CHAPITRE VII WHAT ARE THE EFFECTS OF APPREHENDING A POSTURAL PERTURBATION ON THE CONTROL OF UPRIGHT STANCE?

(Soumis à Cognitive Brain Research)

Philippe Corbeil1-2, Julie Marcotte1, Olivier Martin2, Vincent Nougier2 &

Normand Teasdale1

1 Université Laval, Division de kinésiologie, Faculté de Médecine, Québec, Canada.

2 Université Joseph-Fourier, Laboratoire Sport et Performance Motrice, Grenoble, France.

Correspondence to:

Philippe Corbeil

PEPS, Division de kinésiologie, Université Laval

Québec (Qc), Canada, G1K 7P4

Phone: 418-656-4920; Fax: 418-656-2441

E-MAIL: Philippe.Corbeil@kin.msp.ulaval.ca

RÉSUMÉ

Les objectifs de cette étude est d’étudier le comportement postural orthostatique de sujets qui sont en attente de recevoir une perturbation externe pouvant créer un déséquilibre. Dans un premier temps, douze sujets devaient maintenir leur équilibre en station debout sans risque qu’ils soient perturbés par une force externe (nombre d’essais égal à cinq). Dans l’essai suivant, tous les sujets subissaient une perturbation externe inattendue produite par le relâchement d’un système mécanique (pendule inversé). Ensuite, tous les sujets étaient informés qu’ils pourraient recevoir une perturbation dans le tiers des 90 essais suivants. Ce protocole a permis d’étudier les ajustements posturaux des sujets lorsqu’ils sont en attente d’une perturbation en comparant : (1) les premiers essais sans perturbation avant et après la consigne, et (2) le premier essai perturbé avec les suivants. Après le premier essai perturbé, l’étendue et l’amplitude des déplacements du centre de pression augmentent largement dans les essais non perturbés. Une diminution de l’électromyographie du jumeau médial a également été mesurée après le premier essai perturbé. La réponse à la première perturbation est probablement construite à partir des rétroactions. Pour les essais perturbés suivants, la stratégie posturale adoptée par les sujets était caractérisée par une flexion à la hanche plus importante et une vitesse de rotation à la hanche plus élevée. Ainsi, cette étude démontre que des facteurs cognitifs reliés à l’incertitude d’une perturbation peuvent grandement influencer la mise en place des ajustements posturaux nécessaires au maintien de l’équilibre orthostatique.

Abstract

It has been suggested that a postural threat (and to some extent, fear of falling) modifies the control strategy used to maintain an upright standing. In some experiments, subjects adopted a stiffening-like strategy to reduce their postural oscillations. In the present study, we first submitted subjects to five quiet standing trials without perturbation. Then, a postural perturbation (forward push to the subject’s upper back) was applied unexpectedly. Thereafter, subjects were informed that they would be perturbed in 33% of the following trials (30/90 trials). This procedure allowed us to specifically examine the effects of apprehending a postural perturbation by comparing: (1) the postural control for quiet standing trials with and without a postural threat and 2) the first-perturbation trial with the following perturbed trials. The results clearly demonstrated that, after the first-perturbation trial, when no perturbation was given, subjects showed an increased amplitude and variability of the center of pressure displacement. We also observed a decrease of the total electromyographic activity for the gastrocnemius medialis. The response to the first perturbation was feedback based. For the subsequent perturbed trials, subjects adopted a postural strategy characterized by a more important hip flexion and a greater rate of change of hip angular position. Hence, with repeated exposure to the perturbation, a mixture of feedforward and feedback control strategies was adopted. Altogether, these results do not support the suggestion that subjects adopted a stiffening ankle strategy to face the postural threat of a forward destabilisation. Environmental and cognitive factors can alter the postural response in multiple ways.

Theme : Motor systems and sensorimotor integration

Topic: Control of posture and movement

Keywords: Apprehension, Postural control, Center of pressure, Center of mass, Perturbation

1. Introduction

In everyday life, we have to adapt our motor behavior to various environmental contexts. For example, when submitted to a threatening postural context, Maki and colleagues [10] hypothesized that elders adopt a stiffening strategy compared to a more “relaxed” strategy for unperturbed standing tests. This was observed also by Carpenter et al. [5] with young adults; when standing was threatened by elevating the support surface, a tighter control of posture characterized by a smaller amplitude of center of pressure (CP) oscillations and a higher frequency of postural sway was observed. In another experiment, Adkin et al. [1] showed that the tighter control strategy adopted by the subjects also was scaled with respect to the height of the support surface. More specifically, the amplitude of the oscillations decreased and the mean power frequency increased with an increasing height of the standing surface. This suggests that a postural threat (and to some extent, fear of falling) modifies the control strategy used to maintain an upright standing. In the above experiments, however, subjects knew that they would not be perturbed during the standing tests. The expectation of a postural perturbation could yield a different strategy for preventing destabilisation and/or facilitating the postural reaction to the upcoming perturbation. Indeed, several authors have reported that expectancy and prior information regarding the magnitude or the nature of an upcoming perturbation can influence the postural strategy adopted by the subjects [2,3,7,8,12,13].

