Brain Advance Access originally published online on May 25, 2005
Brain 2005 128(7):1511-1524; doi:10.1093/brain/awh504
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Torsional optokinetic nystagmus after unilateral vestibular loss: asymmetry and compensation
1 Laboratoire de Neurobiologie Intégrative et Adaptative, Université de Provence–CNRS and 2 Service d'Oto-rhino-laryngologie et Chirurgie Cervico-faciale, Hôpital Nord, Marseille, France
Correspondence to: Liliane Borel, Laboratoire de Neurobiologie Intégrative et Adaptative, UMR 6149, Université de Provence–CNRS, Pôle 3C Case B, Centre de Saint-Charles 3, Place Victor Hugo, F-13331 Marseille Cedex 3, France E-mail: borel{at}up.univ-mrs.fr
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsReferences |
|---|
The aim of this study was to analyse torsional optokinetic nystagmus (tOKN) in 17 patients with Menière's disease before and after (1 week, 1 month and 3 months) a curative unilateral vestibular neurotomy (UVN). The tOKN was investigated during optokinetic stimulations around the line of sight directed towards either the lesioned or the healthy side, at various constant angular velocities. Dynamic properties of tOKN and static ocular cyclotorsion were analysed using videonystagmography. Patients' performances were compared with those of 10 healthy subjects. The results indicate that, in the acute stage after UVN, patients exhibited drastic impairment of tOKN velocity that depended on the direction of stimulation: tOKN velocity increased for ipsilesional stimulations and decreased for contralesional stimulations. These changes were responsible for a dramatic tOKN asymmetry, with ipsilesional directional preponderance of torsional slow-phase eye velocity. The changes were associated with static ocular cyclotorsion towards the operated side. Despite progressive compensation of tOKN deficits over time, tOKN velocity still differed from that recorded preoperatively, and tOKN asymmetry remained uncompensated 3 months after UVN. A static ocular cyclotorsion remained up to 3 months after lesion. These results are the first description of tOKN deficits and recovery after unilateral vestibular loss. They show that vestibular cues contribute to gaze stabilization during optokinetic stimulation around the line of sight. They also strongly suggest that tOKN impairment could be part of the long-term asymmetrical functions reported after unilateral loss of vestibular functions.
Key Words: gaze stabilization; roll optokinetic stimulation; static ocular cyclotorsion; unilateral vestibular neurotomy; functional recovery
Abbreviations: CCW = counterclockwise; CW = clockwise; D – 1 = tests carried out one day before neurotomy; D + 7 = tests carried out one week after neurotomy; D + 30 = tests carried out one month after neurotomy; D + 90 = tests carried out three months after neurotomy; IDP = index of directional preponderance; OKN = optokinetic nystagmus; tOKAN = torsional optokinetic afternystagmus; tOKN = torsional optokinetic nystagmus; UVN = unilateral vestibular neurotomy
Received December 6, 2004. Revised February 22, 2005. Accepted March 15, 2005.
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsReferences |
|---|
When a large visual pattern is moved in front of a subject, an optokinetic nystagmus (OKN) is elicited that consists of alternating quick and slow phases of eye movements. Torsional OKN (tOKN) involves ocular rotation around the visual axis, which contributes to stabilization of the images on the retina when a visual pattern is rotated around the line of sight. The dynamic properties of tOKN have been studied with rotation of a visual pattern at steady velocities (Collewijn et al., 1985
; Cheung and Howard, 1991; Thilo et al., 1999) and with sinusoidal oscillations (Cheung et al., 1995) in healthy humans.
In natural conditions, OKN operates in conjunction with vestibular eye movements to optimize gaze stabilization in a wider range of frequencies. In the case of torsional eye movements, OKN combines with ocular counter-rolling in response to static roll tilt of the head (Diamond and Markham, 1983), and with torsional vestibulo-ocular reflex elicited either by linear acceleration along the interaural axis (Lichtenberg et al., 1982) or by dynamic head-roll rotation (Collewijn et al., 1985). The interaction between optokinetic and vestibular eye movements has been described largely in behavioural, neurophysiological and neuroanatomical studies. So what are the effects of vestibular cues on OKN?
In healthy subjects, the effects of vestibular cues on optokinetic responses have been described particularly for horizontal and vertical OKN. There is evidence that dynamic rotatory (e.g. Demer, 1996) or linear (Buizza et al., 1980) vestibular stimulations increase horizontal OKN velocity. Moreover, experiments performed during orbital space flight support the contribution of otolithic inputs on OKN since horizontal and vertical OKN velocity increased in astronauts exposed to weightlessness (Clément et al., 1993). By contrast, when tonic vestibular cues are modified by static roll tilt of the head, horizontal OKN velocity decreased (Clément and Lathan, 1991). Studies dealing with the effects of vestibular cues on tOKN have also shown some discrepancies in results. Morrow and Sharpe (1993) found that tOKN was not modified by otolithic changes caused by the supine position, whereas Thilo et al. (1999) reported enhanced tOKN gain. Finally, microgravity phases of parabolic flights resulted in a tOKN gain increase (Cheung et al., 1995).
Loss of the vestibular input changes the dynamic properties of optokinetic eye movements. In bilateral vestibular-defective patients, horizontal OKN gain was drastically reduced (Zee et al., 1976). These findings were corroborated by data collected in bilateral labyrinthectomized rabbits (Barmack et al., 1980), cats (Fukushima and Fukushima, 1991) and monkeys (Cohen et al., 1973). The effect of unilateral labyrinthine dysfunction on OKN has not been widely investigated; studies have reported an ipsilesional directional preponderance of horizontal OKN in such patients (Coats, 1968; Abel and Barber, 1981). Congruently, unilateral loss of vestibular inputs in cats resulted in a strong asymmetry of horizontal OKN (Precht et al., 1981; Maioli et al., 1982). On the contrary, vertical OKN remained unchanged in hemilabyrinthectomized cats (Gustave Dit Duflo et al., 1998). However, no information is available on tOKN after unilateral vestibular loss.
The present study was designed firstly to analyse how acute unilateral vestibular loss affects the torsional optokinetic responses in humans. For this, we investigated patients with Menière's disease who underwent a curative unilateral vestibular neurotomy (UVN) for their intractable vertigo. We focused on the dynamic properties and symmetry of tOKN recorded during constant optokinetic stimulation around the line of sight. A second aim of this study was to analyse the time-course of the recovery of the torsional optokinetic responses after UVN. Patients were examined during the week before surgery and at three postoperative times: 1 week, 1 month, and 3 months after UVN. Patients' performances were compared with those of healthy control participants.
