|Home | About | Journals | Submit | Contact Us | Français|
While active dynamic visual acuity (DVA) has been shown to improve with gaze stabilization exercises, we sought to determine whether DVA during passive head impulses (pDVA) would also improve following a rehabilitation course of vestibular physical therapy (VPT) in patients with unilateral and bilateral vestibular hypofunction. VPT consisted of gaze and gait stabilization exercises done as a home exercise program. Scleral search coil was used to characterize the angular vestibulo-ocular reflex (aVOR) during pDVA before and after VPT. Mean duration of VPT was 66 ± 24 days, over a total of 5 ± 1.4 outpatient visits. Two of three subjects showed improvements in pDVA with a mean reduction of 43% (LogMAR 0.58 to 0.398 and 0.92 to 0.40). Our data suggest improvements in pDVA may be due in part to improvements in aVOR velocity and acceleration gains or reduced latency of the aVOR. Each subject demonstrated a reduction in the ratio of compensatory saccades to head impulses after VPT. Preliminary data suggest that active gaze stability exercises may contribute to improvements in pDVA in some individuals.
Individuals with vestibular hypofunction commonly complain of oscillopsia or “blurring” of the visual world with head movements. In this case, oscillopsia is due to the inability of the vestibulo-ocular reflex (VOR) to maintain gaze stability during head movement. Retinal slip (i.e. images “slipping” off of the fovea) is widely regarded as the primary mechanism underlying degraded visual acuity in people with vestibular pathology [12,20]. Functional activities that involve unpredictable (passive) head perturbations (e.g. unexpectedly stepping off of a curb, vehicular travel, ambulation) can present significant gaze stabilization challenges to people with vestibular hypofunction and have been shown to reduce visual acuity [8,13,22,31].
Dynamic visual acuity (DVA) is a behavioral measure of vestibular function that quantifies one’s ability to see clearly during predictable (active) or unpredictable (passive) head movements. While individuals with a normal VOR will likely demonstrate little or no degradation in visual acuity during head movements, those with vestibular hypofunction typically experience reduced visual acuity during head motion relative to their static visual acuity. Herdman et al demonstrated the reliability and validity of computerized DVA to identify the side of vestibular dysfunction during active head motion in patients with unilateral (UVH) and bilateral (BVH) vestibular hypofunction . DVA during passive head motion (pDVA) was recently validated with the scleral search coil technique for measuring individual semi-circular canal function in canal planes .
DVA has been shown to improve during active and passive head rotations in patients with UVH and BVH following vestibular physical therapy (VPT) that incorporates gaze stability exercises. Recent reports suggest improved DVA during active head rotations is due to increased slow component eye velocity and increased use of compensatory saccades . Although a significant improvement in passive dynamic visual acuity (pDVA) has been reported in subjects with UVH, regression analysis did not identify exercise type, age, time since onset, presence of oscillopsia, weeks of exercise, or initial pDVA score as significant contributors to the variance in the model or causative for this improvement. Thus, despite documented recovery of pDVA , the underlying mechanism for improvement among patients with UVH remains elusive.
pDVA is believed to provide an assessment of peripheral vestibular function largely independent of other systems that may augment gaze stability ; compensatory saccades [19,35–37],anticipatory slow phase eye movements , potentiation of the cervico-ocular reflex [23,32], or increased smooth pursuit eye velocities . Like the passive head impulse test described by Halmagyi et al. , data suggest pDVA testing similarly assesses the peripheral vestibular system by exciting the neural pathways of one semi circular canal while inhibiting contributions from the co-planar canal . These two tests are distinct however in the nature of information obtained; head impulse tests indicate the presence or absence of pathology and the degree of peripheral and/or central recovery; while the pDVA documents a more functional outcome given the nature of the task (i.e. identify the orientation of a visual target during head movement).
To our knowledge, no study has recorded eye movements during pDVA testing in patients with vestibular hypofunction undergoing VPT. The purpose of this pilot study was to investigate the effect of gaze stability exercises on pDVA in subjects with vestibular hypofunction.
