Abstract
Video head impulse testing, an instrumented expansion of the bedside head impulse test using video eye tracking and quantification of eye and head movement, is a relatively new objective measure of dynamic semicircular canal function. This article provides an overview of both the bedside and video head impulse tests. The anatomical and physiological bases of the test are briefly reviewed, along with test administration and response characteristics in healthy individuals and patients with vestibular loss. Literature examining the use of the video head impulse test in several common vestibular disorders, in children, and in patients with acute vestibular syndrome is summarized.
Keywords
Head impulse test, Semicircular canal, Vestibular, Vestibulo-ocular reflex, vHIT, Video head impulse test, VOR
Acknowledgement
This work was supported by a Merit Review (F1540-R) and by the Auditory and Vestibular Research Enhancement Award Program (C4339-F), both sponsored by the Rehabilitation Research and Development Service, Department of Veterans Affairs, Washington, D.C. The contents of this chapter do not represent the views of the Department of Veterans Affairs or the United States Government.
Introduction
Many techniques are available for the assessment of the vestibular system, particularly the assessment of horizontal semicircular canal (SCC) function. The head impulse test (HIT) was first described by Halmagyi and Curthoys in 1988 and is the most widely used bedside test of SCC function. Since that time, a relatively limited number of vestibular research laboratories have used the magnetic field scleral search coil technique, the gold standard for recording eye movements, to record three-dimensional (3D) eye movement during the HIT to validate its use as a test of SCC function and to better understand the clinical utility of the test. The search coil technique is generally considered too invasive, expensive, and time-intensive for routine clinical use, and those limitations were the impetus for the development of two-dimensional (2D) eye-movement recording via high-speed video cameras. The initial studies using high-speed digital video cameras to record eye movement during the HIT (now referred to as the video head impulse test or vHIT) were published by several independent research laboratories, with the critical experiments demonstrating comparable results for simultaneous video and scleral search coil recordings for head impulses in horizontal and vertical planes in normal controls and in selected patients with well-defined vestibular losses. Since those initial studies, at least four vHIT devices have been developed, and the routine clinical use of the vHIT has expanded rapidly. The purpose of this chapter is to provide an overview of the HIT with an emphasis on the vHIT. The reader is also directed to Halmagyi et al. for a comprehensive review of this topic.
Background of the Head Impulse Test
The angular vestibulo-ocular reflex (VOR) ensures gaze stability during head rotations by generating eye movements that are equal and opposite to head rotation. The gain of the VOR (eye velocity divided by head velocity) for natural head movements, therefore, approaches unity in healthy individuals. The VOR has three main anatomic components: (1) the SCCs and the superior vestibular nerve and inferior vestibular nerve afferents in the peripheral vestibular system, (2) the vestibular and ocular motor nuclei in the brainstem, and (3) the extraocular muscles. The SCCs are positioned in three nearly orthogonal planes within the head, allowing for the detection of head rotation in 3D space. The SCCs function as angular accelerometers in a push-pull fashion with two coplanar canals on each side of the head working together, i.e., the left and right horizontal SCCs, the right anterior and left posterior SCCs or “RALP,” and the left anterior and right posterior SCCs or “LARP.” For example, during rightward head rotation in the horizontal plane, the discharge rate of the right horizontal SCC afferents increases and, at the same time, the discharge rate of the left horizontal SCC afferents decreases relative to the resting discharge rate. The difference in output between the right and left horizontal SCCs drives the leftward compensatory eye movement of the VOR so that the eyes remain still in space during head rotation and enable stable vision. The observation or measurement of eye movement, therefore, can aid in the detection and localization of vestibular pathology because of the relationship between the function of vestibular sensory receptors in the inner ear and the compensatory eye movements produced by the VOR. The majority of bedside and laboratory tests of vestibular function involve the observation or measurement of horizontal eye movements (i.e., horizontal VOR) produced by stimuli that activate the horizontal SCCs and the superior vestibular nerve.