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsReferences |
|---|
Subjects
Experiments were carried out on 17 patients suffering from unilateral Menière's disease. Patients included seven men and 10 women aged 31–63 years (mean ± SD, 45 ± 10 years). For all patients, the neuro-otological examination revealed pure unilateral deficit, with attacks of vertigo, hearing loss, and tinnitus causing them to stop their professional activity. Preoperative clinical status is reported for each patient in Table 1. The unilateral vestibular deficit, determined by bithermal caloric irrigation with cold (30°C) and warm (44°C) water, averaged 30 ± 16%. Hearing loss, assessed using a clinical audiometer calibrated in 1 dB steps (Madsen OB822; Madsen Electronics), averaged 53 ± 25 dB in the affected ear (one patient was cophotic). Videonystagmography revealed horizontal spontaneous nystagmus in darkness in eight patients with slow-phase eye velocity averaging 1.5 ± 1.1°/s (Table 2). The history of the disease ranged from 1 to 20 years (mean 6 ± 6 years) and the attacks of vertigo happened daily (20% of patients), weekly (27%) or monthly (53%). Because the patients became pharmacoresistant to antivertigo substances, they underwent unilateral neurotomy to abolish intractable vertigo without affecting hearing acuity. Surgical treatment consisted of retrosigmoid vestibular neurotomy (Magnan et al., 1991
) on the right side for nine patients and on the left side for the eight others. All patients were able to stand straight and motionless 1 week after surgery. They received no drug that could have influenced their performance. The caloric test showed a total lack of responses on the operated side acutely after UVN as well as later on.
|
|
Patients' performances were compared with those of 10 healthy control subjects. Controls included five men and five women aged 23–53 years (mean 40 ± 10 years). They were selected because they had normal vestibular, visual and oculomotor functions. Each subject gave informed consent according to the Declaration of Helsinki and the study was approved by the local Ethics Committee.
Sessions
Patients were examined 1 day before surgical treatment (D – 1), when they did not report attacks of vertigo during the preceding week and had not received antivertigo medications. All of them were tested at three postoperative times: during the acute stage (the first week after surgery; D + 7) and during the compensatory stage (1 month and 3 months after surgery; D + 30 and D + 90). Controls were tested four times with the same intervals between the sessions to prevent the effect of familiarity with and habituation to optokinetic stimulation and experimental procedures.
Optokinetic stimulation
Torsional OKN was provided by rotation of a large disc (1.5 m in diameter) positioned 1.3 m in front of the subjects. The disc was covered with a pseudorandom-dot visual pattern. Dots were 1 cm in diameter and subtended 0.44° of the visual field with a density of 1175 dots/m2. A black circular target was located at the centre of the disc on the rotation axis (Fig. 1A). For all the subjects, we had previously aligned the target with their own direction of gaze. The disc was motorized and rotated clockwise (CW) and counterclockwise (CCW) with constant angular velocities (5, 20, 80 and 120°/s).
|
Experimental procedures
Examinations were done with subjects wearing goggles and instructed to stand straight and motionless in front of the disc. The goggles supported a camera in front of one eye and narrowed the visual field to the intended visual scene (40° in the horizontal and vertical planes). Each experimental session was composed of static ocular cyclotorsion recordings and tOKN recordings. Each recording lasted 51.2 s.
Static ocular cyclotorsion and spontaneous nystagmus exhibited by patients were recorded in light and in darkness. In the light, the subjects were instructed to look at the central target of the stationary visual pattern. This request was aimed at preventing exploratory saccades. In darkness, the goggles were manually closed so that they were perfectly opaque, and subjects were asked to stabilize their gaze on the memorized target.
Torsional OKN was recorded during optokinetic stimulations at each stimulation velocity (5, 20, 80 and 120°/s) and for each direction (ipsilesional and contralesional for the patients, and CCW and CW for the controls). During the first 5 s of the recording period, the disc remained stationary. During the following 30 s, disc rotation elicited tOKN. Finally, the disc remained stationary during the last 16.2 s, allowing torsional optokinetic afternystagmus (tOKAN), if any, to be measured (Fig. 1B). Again, subjects were asked to gaze at the central target throughout the recording period. Sequences of optokinetic stimulations were randomized across sessions and for each subject. They were triggered after a few minutes of rest to eliminate possible perturbations due to transition from a condition to another.
Eye movement recording
Eye movements were sampled at 25 Hz with videonystagmography. A CCD camera was fixed on the goggles and lit with infrared LEDs. We recorded the ipsilesional eye in the patients since it was impossible for them to undergo two complete trials of optokinetic stimulations, especially in the acute stage after UVN. In addition, no difference in static ocular cyclotorsion between the ipsilesional and contralesional eyes was reported in unilateral vestibular-defective patients (Curthoys et al., 1991). For healthy subjects, we had previously verified that the characteristics of the tOKN responses did not differ whichever eye was recorded. We recorded the non-dominant eye to maximize ease on the eye. We assumed that stimulating one eye and measuring the response of the other did not affect the changes in the pathological population with respect to controls. In fact, there is a high degree of conjugacy of the right and left eyes during torsional eye movements (Diamond and Markham, 1981, 1983; Collewijn et al., 1985; Seidman et al., 1995; Merfeld et al., 1996). In addition, tOKN elicited in binocular viewing showed dynamic properties that were similar for the two eyes when they were recorded simultaneously (Young et al., 1981) or successively (Collewijn et al., 1985; Cheung and Howard, 1991). Finally, stimulating one eye and measuring the responses of the other did not affect the horizontal and vertical OKN symmetry described for binocular stimulation (van den Berg and Collewijn, 1988).
