Walking performance of vestibular-defective patients before and after unilateral vestibular neurotomy

Transcription

1 Behavioural Brain Research 150 (2004) Research report Walking performance of vestibular-defective patients before and after unilateral vestibular neurotomy Liliane Borel a,, Françoise Harlay a, Christophe Lopez a, Jacques Magnan b, André Chays b, Michel Lacour a a UMR 6149, Neurobiologie Intégrative et Adaptative, Université de Provence/CNRS, 52, Faculté de St. Jérôme, Case 361, Marseille Cedex 20, France b Service d ORL et de Chirurgie Cervicofaciale, Chemin des Bourrelly, CHU Marseille Nord, Marseille Cedex 20, France Received 24 February 2003; received in revised form 18 July 2003; accepted 18 July 2003 Abstract The present study investigated goal-directed linear locomotion in nine Menière s patients before and after (1 week, 1 and 3 months) a curative unilateral vestibular neurotomy (UVN). Experiments were done using a 3D motion analysis system in subjects walking eyes open (EO) and eyes closed (EC) towards a real or memorized target, respectively. Locomotor pattern (velocity, step length, step frequency and walk ratio) and walking trajectory deviations were evaluated for normal and fast speeds of locomotion and compared to those recorded in 10 healthy subjects. Before UVN, patients showed no walking deviation but gait pattern changes characterized by slower walks compared to the controls, mainly due to step length and step frequency reductions for both visual conditions and locomotion speeds. In the acute stage after UVN, locomotor pattern impairments were significantly accentuated. On the other hand, patients showed strong walking deviations towards the lesioned side with EC. Opposite lateral deviation towards the intact side were observed with EO for normal speed only. Recovery from impaired locomotor pattern was achieved within 1 month for normal speed but remained uncompensated 3 months post-lesion for fast speed particularly in EC condition. Finally, the walking trajectory deviation towards the lesioned side in the dark was maintained up to 3 months after UVN. The results show that central processing of visual and vestibular cues contributes to an accurate locomotor pointing. They argue for an increased weight of visual reference frame on locomotor functions when vestibular function is unilaterally impaired Elsevier B.V. All rights reserved. Keywords: Locomotion pattern; Walking trajectory; Kinematics; Unilateral vestibular-defective patients; Visual reference frame; Locomotion speed; Self-motion perception; Gait recovery 1. Introduction Locomotion is a highly automated behaviour that is predominantly generated at the spinal level. Adaptation of gait to the environmental conditions, when required, is achieved by multisensory feedback to postural reflexes interacting with the spinal pattern generator [18]. Supraspinal influences, however, are necessary in circumstances such as obstacle avoidance or voluntary control of gait when walking on uneven terrain. The corticospinal control of gait has been evidenced in the cat during visually guided stepping tasks [14] and its implications for human gait were recently investigated by transcranial magnetic stimulation [38]. On the other hand, the vestibulospinal pathways are involved in the supraspinal Corresponding author. Tel.: ; fax: address: borel@up.univ-mrs.fr (L. Borel). control of gait. Activity of the vestibulospinal neurons is modulated during treadmill locomotion in the cat [34]. Natural stimulation of the vestibular system exerts intense modulatory effects on locomotor muscular activity in the guinea pig [30]. Adaptive plasticity of locomotor trajectory was found in normal subjects trained to walk for 2 h on the perimeter of a horizontally rotating disc; these subjects generated curved walking trajectories when asked thereafter to walk blindfolded straight ahead on firm ground [17]. Vestibular loss affects both gait parameters and walking trajectory. Patients with vestibular loss are primarily concerned with their balance and gait problems [19]. During the acute stage after total bilateral vestibular loss, patients show balance problems when standing or walking, i.e. in both static and dynamic conditions, but they may feel off-balance even when lying down or sitting. Goal-directed linear locomotion was investigated in normal individuals and bilateral labyrinthine-defective patients [16]. The authors showed /$ see front matter 2003 Elsevier B.V. All rights reserved. doi: /s (03)

