Timothy C. Hain, MD • Page last modified: June 3, 2018 •
|Rotatory chair with optokinetic stripes projected on the background. This particular chair was built by Micromedical Technology.||Rotatory Chair illustration from ICS, another rotatory chair vendor.|
The purpose of rotational testing is to determine whether or not dizziness may be due to a disorder of inner ear or brain, and particularly to determine whether or not both inner ears are impaired at the same time.
There are three parts to the test.
The chair test measures dizziness (well jumping of the eyes really -- called nystagmus) while being turned slowly in a motorized chair (see rightward illustration above). Persons with inner ear disease become less dizzy than do normal persons.
The optokinetic test measures dizziness caused by viewing of moving stripes (see leftward illustration above). Optokinetic testing is sometimes useful in diagnosis of bilateral vestibular loss and central conditions.
The fixation test measures nystagmus while the person is being rotated, while they are looking at a dot of light that is rotating with them. Fixation suppression is impaired by central nervous system conditions and improved by bilateral vestibular loss.
The author of this page has had extensive experience with the Micromedical rotatory chair system. My opinion is that the rotatory chair is occasionally useful in diagnostic testing. Because of it's high cost (typically about $100,000) and limited usefulness, it is best used in a diagnostic setting where multiple clinicians can use it, such as a hospital laboratory. At Chicago Dizziness and Hearing we use a Micromedical rotatory chair. Click here for a movie of our rotatory chair in motion. Most of the illustrations on this page are taken directly from our rotational chair.
Rotatory chair tests are usually obtained in addition to ENG (caloric) testing, or alternatively, VHIT testing. Why get both when both test the same part of the ear (lateral semicircular canal) ? The reason is to add accuracy. ENG tests by themselves may be falsely positive or falsely negative. They can be falsely positive when wax blocks one ear canal. Rotatory chair testing is not affected by mechanical obstructions of the ear. ENG testing can be falsely negative particularly in situations where there is damage to each ear. VHIT testing is commonly falsely negative, when both ears are affected at the same time.
In a survey of 85 persons who used rotatory chairs, by ICS, the main indications for using rotatory chair testing was felt to be:
- Indications for rotatory chair testing
- Bilateral canal paresis
- Inconclusive/equivocal ENG reults
- Testing of special populations (pediatric, handicapped)
- Evaluation of vestibular compensation
- Ototoxicity management
The author of this page has had extensive experience with the Micromedical rotatory chair system and offers rotatory chair testing in our clinic in Chicago. My opinion is that the rotatory chair is occasionally useful in diagnostic testing. Because of it's high purchase cost (typically about $100,000) and limited usefulness, it is best used in a diagnostic setting where multiple clinicians can use it, such as a hospital laboratory.
Rotatory chair testing is a type of "systems identification" -- engineers use this word to describe the process of attempting to figure out what a "black box" is doing, but giving it an known input, and measuring the the output. The ratio of the output to input is called the "transfer function". There are many reasonable protocols for the input. For a linear system, any protocol that includes a reasonable selection of frequency components should result in the same result -- a gain and time constant. As there are nonlinear processes in the vestibular system (such as prediction), the various methods may not always produce the same results. At present, most laboratories use either sinusoidal testing or step-testing.
In a survey of 85 persons who used rotatory chairs, by ICS, the most important subtests for using rotatory chair testing were felt to be:
- Most important subtests
- Slow harmonic (sinusoidal testing) in darkness
- High frequency or high-velocity sinusoidal rotation
- Rotation with fixation on head-fixed targets (i.e. suppression)
- Optokinetic afternystagmus
- Rotation with fixation on earth-fixed targets
- Velocity steps in darkness
- Tilted-axis rotation (OVAR)
The sinusoidal test protocol involves rotating the chair so that it moves sinusoidally. Because the derivative of a sine is another sinusoid, chair position, velocity and acceleration all change sinusoidally. Ordinarily one chooses a desired peak chair velocity, such as 60 deg/sec, and one also picks a series of frequencies to test covering about 0.01 to 1 hz. These frequencies cover the range of responses where gain and phase show their greatest variability when there is disease.
