Ocular vestibular evoked myogenic potentials in superior canal dehiscence.

OBJECTIVE: Patients with superior canal dehiscence (SCD) have large sound-evoked vestibular reflexes with pathologically low threshold. We wished to determine whether a recently discovered measure of the vestibulo-ocular reflex-the ocular vestibular evoked myogenic potential (OVEMP)-produced similar high-amplitude, low-threshold responses in SCD, and could differentiate patients with SCD from normal control patients. METHODS: Nine patients with CT-confirmed SCD and 10 normal controls were stimulated with 500 Hz, 2 ms tone bursts and 0.1 ms clicks at intensities up to 142 dB peak SPL. Conventional VEMPs were recorded from the ipsilateral sternocleidomastoid muscle to determine threshold, and OVEMPs were recorded from electrode pairs placed superior and inferior to the eyes. Three-dimensional eye movements were measured with scleral dual-search coils. RESULTS: In patients with SCD, OVEMP amplitudes were significantly larger than normal (p<0.001) and thresholds were pathologically low. The n10 OVEMP in the contralateral inferior electrode became particularly large with increasing stimulus intensity (up to 25 microV) and with up-gaze (up to 40 microV). Sound-evoked (slow-phase) eye movements were present in all patients with SCD (vertical: upward; torsional: upper pole away from the affected side; and horizontal: towards or away from the affected side), but began only as the OVEMP response became maximal, which is consistent with the surface potentials being produced by activation of the extraocular muscles that generated the eye movements. CONCLUSIONS: OVEMP amplitude and threshold (particularly the contralateral inferior n10 response) differentiated patients with SCD from normal controls. Our findings suggest that both the OVEMPs and induced eye movements in SCD are a result of intense saccular activation in addition to superior canal stimulation.

Vestibular responses to sound.

Research into vestibular responses to sound has evolved in four stages. The first, largely the work of Tullio in the 1920s, involved inspection of the eye, head, and postural responses to sound of alert animals with surgical fenestrae into various parts of the bony labyrinth. The second, begun in 1964 by Bickford and his group and continued by our group and then by others in the last 10 years, involves the measurement of evoked myogenic potentials to air-conducted and bone-conducted clicks and tones in normal humans. The third, begun by Mikaelian at about the same time as Bickford and continued by McCue, our group, and others, involves electrophysiological recordings of primary vestibular afferent neuron responses to sound in anesthetized animals. The fourth involves measurements of vestibulo-ocular responses to sound in humans with the Tullio phenomenon. It was begun by Minor and his group in 1998 with the observation that sound-induced nystagmus in humans, the Tullio phenomenon, aligned with the rotation axis of the superior semicircular canal. They then showed a defect in the temporal bone between the apex of the superior semicircular canal and the middle cranial fossa, which was the cause of most, if not all, cases of sound-induced nystagmus. Here some of the key observations made in each of these four stages are reviewed.

Individual semicircular canal function in superior and inferior vestibular neuritis.

OBJECTIVE: To examine the concept of selective superior and inferior vestibular nerve involvement in vestibular neuritis by studying the distribution of semicircular canal (SCC) involvement in such patients. BACKGROUND: Vestibular neuritis was traditionally thought to involve the superior and inferior vestibular nerves. Recent work suggests that in some patients, only the superior nerve is involved. So far there are no reported cases of selective involvement of the inferior vestibular nerve. METHODS: The authors measured the vestibuloocular reflex from individual SCC at natural head accelerations using the head impulse test. The authors studied 33 patients with acute unilateral peripheral vestibulopathy, including 29 with classic vestibular neuritis and 4 with simultaneous ipsilateral hearing loss, 18 healthy subjects and 15 surgical unilateral vestibular deafferented patients. RESULTS: In patients with preserved hearing, eight had deficits in all three SCC, suggesting involvement of the superior and inferior vestibular nerves. Twenty-one had a lateral SCC deficit or a combined lateral and anterior SCC deficit consistent with selective involvement of the superior vestibular nerve. Two patients with ipsilateral hearing loss had normal caloric responses and an isolated posterior SCC deficit on impulsive testing. The authors propose that these two patients had a selective loss of inferior vestibular nerve function. CONCLUSION: Vestibular neuritis can affect the superior and inferior vestibular nerves together or can selectively affect the superior vestibular nerve.