2 192 L. Borel et al. / Behavioural Brain Research 150 (2004) these subjects performed as well as normal ones in terms of the distance error in reaching the target, and concluded the vestibular system was not necessary for active linear path integration. However, they reported larger lateral errors, modifications of gait parameters, and instability when patients walked without vision. Posturolocomotor parameters as well as walking trajectory impairments were reported also in unilateral vestibular neurotomy (UVN) patients. In static posturographic recordings, the patients asked to stand quietly showed body sway modifications [27]. Circular locomotion tasks showed walking trajectory deviations pointing to vestibulospinal asymmetries [22]. Deviation from a linear trajectory was observed for subjects with chronic vestibulopathy and for those with resection of acoustic neuromas during the acute stage who were asked to walk eyes closed (EC) [10]. In addition, the vestibular loss-induced deviations from the walking direction depend on the locomotion speed [7]. Moreover, we recently showed the role of the visual reference frame in compensating the vestibular deficits in standing and moving vestibular patients. The changes observed in spatial reference frames depend on whether the subjects were tested with or without vision. They switched from egocentric (eyes closed condition) to exocentric (eyes open condition: EO) reference frames for posture control [5] and knee-bends task [6]. The sole presence of vertical references in the visual environment induced such a shift in spatial head or trunk orientation, from the lesioned side (EO without vertical references and EC conditions) to the intact side (EO with vertical references) [5]. Because normal balance while walking requires the combined use of visual and vestibular (plus somatosensory) inputs, patients with unilateral loss of vestibular function would likely exhibit changes in their locomotor pattern and/or walking trajectory depending on whether they are examined in light or in darkness. In addition, the changes in reference frame could differentially affect the locomotor pattern and the walking trajectory. To test these hypotheses, we used 3D motion analysis to investigate Menière s patients before and after curative surgery of their disease (UVN) in a locomotion task performed with (EO) and without (EC) vision. The walking performance of the patients (step length and frequency, deviation of locomotion) was tested for different locomotion speeds as a function of the postoperative time and compared to that of healthy volunteers. 2. Methods 2.1. Subjects Data from nine unilateral vestibular-defective subjects with Menière s disease (four women, five men aged years, mean 45 years) were compared to those of 10 healthy subjects (five women, five men aged years, mean 40 years) selected on the basis of normal vestibular and visual functions. Of the nine patients, six had a right-side disease and three had a left-side disease. All the patients complained about recurrent vertigo, hearing loss, and tinnitus. Neurotomy was used to eliminate vertigo and preserve hearing for patients who became pharmacology resistant to antivertigo drugs and had increased frequency of vertigo attacks. The neuro-otological examination performed before vestibular neurotomy established that patients exhibited pure unilateral vestibular deficit averaging 52% (range %) to the caloric test. Hearing loss averaged 52% (10 85%) on the diseased side while the audiograms on the healthy side were normal. The history of vertigo ranged from 2 to 15 years (mean 10 years). Patients had no additional motor or visual disorders. None was under antivertigo medication. They were examined during 4 experimental sessions: before undergoing unilateral vestibular neurotomy (D 1), when not experiencing vertigo, and postoperatively during the acute (1 week: D + 7) and the compensatory (1 month: D + 30 and 3 months: D + 90) stages. Each subject gave informed consent to the study, which was approved by the local ethics committee Experimental procedures The subjects were instructed to walk straight at two locomotion speeds: their own, preferred velocity (normal speed), and as fast as possible without running (fast speed). Subjects were tested under two visual conditions, eyes open and eyes closed, to evaluate the role of the visual reference frame in the recovery process after UVN. The subjects were tested while walking without shoes on a flat floor. The experimenter walked behind them, without holding them, to prevent any disequilibrium and to avoid fall. In the EC condition, subjects wore glasses that were manually closed allowing vision to be totally excluded. In both conditions, subjects were told to walk straight either towards a visual target 8 m in front of them (EO condition) or towards this memorized target (EC condition). The experimenter asked the subjects