In this investigation, we studied 3 subjects (mean age 55 ± 4 years, range 50–59) with vestibular hypofunction (n = 2 unilateral, UVH, n = 1 asymmetrical BVH) before and after a six-week course of active gaze stability exercises. The diagnosis of vestibular hypofunction was based on history of imbalance, non-positional vertigo or dizziness, physical exam revealing positive head thrust test towards the affected side, and no evidence on MRI for mass enhancing lesions within the internal auditory canals or cerebellopontine angle. Onset of symptoms was determined by patient self-report during the history-taking portion of the exam. Cause of vestibular hypofunction in these subjects included vestibular neuritis of viral etiology for both subjects with UVH, and presumed autoimmune disorder or possible Ménières disease affecting both ears for the patient with BVH. We also studied two normal subjects (mean age 56 ± 6 years, range 10 years) that had no complaints of vertigo, dizziness, or imbalance and had normal DVA for active horizontal head rotation . Healthy control subjects were age-matched to the subjects with vestibular hypofunction in order to compare pDVA scores. All patient subjects underwent passive head impulse testing using the three dimensional scleral search coil recording technique prior and subsequent to VPT to assess initial peripheral function, and the degree of peripheral recovery post intervention [25, 33]. Healthy subjects did not participate in any VPT.
Participation in this study was voluntary and all subjects consented to participate in accordance with a protocol approved by the Johns Hopkins School of Medicine Institutional Review Board.
Methods described here are identical to that previously reported by Schubert et al for this procedure . Briefly, a passive head impulse test (HIT) consists of unpredictable (timing and direction) manual head rotations or “impulses” with peak amplitude ~ 15°, peak velocity ~250°/sec and peak acceleration ~3500°/sec . The HIT was performed prior to pDVA testing on the assessment day. Subjects were tested while sitting centered and upright in a uniform magnetic field with their horizontal semi-circular canals oriented in the horizontal plane. This was achieved by aligning the subject’s head such that Reid’s line (the external anatomic reference for the horizontal canal) was also in the Earth-horizontal plane. Reid’s line is the shortest distance between the superior most point of the bony-cartilaginous junction of the external auditory canal and the lowest point of the cephalic edge of the infra-orbital rim. Prior to each head impulse, the subject’s head was placed in Reid’s Line for 5 seconds, enabling eye and head angular position to be calibrated in vivo while the subject fixated a target light-emitting diode (LED), which was positioned directly in front of the subject at 124 cm along the naso-occipital axis. The room was completely dark except for this LED.
Binocular eye movements were recorded in three rotational dimensions using a pair of search coils embedded in a silicone annulus placed on each eye. A search coil pair embedded in a bite block was used to measure head rotation. Eye and head angular positions were sampled at 500 Hz at 16-bit resolution. Analog signals (pre-sampled) were low-pass filtered with a single-pole analog filter that had a 3-dB bandwidth of 100 Hz .
Detailed description of the DVA test has been reported previously . Subjects in this study were initially assessed at baseline for both active and passive DVA using scleral search coils to quantify oculomotor behavior during testing. DVA of the patient subjects during active head motions was tested during weekly outpatient re-evaluations (without coils) to assess progress with their VPT protocol and as a determinant for re-assessment of pDVA using coils. When active DVA testing had normalized, pDVA was scheduled to be measured at the next clinic visit. During testing, subjects were seated 2 meters directly in front of a high resolution 18.1 viewable-inch monitor with a refresh rate of 85 Hz. Subjects who normally wore glasses or contact lenses for distant viewing were instructed to wear them during all pDVA testing. Static visual acuity was measured first by repeatedly displaying a single optotype (the letter E, randomly rotated each trial by 0, 90, 180 or 270 degrees) on a computer monitor. Subjects viewed five optotypes per acuity level, with optotype size then being decremented in steps equivalent to a visual acuity change of 0.1 LogMAR (log 10 X, where X = the minimum angle resolved, in arcmin, with 1 arcmin = 1/60 degrees) . The better one’s visual acuity, the lower one’s LogMAR score, with Log-MAR = −0.3, 0, 0.3, 0.7, 1.0 and 1.3 corresponding to Snellen visual acuity of 20/10, 20/20, 20/40, 20/100, 20/200 and 20/400, respectively. Static visual acuity was scored when the subject failed to correctly identify five optotypes on an acuity level or reached the LogMAR score of 0.000 (Snellen equivalency of 20/20 acuity).