Data processing
Torsional eye movements were analysed off-line using a method based on a mathematical dynamic neural network; the basic principle is described by Guillemant and colleagues (Guillemant et al., 1995). This approach uses a combination of supervised and dynamic learning to identify the iris pattern and the pupil. A ring-shaped area covering the iris from the external border of the pupil to the sclera is first defined manually. The algorithm calculates synaptic weights of mathematical neurons. It is capable of learning the rotating iris structure in order to separate it from the non-rotating structure (diode reflections, eyelids and so on). Learning was done in two stages. (i) Supervised learning was a static stage using numerical iris rotation around the pupil centre; for each mathematical neuron angle it allowed the rotation value to be proportional to the real iris angle. (ii) The dynamic learning stage was used to separate rotating and non-rotating iris structures and then eliminate the latter. The calculation uses the difference between the angular global result and the individual results to modulate the synaptic weights of the other neurons. Within a few cycles, the output of the network then gives a response to geometric apparent rotation of the iris. Ten images during rotation are enough to obtain a gain equal to one. The calculation time was 25 ms on a Pentium 1.2 GHz PC for the output of the network with 64 000 mathematical neurons. The angular resolution was 0.05° and the sampling rate was 25 Hz. Image reflections of the illumination lights from the cornea and occlusion of the pupil by blinks are implicitly dealt with by the algorithm. Since torsional eye movements cannot be made voluntarily, we first verified the calibration of the system by using a glass eye mounted on a protractor device rotating around the three orthogonal axes. The system was linear between ±20°; the magnitude of torsional eye deviations for the patients never exceeded the linearity range. This method of recording eye movement has been described for measurements of torsional eye position (Borel et al., 2001). The noise (corresponding to the standard deviation of the signal) was measured with the immobile glass eye. In this condition, the noise was 0.02°. A similar value was found in our experimental condition during the fixation period in light after blinks were discarded. The crosstalk measured both using a glass eye mounted on a protractor device and in subjects performing horizontal, vertical and oblique saccades was less than 1% in the torsional signal.
Velocity of the tOKN and of the torsional spontaneous nystagmus, recorded in light and in darkness, was assessed by differentiating the eye position signal. For differentiation of the sampled signal it was not necessary to smooth the position signal or to calculate its derivation with statistical compensation. In fact, the neural network of the torsional algorithm already has a statistical effect by automatically eliminating noisy iris points and reinforcing successful iris points. These successful iris points (see the darkest points in Fig. 1A) correspond to higher grey-level gradients in their neighbourhood. Velocity of the tOKN was expressed as the mean maximal velocity of each torsional slow phase (in degrees per second) during the whole period of disc rotation for each stimulation velocity, and for each direction. To process the tOKN velocity of patients, we subtracted the velocity of the torsional spontaneous nystagmus recorded in the same condition, i.e. in the light, from the tOKN velocity. This correction was necessary since some patients displayed torsional spontaneous nystagmus in the light when looking at the stationary visual pattern (Table 2). In addition, the mean slow-phase eye velocity of the horizontal and vertical components of spontaneous nystagmus and of the nystagmus response to torsional optokinetic stimulation was computed based on the position of the pupil's centre.
To investigate the dynamic sensitivity of the subjects to optokinetic stimulation, the evolution of the tOKN velocity was plotted as a function of the log of the stimulation velocity (Farooq et al., 2004). The dynamic sensitivity was evaluated as the mean slope of the individual linear regression curve best fitting the experimental data.
Moreover, we determined tOKN symmetry by computing an index of directional preponderance (IDP). For the patients, this ratio was expressed as follows:
 |
 |
This data processing concerns the dynamic torsional eye movements. To get an overall picture of the torsional eye movements after unilateral vestibular loss, we also measured static ocular cyclotorsion. To initially determine static ocular torsion, subjects were required to gaze at the central target with their eyes wide open to remove the eyelid from the iris. The initialization of the iris neural network of a subject required less than 10 successive images. The iris signature was then memorized. It represents the reference for the subject. Such a reference was established from the preoperative recordings (D – 1). For further processing, the iris signature of the subject was recognized automatically. The amplitude of static ocular cyclotorsion for patients after vestibular lesion was calculated as the geometric difference between the neural network recorded postoperatively and the neural network memorized before UVN (Fig. 2A). The consistency of torsional position measurement was determined in control subjects in the same conditions and using exactly the same procedures as those described for UVN patients. Data indicate a high degree of consistency of torsional position across the different experimental sessions (Fig. 2A, histogram).
|
) after vestibular lesion was calculated as the geometric difference between the neural network recorded postoperatively (D + 7) and the neural network memorized before UVN (D – 1). The histograms show the mean static ocular cyclotorsion (± confidence interval (CI), vertical bars) for all patients and all sessions after UVN. Data from controls illustrate the consistency of torsional position measurement across the different sessions. (B) The raw data from a typical patient are illustrated for ipsilesional (upper part) and contralesional (lower part) optokinetic stimulations at a constant angular velocity of 80°/s. Results are shown for the different experimental sessions: before UVN (D – 1) and after (D + 7, D + 30, D + 90).Statistical analysis
Each dependent variable described above was analysed using mixed design analysis of variance (ANOVA) with group (patients versus controls) as the between-subject factor and with session (D – 1, D + 7, D + 30, D + 90), direction of stimulation (ipsilesional and contralesional for patients, CW and CCW for controls) and stimulation velocity (5°/s, 20°/s, 80°/s, 120°/s) as within-subject factors. In addition, repeated-measures ANOVAs in patients and in controls were performed separately with session, direction of stimulation and stimulation velocity as within-subject factors. Finally, Pearson correlations were calculated between eye movement measures and each parameter of the clinical evaluation of the preoperative status of the patients. Results were considered statistically significant at P < 0.05.
|
Results |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsReferences |
|---|
Torsional eye movements in healthy subjects
Circular optokinetic stimulation elicited a tOKN that was made up of torsional slow phases directed towards the direction of disc rotation, interrupted by saccades rotating the eyes in the opposite direction. Torsional OKN was characterized by the direction of the slow phases. A typical tOKN for one healthy subject is shown in Fig. 1B. ANOVA of the mean tOKN velocity for the 10 controls revealed that the only significant source of variation was stimulation velocity [F(3,27) = 19.97, P < 0.0001].
Table 3 summarizes mean tOKN velocity in the controls for each stimulation velocity over the four recording sessions and for CCW and CW stimulations. We found that tOKN velocity increased with stimulation velocity up to 80°/s. As a rule, tOKN velocity rarely exceeded 3°/s, confirming that the compensatory function of tOKN is poor. In addition, the present data confirmed that, for control subjects, tOKN velocity was linearly related to the log of the stimulation velocity (r2 = 0.95 and r2 = 0.86 for CCW and CW stimulations, respectively; Fig. 4A). Furthermore, the data emphasized the consistency of oculomotor responses over time, indicating low intra-individual variability. Finally, tOKN velocity did not differ significantly for CW and CCW optokinetic stimulations. These data were corroborated by IDP values close to zero whatever the stimulation velocity (see controls in Fig. 5).
|
After the disc had stopped, tOKAN was sometimes observed. When it existed, tOKAN was composed of two or three torsional slow phases whose direction was similar to that of tOKN (Fig. 1). Since tOKAN was not consistent and not recorded for all subjects, it was discarded in the rest of the study.