A variant of sinusoidal testing is "sum of sines" -- SOS -- one mixes together a group of sine waves to make the input less predictable. Although the SOS appears complex, it can easily be analyzed using standard mathematical methods (i.e. Fourier analysis). A "Bode plot" -- a semilogarithmic plot of vestibular gain and phase, is generally used to present results. A powerful motor is needed to attain the higher frequencies, and for this reason, sometimes testing will only include lower frequencies or the peak velocity will be reduced at the highest frequency.
An example of the expected output from sinusoidal test is shown below. The upper gain plot and lower phase plot depict a normal person. The peak gain is set to 0.8. The lower gain plot and upper phase plot depect expected output from someone with a unilateral vestibular loss. Note that at frequencies above roughly 0.1 hz, the unilateral loss and normal person are indistinguishable, because gain asymptotes at these frequencies. Phase differences are also greatest at low frequencies. Thus, rotatory chair testing is most sensitive to vestibular problems at low frequencies !
Theoretical gain and phase in patient with unilateral vestibular loss (Tc=7) compared to normal (Tc=14.5 and 21). Peak gain is set to 0.8 in all cases. On the top plot, the higher curve is 21. An identical plot, for the 14.5 Tc, can be found in Dimitri et al (1996) -- this is to show we got our math right (:
|Real gain and phase in patient with stable unilateral vestibular loss. More test data from this person are shown here here. Also see section below.|
The step test involves suddenly changing chair velocity (with an impulse of velocity). Step responses provide roughly equivalent gain/phase information as does sinusoidal testing. Step responses have many problems. They require a powerful chair to provide a high acceleration transient. They may be less reliable as well as somewhat more stressful to the patient, and for this reason, sinusoidal testing is generally preferred. Motion sickness is sometimes associated with prolonged vestibular responses (Hoffer et al. 2003), and for this purpose, step responses may be preferable to sinusoids. Practically though, nausea is unusual in sinusoidal testing and this is not a strong consideration.
A variant of the step response is PRBS -- pseudorandom binary. This is a very reasonable protocol, that has not been generally adopted.
Step responses should be largely predictable from sinusoidal testing. High gain on step responses should correlate with high gains on sinusoidal rotatory chair testing. Small time constants should be associated with increased low frequency phase lead. When there is a mismatch, this can be valuable in figuring out whether the patient is consistent as well as one's equipment is functioning.
Common mismatch patterns are:
More commentary about mismatch between step and sinusoids is found in the norms discussion.
- Spontaneous nystagmus - -asymmetrical step response time constants (by direction of nystagmus). Should be easily seen on spontaneous nystagmus test too. Spontaneous nystagmus also makes the low-frequency rotations more asymmetrical.
- Too short rotation time -- asymmetrical step response time constants with too long post-rotatory. Need to rotate for at least 5 minutes to be safe.
- Light leak -- asymmetrical step response time constants with too long per-rotatory. See "blunders" page.
- Mixture of all of the above (this can get pretty nasty !)
Ideally subjects undergoing rotational tests should have no sedating medications for the last 24 hours. Sometimes this is difficult, as for example, when persons are addicted to medications in the Valium family. In this situation, usually 12 hours is sufficient. More data about medication effects is found here.
A much more extensive discussion of OKN is found here.
Optokinetic testing does not actually involve a rotating chair -- instead a large pattern is rotated around the subject. OKN is much less useful than is rotatory chair testing as it is rarely affected substantially by disease. Optokinetic afternystagmus (OKAN) describes the eye movements that occur after the lights are turned out for OKN, and the subject is in complete darkness. OKAN is more sensitive to disease than OKN, but it is variable in normal subjects, which again limits its usefulness.
Contemporary rotatory chairs often implement optokinetic patterns that have no clinical utility. For example, a row of smiley faces. A "real" optokinetic stimulus should include at least 50% of the patient's visual field, and be a pattern rather than a line.
In VVI, a person is rotated with a visual surround or target also present. The most useful variant of this is to have a person look at (fixate) a laser that is fixed to the rotatory chair. VVI is generally a good index of ones CNS's ability to suppress nystagmus, and thus it is a measure of cerebellar and brainstem function.
