This material is intended for clinicians and vestibular scientists. See also: bedside version -- HIT test
Timothy C. Hain, MD Page last modified: April 27, 2020
The VHIT is an instrumented bedside technique used to diagnose reduction in vestibular function in one ear vs. the other.
An examiner abruptly accelerates and then decellerates the head, moving the head in rapidly at high speed and then stopping it.
The graphic above shows a positive VHIT in a patient with unilateral vestibular loss. For head thrusts to the right, head and eye are similar. For head thrusts to the left, there are many large "covert" saccades in the middle of the eye traces. This rather clearly shows a compensated patient who has a complete unilateral vestibular loss. (This device is the Interacoustic VHIT).
The graphic above shows a positive VHIT in a patient with bilateral vestibular loss. For head thrusts to either side the eye is quiet for the first 100 msec, and then there are many large "covert" saccades . This rather clearly shows a compensated patient who has a complete bilateral vestibular loss. (This device is the Interacoustic VHIT).
It should be said that in neither of these patients was the VHIT used to establish the diagnosis. The diagnosis was already well known from other tests (VENG and rotatory chair).
The VHIT test is a commercial version of the HIT test (head impulse test) using sophisticated eye tracking and head velocity transducers. This is presently the most useful version of the HIT test, mainly because it can detect covert saccades.
The VHIT test users are growing very rapidly worldwide, except in the USA where adoption has been greatly limited by the lack of a health insurance mechanism to bill for the test. In our opinion, the VHIT is obviously a useful diagnostic device. We think that the time to do a VHIT and the utility of the test is roughly the same as a caloric test. We are presently dubious that anything but the "horizontal" test has much diagnostic utility. If/when a billing code is developed, we hope that the description of the procedure is narrow enough to prevent fraudulent billing for "fake" VHIT implementations, such as the field has seen in many other vestibular testing procedures.
Different versions of the VHIT
VHIT variants Dr. Hain modelling the GN-otometrics VHIT device. (Branded "Impulse").
Interacoustic VHIT (from http://www.hearingreview.com/2013/08/video-head-impulse-testing-vhit-vor-analysis-of-high-frequency-vestibular-activity/).
Micromedical VHIT device -- looks just like their regular VENG goggles. This is branded "VisualEyes".
That being said, there are several commercial versions of the VHIT -- including (at least) one marketed by Interacoustic, one by GN-otometrics (GNO or ICS), one by Micromedical technology, and one by Synapsis. The first three devices all share similar technological features -- an eye camera worn by the patient, a head velocity sensor. The first two have a head-mounted calibration arrangement. The Micromedical Device is somewhat bulky. The GN-otometric device is asymmetrical having more mass on the right side. The Interacoustic device is also asymmetrical but it is not as bulky and the mass of the camera is closer to the center of head rotation (meaning that it is less likely to be jerked around (pun intended) during rapid head acceleration and decelleration).
A full "test", defined as horizontal only, can be done in about 12 minutes. We would expect a test that included the other 2 planes would take perhaps another 10 minutes, for a total of about 22 min. We have both the Interacoustic and GNO versions in our clinic, and agree that the test can be done quickly, but due to problems with eye tracking, and the overhead in analyzing, realistically this seems to us to be a 30 minute procedure. This compares poorly to the bedside "HIT" which can be done in 2-15 seconds.
These devices differ fairly markedly, and so far, little comparison data is available between the devices. Bansal and Sinha (2016), using the ICS device, reported that in normal subjects the mean VOR gain was greater for every test involving the right side compared to the left (2016). We have so far not noticed this in our use of the Interacoustic VHIT, or previous studies using scleral eye coils. Janky et al (2017), compared the three devices regarding variability -- none of them do very well (see discussion of norms below).
We have had a large experience with the IAC VHIT test, and have encountered a number of technical issues. These are discussed in this page.