Effects of unilateral vestibular loss on torsional eye movements
Torsional OKN velocity
Repeated-measures ANOVAs conducted on patients and controls indicated that unilateral vestibular loss constituted a main effect providing the sources of variation of tOKN velocity. Evidence of changes in tOKN velocity after UVN is that a significant interaction of group x session x direction of stimulation was observed (Table 4). This effect indicated that tOKN velocity was affected differently over time in patients with respect to the controls for ipsilesional and contralesional optokinetic stimulations.
|
Concerning the group of 17 Menière's disease patients, ANOVA performed before and after UVN evidenced that, contrary to the control group, direction of stimulation [F(1,16) = 21.08, P < 0.0005] was a main source of variation. The interaction session x direction of stimulation [F(3,48) = 30.92, P < 0.0001] confirmed postoperative changes in tOKN velocity depended on direction of stimulation. In addition, there was a highly significant effect of stimulation velocity [F(3,48) = 28.64, P < 0.0001]. Figure 2B illustrates raw tOKN data recorded at the stimulation velocity of 80°/s for one patient through the four experimental sessions. It shows that tOKN changed after surgery according to the session and the direction of stimulation. The mean values of tOKN velocity are shown in Fig. 3 for ipsilesional and contralesional stimulations in the group of patients before and after UVN. The comparison of patients' preoperative (D – 1) and acute postoperative (D + 7) data revealed drastic changes in the dynamic properties of tOKN 1 week after UVN. Interestingly, tOKN velocity was altered differently according to the direction of stimulation. Indeed, for ipsilesional optokinetic stimulation, planned comparisons indicated that tOKN velocity was significantly increased (P < 0.0001). The increase typically averaged 123% (P < 0.0001), 107% (P < 0.0001), 142% (P < 0.0001) and 142% (P < 0.0005) for stimulation velocities of 5, 20, 80 and 120°/s, respectively. Moreover, it is noteworthy that tOKN velocity parameters recorded acutely after UVN were higher than those of the controls (P < 0.05). Inversely, for contralesional optokinetic stimulation, tOKN velocity was significantly lower than preoperatively (P < 0.0001). Torsional OKN was either abolished or highly reduced, especially for the lower stimulation velocity. The decrease averaged 86% (P < 0.0005), 70% (P < 0.0001), 72% (P < 0.0001) and 66% (P < 0.005) with respect to the preoperative data for stimulation velocities of 5, 20, 80 and 120°/s, respectively. Again, tOKN velocity was significantly lower than for the controls (P < 0.01).
|
In the later stages after UVN (D + 30, D + 90) and for ipsilesional stimulation, tOKN velocity decreased progressively but remained higher than preoperatively (D + 30, P < 0.01; D + 90, P < 0.05). Mirror image data were obtained for contralesional optokinetic stimulation, with a progressively increased tOKN velocity that did not regain the preoperative values (D + 30, P < 0.05; D + 90, P < 0.05).
Figure 4B shows that tOKN velocity remained linearly related to the log of the stimulation velocity over the experimental sessions for ipsilesional stimulations (D – 1, r2 = 0.71; D + 7, r2 = 0.87; D + 30, r2 = 0.96; D + 90, r2 = 0.88) as well as for contralesional ones (D – 1, r2 = 0.87; D + 7, r2 = 0.92; D + 30, r2 = 0.99; D + 90, r2 = 0.99). However, UVN induced significant changes in the mean slope of the linear regression curves, suggesting changes in sensitivity to the optokinetic stimulation in the acute stage after UVN. Indeed, the mean slope of the linear regression curves differed significantly for stimulations directed towards the ipsilesional and the contralesional side (P < 0.001), with an increase and a decrease in the mean slope, respectively. These data argue for enhanced and reduced sensitivity to the ipsilesional and contralesional stimulation, respectively, restricted to the acute stage after UVN (Fig. 4C).
|
Mean tOKN velocity did not differ for patients tested before surgery and for controls, whatever the direction of stimulation. To specify how the clinical status of the patients relates to eye movement measures, each parameter of the preoperative clinical evaluation was correlated with tOKN velocity. No correlation was found in patients between tOKN velocity and history of the disease (e.g. r2 = 0.01, P = 0.75 for contralesional stimulation at 20°/s), vestibular deficit (e.g. r2 = 0.07, P = 0.33), hearing loss (e.g. r2 = 0.01, P = 0.36) or horizontal spontaneous nystagmus velocity (e.g. r2 = 0.003, P = 0.91), whatever the stimulation velocity. Similarly, no correlation was found between the patients' preoperative clinical status and tOKN velocity for the postoperative sessions (Table 1). In addition, for patients as for controls, tOKAN was not consistently observed over sessions.
Finally, one can ask whether torsional optokinetic stimulation evokes horizontal and vertical components in the nystagmus response before and after UVN. Data indicated that the horizontal and vertical eye positions were correctly stabilized at the central target. Indeed, the horizontal and vertical components of the nystagmus response were very weak (Table 5). In addition, statistical analysis indicated that the velocity of the horizontal and vertical components of the nystagmus response differed neither according to the experimental session [horizontal component, F(3,48) = 2.34, P = 0.10; vertical component, F(3,48) = 1.26, P = 0.31] nor according to the direction of stimulation [horizontal component, F(1,16) = 1.65, P = 0.24; vertical component, F(1,16) = 1.02, P = 0.34].
|
Torsional OKN symmetry
The ANOVAs of the mean IDP in patients and controls are summarized in Table 4. Significant effects of group, session, and a group x session interaction were evidenced. Repeated-measures ANOVA on the 17 Menière's disease patients indicated that tOKN symmetry varied as a function of the experimental session [F(3,48) = 25.74, P < 0.0001] and of the stimulation velocity [F(3,48) = 4.89, P < 0.005].