to stop after a 5.5 m walk. In the EC condition, subjects were guided back to the starting position to avoid any feedback information about their locomotor trajectory deviation. In each experimental session, subjects had to walk three times with both visual conditions and both locomotion speeds; the tasks were randomly presented Data acquisition The kinematic analysis of locomotion was performed on data collected by a video-motion analyser (ELITE system) at a sampling frequency of 100 Hz. Two infrared video cameras, located 7 m from the starting position, were placed in front of the subject. Three infrared markers 8 mm in diameter were placed full-face, one on the sternal fork and one on each distal phalanx of the two big toes. In the present study, the locomotor parameters analysed were the locomotor pat-

3 tern itself (velocity, step length and step frequency) and the locomotion trajectory deviation Data processing L. Borel et al. / Behavioural Brain Research 150 (2004) Experimental data focused on the kinematic characteristics of the locomotor pattern and locomotor trajectory for each speed. Analyses were performed on a walking distance of 3 4 m. Locomotor pattern was defined by three parameters: velocity, step length, and step frequency. Velocity was calculated in m/s from the sternal marker. Feet markers were used to calculate the remaining parameters. Step length was measured as the mean distance of the swing phase for each foot. Frequency was the mean number of steps/s. To determine pathological changes from normal locomotor pattern that were related to the vestibular status, the speed, and the visual condition, we completed the kinematic approach by computing a walk ratio [37]. The walk ratio was evaluated as the ratio step length/step frequency, in m/step/s. Analyses were done on the average of the three trials for each subject. Locomotor trajectory was analysed using lateral deviation of the marker located on the sternum with respect to the theoretical straight line between the sternal marker and the real or memorized target. Locomotor trajectory deviation was calculated from the corresponding regression lines. A 0 angle indicates a lack of trajectory deviation. Positive and negative values refer to deviations towards the operated and intact sides, respectively, for the patients, and deviations towards the left and right sides, respectively, for the controls. To evaluate locomotion pointing accuracy, data from trials with imbalance were included in the data analysis up to the moment the subjects exhibited imbalance or tendency to fall (cf. Fig. 3). Statistical analyses were conducted on using mixed design analyses of variance (ANOVAs) with group (patients versus controls) as between-subject factor, and with locomotion speed (normal versus fast) and visual condition (EO versus EC) as within-subject factors on each of the dependent variables. Moreover, supplementary ANOVAs were performed on the nine Menière s patients with session (D 1, D + 7, D + 30, D + 90), locomotion speed and visual condition as within-subject factors on each of the dependent variables. 3. Results 3.1. Effects of unilateral vestibular loss on the locomotor pattern Comparison of walking performance of controls and Menière s patients before unilateral vestibular neurotomy The ANOVA indicated that patients before UVN had changes in locomotor pattern parameters with respect to the controls (step frequency: F(1, 17) = 8.24, P = 0.01; step length: F(1, 17) = 11.88, P = 0.003; velocity: F(1, 17) = Fig. 1. Effect of vestibular lesion on locomotor pattern during normal speed. From top to bottom: mean velocity, step frequency, step length, and walk ratio for patients before and after the vestibular lesion, and for control group under EO (open areas) and EC (hatched areas) conditions. Vertical bars represent confidence interval (CI). Significantly different from the control data, P< , P = 0.001). Only the walk ratio did not differ for the patients compared to the controls. Mean modifications of the locomotor pattern in both EO and EC conditions are illustrated in Fig. 1 for normal locomotion speed. Planned comparisons revealed that mean walking velocity was significantly lower for the patients before UVN than for the controls in both visual conditions (P <0.001 and P< for the EO and EC conditions, respectively). This result can accounted for by a decrease in both step frequency (P <0.001 for EO and P<0.001 for EC) and step length (P = for EO and P< for EC). In the normal subjects tested EO and EC, walking velocity averaged 1.0 ± 0.17 m/s (S.D.) and 0.96 ± 0.21 m/s, step frequency 1.71 ± 0.14 steps/s and 1.74 ± 0.17 steps/s, and step length 0.6±0.07 m and 0.58±0.09 m, respectively.