Inferior vestibular neuritis.

Sudden, spontaneous, unilateral loss of vestibular function without simultaneous hearing loss or brain stem signs is generally attributed to a viral infection involving the vestibular nerve and is called acute vestibular neuritis. The clinical hallmarks of acute vestibular neuritis are vertigo, spontaneous nystagmus, and unilateral loss of lateral semicircular function as shown by impulsive and caloric testing. In some patients with vestibular neuritis the process appears to involve only anterior and lateral semicircular function, and these patients are considered to have selective superior vestibular neuritis. Here we report on two patients with acute vertigo, normal lateral semicircular canal function as shown by both impulsive and caloric testing, but selective loss of posterior semicircular canal function as shown by impulsive testing and of saccular function as shown by vestibular evoked myogenic potential testing. We suggest that these patients had selective inferior vestibular neuritis and that contrary to conventional teaching, in a patient with acute spontaneous vertigo, unilateral loss of lateral semicircular canal function is not essential for a diagnosis of acute vestibular neuritis.

Impulsive testing of individual semicircular canal function.

In order to test the human angular vestibulo-ocular reflex in the dynamic range of normal head movements, we measured 3-dimensional compensatory eye-movement responses to low-amplitude (10-12 degrees), high-acceleration (3000-4000 degrees/s/s), passive, manually delivered head rotations (head "impulses") in the three planes of the semicircular canals in normal subjects, in subjects who had recovered from surgical unilateral vestibular deafferentation, and in patients after acute unilateral peripheral vestibulopathy, that is, from vestibular "neuritis." We found that canal-plane head impulses away from an intact semicircular canal, that is, toward a lesioned semicircular canal, invariably produce a vestibulo-ocular reflex with permanently low gain, typically less that 0.4 if the lesion is complete. These results are a necessary consequence of primary semicircular canal afferents being driven into inhibitory saturation by rapid angular accelerations. With practice, clinicians can learn to recognize the telltale compensatory saccades that patients with unilateral loss of semicircular canal function will make if asked to look at an earth-fixed target during head impulses in any one of the three semicircular canal planes.

The effects of galvanic stimulation on the human vestibulo-ocular reflex.

We studied the effects of 5 mA bilateral or unilateral, bipolar or monopolar, galvanic stimulation on the horizontal vestibulo-ocular reflex (hVOR) in six normal subjects during 0.01, 0.05, 0.1, 0.5 and 1 Hz yaw rotations and in two subjects during high-acceleration, low-amplitude yaw head rotations (head impulses). Bipolar galvanic stimulation induced horizontal nystagmus in all subjects and an asymmetry of the hVOR only during rotations below 0.1 Hz. Monopolar stimulation had no significant effect. The findings suggest that in humans galvanic stimulation affects those primary horizontal semicircular canal neurons that mediate the hVOR via indirect pathways through the velocity storage mechanism.

Contribution of the vertical semicircular canals to the caloric nystagmus.

Modulation of the caloric nystagmus in response to repositioning the plane of one vertical semicircular canal from gravitational horizontal to vertical during continuous caloric stimulation was used to measure the vertical canal's contribution to the nystagmus. The rationale was to examine the thermovective response from one vertical canal at a time, after a temperature gradient had been established across its two limbs. The nystagmus was measured and analysed in three dimensions using orthogonal head-referenced coordinates. The magnitude of each semicircular canal's contribution to the overall caloric response, the canal vector, was determined in non-orthogonal, contravariant semicircular canal plane coordinates. By using the canal plane reorientation technique and contravariant canal plane coordinates, we were able to measure the proportional thermovective response magnitude generated by each vertical canal during caloric stimulation. We found that the anterior canal contributed about one-third and the posterior canal about one-tenth as much as the lateral canal did to the overall caloric response magnitude when it was reoriented from horizontal to vertical. Comparison of the eye rotation axis before and after each vertical canal plane reorientation, with the geometry of the stimulated semicircular canals, also showed directional modulation of the caloric nystagmus by the vertical canal response. When one vertical canal plane was horizontal during caloric stimulation, the eye rotation axis aligned with the resultant of the other vertical canal and the lateral canal response axes. After vertical canal plane reorientation, the eye rotation axis realigned towards the resultant of the maximally stimulated vertical canal and the lateral canal, by 55.2+/-33.9 degrees (mean+/-SD) after anterior canal plane reorientation and by 32.3+/-21.2 degrees after posterior canal reorientation.