As seen in Fig. 5, mean IDP values revealed drastic changes in tOKN symmetry for patients through the four recording sessions with respect to the control data. One week after UVN, the mean IDP showed a directional preponderance for ipsilesional stimulation, reaching 80 ± 4%. Detailed analyses revealed that IDP was drastically higher than preoperative data (P < 0.001). Despite the marked decrease in mean IDP later on, tOKN asymmetry persisted: at 1 month after UVN the mean was 30 ± 6% (P < 0.001) and at 3 months it was 12 ± 5% (P < 0.05). These data show that recovery of tOKN symmetry was incomplete. For patients tested before surgery, mean IDP did not differ significantly from that of the controls (mean, –8 ± 5%; P = 0.26), indicating that patients displayed symmetrical tOKN velocity for optokinetic stimulations directed towards the diseased and healthy sides. Again, we found no correlation between IDP and patients' preoperative clinical status either preoperatively or for the different postoperative sessions.
|
Static ocular cyclotorsion
ANOVA on static ocular cyclotorsion showed drastic changes after UVN [F(2,32) = 54.62, P < 0.0001]. Indeed, all patients displayed a static ocular cyclotorsion due to a rotation of the upper pole of both eyes towards the operated side. The impairment of static ocular position became obvious at D + 7 (P < 0.0001), with a mean cyclotorsion of 8.7 ± 0.7°. The amplitude of the static cyclotorsion progressively decreased during the three postoperative months, with mean cyclotorsion of 3.6 ± 0.5° (P < 0.0001) and 2.3 ± 0.6° (P < 0.0001) at D + 30 and D + 90, respectively (Fig. 2A). In addition, we verified that static cyclotorsion was similar in light and in darkness (P = 0.47). Finally, to investigate the consequences of static ocular position on tOKN, we analysed the relation between the amplitude of static ocular cyclotorsion and either tOKN velocity or IDP. Amplitude of static ocular cyclotorsion was not significantly related either to tOKN velocity, whatever the stimulation velocity and the direction of stimulation, (e.g. D + 7, r2 = 0.19, P = 0.12; D + 30, r2 = 0.01, P = 0.79; D + 90, r2 = 0.02, P = 0.83, for contralesional stimulation at 20°/s) or with IDP (e.g. D + 7, r2 = 0.18, P = 0.13; D + 30, r2 = 0.09, P = 0.44; D + 90,r2 = 0.15, P = 0.61, for stimulations at 20°/s).
Discussion
This study investigated how patients with unilateral vestibular loss (Menière's disease patients) stabilized their gaze during optokinetic stimulation around the line of sight. Dynamic properties and symmetry of torsional optokinetic responses were analysed before curative unilateral vestibular neurotomy and during the recovery time-course. Patients' performances were compared with those of healthy control participants.
Asymmetrical changes in tOKN velocity after acute unilateral vestibular loss
Our data show that UVN drastically altered the dynamic properties of the torsional optokinetic responses during stimulation around the visual axis. We found that torsional slow-phase eye velocity was strongly decreased for contralesional stimulation, but was increased for ipsilesional stimulation as early as the first postoperative week. These findings resulted in marked asymmetry of tOKN. Note, however, that tOKN velocity remained linearly related to the log of the stimulation velocity, even if the characteristics of the linear relationship were temporarily altered.
The tOKN asymmetry we found for neurotomized patients is consistent with that reported in previous studies dealing with horizontal OKN in unilateral vestibular defective patients (Coats, 1968; Abel and Barber, 1981; Pfaltz and Ildiz, 1982; Vitte et al., 1994). These authors showed reduced slow-phase eye velocity of horizontal OKN when optokinetic stimulation was directed towards the intact side. Such results corroborate the data collected in unilateral neurotomized and labyrinthectomized animals (Precht et al., 1981; Maioli et al., 1982; Hamann et al., 1998). In contrast, the transitory increase in tOKN velocity we observed acutely after UVN for ipsilesional stimulation has never been reported, either for horizontal or for vertical OKN. This finding could be due to the lack of data acutely after the surgery or of comparison between preoperative and postoperative data. We hypothesize that ipsilesional enhanced tOKN velocity could result from reweighting of visual cues. Such an increased weight of visual inputs after UVN has been reported in patients with similar disease as regards body orientation and stabilization in space (Borel et al., 2001, 2002). In support of this view is the observation of enhanced tOKN in humans exposed to microgravity (Cheung et al., 1995). These authors suggested that increased optokinetic effects result from the lack of contradiction of the visual input by the otolith. In contrast with our observations, some animal studies have reported that horizontal OKN was unaffected (Precht et al., 1981; Hamann et al., 1998), and even slightly impaired for ipsilesional stimulation (Maioli et al., 1982). The difference in OKN direction (torsional in our study versus horizontal in others) could account for the discrepancies with data collected in animal models. However, the reported increased weight of visual cues cannot explain the asymmetry of tOKN responses for ipsilesional and contralesional stimulations.
Such asymmetry could be partly imputable to neurophysiological changes in central vestibular structures after UVN. Specifically, the vestibular nuclei, a relay in optokinetic pathways, displayed drastically imbalanced neuronal resting activity after the vestibular inputs had been unilaterally suppressed (for review see Smith and Curthoys, 1989). In particular, the neurons mediating optokinetic stimulation towards the healthy side (mainly type I neurons in the ipsilesional vestibular nuclei) displayed a large decrease in their resting activity after UVN (Precht et al., 1966; Smith and Curthoys, 1988b; Ris et al., 1995). These observations could be taken as the physiological basis of the decrease in tOKN velocity for contralesional optokinetic stimulation. Conversely, there is evidence that resting activity increased in neurons mediating optokinetic stimulations towards the lesioned side (mainly type I neurons in the contralesional side) acutely after UVN (Smith and Curthoys, 1988a; Ris and Godaux, 1998). Such an increase could represent the underlying mechanism for the rise in tOKN velocity we found here for optokinetic stimulations directed towards the lesioned side. Another explanation for lesion-induced changes in the tOKN could be a change in the dynamic sensitivity of second-order vestibular neurons to optokinetic stimulation after the vestibular lesion. In line with this view, Zennou-Azogui and colleagues showed an increase in visual-dependent response of second-order vestibular neurons in the acute stage after UVN (Zennou-Azogui et al., 1994). This enhanced weight of visual cues could be related to the rise in tOKN velocity we observed for stimulations towards the lesioned side. Enhanced weight could also account for the increased sensitivity to optokinetic stimulation evidenced by the rise in the slope of the linear regression curves acutely after the lesion.
Our results indicate that torsional spontaneous nystagmus occurred in conjunction with tOKN asymmetry following neurotomy. Such torsional spontaneous nystagmus has been reported in various neurological patients (Lopez et al., 1992). However, it should be noted that the torsional component of the spontaneous nystagmus recorded in light could not account for the tOKN asymmetry in the present study. Firstly, we subtracted its influence from the mean torsional slow-phase eye velocity. Secondly, whereas the torsional spontaneous nystagmus had vanished in all the patients but one 1 month after UVN, the tOKN remained asymmetrical until D + 90. Furthermore, Brandt and colleagues reported a lack of correlation between horizontal OKN asymmetry and spontaneous nystagmus intensity in unilateral vestibular defective patients (Brandt et al., 1978).