MacDougall et al (2016) recently reported a variant VHIT test where patients were asked to view a laser dot that moved with their head. This protocol then asks patients to turn OFF their VOR, rather than to use it. The authors state that "anticompensatory" saccades are elicited in normal controls, but are less commonly found in persons who have less VOR to suppress. This makes sense. Our thought however is that this paradigm adds another variable to the mix -- how well people can suppress their VOR, and that it is inferior to the ordinary HIT protocol. We think that this procedure adds noise to the HIT test.
One would EXPECT that the VHIT gain in normal people should be 1.0, and there should be few if any saccades during or just after the head thrust. The VHIT gain assumption of 1.0 is wrong because VHIT gains are "all over the place".
The graph above shows all VHIT tests through 4/2020, from more than 3000 patients tested at Chicago Dizziness and Hearing (using the interacoustic device). There are many patients with gains > 1. It is almost certain that these values are not "real" but rather are due to measurement error. Thus the VHIT test is not a "clean" test with a normal gain of 1.0, but rather it is "all over the place". Notice that there are also some "real" abnormal VHIT tests shown here -- with low gain on one or both sides. There is useful diagnostic information in VHIT testing, but it is obscured by a cloud of noise, like the problem with almost all other vestibular tests. It is curious that there seem to be some "forbidden areas" on this plot -- we just don't have many individuals with gains of 1.0 on one side, and gains < 0.25 on the other. Presumably there are some physiological constrants with superimposed noise.
Figure from Jankey et al (2017) showing scatter in VHIT gains in normal subjects. FIgure from van Dooren et al (2018), showing trend for higher VHIT gain with refractive error.
Pursuing this idea, Jankey et al (2017) wrote about "effects of device on .. vHIT gain". They observed that three different VHIT devices all do things a bit differently. They plotted this out in their figure 2, shown above. PG is position gain, IG is instantaneous gain, and AG is area under the curve gain (similar to position gain). Or another way to say this is that position gain is the Visual Eyes device from Micromedical, Instantaeous gain is the EyeSeeCam device from Interacoustic, and Area under curve is the "Impulse" device from GN Otometrics.
This plot shows very clearly that VHIT gains average about 1, but their variability is astounding. About half of normals have VHIT gains greater than 1 ! If this were REALLY TRUE, the eyes would have to go backwards after getting to the target. This doesn't happen often so it isn't true. This means that this variability is NOISE.
So the take home message here is that the 95% limits on VHIT gain looks to be roughly 15% in either direction. We have also observed a similar variability in our clinic population (we use the Interacoustics -- instantaneous gain -- device).
Why are things so variable ? Well it must be that there are lots and lots of technical issues that haven't been worked out. We think ourselves that goggle slippage (i.e. inertia of the camera/google) is the big one. There may also be an effect of the point of regard -- i.e. it has been shown previously using other devices that people who are looking at near targets have higher gains. Finally, perhaps the calibration procedures are not so accurate on these devices. We also wonder whether the video tracking on these devices can be relied upon. Whatever it is, there is a lot of room for improvement.
Rey-Martinez, Burgess and Curthoys (2018) suggested in a case report in an open access journal that "we should consider the hypothesis of a physio-pathologic cause for the enhanced eye responses to vHIT testing of some patients with vestibular function. It is easy to suggest considering hypotheses when one is dealing with one or two cases, but the simplest and most likely assumption, given the wild variability shown above, should just be that the VHIT is a messy test, and that little changes in gain are meaningless.
Van Dooren et al (2018) examined the question as to whether or not use of contacts or spectacles might affect VOR gain. The figure above, from their paper, illustrates their results. In essence, the small effect of refraction was drowned out by the very large variability of VHIT gain. They concluded that it was not necessary to correct for spectacle wearing, arguing that lack of a significant effect was the same as no effect. Of course, the problem with this argument is that had they studied more subjects, it seems highly likely that significance would have been achieved. The difference in gain, from their trend line, looks to be roughly 10% at 5 diopters. Note that these are normal subjects, but their gains are "all over the place".
Covert saccades, seen on both sides, in patient with complete vestibular loss on R.