For the dynamic component of the test, a single-axis Watson rate sensor (Micromedical Technologies, Inc., Chatham, IL, U.S.A.) was positioned on the subject’s head so that the sensor’s axis of maximum sensitivity approximately aligned with that of the horizontal semicircular canal . All subjects performed an initial practice trial for passive horizontal head rotations in order to control for practice effect . During each head rotation, an optotype ‘E’ randomly oriented in one of four directions was displayed on the monitor 2 m in front of the subject when head velocity, sensed by the Watson rate sensor, was between 120 and 180 deg/sec (for right side DVA testing, pDVA R) or between −180 to −120 deg/sec (for left side DVA testing, pDVA L) for more than 40 ms. For the pDVA R testing condition, the optotype only flashed for passive head impulses directed toward the right side. During the pDVA L condition, the optotype would only flash on the screen with leftward impulses. This method enables us to separately test horizontal semicircular canal contributions to gaze stability. Within each condition however, there were passive head impulses directed toward the non-optotype flashing side.
The pDVA test score was calculated by subtracting the static visual acuity LogMAR score from the dynamic visual acuity LogMAR score. Additional information about LogMAR computation has been published elsewhere .
Angular positions for eye and head with respect to space coordinates and eye with respect to head coordinates were represented by rotation vectors [15,26]. Head-in-space, eye-in-space, and eye-in-head velocity vectors were calculated from the corresponding rotation vectors . Head velocity was calculated and reported with reference to a head-fixed, right-handed coordinate frame (superior-inferior axes corresponds with yaw axes), so that eye-in-head and head velocities were expressed with reference to exactly the same coordinate frame .
Data were calculated for only those trials in which pDVA and HIT velocities were within the range of 120–180 deg/sec. The onset of each head thrust was identified with curve fitting. The time at which the magnitude of the fitted curve became greater than 2% of the curve’s peak magnitude (typically this threshold was ≈4°/s) was defined as the onset. A similar approach was used to identify the onset of the eye movement responses. Horizontal aVOR gain during pDVA testing and HIT was calculated by dividing peak horizontal eye velocity by peak horizontal head velocity. aVOR latency was determined by establishing the difference in time (milliseconds) for the head and eye velocities to reach 10 deg/sec.
We define compensatory saccade (CS) as a saccade occurring during the head rotation in the direction of the vestibular slow component. Peak CS velocity and amplitude were determined from velocity and position traces. CS latency was calculated by fitting a line to the saccade and subtracting it’s onset from head velocity onset (Fig. 1). To help distinguish CS traces from aVOR traces, we established an acceleration threshold criterion: CS were counted only when the acceleration exceeded the mean peak aVOR acceleration plus two standard deviations.
Subjects with vestibular hypofunction were asked to perform active gaze stability exercises 4–5 times per day, for a total of 20–30 minutes as part of a home exercise program. This exercise protocol illustrated in Table 1 was similar to the protocol, which has previously been established to improve DVA to active head rotations . Patient subjects were also instructed in static and dynamic balance exercises. Mean duration of VPT was 6 6± 24 days, over a total of 5 ± 1.4 out-patient visits. Subject compliance was established per self-report during scheduled outpatient visits. Patients were progressed in their rehabilitation program based on their ability to perform the previously performed set of exercises at the desired intensity level (head velocity and duration) for the specified duration.
The study design was a prospective, non-blinded, repeated-measures design. Individual differences between aVOR velocity and acceleration gains, aVOR latency, CS velocity, C S amplitude, CS duration, CS latency and CS frequency during DVA were assessed using independent T-tests assuming equal variance when indicated. Unequal variance was assumed when F-test results indicated that this was appropriate. Alpha was set at 0.05 for all tests. For the normal controls, statistical comparisons for the CS were performed only when the number of CS generated was ten or more. We described the number of CS per head rotation (CS/head rotation) as a ratio to determine change for the patient subjects, before and after VPT. DVA scores, aVOR gain values, and CS metrics in subjects with UVH and BVH were compared with age-matched normal controls.