4 194 L. Borel et al. / Behavioural Brain Research 150 (2004) Table 1 ANOVA conducted on Menière s patients before and after UVN showing that session (D 1, D + 7, D + 30, D + 90), locomotion speed (normal vs. fast) and visual condition (EO vs. EC) are the main effects providing the sources of variations of step frequency, step length, step velocity, and walk ratio Variable Source of variation F P Step frequency Session F(3, 24) = < Locomotion speed F(1, 8) = < Visual condition F(1, 8) = Step length Session F(3, 24) = Locomotion speed F(1, 8) = < Visual condition F(1, 8) = < Step velocity Session F(3, 24) = < Locomotion speed F(1, 8) = < Visual condition F(1, 8) = < Walk ratio Session F(3, 24) = Locomotion speed F(1, 8) = Visual condition F(1, 8) = < F-statistics with degrees of freedom, and probability level (P) are reported. Significant differences between variables. and P = 0.02 for EC) and step length (P < for EO and P< for EC). Fig. 2. Effect of vestibular lesion on locomotor pattern during fast speed. From top to bottom: mean velocity, step frequency, step length, and walk ratio for patients before and after the vestibular lesion, and for control group under EO and EC conditions. Same conventions as in Fig. 1. Compared with controls, patients tested before UVN showed decreased values in both EO and EC conditions (for velocity: 19 and 29%, respectively, step frequency: 12 and 12%, step length: 10 and 24%). These data indicate that walking velocity and step length were more reduced in the EC than in the EO condition. The effects of increased locomotion speed are illustrated in Fig. 2. In the normal subjects, fast speed of locomotion with and without vision led to mean velocities of 1.65 ± 0.23 m/s and 1.61 ± 0.26 m/s, mean frequencies of 2.24 ± 0.20 steps/s and 2.22 ± 0.18 steps/s, and mean step length of 0.8 ± 0.12 m and 0.79 ± 0.13 m, respectively. Compared to these control values, the preoperative data showed modifications similar to those described for normal speed: significant reductions of walking velocity (P < for EO and P< for EC), step frequency (P <0.01 for EO Changes in walking performance of Menière s patients after unilateral vestibular neurotomy Repeated measures ANOVAs conducted on before and after UVN for the nine patients indicated that unilateral vestibular loss constituted a main effect providing the sources of variation of the locomotor pattern parameters (step frequency, step length, step velocity, and walk ratio). In addition, significant effects of visual condition and locomotion speed were evidenced (see Table 1). Fig. 1 illustrates the mean modifications of the locomotor pattern in the group of patients before and after UVN in both EO and EC conditions for the normal locomotion speed. The comparison of the patients preoperative (D 1) and acute postoperative (D + 7) data showed a significant decrease of the walk ratio in both EO (P <0.05) and EC (P <0.0001) conditions as a result of UVN. Therefore, the walk ratio constitutes a reliable parameter for pathological changes in locomotor pattern. This decrease was mainly due to an increase in step frequency since both walking velocity and step length did not differ before and in the acute stage after UVN. As a rule, the walking performance of the patients regained near normal values as early as 1 month after their vestibular nerve surgery. As concerns the fast locomotion speed, the acute stage was characterized by a significant decrease in the walk ratio in the EC condition (P <0.05) related to an increase in step frequency (P = 0.01), while walking velocity and step length did not differ significantly from those recorded preoperatively (Fig. 2). Later on, walking velocity, step length and walk ratio remained reduced up to D + 90 with respect to the data from the controls.

5 L. Borel et al. / Behavioural Brain Research 150 (2004) Fig. 3. Examples of raw walking trajectories during a normal speed locomotion under EO and EC conditions for a control subject (A) and patients tested 1 week after right (B) and left (C) UVN Effects of unilateral vestibular loss on locomotion trajectory Fig. 3 illustrates three consecutive raw locomotion trajectories, during normal walking speed, in one normal subject (Fig. 3A) and in two Menière s patients 1 week after UVN on the side of their affected ear, the right and the left sides for Fig. 3B and C, respectively. Locomotion trajectory showed no deviation in the control subjects either EC or EO. As a rule, the healthy volunteers walked straight ahead towards the target. By contrast, the patients with total unilateral vestibular loss examined during the acute stage (1 week after surgery) exhibited large deviations of their locomotion trajectory towards their operated side when tested in the EC condition and showed slight deviations towards the intact side in the EO condition. Imbalance were often observed in these patients tested in the acute stage after UVN in the EC condition (cf. Fig. 3B). The mean regression curves calculated from the raw trajectories during normal walking speed are plotted in Fig. 4