The HIT is used to assess dynamic function of the SCCs and was initially described as a bedside test used to measure the function of each horizontal SCC. The HIT is based on two principles or laws (Ewald’s laws) in vestibular physiology: (1) eye movements evoked by stimulation of a single SCC occur in the plane of that canal, and (2) excitatory responses have a larger dynamic range than inhibitory responses. Specifically, because the three SCC pairs (horizontal, RALP, and LARP) are nearly orthogonal to each other, a head impulse delivered in the plane of one pair will stimulate mainly that pair and not the other two SCC pairs. In addition, the VOR during a canal-plane impulse toward a particular SCC is driven largely by that SCC and not by its coplanar counterpart because of the asymmetric response (excitatory > inhibitory) of primary vestibular afferents (see Fig. 8.1 ). The HIT, therefore, can assess the function of each SCC separately and, in a patient with unilateral vestibular hypofunction, the gain of the VOR during ipsilesional head impulses (i.e., head rotation toward the side of the vestibular loss) will be lower than the gain during contralesional head impulses.
Bedside Head Impulse Test
The bedside HIT is the most widely used bedside test of SCC function and has largely been used to assess the function of horizontal SCCs. To perform the bedside HIT, the clinician sits in front of the patient, holds the patient’s head, and instructs the patient to keep staring at an earth-fixed target (e.g., the clinician’s nose). The clinician turns the patient’s head abruptly and unpredictably to the left or right, through a small angle (10–20 degrees). Patients with normal VOR function will be able to maintain their gaze on the target during head rotation to either side (i.e., the head rotation produces a short-latency compensatory eye movement [the VOR] that is equal and opposite of the head rotation). In contrast, patients with unilateral vestibular loss (UVL) will be unable to maintain their gaze on the target during ipsilesional head rotation. Instead, the eyes will move with the head (because of the reduction in VOR gain) and are taken off target so that at the end of the head rotation the patient must make a voluntary corrective saccade back to the target. The corrective or “catch-up” saccade is visible to the clinician and is, therefore, called an overt saccade. The observation of an overt saccade is an indirect sign of horizontal SCC hypofunction on the side toward which the head was rotated. The bedside HIT depends on the timing and size of the corrective saccades and on the ability of the clinician to accurately observe the corrective saccades. Corrective saccades are observed at the end of both rightward and leftward head impulses in patients with bilateral vestibular loss (BVL).
Advantages of the bedside HIT include the ability to detect UVL and BVL, no equipment cost, short test time, portability, and the ability to assess horizontal SCC function at frequencies of head rotation that are representative of head movements that occur during activities of daily living. In addition, the bedside HIT is less likely to elicit the vertigo and occasional nausea associated with caloric stimulation.
There are, however, a number of significant limitations related to the bedside HIT. Specifically, the bedside HIT is a subjective test, and there is no objective measure of the corrective saccades or VOR gain; the outcome of the test is based on the clinician’s subjective visual observation of the presence or absence of overt saccades; and the interpretation of the bedside HIT depends on the experience of the clinician. Furthermore, during the bedside HIT, the magnitude of the head acceleration is unknown and likely varies within a single clinician/patient and between individual clinicians/patients. Corrective saccades that occur during the head rotation are called covert saccades, and these cannot be observed by the clinician.
The false-negative rate of the bedside HIT in patients with peripheral vestibular disorders has been estimated at 14% based on the rate of occurrence of isolated covert saccades detected with the vHIT. The clinical application of the bedside HIT has been limited to the evaluation of horizontal SCC function and has not been used routinely for the assessment of vertical SCCs.
Magnetic Field Scleral Search Coil Head Impulse Test
The magnetic field scleral search coil technique has been used to record 3D eye movement during the HIT, and has demonstrated that covert saccades occur during the HIT in patients with UVL and BVL and that the presence of covert saccades can produce false-negative results for the bedside HIT even in patients with total UVLs. The 3D search coil technique has also been used to measure the function of individual vertical SCCs using head impulses delivered in the vertical plane. Importantly, it was shown that 3D and 2D (horizontal and vertical eye movement) scleral search coil techniques were equally accurate in detecting isolated hypofunction in horizontal and vertical SCCs, indicating that 2D methods (i.e., video pupil tracking) are capable of assessing all six SCCs independently.