Head impulses reveal loss of individual semicircular canal function.

We studied individual semicircular canal responses in three dimensions to high-acceleration head rotations ("head impulses") in subjects with known surgical lesions of the semicircular canals, and compared their results to those of normal subjects. We found that vestibular-ocular reflex (VOR) gains at close to peak head velocity in response to yaw, pitch and roll impulses were reliable indicators of semicircular canal function. When compared to normals, lateral canal function showed a 70-80% decrease in VOR gain at peak of yaw head velocity during ipsilesional yaw impulses. After the loss of one vertical canal function there was a 30-50% decrease in vertical and torsional VOR gain in response to ipsilesional pitch and roll impulses respectively. Bilateral deficits in anterior or posterior canal function resulted in a 80-90% decrease in vertical VOR gain during ipsilesional pitch impulses, while the loss of ipsilateral anterior and posterior canal functions will result in a 80-90% decrease in torsional VOR gain in response to ipsilesional roll impulses. Three-dimensional vector analysis and animation of the VOR responses in a unilateral vestibular deafferented subject to yaw, pitch and roll impulses further demonstrated the deficits in magnitude and direction of the VOR responses following the loss of unilateral lateral, anterior and posterior canal functions.

Horizontal head impulse test detects gentamicin vestibulotoxicity.

BACKGROUND: Parenteral antibiotic therapy with gentamicin, even in accepted therapeutic doses, can occasionally cause bilateral vestibular loss (BVL) due to hair cell toxicity. OBJECTIVE: To quantify in patients with gentamicin vestibulotoxicity (GVT) the extent of acceleration gain deficit of the horizontal vestibulo-ocular reflex at different accelerations with a graded head impulse test (HIT) in comparison with standard caloric and rotational testing. To characterize the corresponding HIT catch-up saccade pattern to provide the basis for its salience to clinicians. METHODS: Horizontal HIT of graded acceleration (750 degrees-6,000 degrees/sec2) was measured with binocular dual search coils in 14 patients with GVT and compared with 14 normal subjects and a control subject with total surgical BVL. RESULTS: Patients showed mostly symmetric HIT gain deficits with a continuous spectrum from almost normal to complete BVL. Gain deficits were present even at the lowest head accelerations. HIT gain correlated better with caloric (Spearman rho = 0.85, p = 0.0001) than rotational testing (rho = 0.55, p = 0.046). Cumulative amplitude of overt saccades after head impulses was 5.6 times larger in patients than in normal subjects. Compared with previously published patients after unilateral vestibular deafferentation, GVT patients with BVL generated only approximately half the percentage of covert saccades during head rotation (23% at 750 degrees/sec2 to 46% at 6,000 degrees/sec2). CONCLUSIONS: Head impulse testing is useful for early bedside detection of gentamicin vestibulotoxicity because most patients, even those with partial bilateral vestibular loss (BVL), have large overt saccades. Covert saccades, which can conceal the extent of BVL, are only approximately half as frequent as in unilateral patients, but may be present even in total BVL.

Head impulse test in unilateral vestibular loss: vestibulo-ocular reflex and catch-up saccades.