6 196 L. Borel et al. / Behavioural Brain Research 150 (2004) Fig. 4. Examples of regression lines for one patient tested during a normal speed locomotion under EO and EC conditions before UVN (D 1) and at different postoperative times: 1 week (D + 7), 1 month (D + 30), and 3 months (D + 90). for one patient before and after UVN, for the two visual conditions and for all the post-lesion times. The patient tested before neurotomy showed no deviation while walking and behaved as the controls in both visual conditions. After UVN and without vision, he deviated significantly towards the operated side during both the acute (D + 7) and compensatory (D + 30, D + 90) stages. By contrast, in the EO condition, walking trajectory deviations were observed towards the intact side, but during the acute stage only. The ANOVA on the locomotion trajectory deviation for patients tested after UVN and controls showed a significant main effect of group (F(1, 17) = 5.78, P = 0.03) and of visual condition (F(1, 17) = 8.64, P = 0.009). The deviation of locomotion trajectory is shown in Fig. 5 for both normal and fast speeds of walking. The mean values of walking trajectory deviation for the nine patients are expressed as angular deviation in degrees. No significant differences were seen between the controls and the patients before UVN. One week after UVN, significant differences appeared when walks were performed without vision: the patients deviated to their operated side and showed similar deviations for both normal (P <0.0001) and fast (P <0.001) speeds of lo-

7 L. Borel et al. / Behavioural Brain Research 150 (2004) locomotor pattern with respect to the controls, with reductions of velocity, step length, step frequency, and walk ratio in both EO and EC conditions, particularly accentuated without vision and when fast locomotion is required. On the other hand, they behave like normal subjects in a simple linear locomotor task regarding their locomotor pointing performance; no deviation of their locomotor trajectory was observed. Relative to controls, the acute effects of UVN-induced total unilateral loss of vestibular input (D+7) consisted of a decrease in velocity and step length in EC condition, whatever the speed of locomotion. Relative to patients tested before UVN, data showed a decrease of the walk ratio in both EO and EC conditions. Moreover, UVN in the acute stage induced strong lateral walking deviations whose direction (lesioned versus intact sides) depends on the visual condition: locomotion trajectory is deviated towards the lesioned side when walks are performed without vision while opposite effects are seen with vision. In the later stages after UVN (D + 30, D + 90), the locomotor pattern changes depend on the locomotion speed; parameters were similar to those of the controls during normal speed while velocity, step length and walk ratio were impaired for the fast speed. In addition, such impairments were higher with EC than with EO. The locomotor trajectory direction also greatly depends on the visual condition; it became normal again in the EO condition but remained deviated to the lesioned side up to D + 90 in the EC condition Locomotor pattern changes Fig. 5. Effect of vestibular lesion on locomotor trajectory. Mean lateral deviation with EO and EC at the normal (upper part) and fast (lower part) speed of locomotion for the patients at the different preoperative and postoperative times. The hatched area shows mean data for the controls. Vertical bars represent the S.E.M. Significantly different from the control data, P<0.05. comotion. Angular deviations calculated 1 week after UVN averaged 3.7±2.5 and 3.0±2.0 for the normal and fast locomotion speeds, respectively. Roughly similar values were recorded up to D + 90 indicating a lack of compensation of locomotion trajectory in the EC condition. In the EO condition and for normal speed, patients deviated significantly to their intact side during the acute stage only (P <0.05). By contrast, when increasing locomotion speed, patients were able to maintain the straight ahead direction of locomotion, whatever the postoperative time. 4. Discussion The present investigation clearly shows that unilateral vestibular-defective patients (Menière s patients) tested before curative UVN exhibit significant modifications of their Relative to the controls, unilateral vestibular-defective patients suffering from Menière s disease and tested before UVN in a goal-directed locomotion task always showed locomotor impairments in both EO and EC conditions. Moreover, these locomotor impairments were significantly greater in the EC condition. The patients have decreased step frequency and step length for both normal and fast speeds. In addition, their normal and fast locomotion speeds are slower than for normal subjects. The data corroborate those reported for bilateral vestibular-defective patients in similar experimental conditions and kinematic recordings [16]. Decreased walking velocity can be considered as a behavioural strategy to avoid imbalance and fall, as reported by some subjects and as already seen in behavioural investigations in our cat model [28]. In addition, slow or reduced locomotion velocity by taking shorter steps was also reported for non-pathological subjects over 65 years [15], as well as in pathological cases including Parkinson s disease [2] and neuropathic patients [13]. Taken together, the similar changes in gait parameters found both with aging and in different pathologies strongly suggest that they are unlikely related directly to a specific sensory loss (unilateral vestibular deficit, herein) but rather to a reduction in self-selected walking speed. This general locomotion strategy, as shown particularly when visual cues are not available, would ensure safety during the locomotor task. The lesion induced significant locomotor pattern