Video Head Impulse Test
Based on the bedside HIT first described by Halmagyi and Curthoys, the vHIT is a relatively new clinical test of dynamic SCC function that uses high-speed digital video camera(s) to record eye movement during and immediately after head impulses in horizontal and vertical planes. The stimulus for the vHIT is the same stimulus used for the bedside HIT and consists of manual, passive (clinician moves patient’s head), unpredictable, brisk head rotations with peak angular velocity of ∼100 to ∼400 degrees/second and a peak angular acceleration of ∼1000 to ∼4000 degrees/second. The vHIT instrumentation consists of high-speed (∼250 frames/second) monocular or binocular digital infrared video camera(s), a laptop computer, and software. The video camera uses pupil detection methods to record 2D eye movements. The vHIT detects and records abnormal eye movements (i.e., overt and covert saccades) and provides measures of VOR gain. Depending on the vHIT device, the camera is either embedded in head-worn goggles or mounted on a tripod facing the patient. Head movement is recorded by an inertial measurement unit (triaxial linear accelerometer and gyroscopes) mounted on the head-worn goggles or by the change in the angle of head position during the head impulse as recorded by an external camera. Notably, prototypes of at least two commercially available vHIT devices have been validated with comparable results obtained for simultaneous video and magnetic field scleral search coil recordings for head impulses in horizontal and vertical planes in normal controls and in selected patients with well-defined vestibular losses.
Video Head Impulse Test Technique
To perform the horizontal vHIT, patients are seated and eye position is calibrated immediately before testing. Patients are then instructed to maintain their gaze on an earth-fixed visual target located at a distance of ∼1 meter straight ahead at eye level. The clinician stands behind the patient and manually rotates the head abruptly and unpredictably to the left or right through a small angle (10–20 degrees) in the horizontal plane to stimulate the left or right horizontal SCC (see Fig. 8.2 ). In general, two different hand placements have been used to perform horizontal head impulses: (1) hands on top of the patient’s head or (2) hands placed on each side of the face at the jaw line. In a sample of 40 healthy adults, higher average VOR gain values were obtained for the hands-on-head technique than for the hands-on-jaw technique. In contrast, higher average head velocities were obtained for the hands-on-jaw technique than for the hands-on-head technique ; and the jaw technique was associated with more frequent vHIT artifacts than the head technique in a group of patients with acute vestibular syndrome.
The effect of initial head position on the vHIT has also been examined (i.e., initial head position at midline with impulses directed laterally [outward impulses] vs. initial eccentric head position with impulses directed toward the midline [inward impulses]). In healthy subjects, there were no differences in either peak head velocity or amplitude for outward and inward head impulses, whereas there was a small but significant VOR gain difference noted, with higher VOR gain for outward impulses than for inward impulses. There was no difference in ipsilesional VOR gains for inward and outward head impulses in patients with acute UVL. Patients are probably less likely to anticipate the direction of the head impulse when the initial head position is at midline with impulses directed laterally.
To test either of the coplanar vertical canal pairs, the patient’s head is first turned either to the right (LARP) or to the left (RALP) ∼30–40 degrees relative to the trunk (i.e., ∼30–40 degrees relative to the central earth-fixed visual target), which aligns the vertical canal pair with the sagittal plane of the trunk. The patient is instructed to maintain gaze on the central target by “looking out” of the left corner (LARP) or right corner (RALP) of the eye. The clinician places one hand on top of the head and the other hand under the chin and rotates the head either forward and down toward the central fixation target (stimulates the anterior canal) or back and away from the fixation target (stimulates the posterior canal) (see Fig. 8.2 ). Alternatively, both the head and body can be turned ∼30–40 degrees relative to the central fixation target (gaze remains on the central target) before delivering the head impulses, and this position may minimize patient discomfort and neck strain relative to turning only the head. McGarvie et al. demonstrated the importance of maintaining gaze along the plane of the stimulated SCC during the vertical canal vHIT. Specifically, they showed a substantial decrease in VOR gain as the direction of horizontal gaze shifted from 40° (gaze direction aligned with the plane of the stimulated vertical canal) to 0° (gaze direction is straight ahead and not aligned with the plane of the vertical canal so the response is a combined vertical and torsional eye movement).