Trials of pDVA and HIT data that included blinks or in which the subject did not fix on the target with both eyes at the onset of head rotation were not included in the analysis. pDVA scores, age, and CS ratio are listed in Table 2. Data is presented for ipsi and contralesional rotations.
pDVA scores for the control subjects were (Subject 1: L 0.00, R 0.120) and (Subject 2: L 0.398, R 0.301). aVOR velocity and acceleration gains during pDVA (pDVA L, pDVA R) were near unity in all cases (Table 3). Velocity and acceleration gains during passive HIT were similar for rotations to both directions (Table 4).
Both subjects used CS during pDVA testing. One subject demonstrated asymmetry in the ratio of CS to pDVA (pDVA R right: 0.26, left 0.67) (pDVA L right: 0.00 left: 0.33) while the other subject did not (pDVA R: right 0.15, left 0.00) (pDVA L: right 0.26 left: 0.15) (Table 2).
Two of the three patient subjects’ demonstrated improvement in pDVA upon re-assessment after six weeks of gaze stability exercises with a mean reduction of 44% (Table 2). The remaining patient subject did not show any change. Tables 2 and and33 illustrate aVOR gain and latencies across all patient subjects for each testing condition.
We did not find any trends in the velocity, amplitude, or latency of the CS between initial and final measurements. However, the ratio of CS/HIT uniformly decreased across all conditions in all subjects (Table 2).
The subject with BVH demonstrated a 32% improvement in pDVA towards the more involved, left side (0.582 to 0.398) and a 15 % improvement in pDVA towards the less involved right side (0.398 to 0.337). While aVOR velocity and acceleration gains remained unchanged for ipsilesional impulses during pDVA L, contralesional aVOR velocity and acceleration gains increased significantly during pDVA R in addition to HIT conditions (p < 0.05). Latency during pDVA R also reduced (p = 0.02) for both ipsi and contralesional head impulses (p < 0.05) (Tables 3 and and4).4). The ratio of CS to head impulses decreased during both pDVA conditions (Fig. 2).
The 56-year-old subject with UVH demonstrated improvement in pDVA to both left (0.919 to 0.4) (56% improvement) and right (0.453 to 0.4) (12%) optotype flashing conditions (Table 2). aVOR acceleration gains increased significantly in both pDVA L and pDVA R conditions and with ipsilesional HIT. There was no difference in aVOR velocity gains between initial and final pDVA test. Interestingly, a VOR latencies increased significantly in the HIT condition to both ipsi and contralesional impulses (p < 0.05) (Table 4). The CS ratio decreased during pDVA L from pre: 1.19 to post: 1.07. During pDVA R, the ratio decreased from pre: 0.11 to post: 0.05 (Table 2).
The 58-year-old subject with UVH demonstrated a slight worsening in pDVA as indicated by a LogMAR drop from 0.479 (pre) to 0.58 (post) on the ipsilesional side (21% decrement). No significant changes in aVOR gains or latency were observed during the pDVA L testing condition (see Table 3). Data was not collected during the pDVA R condition because the scleral coil was removed from the subject’s eye per the patient’s request. aVOR velocity gains did however, increase during the HIT condition during both ipsi and contralesional impulses (p < 0.05) (Table 4). The CS ratio decreased from 1.24 to 1.08 for ipsilesional impulses and 1.00 to 0.58 for contralesional impulses during the pDVA L condition.
While the effects [19,20,33] and mechanism of recovery  for active gaze stability exercises on predictable DVA have been demonstrated, this pilot study identifies possible effects of such a protocol on gaze stability during unpredictable head movements. Though not all study participants showed an improvement in pDVA, improvement in 2 out of 3 subjects suggests active gaze stability exercises may also provide benefit for unpredictable head rotations, potentially broadening the scope of treatment parameters beyond what has been previously reported. Our findings suggest that active gaze stability exercises may have carry-over effects on passive dynamic visual acuity. Possible mechanisms for pDVA improvement may include: 1) peripheral vestibular system recovery, 2) central compensation, or 3) modification of compensatory saccades. Improved aVOR velocity or acceleration gains and latency may be the physical manifestation of the peripheral vestibular recovery and/or central compensation associated with improved pDVA.