More about VHIT and compensation is here. The ability to detect "covert" saccades, as illustrated above, is what makes the VHIT device better than just doing the HIT test by hand. On the other hand, acutely in unilateral loss, there are no covert saccades and the bedside "HIT" test works very well and also doesn't require one to master a new device. Or to say this a different way, you don't need the VHIT test for acute dizziness -- the HIT test does perfectly well. We have observed patients with "covert" saccades as early as 1 week post vestibular neuritis, while on large amounts of meclizine. We have also observed "overt" saccades in patients more than 1 year "out" from unilateral loss. Thus, it would seem that there is a lot to learn about the supposed ability of covert/overt saccades to detect compensation.
Concerning the sensitivity to acoustic neuromas, Fujiwara et al (2018) reported on VHIT tests in 15 patients with acoustic neurinomas. They reported that "Dysfunction of anterior and posterior semicircular canals was detected by vHIT in 26.7% and 60.0%, respectively. Six patients (40.0%) demonstrated abnormalities referable to both vestibular nerve divisions. Abnormalities referable to the superior vestibular nerve were identified in 3 patients (20.0%), while 3 patients (20.0%) demonstrated a pattern indicative of inferior vestibular nerve involvement. Anterior semicircular canal vHIT produced fewer abnormalities than did either horizontal or posterior semicircular canal vHIT." This doesn't make much sense as one would think that the vestibular nerve damage would be grouped according to the superior and inferior division. Perhaps the AC type VHIT just doesn't work very well as there are many technical problems. We are not so sure that this is a report about how well the VHIT works, or where the lesion is in acoustic neuromas. Perhaps it cannot be said right now.
Some enthusiasts about the VHIT test, suggest that there is no longer any need to do rotatory chair or VENG tests. This is a naive viewpoint as these older and time-tested vestibular tests provide information that the VHIT lacks. The rotatory chair provides low-frequency information. We have encountered patients with very clear abnormal rotatory chair tests, but very normal VHIT tests. This makes perfect sense from a physiological perspective. This means that the two tests are not identical.
The ENG provides excellent side of lesion information. The VHIT test provides very high frequency information -- part of the whole picture, but it doesn't replace the other vestibular tools.
VHIT after 50 years. This is a VHIT of a patient who had an acoustic neuroma removed more than 50 years ago, as well as substantial ipsilateral cerebellar damage. Her 8th nerve is not connected up to her brainstem. Note that there has been no recovery at all of VOR gain on the left, and also that there are no predictive saccades.
Bottom line regarding what is VHIT good for:
At this writing (6/2019), we think that the main use of the VHIT device is to detect vestibular neuritis. Our thought is that it is very good at picking up acute vestibular nerve lesions. We also see it as a good "tie breaker" when there is disagreement between ENG and rotatory chair. Something to add on, not a replacement. We also think that the VHIT is an excellent method of diagnosing bilateral vestibular loss, perhaps superior to the rotatory chair test, which has suppression as a problem. The problem comes up however when the VHIT is normal and the rotatory chair is abnormal. Usually we will prefer the rotatory chair.
We think that the initial diagnosis of unilateral vestibular loss will likely be made by the time-tested ENG device in the lab, and by simple clinical signs (vibration, HSN) at the bedside. Bilateral loss will continue to be diagnosed with the rotatory chair, as VHIT can miss well compensated, partial, bilateral weakness.
The VHIT could potentially allow clinicians to determine whether persistent symptoms are related to poor compensation.
We are not so convinced it is quite as good for hair cell causes of vestibular damage (such as gentamicin or Meniere's disease).
Various central (brainstem/cerebellar) disturbances.
