" that is abundant in the VEMP literature.

Schaaf et al (2013) attempted to correlate oVEMP tests with several other maneuver ("Turn over" test, and Subjective vertical). They reported about 63% agreement among these three maneuvers.


References: Journals that have links are those with low scores for quality or other issues concerning their reliability

Copyright (c) Timothy C. Hain, MD 2002-2007

Copyright June 5, 2019 , Timothy C. Hain, M.D. All rights reserved. Last saved on June 5, 2019