While partial to complete resolution of aVOR gain deficits within weeks of onset of vestibular neuritis (VN) to low frequency assessments (i.e. calorics or rotary chair testing) is well established [4,27]; persistent deficits (years) have also been documented when subjects are assessed with high frequency testing methods . This suggests that the recovery of vestibular function is frequency and/or velocity dependent . In a non-intervention study looking at the recovery of high acceleration aVOR gains after VN, Palla and Straumann reported early gain increases with a mean magnitude of 0.15 during ipsilesional head impulse testing within 3–4 weeks of condition onset. The authors suggested that such increases in the gain may have been due to either “incremental recovery of peripheral vestibular function” or possibly due to the effects of “unchanged residual peripheral input driving an emerging central compensation mechanism” . Like the authors in the aforementioned study, we acknowledge an inability to infer causation with respect to gain improvement during passive head impulses in our subjects. However, it seems likely that early, moderate to large increases in ipsilesional gains in patients with VN may be due to peripheral recovery from partial resolution of their neuritis based on what is known of the condition’s natural history.
While subjects in the Palla and Straumann study demonstrated significant gain increases during the acute phase of recovery (i.e. the first month after onset), the mean time from onset to initiation of VPT in our study was 6.3 months with those demonstrating the greatest responses to VPT (subject 1 and subject 2) initiating the protocol at 6 and 9 months respectively. Given what has been reported in the literature about the natural course and early recovery following onset of VH [1,27,30], it seems reasonable to surmise that the improved high frequency content velocity and acceleration gains reported in our subjects may be attributed, at least in part, to augmentation of central compensatory mechanisms and/or adaptation due to a treatment effect from the gaze stabilization protocol rather than simply a function of acute recovery.
Prior work by Schubert et al demonstrated that improvement in aVOR gain lead to improved aDVA independent of peripheral recovery . In this study, although two of three subjects did demonstrate improvement in aVOR velocity and acceleration gains after a six-week course of VPT, these gains occurred concurrently with increases in HIT gains. This makes it difficult to attribute improvement in pDVA scores to VPT independent of peripheral recovery or central compensation effects.
Reduced latency may have contributed to improved pDVA scores. In this study, we showed significant reduction in aVOR latency (after VPT), which corresponded with pDVA improvement.
The role of pre-programmed eye movements and efference copy (EC) is well documented in the rehabilitation literature [9,18–20,33,35]. Prevailing theory suggests that superior aVOR gains and LogMAR scores during DVA testing with active head rotations relative to passive head rotations may be due in part to the effects of efference copy or pre-programmed eye movements (compensatory saccades). While EC is unlikely to play a significant role during pDVA testing given the lack of an efferent motor command generated by the subject, it is possible that afferent information from retinal sources in the form of a position signal , oculomotor proprioceptors , or neck proprioceptors  could augment vestibular contributions to gaze stability during passive head movements .
Schubert et al. found that subjects with UVH and BVH who performed vestibular rehabilitation demonstrated an increase in the recruitment of CS with active head movements during DVA after rehabilitation was completed . It has also been suggested that the recruitment of CS is inversely correlated with aVOR gain [33,35,37] such that as peripheral aVOR gain improves, the frequency of CS decreases. This finding was corroborated in 2/3 of our patient subjects suggesting a decreased reliance on CS as an extra-vestibular mechanism for gaze stability as vestibular function improved with increased velocity or acceleration gains.
The three patient subjects in this study represent a limited cross section of vestibular pathology, ages and etiology. Another limitation of this study was that we assessed change in pDVA when change in active DVA occurred. It is possible that despite significant improvements in active DVA, subjects may not have had enough time for the rehabilitation to have a more beneficial effect on pDVA.
Preliminary data suggests gaze stability during passive head rotations can be improved following an active gaze stability program. Possible explanations for this improvement include increased aVOR velocity and acceleration gains, and reduced aVOR latency either in conjunction with, or independent of peripheral recovery.
This study was supported by the National Institute on Deafness and Other Communication Disorders (MCS K23-007926, AAM R03-DC007346) and the Foundation for Physical Therapy, American Physical Therapy Association.