Choi et al (2018) published an article concerning "recent advances in head impulse test findings in central vestibular disorders", in the "views and reviews" section of a neurology journal (2018). They state: "The horizontal head impulse VOR gains can be significantly reduced unilaterally or bilaterally (positive HITs) in lesions involving the vestibular nucleus, nucleus prepositus hypoglossi, or flocculus. In diffuse cerebellar lesions, the VOR gain during horizontal head impulses may increase (hyperactive) with corrective saccades directed the opposite way. The presence of cross-coupled vertical corrective saccades during horizontal HITs is also suggestive of diffuse cerebellar lesions. Lesions involving the vestibular nucleus, medial longitudinal fasciculus, and cerebellum may show decreased or increased gains of the VOR during vertical HITs. "
Practically, in our Neurology setting at Chicago Dizziness and Hearing in Chicago, we have rarely observed any of these unusual conditions, with the exception of backup saccades (see below). It is impossible to see cross-couple saccades because clinical equipment is just not designed to show cross-axes (i.e. vertical for a horizontal HIT). Perhaps it is there, but we need to wait for better analysis software. We also don't think that vertical HIT testing is reliable enough for use. We suspect that "Cross-torsion" might be a strong sign, but equipment in 2019 doesn't track torsion.
Abnormal VHIT test with "backup" saccades. Patient with migraine and fluctuating bilateral hearing loss. This may be a technical artifact rather than a "real" finding. It is implausible that anyone should have backup saccades, unless they are not following instructions (see below)
Central -- Backup saccades ?
Back-up saccades for a tracking task (such as VOR or pursuit) are almost always a central sign, suggesting brain disease rather than ear disease. The reason is that the VOR gain is tightly controlled by the brain. Should one's VOR gain be too high, the brain would rapidly suppress it. Thus back-up saccades should mean that there is a cerebellar disturbance. Backup saccades are seen in "latent nystagmus", which is a central congenital nystagmus with tracking issues.
Detecting backup saccades pattern in the VHIT test requires that one check the velocity profile (as shown above), for spikes of eye velocity that occur during the head thrust, and go in the opposite direction as the VOR. This pattern is seen above. The velocity regression plot (not shown) is normal.
This observation is potentially very important. Small "backup" saccades are still saccades - -they are very fast. The VHIT device can "see" tiny backup saccades that one cannot observe with one's naked eyes. Thus finding "backup" covert saccades should be a unique capability of the VHIT, that cannot be duplicated by either the VENG or Rotatory chair test.
We have seen this pattern once in a patient, where we thought it was real. This individual had both a high VOR gain as well as clear backup saccades.
We have observed this pattern in people who don't look where they are instructed. If a normal person is told to look away from the target, many backup saccades are generated. This suggests to us that backup saccades may be a sign of uncooperative patients.
We think that this observation needs more research. We are somewhat doubtful that finding this in one of 25 trials has much meaning, as suggested by Hueberger et al (2015), but it is very much worth checking as the VHIT device can provide data about this potential central sign unavailable through any other method.
- Uncooperative patients not looking at the target. This is an obvious problem as the current system software does not check for eye position prior to head thrusts. This is a flaw.
- Choi et al (2014) observed a similar pattern in two patients with "acute cerebellar dysfunction". We have seen this once ourselves.
- One would also think that patients with myasthenia gravis, after taking their medication, might also show backup saccades.
- One might also conjecture that in Meniere's disease with hydrops, vestibular function might improve due to larger "pipes" in the inner ear, and result in a transient backup saccade pattern. As Meniere's is usually unilateral, one might expect some asymmetry. Rey-Martinez et al (2020) offered some theoretical support for the idea that VOR gain might be higher in hydrops. However, as back-up saccades are almost never encountered, if this is true, it doesn't have any practical meaning.
Meniere's disease ?
One would think that fluctuating gain in Meniere's disease might be easily detected with the VHIT. Theoretically, gain should go up in hydrops, and drop after hydrops is resolved (i.e. during an attack). In this regard, Lee et al (2016) reported widely variable results in 14 patients during attacks of MD. They tested all 6 canals in many patients, providing a profusion of results. The gain was either normal (67%), decreased (19%)or increased (14%). Results were not very different towards the good ear either. On the other hand, caloric responses were decreased in 64% towards the "bad" ear. Yacovino et al (2017, 2018) have reported similar findings.
While at first look, these reports suggests VHIT is not an effective method in a Meniere's attack, we don't think the story is over here. Looking for a change in vestibular function, rather than an absolute value might be more effective. Additionally, using the vertical canal VHIT may have diluted the data with random error, as vertical canal VHIT techology is still more primative (lacking torsional eye movement processing).