Few experimental studies, however, investigated the effects due to the apprehension of an external perturbation on the control of posture. In the experiment of Maki and Whitelaw [12], subjects showed a strong tendency to lean forward just prior to perturbation onset (perturbation direction was randomized). This effect was influenced by experience (repetition of trials) and prior information and lasted about 4 trials. In another experiment, Brown and Frank [3] showed that, when standing at the edge of an elevated platform compared to ground level, subjects adopted a more posterior starting position just prior to perturbation onset (one direction perturbation). In both experiments, however, subjects knew that they would be submitted to an external perturbation at each trial and no measurements have been made to quantify the performance of quiet standing control prior to perturbation onset.

In the present study, we first submitted subjects to five quiet standing trials without perturbation. Then, a postural perturbation was applied unexpectedly. The perturbation was a forward push in the upper back created by releasing a pendulum mechanism. Thereafter, subjects were informed that they would be perturbed in some of the following trials. This procedure allowed us to specifically compare the postural control for quiet standing trials with and without a postural threat. Also, for the first perturbed trial, there was no postural threat since subjects were completely unaware that they would be perturbed. The comparison of this first-perturbation trial with the following perturbed trials also allowed a more direct examination of the effect of a postural threat upon the postural strategy. Accordingly, the specific purpose of this experiment was twofold: a) to determine whether a postural threat yielded subjects to adopt a stiffening strategy during unperturbed quiet standing trials and b) to determine how such a strategy may modify postural reactions to external perturbations.

2. Materials and methods

2.1. Subjects

Twelve healthy male subjects participated in the study (age: 27.1 ± 5.2 years, height: 178.5 ± 5.6 cm, body weight: 73.6 ± 5.9 kg). Subjects were recruited at Joseph Fourier University and had no evidence of gait, postural or musculo-skeletal abnormalities. Informed consent was obtained from each subject according to the local research ethics committee.

2.2. Apparatus and procedure

Perturbations to balance were delivered by a mechanical pendulum. As shown in Fig. 1, the perturbation setup was mounted behind the subjects. The setup was composed of a foamy mass fixed on a rod. The pivot was fixed on the ceiling. A potentiometer was mounted on the pivot to measure the angular displacement of the mechanical pendulum. The height of the foamy mass was initially centered with respect to the height of the subject’s sternum. When in its vertical position, the pendulum was aligned with the position of the subjects’ heels. Two electromagnets initially kept the pendulum from moving at 7.5˚ from the vertical. When releasing the electromagnets, the pendulum accelerated relative to the subjects’ sagittal plane and the foamy mass of the pendulum collided with the subject’s upper back yielding a forward push. Small and large postural perturbations were set by adding a 1-kg or 4-kg mass to the pendulum (mean peak magnitude of 76.5 N and 101.1 N, respectively).

Figure 1 : Perturbation setup used to perturb balance.

A mechanical pendulum was mounted behind the subject. When in its vertical position, the pendulum was aligned with the position of the subjects’ heel. The starting position of the pendulum was set at 7.5 degrees from the vertical. When released, the pendulum accelerated relative to the subject’s sagittal plane and the foamy mass of the pendulum (31.5 cm width and 15.5 cm height) collided with the subject’s upper back yielding a forward push.

Postural stability was evaluated by means of a force platform (AMTI OR6-5-1). The force platform was leveled with the floor. Amplified force and moment components were sampled at 200 Hz (12 bit A/D conversion) and filtered with a Butterworth filter (7 Hz, second order, low pass cut-off frequency with dual-pass to remove phase shift). The antero-posterior (A-P) and medio-lateral (M-L) coordinates of the CP were derived from filtered data. Passive reflective markers were placed on the left side: foot (fifth metatarsal phalangeal joint), ankle (external malleolus), knee (lateral femoro-tibial joint), hip (greater trochanter and antero-superior iliac spine), shoulder (acromio-clavicular joint), elbow (lateral epicondyle), and on the head (zygomatic process and glabella). Kinematics of postural movements were obtained by filming subjects with three digital video cameras (25 Hz). A direct linear transformation was used to obtain the coordinates of the markers. Data were filtered at 7 Hz using a dual-pass Butterworth second-order filter. The 2D position of the total body center of mass (CM) was estimated using a 5-segment anthropometric model (foot, shank, thigh, trunk, neck and head, and arm) based on Dempster’s estimates of the segment weight and segment mass-center location [6]. Surface electromyographic activities (EMGs) were recorded on the right side of the subject for the tibialis anterior (TA) and gastrocnemius medialis (GM). Both electrodes were placed parallel to the muscle fibers with an interelectrode distance of 2 cm. EMG signals were pre-amplified at the source (×600) prior to be filtered (bandwidth of 10-350 Hz) and recorded at a rate of 800 Hz.