BACKGROUND: Quantitative head impulse test (HIT) measures the gain of the angular vestibulo-ocular reflex (VOR) during head rotation as the ratio of eye to head acceleration. Bedside HIT identifies subsequent catch-up saccades after the head rotation as indirect signs of VOR deficit. OBJECTIVE: To determine the VOR deficit and catch-up saccade characteristics in unilateral vestibular disease in response to HIT of varying accelerations. METHODS: Eye and head rotations were measured with search coils during manually applied horizontal HITs of varying accelerations in patients after vestibular neuritis (VN, n = 13) and unilateral vestibular deafferentation (UVD, n = 15) compared to normal subjects (n = 12). RESULTS: Normal VOR gain was close to unity and symmetric over the entire head-acceleration range. Patients with VN and UVD showed VOR gain asymmetry, with larger ipsilesional than contralesional deficits. As accelerations increased from 750 to 6,000 degrees /sec(2), ipsilesional gains decreased from 0.59 to 0.29 in VN and from 0.47 to 0.13 in UVD producing increasing asymmetry. Initial catch-up saccades can occur during or after head rotation. Covert saccades during head rotation are most likely imperceptible, while overt saccades after head rotation are detectable by clinicians. With increasing acceleration, the amplitude of overt saccades in patients became larger; however, initial covert saccades also became increasingly common, occurring in up to about 70% of trials. CONCLUSIONS: Head impulse test (HIT) with high acceleration reveals vestibulo-ocular reflex deficits better and elicits larger overt catch-up saccades in unilateral vestibular patients. Covert saccades during head rotation, however, occur more frequently with higher acceleration and may be missed by clinicians. To avoid false-negative results, bedside HIT should be repeated to improve chances of detection.

Gentamicin vestibulotoxicity impairs human electrically evoked vestibulo-ocular reflex.

BACKGROUND: Electrical vestibular stimulation is believed to directly activate the vestibular afferents to mediate an electrically evoked vestibulo-ocular reflex (eVOR). Gentamicin, an aminoglycoside antibiotic, induces vestibulotoxicity by hair cell damage and death. OBJECTIVE: To determine if human eVOR is impaired by hair cell damage and death in systemic gentamicin vestibulotoxicity (GV). METHODS: Three-dimensional binocular eye movements evoked by bilateral, bipolar, 100 msec direct current-step at intensities of 0.9, 2.5, 5.0, 7.5, and 10.0 mA were recorded with dual-search coils in 12 GV patients, and the results were compared to 13 healthy subjects. RESULTS: Normal eVOR was predominantly torsional, comprising phasic eVOR initiation and cessation acceleration pulses at 9 msec latency after current onset and offset, with a tonic eVOR velocity-step during the 100 msec intervening period of maintained current. Normal phasic eVOR increased, while tonic eVOR scaled linearly, with current intensity. GV impaired phasic eVOR more severely than tonic eVOR, and prolonged the latency to 12-13 msec. In patients without mechanical response to vestibular tests, phasic eVOR was reduced to one-fifth of normal amplitude, doubled in duration, had reduced ability to vary with current intensity, and threshold was increased. Tonic eVOR was reduced to one-third of normal, but still scaled linearly with current intensity. Patients, who retained partial mechanical responses to vestibular tests, had phasic eVOR impairment without tonic eVOR abnormality. CONCLUSION: Impairment of evoked vestibulo-ocular reflex (eVOR) in gentamicin vestibulotoxicity (GV) suggests that vestibular hair cells, activated by electrical stimulation, mediate the eVOR. Abnormalities of the eVOR, especially the phasic component, might be a marker of vestibular injury in GV.

Click-evoked vestibulo-ocular reflex: stimulus-response properties in superior canal dehiscence.

BACKGROUND: An enlarged, low-threshold click-evoked vestibulo-ocular reflex (VOR) can be averaged from the vertical electro-oculogram in a superior canal dehiscence (SCD), a temporal bone defect between the superior semicircular canal and middle cranial fossa. OBJECTIVE: To determine the origin and quantitative stimulus-response properties of the click-evoked VOR. METHODS: Three-dimensional, binocular eye movements evoked by air-conducted 100-microsecond clicks (110 dB normal hearing level, 145 dB sound pressure level, 2 Hz) were measured with dual-search coils in 11 healthy subjects and 19 patients with SCD confirmed by CT imaging. Thresholds were established by decrementing loudness from 110 dB to 70 dB in 10-dB steps. Eye rotation axis of click-evoked VOR computed by vector analysis was referenced to known semicircular canal planes. Response characteristics were investigated with regard to enhancement using trains of three to seven clicks with 1-millisecond interclick intervals, visual fixation, head orientation, click polarity, and stimulation frequency (2 to 15 Hz). RESULTS: In subjects and SCD patients, click-evoked VOR comprised upward, contraversive-torsional eye rotations with onset latency of approximately 9 milliseconds. Its eye rotation axis aligned with the superior canal axis, suggesting activation of superior canal receptors. In subjects, the amplitude was less than 0.01 degrees, and the magnitude was less than 3 degrees/second; in SCD, the amplitude was up to 60 times larger at 0.66 degrees, and its magnitude was between 5 and 92 degrees/second, with a threshold 10 to 40 dB below normal (110 dB). The click-evoked VOR magnitude was enhanced approximately 2.5 times with trains of five clicks but was unaffected by head orientation, visual fixation, click polarity, and stimulation frequency up to 10 Hz; it was also present on the surface electro-oculogram. CONCLUSION: In superior canal dehiscence, clicks evoked a high-magnitude, low-threshold, 9-millisecond-latency vestibulo-ocular reflex that aligns with the superior canal, suggesting superior canal receptor hypersensitivity to sound.