Cryptic bilateral vestibular neuropathies such as Charcot Marie Tooth ?
Abnormal VHIT in CMT Abnormal VHIT in a patient with a slowly progressive neuropathy attributed to CMT. Both sides have a gain of about 0.6 Rotatory chair test also showing bilateral weakness. The Gain-TC product is 4.3, suggesting roughly half of vestibular function is lost.
A group of bilateral patients that can go "under the radar screen" are individuals with vestibular neuropathies. Neuropathies that affect the limbs can sometimes affect the inner ear as well. Poretti et al (2013) reported reduction of VHIT gain and increased cVEMP latencies in 15 patients with CMT (Charcot Marie Tooth disease). We have seen this as well.
|Normal VHIT test in extremely ataxic patient with downbeating nystagmus -- VHIT doesn't work for this group.|
Central vestibular disorders:
VHIT is not useful for strokes, brain tumors (of anything but the 8th nerve), cerebellar degenerations, nystagmus disorders (such as downbeating nystagmus), Chiari malformations -- basically anything other than unilateral or bilateral vestibular loss. This lack of sensitivity to anything other than unilateral vestibular loss has been used -- basically when people fail the VHIT, they are unlikely to have any of the above. The main potential exception to this general rule is longstanding cerebellar disturbances (see above).
Another potential exception is when one observes "perverted" VHIT tests -- basically meaning the head goes one way and the eyes the other. One would think that this would mainly be due to central lesions such as lateral medullary syndrome where vertical saccades can provoke horizontal eye movements. Jeong et al (2013), reported a case of cerebellar ataxia where horizontal head-impulses produced vertical eye movements. In cerebellar disorders, one's directional calibration can get disturbed. As vertical eye velocity is not generally monitoried for horizontal impulses, seeing this phenomenon would not be easy. It seems somewhat plausible however.
We do think that interpreters of VHIT should be watching out carefully for backup saccades (see above), as this is a central sign similar to overshoot saccadic dysmetria. Stay tuned.
The VHIT test just tests the high-frequency VOR. It doesn't test the entire spectrum. Would you test someone's color vision using just "blue", and ignore red and green ? Unsurprisingly then, the VHIT can fail in partial lesions.
Normal VHIT in man with acoustic for about 20 years Large acoustic neurinoma in man with normal VHIT.
Even worse, we have encountered patients with large acoustic neurinoma's, with absolutely normal VHIT tests. This man above has only a small amount of clinical findings -- and in spite of this substantial size mass (clearly a tumor), gradually enlarging over 2 decades. The images above show that VHIT is not a "screening" test for acoustic neurinomas (sadly). I suppose nobody is perfect, and that is certainly not true for VHIT. VHIT can mislead you by being normal, when there are huge structural lesions. In other words, it is not 100% sensitive.
This VHIT is done 4 months post labyrinthectomy for a left sided acoustic. The gain to L is 0.49, and there are very consistent covert saccades. Preop MRI of acoustic neuroma.
On the other hand, after an acoustic neuroma is removed with a labyrinthectomy, the VHIT is very positive. So in other words, VHIT does NOT diagnose all acoustic neuromas. It is reliably positive however after the acoustic neuroma has been removed with a destructive procedure.
Mysterious vestibular disorders
We have changed our mind about this topic with experience. VHIT is a good way to evaluate an unknown source of dizziness, because it is largely specific for disturbances of the vestibular nerve. A VHIT test is a very efficient method of diagnosing a common condition, vestibular neuritis. When the VHIT is normal, one can tell other providers with reasonable certainty, that one kind of vestibular physical therapy (i.e. gaze stabilization) is a waste of time. This saves patients and their insurance companies time and money.
We are simply unconvinced that the oblique VHIT testing is generally useful in the clinic. Without measuring the full vector of eye movement, i.e. torsion, there are simply too many things that can go wrong. We would be a little more trustful about research studies where the investigator is very proficient. We are waiting for the technology to improve. This may happen -- but we have been waiting for about 5 years.