Participants stood barefoot on the force platform with their arms comfortably crossed at their chest and their feet 16 cm apart. They were instructed to stand still and to fixate a point located 2 m in front of them. For all trials, subjects also were asked to respond verbally as fast as possible to an auditory stimuli. The task was discriminatory: subjects had to respond by a verbal “top” signal only when two consecutive signals were presented. Two signals could be presented: a 1 kHz 100-ms signal or two consecutives 1 kHz 100-ms signals spaced by 100 ms. The auditory stimuli were given every 1.5 s throughout each trial, but only 30% of the auditory stimuli required a verbal response. The purpose of this task was not to measure attentional demands. Rather, it only served to induce subjects not to focus to the perturbation only. Reaction times to the auditory stimuli were obtained from the digital waveform of the verbal signal recorded from a piezo-electric sensor (sampling frequency of 200 Hz). A recording of white noise was played through headphones to mask possible auditory cues from the mechanical pendulum but subjects were able to hear clearly the auditory stimuli. Each trial lasted 15 s. For a first bloc of trials (n=6), subjects were informed that no external perturbation would occur. After the fifth trial, however, the movement of the pendulum was triggered (by turning off the electromagnets) and the equilibrium of the subject was perturbed by the impact of the foamy mass of the pendulum (large perturbation) with their back. Thus, for this first-perturbation trial, all subjects were perturbed unexpectedly. They were also unaware of the magnitude of the perturbation. Then, subjects performed a second bloc of 90 trials. For these trials, subjects knew they would be submitted to a postural perturbation for one third of all trials (15 small and 15 large perturbation trials randomly distributed across 60 quiet standing trials). The onset of the perturbation varied randomly within the time period between seconds 6 and 12.The onset and the magnitude of the perturbation were kept unknown to the subjects. Rest periods of 20 s were provided between each trial; five-minute rest periods were provided after 30 consecutive trials. For the sake of brevity and because the first perturbation was always a large perturbation, data for the small magnitude perturbation are not presented in this manuscript.

2.3. Data and statistical analyses

In the present paper, the initial analyses focused on the effects of apprehending an external perturbation on the regulation of quiet standing. For this first purpose, the quiet standing behavior before and after the first-perturbation trial were evaluated. Four blocs of five trials were selected for the analyses: the first five trials before the first-perturbation trial (control), the first and last five quiet standing trials following the first-perturbation trial (initial and last), and five quiet standing trials midway through the 90 trials (middle). For the different blocs of trials, the mean position, range, standard deviation and mean velocity of the CP fluctuations along the antero-posterior and the medio-lateral axes were calculated. Under quasi static conditions, the mean CP position is approximately equal to the center-of-mass location and thereby provides information regarding the standing posture [12]. For all subjects, the position of the feet was standardized with a jig lying flat on the force platform. The range of the CP displacement indicates the maximal deviation of the CP in any direction. The calculation of the standard deviation of the signal gives the variability of the balance reference over the sampled period. The mean velocity represents the total distance covered by the CP (total sway path) divided by the duration of the sampled period and constitutes a good index of the amount of activity required to maintain stability. EMG signals for the TA and the GM were full-wave rectified and filtered using a moving-window algorithm with a time constant of 100 ms. Then, EMG activity of the postural muscles was integrated over the entire duration of each trial. Hence, the analysis of EMG focused on the total activity of the postural muscles needed to regulate the upright stance. Mean reaction times for all blocs were also computed. For all these dependent variables (mean position, range, standard deviation and mean velocity of the CP fluctuations, EMG activity of the GM and TA and mean reaction times), within-subject means of each bloc of trials were computed. Then, the data were submitted to one-way analyses of variance (ANOVAs) with repeated measures (bloc as a factor). When a significant main effect of Bloc was observed, a post hoc analysis was performed using a planned comparison to compare data for the control bloc (before the first-perturbation trial) with all other blocs of trials (after the first-perturbation trial). Statistical significance was set at p < 0.05. Data for one subject were excluded from the analyses because of an intermittent cable problem that was not detected during acquisition.