8 198 L. Borel et al. / Behavioural Brain Research 150 (2004) changes even though patients did not behave like the controls before UVN. Thus, the influence of visual cues was enhanced, as shown by the walk ratio (step length/step frequency) decrease in the acute stage after UVN as compared to the preoperative data. Therefore, our results confirm the concept developed by Sekiya and Nagasaki [37]: the walk ratio is a speed-independent index of walking patterns describing temporal and spatial coordination and seems to be a reliable measure for evaluating both aging and pathological walks. After neurotomy, the locomotor pattern regained basically normal values for normal speed while for fast speed it was all the more impaired since visual cues were lacking. That the walk ratio was significantly decreased after UVN suggests that vestibular input might regulate also the locomotor pattern. Taken together, the results suggest that vision as a regulator factor of gait parameters is crucial in unilateral vestibular-defective patients before as well as after UVN, and particularly during the acute stage after the lesion. They confirm the substitution role of static and dynamic visual cues in compensating unilateral vestibular loss in both animal models [24,26,41] and human beings [9,23]. Vision serves UVN patients as a reference frame for goal-directed locomotion and control of walking trajectory. This corroborates the general statement on the role of vision in the control of human locomotion [35]: Visual cues are used to regulate locomotion, i.e. for maintaining dynamic stability of the body, modulating the gait pattern with respect to the environmental conditions, and guiding locomotion towards endpoints, i.e. in cases of goal-directed linear locomotion and route planning. From a clinical point of view, the results also show that surgical treatment of Menière s patients by unilateral vestibular neurotomy allowed locomotor parameters to be improved in the chronic stage for normal speed walking. This suggests that non-fluctuating vestibular input is required to regain normal locomotion Changes in locomotor trajectory deviation That patients before UVN and controls behave roughly similarly with EO confirms the results on spatial orientation in deaf vestibularly deficient subjects [40] and in patients with acoustic neuromas [10]. Compensatory mechanisms have very likely developed in Menière s patients before UVN, including increased functional weight of visual and proprioceptive cues. This hypothesis was already supported by posturographic recordings we performed in such pathology [27]. Adaptive plasticity in the control of locomotor trajectory has also been reported in normal subjects exposed to 2 h of walking on the perimeter of a horizontally rotating disc and asked thereafter to walk blindfolded straight ahead on firm ground [17]. They generated curved walking trajectories although they perceived themselves as walking linearly. The authors hypothesised a vestibular somatosensory/motor rearrangement. In the present study, patients tested after UVN in the EC condition deviate to the lesioned side. Are the patients driven to that side or is the deviation an active correction? In vestibular neuritis, two sensations opposite in direction were reported: a pure subjective sense of self-motion in the direction of the nystagmus fast phase, and a compensatory vestibulospinal reaction resulting in gait deviation [7]. In our experiment, the UVN Menière s patients always reported to be driven to their lesioned side. In addition, they did not exhibit any spontaneous nystagmus in darkness 3 months after UVN, but still deviated significantly towards the operated side when vision was excluded (EC) in both normal and fast walking speed. Unilaterally impaired vestibular input-induced walking deviation can be explained by tone asymmetries in leg muscles [1,20,25] since the vestibulospinal influences interfere with the execution of the walking trajectory (see reference [11] for a review). Another non-exclusive hypothesis is that total unilateral loss of vestibular inputs induces an erroneous perception of self-motion and computation of subject s position with respect to the memorized target when visual references are lacking. Such data have been reported for patients after unilateral acoustic neuroma resection, who exhibited lateral errors in a goal-directed linear locomotion task [10]. Moreover, an asymmetrical perception of space orientation after unilateral loss of vestibular function was evidenced through matching perceptual estimates of self-rotation [8]. Indeed, patients had hypometric responses when rotated towards the lesioned side but normometric to the intact side. The authors suggested that such asymmetrical perceptual responses were due to the inability of the remaining labyrinth to signal off direction. The walking trajectory deviation towards the lesioned side recorded in the present study could refer to the same mechanism. Such data also corroborate the notion of impaired spatial performance in unilateral vestibular-defective patients tested at similar post-lesion times in a non-visual navigation task [36]. The vestibular system plays a significant role in spatial memory for both whole-body angular [3,31] and linear [4,21] displacements. Altogether, the data point to the role of vestibular cues in the elaboration of an accurate internal representation of the environment, confirming more generally the role of the vestibular system for idiothetic navigation and route planning [32,33]. However, the role of the vestibular system in estimating distance has not been tested here. This function could be more independent of the vestibular cues [16]. The data also point to the permanent role of vision in compensating the vestibularly induced deficits (see references [12,29,39] for reviews), including the spatial orientation impairments. The lateral deviation towards the lesioned side in patients tested EC after UVN was higher for the normal locomotion speed. In patients with acute vestibulopathy, increased walking speed also improved their walking direction [7]. These data corroborate the task-dependent locomotion strategies we described in UVN cats for preserving their dynamic equilibrium: adaptive increase in locomotion speed reduces their balance and gait problems in the rotating beam task [28].