Interestingly, tOKN asymmetry was associated with static ocular cyclotorsion towards the operated side. We found this torsion in both light and darkness, with the same amplitude of about 8° during the acute stage. Our data corroborate previous observations of static ocular cyclotorsion in humans after unilateral peripheral vestibular lesion (Halmagyi et al., 1979) and neurotomy (Curthoys et al., 1991; Riordan-Eva et al., 1997; Borel et al., 2001). It has been proposed that static ocular cyclotorsion is partly due to imbalanced otolithic cues (Curthoys et al., 1991). Indeed, otolithic stimulations induce ocular torsions (Lichtenberg et al., 1982; Zink et al., 1998). However, the contribution of imbalanced semicircular canal cues cannot be excluded since Strupp et al. (2003) reported 17° ocular cyclotorsion after semicircular canal lesion. To date, no one has investigated the influence of static ocular cyclotorsion on tOKN symmetry. At first, one might expect that static ocular cyclotorsion, which rolled the upper pole of both eyes towards the lesioned side, could have limited the torsional eye movements during ipsilesional stimulation, which also induces slow phases towards the lesioned side. Congruently, static ocular cyclotorsion might have facilitated tOKN during contralesional optokinetic stimulation, which induces torsional slow phases away from the lesioned side. On the contrary, our data show that ipsilesional stimulation produces torsional eye movements towards the lesioned side beyond the static ocular cyclotorsion. Thus, the excyclotropia induced on the ipsilesional eye by the torsional slow phases during ipsilesional optokinetic stimulation, for example, are to be added to excyclotropia induced by the static ocular cyclotorsion. Torsional OKN does not seem mechanically limited by static ocular cyclotorsion. In addition, it stays within the range of free torsional eye movements. Finally, we found no correlation between the amplitude of static ocular cyclotorsion and tOKN asymmetry after UVN. The overall data indicate that tOKN asymmetry is probably due to neurophysiological changes in central vestibular structures.
The present data indicate that tOKN measured in patients tested before they were operated did not differ from data collected in healthy subjects. We found symmetrical tOKN for stimulations directed towards the affected and healthy side in patients preoperatively and for CCW and CW stimulations in controls. In addition, tOKAN was very weak and not consistently quantifiable. Healthy subjects' performances were similar to those of subjects in previous studies (Collewijn et al., 1985; Cheung and Howard, 1991; Morrow and Sharpe, 1993; Cheung et al., 1995; Thilo et al., 1999; Farooq et al., 2004). Our data confirm that the functions of the tOKN are poorly compensatory compared with those of horizontal and vertical OKN, which saturate when they have reached 60–80°/s. In the present study, tOKN velocity reached 3°/s in healthy subjects. Nevertheless, higher tOKN velocity has been observed by some authors using larger fields of view; tOKN velocity increases with the size of the visual field (Farooq et al., 2004). That tOKN is weaker than horizontal and vertical OKN has been attributed to both behavioural (amplitude and frequency of head movements in roll are weak; Cheung and Howard, 1991) and morphological reasons (velocity of image slip across the central retina is weaker for tOKN than for horizontal and vertical OKN; Morrow and Sharpe, 1993; Seidman et al., 1995).
Compensation of tOKN deficits after unilateral vestibular loss
The present study provides the first description of the recovery time-course of tOKN after unilateral vestibular loss. Despite progressive compensation over time, tOKN symmetry is not totally regained 3 months after UVN. It appears that tOKN velocity was still higher for ipsilesional than for contralesional stimulation at this postoperative time. In humans, studies dealing with the effects of unilateral vestibular loss on horizontal OKN have focused on the compensatory stage only (Abel and Barber, 1981; Vitte et al., 1994). Moreover, data were recorded from periods ranging from 1 to 52 months after surgery, which could have led to large variability. By contrast, the recovery time-course has been described in unilateral vestibular lesioned animals with species-dependent effects. Although the horizontal OKN symmetry was regained 8 days after lesion in rats (Farhat et al., 1995), compensation of OKN was not entirely achieved up to 1 year after lesion in cats (Maioli et al., 1982; Precht et al., 1981).
There is a large body of evidence that restoration of balanced neuronal activity within the vestibular nuclei could partially account for the functional recovery reported after UVN (for reviews see Precht and Dieringer, 1985; Lacour et al., 1989; Smith and Curthoys, 1989; Curthoys and Halmagyi, 1995). Thus, progressive recovery of both resting activity and sensitivity in vestibular neurons can be taken as the physiological evidence of the decrease in tOKN asymmetry and of the disappearance of the torsional spontaneous nystagmus we described after UVN. In animal models, however, recovery of the neuronal activity of vestibular neurons on the lesioned side can be incomplete long after the lesion. Smith and Curthoys (1988b) found that the number of responsive neurons in vestibular nuclei and neuronal sensitivity were affected up to 1 year after vestibular loss. Taken together, these results support our observation of incompletely compensated tOKN 3 months after surgery.
That the dynamic properties of tOKN were not totally compensated whereas torsional spontaneous nystagmus had vanished within the first three postoperative months supports the idea that compensation of dynamic deficits is subtended by long-term adaptive mechanisms. The tOKN asymmetry could therefore be part of such dynamic functions (one of the dynamic functions) that have not totally recovered after unilateral vestibular loss in humans (see Curthoys and Halmagyi, 1995). In support of this view is the observation that horizontal optokinetic afternystagmus remained asymmetrical up to 4 years after UVN (Brantberg et al., 1996). Congruently, it has been reported that the horizontal vestibulo-ocular reflex (Halmagyi et al., 1990) and dynamic ocular counter-rolling (Diamond and Markham, 1983) displayed asymmetrical patterns several years after unilateral neurotomy. Similarly, neurotomized Menière's disease patients tested in a dynamic postural task of knee-bends exhibited asymmetrical body orientation and impaired body stabilization that remained uncompensated 3 months after lesion (Borel et al., 2002). We have also reported impairments in walking performance during goal-directed locomotion up to 3 months after UVN for patients with similar disease (Borel et al., 2004).
Finally, significant static ocular cyclotorsion remained 3 months after the unilateral neurotomy. These results fit with the observations of Curthoys and colleagues for neurotomized patients, who reported residual cyclotorsion up to 16 weeks after surgery (Curthoys et al., 1991). In fact, cyclotorsion has been described as a permanent otolithic symptom. Interestingly, other symptoms assumed to be of otolithic origin, such as subjective visual vertical and horizontal perceptions (Vibert and Häusler, 2000) and body roll-tilt perception (Dai et al., 1989), have been reported never to be totally compensated or to take a long time to improve.