Video Head Impulse Test and Vestibulo-Ocular Reflex Gain
vHIT VOR gain is usually quantified as either position gain or velocity gain that is measured over a wide response interval or over limited time intervals usually associated with peak head velocity or peak head acceleration. Specifically, at least one vHIT device quantifies gain over a wide response interval, from 60 ms before the peak head velocity to the next zero-crossing of head velocity; position gain is then calculated as the ratio of the area under the desaccaded eye velocity waveform to the area under the head velocity waveform during the same time interval. Quantifying VOR gain over a wide response interval increases the likelihood that camera movement artifacts (e.g., goggle slippage) in the eye velocity traces will be reduced if the artifact is biphasic. In contrast, another commercially available vHIT device uses an interval of 100 ms post-head impulse onset, and gain is calculated over a ±20 ms interval around fixed head velocity times of 40, 60, 80, and 100 ms; gain is also calculated as the slope of the regression between eye and head velocities over the first 100 ms. A recent study has shown no difference in the mean gain and variance obtained with two different commercially available vHIT devices when the same gain calculation technique was used; however, lower variances were obtained for area gain than for regression gain.
Video Head Impulse Test Artifacts
Current commercially available vHIT devices provide the examiner with immediate visual and/or audio feedback regarding the adequacy of each head impulse. If the head impulse and eye movement response meet the criteria of the manufacturer’s data collection/processing algorithm, then the eye movement response is accepted and the gain of the VOR is calculated. In general, vHIT data collection/processing algorithms include a minimum velocity/acceleration for the head impulse and maintenance of eye tracking during the head impulse. It is important, however, for clinicians to manually inspect the “accepted” waveforms for the presence of artifacts.
Mantokoudis et al. used laboratory simulations in a healthy subject to deliberately induce artifacts that were reliably reproduced and “accepted” by the data collection/processing algorithm used by a commercially available vHIT device. The simulations were used to develop a classification scheme of horizontal vHIT artifacts: (1) delay/phase shift, e.g., eye movement leads head movement (caused by loose goggle strap); (2) high gain (calibration error); (3) pseudosaccades (eye blinks); (4) double peaks in eye trace (examiner touching goggles or mini eye blinks); (5) head impulse overshoot or bounce (examiner error produced by head direction reversal following deceleration of the head impulse); (6) eye trace goes in wrong direction (patient inattention); and (7) eye trace oscillations (loss of pupil tracking). Fig. 8.3 illustrates vHIT artifacts obtained from a sample of clinic patients according to the classification scheme of Mantokoudis et al.
Mantokoudis and colleagues used the artifact classification scheme to evaluate vHIT records obtained from a group of patients with acute vestibular syndrome and found that 44% of the vHIT records contained at least one artifact. The two most common artifacts that resulted in uninterpretable recordings were noisy eye movement traces with multiple peaks and eye movement traces with oscillations due to loss of pupil tracking. In a subsequent study that used the same data set, however, there was no clinically relevant difference between the unfiltered (artifacts included) and filtered (artifacts removed) VOR gain values in terms of sensitivity and specificity for the detection of posterior inferior cerebellar artery stroke.
Recently, Halmagyi et al. described artifacts caused by the eyelid briefly obscuring the pupil image during the vertical canal vHIT. The artifactual eye movement response during the anterior SCC head impulse is characterized by a biphasic eye movement trace, whereas the eye movement artifact during posterior SCC impulses has the appearance of a covert saccade that overlays the vertical eye velocity trace. The eyelid artifact is observed more frequently during anterior SCC impulses than during posterior SCC impulses, and it is generally recommended that the eyelids are taped up using medical-grade tape in order to avoid this artifact.
Video Head Impulse Test Results From Healthy Subjects and the Effects of Age
Several recent studies have reported normative VOR gain data for the horizontal vHIT and the vertical vHIT in relatively large samples of healthy community-dwelling individuals across a wide age span (∼10–90+ years of age) and over a range of peak head velocities (∼70 to ∼250 degrees/second). Overall, results of these studies indicated that VOR gain was constant through 70–80 years of age, and that there was a small decrease in VOR gain at relatively high peak head velocities at all ages. The average VOR gains collapsed across age ranged from ∼0.94 to 1.06 for horizontal SCCs with slightly lower average VOR gain values for vertical SCCs. The lower limit for normal VOR gain (mean VOR gain minus 2 SD) was ∼0.7 to ∼0.80 for horizontal canals and ∼0.60 to ∼0.70 for vertical canals. Fig. 8.4 shows the vHIT head and eye velocity waveforms obtained from a normal adult subject.