Benign positional nystagmus: a study of its three-dimensional spatio-temporal characteristics.

OBJECTIVE: To describe the spatial and temporal characteristics of benign positional nystagmus (BPN) subtypes in benign positional vertigo (BPV) due to vestibular lithiasis affecting one or more semicircular canals (SCCs). BACKGROUND: Activation of SCC receptors by sequestered otoconia, either freely moving (canalithiasis) or cupula-adherent (cupulolithiasis) during head position changes with respect to gravity, is the accepted cause of BPV. Although accurate identification and interpretation of BPN is critical to BPV therapy, no rigorous, kinematically correct three-dimensional spatio-temporal analysis of BPN in all its forms exists. METHODS: Using dual-search scleral coils, the authors recorded BPN provoked by Dix-Hallpike or supine ear-down test in a two-axis whole-body rotator in 44 patients with refractory BPV. To localize the SCC affected, BPN rotation axes were compared to SCC axes, axes orthogonal to average SCC planes. RESULTS: Sixteen patients had upbeat, geotropic-torsional BPN in the Dix-Hallpike test to one side and five to both sides, with BPN rotation axes clustered around the lowermost posterior SCC axis. Seven had direction-changing horizontal BPN, three geotropic (canalithiasis) and four apogeotropic (cupulolithiasis), with rotation axes around the lowermost and uppermost horizontal SCC axis. Seven had predominantly downbeating BPN with rotation axes clustered around one superior SCC axis. Nine had upbeat, horizontal-torsional BPN with rotation axes located between posterior and horizontal SCC axes of the lowermost ear suggesting simultaneous lithiasis in both SCCs. BPN vector-guided repositioning therapy was successful in 43 patients. CONCLUSION: Benign positional vertigo can affect one or more semicircular canals and three-dimensional recording with vector analysis of the benign positional nystagmus (BPN) can guide canalith repositioning therapy especially in refractory cases with atypical BPN.

Three-dimensional spatial characteristics of caloric nystagmus.

We investigated the three-dimensional spatial characteristics of caloric nystagmus during excitation and inhibition of the lateral semicircular canal in five normal human subjects. Each subject was repositioned in 45 degrees steps at 1-min intervals such that the right lateral semicircular canal plane was reoriented in pitch, from 135 degrees backwards from the upright position to 135 degrees forwards, while the right ear was continuously stimulated with air at 44 degrees C. In orientations in which caloric stimulus resulted in excitation of the right lateral semicircular canal, the eye velocity axis was orthogonal to the average orientation of the right lateral semicircular canal plane. However, in orientations in which caloric stimulus resulted in inhibition of the right lateral semicircular canal, the eye velocity axis was orthogonal to the average orientation of the left and not the right lateral semicircular canal plane. These findings suggest that velocity and direction of caloric nystagmus depend not only on the absolute magnitude of vestibular activity on the stimulated side but also on the differences in activity between the left and right vestibular nuclei, most probably mediated centrally via brainstem commissural pathways.

Jerk-waveform see-saw nystagmus due to unilateral meso-diencephalic lesion.