OVAR is obtained by tilting the axis of chair rotation with respect to the gravitational axis. OVAR is largely a test of otolith function. While this is certainly of interest, OVAR is very nauseating and for this reason has been used little in clinical settings. In our opinion, VEMP testing is a much more practical method of assessing otolith function.
|Unilateral Loss sinusoidal responses. This image is from a patient who had a right sided vestibular nerve section. Note that although the gain and phase are abnormal, there is no asymmetry.|
In persons with unilateral vestibular loss, such as after a nerve section, there is also a typical pattern of rotatory chair testing in which the time constant is reduced and phase lead is increased. (Koizuka et al, 1995). See figure above. Rotatory chair testing is thus a valuable adjunct to ENG testing as well as VEMP testing by confirming an abnormality. Rotatory chair testing should not be used, by itself, for unilateral vestibular loss, as it may not be accurate in determining the side of lesion.
Rotatory chair tests are the "gold standard" for diagnosis of bilateral vestibular loss. One expects to see the following pattern on rotatory chair testing after a process that reduces vestibular function.
Complete bilateral vestibular loss. There is no response to rotation. There is no phase because there is nothing to quantify.
Expected rotatory chair results for acute bilateral impairment of vestibular function
Acute Bilateral Impairment of Vestibular Function
Rotatory Chair Gain Rotatory Chair Phase Mild Normal gain at all frequencies Mild phase lead, Moderate Less than 0.4 gain at 0.32 hz. Moderate phase lead Severe 0 to 0.1 gain at all frequencies Unobtainable
Chronically, gain recovers at mid frequencies. A lack of recovery seen on rotatory chair testing after 2 years suggests that the test was not done properly. There are several possible reasons -- the tested individual might be taking vestibular suppressants (such as a benzodiazepine or anticholinergic), or the person might be purposefully suppressing their vestibular responses (this possibility mainly occurs in legal cases where there is a benefit to an individual in pretending to be more ill than they really are). Optokinetic afternystagmus is sensitive to bilateral vestibular loss and should be absent both acutely and chronically.
Expected rotatory chair results for chronic bilateral impairment of vestibular function
Chronic Bilateral Impairment
of Vestibular Function
Gain Phase Gain-TC product Mild Normal Lead < 4 Moderate Less than 0.4 Lead 2-5 Severe Low normal at highest frequency (.32 hz), less than 0.1 at lower frequencies High at high frequency < 2
With respect to the Gain-TC product, we have recently published on this subject (Hain, Cherchi and Perez, 2018). One can follow vestibular function using this method.
There are several commercial devices, not incorporating motorized chairs, that provide data that overlaps in part with the data provided by rotational chairs. These devices are called "active head" devices, and compare eye movements induced by active motion of the head (rather than the passive movement induced by a motorized chair). Both of these tests measure -- the contribution of the inner ear, intentional cognitive input, and neck inputs to nystagmus rather than the contribution of the inner ear alone (Dell Santina et al, 2002). Brand names for these devices include the "Vorteq" and "VAT" among others.
|Overlap between AHR device and conventional rotatory chair.|
As can easily be seen from the plot above, these devices assess the high frequency range, but do not assess the low frequency range. They are not equivalent to rotatory chairs because the low-frequencies are missing. See this page for more details and examples of output.
|Typical picture of suppression of rotatory chair responses. Phase is normal at all frequences. Gain is reduced at high frequencies. For more about this, see this page.|
There are many ways that rotatory chair testing can be made inaccurate through technical errors. Fortunately however, there are less chances go make errors on a rotatory chair than the much more complicated ENG test.
The plot above shows the most common blunder -- a lack of "tasking" -- in other words, the technician did something else than keep the subject distracted during the test (? checked Facebook perhaps ?). Following is a brief list of chair blunders. See the "blunders" page if you are interested in more.
- Poor instruction of patient -- who suppressses responses
- Wrong value for distance to eye calibration array. When there is a mismatch, there can be a peculiar pattern where mid-frequency VOR gain is too high.
- Light leak -- this causes too high gain at all frequencies, including low frequency. Phase is high or 0 at all frequencies.
- Slip ring problem -- causes noise.
- Uncalibrated chair velocity, or uncalibrated OKN velocity. This can be spotted with the VVOR test as gain should be 1.0 at low frequencies.
We thank ICS, Interacoustic, and Micromedical Technology for use of figures of their equipment to illustrate this page.