Thinking about technology, one needs to analyze a LOT of data accumulated over about a second. There is no reason why offline processing couldn't work for VHIT. Still, there is a lot of video data flying by. We are not sure with contemporary cameras that it really possible to stream 60 hz of things changing rapidly at high resolution. We would not think compression would work so well because compression relies upon just keeping the changes, and here the eye is moving very quickly. Perhaps using some fancy hardware with a big video buffer (i.e. not a standard USB camera), or perhaps a buffered USB interface. It does appear that composite video can be acquired via USB at rather high resolutions.
Another issue with vertical canal testing is proving that it works. It is nearly impossible to identify single canal surgical lesions in humans, the main exception being canal plugging patients, such as are created in SCD repairs. One wonders in these patients if the plug is really 100% effective -- meaning that high accelerations might have some VOR anyway. It would seem difficult to be sure that the device really works.
This device looks as if it should be easy, but it can be very difficult to get it right. The frames of the Interacoustic device need to be very tight. While the ICS/GNO device is easier to secure to the head, certain types of faces (with smaller noses) seem to result in left-right asymmetry in VOR gain. This can be as much as 20%. So this adds on to the +-15% of variability in gain - -thats a lot of variability !
For all devices, the head has to be moved in the right direction, right speed, and with little wobble at the end. Wobble is called "backlash" -- an inexperienced VHIT provider can sometimes cause backlash to be as big as the response itself. One cannot hold onto the goggles by their frame or their straps. Any of these errors make the procedure useless. We think that one should plan to be "trained" for several sessions over a month -- ideally, about four 1-hour sessions, for most to learn how to do the VHIT.
When one does not move the head with high acceleration, the VHIT test becomes less sensitive. Weber et al (2008) reported that as accelerations increased from 750 to 6000 deg/sec**2, ipsilesional gains decreased greatly, increasing sensivitiy. This is very logical, as if one doesn't do the test fast enough, it shouldn't work. Of course, to do the test with the highest sensitivity, one must turn the head with the highest acceleration, which could lead to safety issues. Here we are thinking about the neck.
There are at least 4 different VHIT devices. The most frequently encountered devices are the Interacoustics (eye-see-cam), and the GN-otometrics devices.
The Interacoustic and GN-otometrics VHIT devices differ substantially. The GN-otometric VHIT device (ICS) is a goggle that measures the right eye alone. If you have a false right eye, or a droopy right eye, it doesn't work. The Interacoustic device is more adjustable, as the camera can be positioned on either eye. Both devices are somewhat fragile as they both carry a mirror attached to the head, which can be broken. Practically however, this does not seem to matter much and we (so far) have not had any breakage.
The Micromedical VHIT uses two cameras placed within the same goggles as their VENG system. The heavy goggle limits the acceleration that one can apply, and thus limits the utility. See the comments above about the drop in sensitivity with reduced acceleration.
The GN otometric device uses "area under the curve" gain. Area under the curve is similar to position as the integral of velocity is position. The EyeSeeCam uses instantaneous velocity gain. The VisualEyes device from Micromedical uses position gain.
The Synapsis device is very different in that there is nothing attached to the head -- it is perhaps best to test small children as otherwise the performance degradation would be unacceptable. We have not had any direct experience with the Micromedical or Synapsis implementations of the VHIT.
VHIT testing is currently not widespread. In Chicago, we offer it at Chicago Dizziness and Hearing. We have had rumors that it is done at "Shirley Ryan", as well at Northwestern Memorial Hospital.
At CDH, it is both available as part of a new-patient evaluation as well as "a la carte". In other cities, it should be possible to find practices that offer it by asking the vendors of these devices.
The VHIT appears to be most useful as part of a rapid battery of bedside tests. (Mandala et al. 2008). It provides unique information about compensation (covert saccades). It does not replace standard testing (i.e. calorics and R-chair) as it has no low-frequency data and also does not isolate the stimulus to one ear (such as does calorics).