See-saw nystagmus is an uncommon but highly characteristic eye movement disorder comprising intorsion and elevation of one eye, with synchronous extorsion and depression of the other. It generally has a pendular waveform and is due to a midline, extrinsic, suprasellar mass lesion compressing or invading the brainstem bilaterally at the meso-diencephalic junction. This report deals with the clinical and MRI findings in three patients (and binocular three-dimensional quantitative oculographic findings in one patient) with a jerk waveform see-saw nystagmus due in each case to a unilateral meso-diencephalic lesion. In each patient the torsional component of the nystagmus fast phases rotated the upper poles of the eyes toward the side of the lesion. Jerk see-saw nystagmus can be clinically indistinguishable from pendular see-saw nystagmus and from the torsional-vertical nystagmus which occurs with medullary lesions. We propose that jerk see-saw nystagmus is due to unilateral inactivation of the torsional eye-velocity integrator, thought to be in the interstitial nucleus of Cajal, with sparing of the torsional fast-phase generator, thought to be in the adjacent rostral interstitial nucleus of the medial longitudinal fasciculus.

Three-dimensional vector analysis of the human vestibuloocular reflex in response to high-acceleration head rotations. I. Responses in normal subjects.

1. The kinematics of the human angular vestibuloocular reflex (VOR) in three dimensions was investigated in 12 normal subjects during high-acceleration head rotations (head "impulses"). A head impulse is a passive, unpredictable, high-acceleration (3,000-4,000 degrees/s2) head rotation of approximately 10-20 degrees in roll, pitch, or yaw, delivered with the subject in the upright position and focusing on a fixation target. Head and eye rotations were measured with dual search coils and expressed as rotation vectors. The first of these two papers describes a vector analysis of the three-dimensional input-output kinematics of the VOR as two indexes in the time domain: magnitude and direction. 2. Magnitude is expressed as speed gain (G) and direction as misalignment angle (delta). G is defined as the ratio of eye velocity magnitude (eye speed) to head velocity magnitude (head speed). delta is defined as the instantaneous angle by which the eye rotation axis deviates from perfect alignment with the head rotation axis in three dimensions. When the eye rotation axis aligns perfectly with the head rotation axis and when eye velocity is in a direction opposite to head velocity, delta = 0. The orientation of misalignment between the head and the eye rotation axes is characterized by two spatial misalignment angles, which are the projections of delta onto two orthogonal coordinate planes that intersect at the head rotation axis. 3. Time series of G were calculated for head impulses in roll, pitch, and yaw. At 80 ms after the onset of an impulse (i.e., near peak head velocity), values of G were 0.72 +/- 0.07 (counterclockwise) and 0.75 +/- 0.07 (clockwise) for roll impulses, 0.97 +/- 0.05 (up) and 1.10 +/- 0.09 (down) for pitch impulses, and 0.95 +/- 0.06 (right) and 1.01 +/- 0.07 (left) for yaw impulses (mean +/- 95% confidence intervals). 4. The eye rotation axis was well aligned with head rotation axis during roll, pitch, and yaw impulses: delta remained almost constant at approximately 5-10 degrees, so that the spatial misalignment angles were < or = 5 degrees. delta was 9.6 +/- 3.1 (counterclockwise) and 9.0 +/- 2.6 (clockwise) for roll impulses, 5.7 +/- 1.6 (up) and 6.1 +/- 1.9 (down) for pitch impulses, and 6.2 +/- 2.2 (right) and 7.9 +/- 1.5 (left) for yaw impulses (mean +/- 95% confidence intervals). 5. VOR gain (gamma) is the product of G and cos(delta). Because delta is small in normal subjects, gamma is not significantly different from G. At 80 ms after the onset of an impulse, gamma was 0.70 +/- 0.08 (counterclockwise) and 0.74 +/- 0.07 (clockwise) for roll impulses, 0.97 +/- 0.05 (up) and 1.09 +/- 0.09 (down) for pitch impulses, and 0.94 +/- 0.06 (right) and 1.00 +/- 0.07 (left) for yaw impulses (mean +/- 95% confidence intervals). 6. VOR latencies, estimated with a latency shift method, were 10.3 +/- 1.9 (SD) ms for roll impulses, 7.6 +/- 2.8 (SD) ms for pitch impulses, and 7.5 +/- 2.9 (SD) ms for yaw impulses. 7. We conclude that the normal VOR produces eye rotations that are almost perfectly compensatory in direction as well as in speed, but only during yaw and pitch impulses. During roll impulses, eye rotations are well aligned in direction, but are approximately 30% slower in speed.