In summary, our results emphasize the role of vestibular cues for stabilizing gaze during optokinetic stimulations around the line of sight. We showed that unilateral loss of vestibular function impaired tOKN parameters (velocity and symmetry) in humans. Torsional slow-phase eye velocity displayed obvious asymmetry after UVN. The asymmetry resulted from a contralesional decrease in tOKN velocity. Interestingly, it also resulted from an ipsilesional increase in tOKN velocity. Three months after lesion, tOKN velocity still differed from that recorded preoperatively for both directions of stimulation. Taken together, these deficits could reflect an asymmetrical pattern of activity within vestibular nuclei. Finally, our results indicate that tOKN impairment and associated static ocular cyclotorsion represent long-term symptoms of unilateral vestibular loss.
|
Acknowledgements |
|---|
This study was supported by grants from CNRS and Ministère de l'Enseignement Supérieur et de la Recherche (UMR 6149 CNRS—Université de Provence). We acknowledge Dr P. Guillemant for his helpful comments on the material and methods section. We thank all participants for their cooperation.
|
References |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsReferences |
|---|
Abel SM, Barber HO. Measurement of optokinetic nystagmus for otoneurological diagnosis. Ann Otol Rhinol Laryngol Suppl 1981; 90: 1–12.[Medline]
Barmack NH, Pettorossi VE, Erickson RG. The influence of bilateral labyrinthectomy on horizontal and vertical optokinetic reflexes in the rabbit. Brain Res 1980; 196: 520–4.[CrossRef][ISI][Medline]
Borel L, Harlay F, Magnan J, Lacour M. How changes in vestibular and visual reference frames combine to modify body orientation in space. Neuroreport 2001; 12: 3137–41.[CrossRef][ISI][Medline]
Borel L, Harlay F, Magnan J, Chays A, Lacour M. Deficits and recovery of head and trunk orientation and stabilization after unilateral vestibular loss. Brain 2002; 125: 880–94.
Borel L, Harlay F, Lopez C, Magnan J, Chays A, Lacour M. Walking performance of vestibular-defective patients before and after unilateral vestibular neurotomy. Behav Brain Res 2004; 150: 191–200.[CrossRef][ISI][Medline]
Brandt T, Allum JH, Dichgans J. Computer analysis of optokinetic nystagmus in patients with spontaneous nystagmus of peripheral vestibular origin. Acta Otolaryngol 1978; 86: 115–22.[Medline]
Brantberg K, Fransson PA, Bergenius J, Tribukait A. Tilt suppression, OKAN, and head-shaking nystagmus at long-term follow-up after unilateral vestibular neurectomy. J Vestib Res 1996; 6: 235–41.[CrossRef][ISI][Medline]
Buizza A, Leger A, Droulez J, Berthoz A, Schmid R. Influence of otolithic stimulation by horizontal linear acceleration on optokinetic nystagmus and visual motion perception. Exp Brain Res 1980; 39: 165–76.[ISI][Medline]
Cheung BS, Howard IP. Optokinetic torsion: dynamics and relation to circular vection. Vision Res 1991; 31: 1327–35.[CrossRef][ISI][Medline]
Cheung BS, Money KE, Howard IP. Dynamics of torsional optokinetic nystagmus under altered gravitoinertial forces. Exp Brain Res 1995; 102: 511–8.[ISI][Medline]
Clément G, Lathan CE. Effects of static tilt about the roll axis on horizontal and vertical optokinetic nystagmus and optokinetic after-nystagmus in humans. Exp Brain Res 1991; 84: 335–41.[ISI][Medline]
Clément G, Popov KE, Berthoz A. Effects of prolonged weightlessness on horizontal and vertical optokinetic nystagmus and optokinetic after-nystagmus in humans. Exp Brain Res 1993; 94: 456–62.[ISI][Medline]
Coats AC. Central and peripheral optokinetic asymmetry. Ann Otol Rhinol Laryngol 1968; 77: 938–48.[ISI][Medline]
Cohen B, Uemura T, Takemori S. Effects of labyrinthectomy on optokinetic nystagmus (OKN) and optokinetic after-nystagmus (OKAN). Int J Equilib Res 1973; 3: 88–93.[Medline]
Collewijn H, Van der Steen J, Ferman L, Jansen TC. Human ocular counterroll: assessment of static and dynamic properties from electromagnetic scleral coil recordings. Exp Brain Res 1985; 59: 185–96.[ISI][Medline]
Curthoys IS, Halmagyi GM. Vestibular compensation: a review of the oculomotor, neural, and clinical consequences of unilateral vestibular loss. [Review]. J Vestib Res 1995; 5: 67–107.[CrossRef][Medline]
Curthoys IS, Dai MJ, Halmagyi GM. Human ocular torsional position before and after unilateral vestibular neurectomy. Exp Brain Res 1991; 85: 218–25.[ISI][Medline]
Dai MJ, Curthoys IS, Halmagyi GM. Linear acceleration perception in the roll plane before and after unilateral vestibular neurectomy. Exp Brain Res 1989; 77: 315–28.[CrossRef][ISI][Medline]
Demer J. How does the visual system interact with the vestibulo-ocular reflex? In: Baloh RW, Halmagyi GM, editors. Disorders of the vestibular system. New York: Oxford University Press; 1996. p. 73–84.
Diamond SG, Markham CH. Binocular counterrolling in humans with unilateral labyrinthectomy and in normal controls. Ann N Y Acad Sci 1981; 374: 69–79.[ISI][Medline]
Diamond SG, Markham CH. Ocular counterrolling as an indicator of vestibular otolith function. Neurology 1983; 33: 1460–9.
Farhat F, Reber A, Leroy MH, Messedi M, Courjon JH. Rapid compensation of horizontal optokinetic nystagmus in hemilabyrinthectomized rats: a fast return to symmetry. Arch Ital Biol 1995; 133: 251–61.[ISI][Medline]
Farooq SJ, Proudlock FA, Gottlob I. Torsional optokinetic nystagmus: normal response characteristics. Br J Ophthalmol 2004; 88: 796–802.
Fukushima K, Fukushima J. Otolith-visual interaction in the control of eye movement produced by sinusoidal vertical linear acceleration in alert cats. Exp Brain Res 1991; 85: 36–44.[ISI][Medline]
Guillemant P, Ulmer E, Freyss G. 3-D eye movement measurements on four Comex's divers using video CCD cameras, during high pressure diving. Acta Otolaryngol Suppl 1995; 520 Pt 2: 288–92.