9 L. Borel et al. / Behavioural Brain Research 150 (2004) When the locomotion task was performed with vision (EO), the patients showed deviations of their walking trajectory towards the intact side only for normal speed in the acute stage (1 week) after UVN. This result shows vision is capable of reversing the vestibular lesion-induced deviations. Similar data were observed in another population of Menière s patients we tested with posturography methods and motion analysis system in static conditions [5]. Body spatial orientation and the centre of foot pressure were deviated to the lesioned side in the EC condition while they were reversed to the intact side with EO, in visual contexts where the reference frame contained vertical and horizontal visual coordinates. Interestingly, such a complete inversion of head and trunk orientation, from the lesioned side without visual cues towards the intact side with vision, was also observed for Menière s patients tested in a more dynamic task of knee-bends for angular rotation patterns in both the roll and yaw planes [6]. Similarly, the deficits remained uncompensated 3 months post-lesion. In summary, we have shown that goal-directed linear locomotion is permanently affected after total unilateral vestibular loss. The dynamic vestibular information elicited during active motion would be used to modulate the locomotor pattern and to elaborate an accurate spatial orientation. In addition, the results suggest the visual condition provides a regulatory factor and an exocentric reference frame able to improve the locomotor pattern and to compensate, up to over-compensation, the impaired central representation of body in space. Acknowledgements This study was supported by grants from CNRS and Ministère de l Enseignement Supérieur et de la Recherche (UMR 6149). We thank patients and controls who participated in the experiments. References [1] Allum JHJ, Pfaltz CR. Postural control in man following acute unilateral peripheral vestibular deficit. In: Igarashi M, Black FO, editors. Vestibular and visual control on posture and locomotion equilibrium. Basel: Karger; p [2] Azulay JP, Mesure S, Amblard B, Blin O, Sangla I, Pouget J. Visual control of locomotion in Parkinson s disease. Brain 1999;122: [3] Bloomberg J, Melvill Jones G, Segal BN, McFarlane S, Soul J. Vestibular-contingent voluntary saccades based on cognitive estimate of remembered vestibular information. Adv Otorhinolaryngol 1988;41:71 5. [4] Borel L, Le Goff B, Charade O, Berthoz A. Gaze strategies during linear motion in head free humans. J Neurophysiol 1994;72: [5] 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: [6] 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: [7] Brandt T. Vestibulopathic gait. Walking and running. In: Ruzicka E, Hallett M, Jankovic J, editors. Gait disorders. Advances in neurology, vol. 87. Philadelphia: Lippincott Williams & Wilkins; p [8] von Brevern M, Faldon ME, Brookes GB, Gresty MA. Evaluating 3D semicircular canal function by perception of rotation. Am J Otol 1997;18: [9] Bronstein A, Guerraz M. Visual-vestibular control of posture and gait: physiological mechanisms and disorders. Curr Opin Neurol 1999;12:5 11. [10] Cohen HS. Vestibular disorders and impaired path integration along a linear trajectory. J Vest Res 2000;10:7 15. [11] Curthoys IS, Halmagyi GM. Vestibular compensation: a review of the oculomotor, neural, and clinical consequences of unilateral vestibular loss. J Vest Res 1995;5: [12] Dieringer N. Vestibular compensation: neural plasticity and its relation to functional recovery after labyrinthine lesions in frogs and other vertebrates. Prog Neurobiol Rev 1995;46: [13] Dingwell JB, Cavanagh PR. Increased variability of continuous overground walking in neuropathic patients is only indirectly related to sensory loss. Gait Posture 2001;14:1 10. [14] Drew T. Visuomotor coordination in locomotion. Curr Opin Neurobiol 1991;1: [15] Ferrandez AM, Pailhous J, Serratrice G. Locomotion in the elderly. In: Amblard B, Berthoz A, Clarac