Three-dimensional vector analysis of the human vestibuloocular reflex in response to high-acceleration head rotations. II. responses in subjects with unilateral vestibular loss and selective semicircular canal occlusion.

1. We studied the three-dimensional input-output human vestibuloocular reflex (VOR) kinematics after selective loss of semicircular canal (SCC) function either through total unilateral vestibular deafferentation (uVD) or through single posterior SCC occlusion (uPCO), and showed large deficits in magnitude and direction in response to high-acceleration head rotations (head "impulses"). 2. A head impulse is a passive, unpredictable, high-acceleration (3,000-4,000 degrees/s2) head rotation through an amplitude of 10-20 degrees in roll, pitch, or yaw. The subjects were tested while seated in the upright position and focusing on a fixation target. Head and eye rotations were measured with the use of dual search coils, and were expressed as rotation vectors. A three-dimensional vector analysis was performed on the input-output VOR kinematics after uVD, to produce two indexes in the time domain: magnitude and direction. Magnitude is expressed as speed gain (G) and direction as misalignment angle (delta). 3. G. after uVD, was significantly lower than normal in both directions of head rotation during roll, pitch, and yaw impulses, and were much lower during ipsilesional than during contralesional roll and yaw impulses. At 80 ms from the onset of an impulse (i.e., near peak head velocity), G was 0.23 +/- 0.08 (SE) (ipsilesional) and 0.56 +/- 0.08 (contralesional) for roll impulses, 0.61 +/- 0.09 (up) and 0.72 +/- 0.10 (down) for pitch impulses, and 0.36 +/- 0.06 (ipsilesional) and 0.76 +/- 0.09 (contralesional) for yaw impulses (mean +/- 95% confidence intervals). 4. delta, after uVD, was significantly different from normal during ipsilesional roll and yaw impulses and during pitch-up and pitch-down impulses. delta was normal during contralesional roll and yaw impulses. At 80 ms from the onset of the impulse, delta was 30.6 +/- 4.5 (ipsilesional) and 13.4 +/- 5.0 (contralesional) for roll impulses, 23.7 +/- 3.7 (up) and 31.6 +/- 4.4 (down) for pitch impulses, and 68.7 +/- 13.2 (ipsilesional) and 11.0 +/- 3.3 (contralesional) for yaw impulses (mean +/- 95% confidence intervals). 5. VOR gain (gamma), after uVD, were significantly lower than normal for both directions of roll, pitch, and yaw impulses and much lower during ipsilesional than during contralesional roll and yaw impulses. At 80 ms from the onset of the head impulse, the gamma was 0.22 +/- 0.08 (ipsilesional) and 0.54 +/- 0.09 (contralesional) for roll impulses, 0.55 +/- 0.09 (up) and 0.61 +/- 0.09 (down) for pitch impulses, and 0.14 +/- 0.10 (ipsilesional) and 0.74 +/- 0.06 (contralesional) for yaw impulses (mean +/- 95% confidence intervals). Because gamma is equal to [G*cos (delta)], it is significantly different from its corresponding G during ipsilesional roll and yaw, and during all pitch impulses, but not during contralesional roll and yaw impulses. 6. After uPCO, pitch-vertical gamma during pitch-up impulses was reduced to the same extent as after uVD; roll-torsional gamma during ipsilesional roll impulses was significantly lower than normal but significantly higher than after uVD. At 80 ms from the onset of the head impulse, gamma was 0.32 +/- 0.13 (ipsilesional) and 0.55 +/- 0.16 (contralesional) for roll impulses, 0.51 +/- 0.12 (up) and 0.91 +/- 0.14 (down) for pitch impulses, and 0.76 +/- 0.06 (ipsilesional) and 0.73 +/- 0.09 (contralesional) for yaw impulses (mean +/- 95% confidence intervals). 7. The eye rotation axis, after uVD, deviates in the yaw plane, away from the normal interaural axis, toward the nasooccipital axis, during all pitch impulses. After uPCO, the eye rotation axis deviates in same direction as after uVD during pitch-up impulses, but is well aligned with the head rotation axis during pitch-down impulses.