Gustave Dit Duflo S, Borel L, Harlay F, Leonard J, Lacour M. Short-term changes in neck muscle and eye movement responses following unilateral vestibular neurectomy in the cat. Exp Brain Res 1998; 120: 439–49.[CrossRef][ISI][Medline]
Halmagyi GM, Gresty MA, Gibson WP. Ocular tilt reaction with peripheral vestibular lesion. Ann Neurol 1979; 6: 80–3.[CrossRef][ISI][Medline]
Halmagyi GM, Curthoys IS, Cremer PD, Henderson CJ, Todd MJ, Staples MJ, et al. The human horizontal vestibulo-ocular reflex in response to high-acceleration stimulation before and after unilateral vestibular neurectomy. Exp Brain Res 1990; 81: 479–90.[CrossRef][ISI][Medline]
Hamann KF, Reber A, Hess BJ, Dieringer N. Long-term deficits in otolith, canal and optokinetic ocular reflexes of pigmented rats after unilateral vestibular nerve section. Exp Brain Res 1998; 118: 331–40.[CrossRef][ISI][Medline]
Lacour M, Toupet M, Denise P, Christen Y, editors. Vestibular compensation. Facts, theories and clinical perspectives. Paris: Elsevier; 1989.
Lichtenberg BK, Young LR, Arrott AP. Human ocular counterrolling induced by varying linear accelerations. Exp Brain Res 1982; 48: 127–36.[ISI][Medline]
Lopez L, Bronstein AM, Gresty MA, Rudge P, du Boulay EP. Torsional nystagmus. A neuro-otological and MRI study of thirty-five cases. Brain 1992; 115: 1107–24.
Magnan J, Bremond G, Chays A, Gignac D, Florence A. Vestibular neurotomy by retrosigmoid approach: technique, indications, and results. Am J Otol 1991; 12: 101–4.[CrossRef][ISI][Medline]
Maioli C, Precht W, Ried S. Vestibuloocular and optokinetic reflex compensation following hemilabyrinthectomy in the cat. In: Roucoux A, Crommelinck M, editors. Physiological and pathological aspects of eye movements. The Hague: Jung; 1982. p. 201–8.
Merfeld DM, Teiwes W, Clarke AH, Scherer H, Young LR. The dynamic contributions of the otolith organs to human ocular torsion. Exp Brain Res 1996; 110: 315–21.[ISI][Medline]
Morrow MJ, Sharpe JA. The effects of head and trunk position on torsional vestibular and optokinetic eye movements in humans. Exp Brain Res 1993; 95: 144–50.[ISI][Medline]
Pfaltz CR, Ildiz F. The optokinetic test: interaction of the vestibular and optokinetic system in normal subjects and patients with vestibular disorders. Arch Otorhinolaryngol 1982; 234: 21–31.[CrossRef][Medline]
Precht W, Dieringer N. Neuronal events paralleling functional recovery (compensation) following peripheral vestibular lesions. In: Berthoz A, Mellvill Jones G, editors. Adaptative mechanisms in gaze control. Amsterdam, New York, Oxford: Elsevier; 1985. p. 251–68.
Precht W, Shimazu H, Markham CH. A mechanism of central compensation of vestibular function following hemilabyrinthectomy. J Neurophysiol 1966; 29: 996–1010.
Precht W, Maioli C, Dieringer N, Cochran S. Mechanisms of compensation of the vestibulo-ocular reflex after vestibular neurotomy. In: Flohr H, Precht W, editors. Lesion-induced neuronal plasticity in sensorimotor systems. Berlin: Springer-Verlag; 1981. p. 221–30.
Riordan-Eva P, Harcourt JP, Faldon M, Brookes GB, Gresty MA. Skew deviation following vestibular nerve surgery. Ann Neurol 1997; 41: 94–9.[CrossRef][ISI][Medline]
Ris L, Godaux E. Neuronal activity in the vestibular nuclei after contralateral or bilateral labyrinthectomy in the alert guinea pig. J Neurophysiol 1998; 80: 2352–67.
Ris L, de Waele C, Serafin M, Vidal PP, Godaux E. Neuronal activity in the ipsilateral vestibular nucleus following unilateral labyrinthectomy in the alert guinea pig. J Neurophysiol 1995; 74: 2087–99.
Seidman SH, Leigh RJ, Tomsak RL, Grant MP, Dell'Osso LF. Dynamic properties of the human vestibulo-ocular reflex during head rotations in roll. Vision Res 1995; 35: 679–89.[CrossRef][ISI][Medline]
Smith PF, Curthoys IS. Neuronal activity in the contralateral medial vestibular nucleus of the guinea pig following unilateral labyrinthectomy. Brain Res 1988a; 444: 295–307.[CrossRef][ISI][Medline]
Smith PF, Curthoys IS. Neuronal activity in the ipsilateral medial vestibular nucleus of the guinea pig following unilateral labyrinthectomy. Brain Res 1988b; 444: 308–19.[CrossRef][ISI][Medline]
Smith PF, Curthoys IS. Mechanisms of recovery following unilateral labyrinthectomy: a review. [Review]. Brain Res Brain Res Rev 1989; 14: 155–80.[CrossRef][Medline]
Strupp M, Glasauer S, Schneider E, Eggert T, Glaser M, Jahn K, et al. Anterior canal failure: ocular torsion without perceptual tilt due to preserved otolith function. J Neurol Neurosurg Psychiatry 2003; 74: 1336–8.
Thilo KV, Probst T, Bronstein AM, Ito Y, Gresty MA. Torsional eye movements are facilitated during perception of self-motion. Exp Brain Res 1999; 126: 495–500.[CrossRef][ISI][Medline]
van den Berg AV, Collewijn H. Directional asymmetries of human optokinetic nystagmus. Exp Brain Res 1988; 70: 597–604.[CrossRef][ISI][Medline]
Vibert D, Haüsler R. Long-term evolution of subjective visual vertical after vestibular neurectomy and labyrinthectomy. Acta Otolaryngol 2000; 120: 620–2.[CrossRef][Medline]
Vitte E, Sémont A, Berthoz A. Repeated optokinetic stimulation in conditions of active standing facilitates recovery from vestibular deficits. Exp Brain Res 1994; 102: 141–8.[ISI][Medline]
Young LR, Lichtenberg BK, Arrott AP, Crites TA, Oman CM, Edelman ER. Ocular torsion on earth and in weightlessness. Ann N Y Acad Sci 1981; 374: 80–92.