F, editors. Posture and gait: development, adaptation and modulation. Paris: Elsevier; p [16] Glasauer S, Amorim MA, Vitte E, Berthoz A. Goal-directed linear locomotion in normal and labyrinthine-defective subjects. Exp Brain Res 1994;98: [17] Gordon CR, Fletcher WA, Melvill Jones G, Block EW. Adaptive plasticity in the control of locomotor trajectory. Exp Brain Res 1995;102: [18] Grillner S. Control of locomotion in bipeds, tetrapods and fish. In: Brooks VB, editor. Handbook of physiology: the nervous system, vol. 2. Motor control. Bethesda, MD: American Physiology Society; p [19] Herdman SJ, Blatt PJ, Schubert MC, Tusa RJ. Falls in patients with vestibular deficits. Am J Otol 2000;21: [20] Igarashi M, Watanabe T, Maxian PM. Dynamic equilibrium in squirrel monkey after unilateral and bilateral labyrinthectomy. Acta Otolaryngol (Stockh) 1970;69: [21] Israel I, Berthoz A. Contribution of the otoliths to the calculation of linear displacement. J Neurophysiol 1989;62: [22] Ito S, Odahara S, Hiraki M, Idate M. Evaluation of imbalance of the vestibulospinal reflex by the circular walking test. Acta Otolaryngol Suppl 1995;519: [23] Kubo T, Kamakura H, Hirokawa Y, Yamamoto K, Imai T, Hirasaki E. 3D analysis of human locomotion before and after caloric stimulation. Acta Otolaryngol (Stockh) 1997;117: [24] Lacour M, Xerri C. Compensation of postural reactions to free-fall in the vestibular neurectomized monkey: role of the visual motion cues. Exp Brain Res 1980;40: [25] Lacour M, Roll JP, Appaix M. Modifications and development of spinal reflexes in the alert baboon (Papio papio) following an unilateral vestibular neurotomy. Brain Res 1976;113: [26] Lacour M, Vidal PP, Xerri C. Visual influences on vestibulospinal reflexes during linear vertical linear motion in normal and hemilabyrinthectomized monkeys. Exp Brain Res 1981;43: [27] Lacour M, Barthélémy J, Borel L, Magnan J, Xerri C, Chays A, et al. Sensory strategies in human postural control before and after unilateral vestibular neurectomy. Exp Brain Res 1997;115: [28] Lacour M, Sun J, Harlay F. Kinematic analysis of locomotion in unilateral vestibular neurectomized cats. J Vest Res 1997;7:

10 200 L. Borel et al. / Behavioural Brain Research 150 (2004) [29] Lacour M, Toupet M, Denise P, Christen Y. Vestibular compensation: facts, theories and clinical perspectives. Paris: Elsevier; p [30] Marlinsky VV. The influence of adequate vestibular stimulation on evoked locomotor muscle activity in the decerebrated guinea pig. Neuroscience 1989;33(3): [31] Metcalfe T, Gresty M. Self-controlled reorienting movements in response to rotational displacements in normal subjects and patients with vestibular disease. Ann NY Acad Sci 1992;656: [32] Mittelstaedt ML, Mittelstaedt H. Homing by path integration in a mammal. Naturwiss 1980;67: [33] Mittelstaedt ML, Glasauer S. Idiothetic navigation in gerbils and humans. Zool J Physiol 1991;95: [34] Orlovsky GN. Activity of vestibulospinal neurons during locomotion. Brain Res 1972;46: [35] Patla AE. Understanding the role of vision in the control of human locomotion. Gait Posture 1997;5: [36] Péruch P, Borel L, Gaunet F, Thinus-Blanc C, Magnan J, Lacour M. Spatial performance of unilateral vestibular-defective patients in nonvisual versus visual navigation. J Vest Res 1999;9: [37] Sekiya N, Nagasaki H. Reproducibility of the walking patterns of normal young adults: test-retest reliability of the walk ratio (step-length/step-rate). Gait Posture 1998;7: [38] Schubert M, Curt A, Jensen L, Dietz V. Corticospinal input in human gait: modulation of magnetically evoked motor responses. Exp Brain Res 1997;115: [39] Smith PF, Curthoys IS. Mechanisms of recovery following unilateral labyrinthectomy: a review. Brain Res Rev 1989;14: [40] Worchel P. The role of the vestibular organs in space orientation. J Exp Psychol 1952;44:4 10. [41] Zennou-Azogui Y, Xerri C, Leonard J, Tighilet B. Vestibular compensation: role of visual motion cues in the recovery of posturo-kinetic functions in the cat. Behav Brain Res 1996;74:65 77.