Responses to the perturbed trials were also compared. Four different blocs were selected for the analyses: the first-perturbation trial (first trial), and three other blocs of five perturbation trials executed after the first-perturbation trial (initial, middle and last bloc). The first trial was analyzed as a specific condition because it is the only trial for which subjects were completely unaware of the upcoming perturbation. Hence, the response to this first trial was presumably feedback based. For all subsequent trials, a mixture of feedforward and feedback control strategies could have been adopted. The mechanical perturbation delivered throughout this experiment was constant. To specifically examine whether subjects modified their response strategy across trials, the A-P position and velocity of the CM and hip angular position and velocity (defined as the angle between the trunk and thigh segments) were evaluated at four moments just before the pendulum made contact with the subjects’ upper back (time 0), and at 160 ms, 320 ms and 480 ms following the perturbation onset. On average, the pendulum was in contact with the subjects’ upper back for 320 ms. Across all subjects, a step was observed before 480 ms for only 12% of all large perturbation trials. These trials were excluded from the subsequent analyses.

Dependent variables (A-P position and velocity of the CM and hip angular position and velocity) at each of the four moments were submitted to one-way repeated measures ANOVAs with Bloc as a factor. When a significant main effect of Bloc was observed, a post hoc analysis was performed using a planned comparison to examine whether, across the blocs of trials, subjects adapted their responses to the perturbations. The level of significance was set at p < 0.05.

3. Results

The purpose of this experiment was twofold: 1) to determine whether a postural threat yielded subjects to adopt a stiffening strategy during unperturbed quiet standing trials, 2) to determine how such strategy may facilitate or modulate the responses to the external perturbations.

3.1. Quiet standing trials

Mean vocal reaction times obtained before and after the first-perturbation trial during quiet standing trials were not different (p > 0.05). On average, reaction times were 348 ms (+/- 39), 364 ms (+/- 62), 352 ms (+/- 62) and 336 ms (+/- 57) for the quiet standing trials before the first-perturbation trial (control bloc) and after the first-perturbation trial (initial, middle and last blocs), respectively. This suggests that subjects allocated similar cognitive ressources to the reaction time task across the experiment.

Fig. 2 shows representative CP oscillations and filtered EMG signal of the GM for one subject for the last trial of the control bloc and the first quiet standing trial following the first-perturbation trial. A greater range of CP oscillations and a decrease of GM activity can be observed for the quiet standing trial following the first perturbation.

Figure 2 : CP displacements and filtered and rectified EMG of the GM.

Data presented are for one subject for the last trial of the control bloc (left side) and the first quiet standing trial after the first-perturbation trial (right side). Vertical and horizontal directions represent the A-P and M-L displacements of the CP, respectively.

For the mean positions of the CP displacement along both axes, the ANOVAs showed no main effect of Bloc (ps > 0.05). On average, the mean position along the A-P axis was 9.78 cm in front of the subjects’ heels for all blocs. The mean position along the M-L axis was 0.08 cm for all blocs. Hence, the apprehension of the postural perturbation did not modify the mean position of the standing posture.

Figure 3 : Range and variability of the CP oscillations.

The range of CP oscillations (left panel) and the variability of the postural oscillations around the mean position of the CP displacement (right panel) along the A-P and M-L directions recorded during quiet standing trials as a function of the four blocs analyzed. Vertical bars denote the 0.95 confidence intervals.

The left panel of Fig. 3 shows the range of CP oscillations along the A-P and M-L axes as a function of the four blocs analyzed. The ANOVAs showed a main effect of Bloc for both directions (F(3,30) = 6.84, p < 0.01 and F(3,30) = 9.99, p < 0.001 for A-P and M-L, respectively). For both directions, a decomposition of the main effect showed a significant increased range for the initial, middle and last blocs compared to the control bloc trials (F(1,10) = 15.62, p < 0.01 and F(1,10) = 21.70, p < 0.001, for the A-P and the M-L directions, respectively). This suggests that subjects showed an increased range of CP displacement along both directions when they were facing uncertainty about a potential postural perturbation. As illustrated on the right panel of the Fig. 3, the variability of the postural oscillations around the mean position of the CP displacement (standard deviation) increased after the first-perturbation trial for both the A-P and M-L directions (F(3,30) = 6.51, p < 0.01 and F(3,30) = 10.23, p < 0.001 for the A-P and the M-L directions, respectively). For both directions, a decomposition of the main effect showed a significant increase of the standard deviation for the initial, middle and last blocs compared to the control bloc (F(1,10) = 15.99, p < 0.01 and F(1,10) = 18.09, p < 0.01, for the A-P and the M-L directions, respectively). This suggests that the variability of the postural oscillations increased after the first-perturbation trial.

For the mean velocity of the CP (Fig. 4), the ANOVAs also showed a main effect of Bloc for both directions (F(3,30) = 11.00, p < 0.001 and F(3,30) = 10.72, p < 0.001, for the A-P and the M-L directions, respectively). A decomposition of the main effects showed that subjects oscillated more after the first-perturbation trial than for the control bloc (F(1,10) = 15.33, p < 0.01 and F(1,10) = 17.06, p < 0.01, for the A-P and the M-L directions, respectively). Overall, the CP data suggest that, following the first-perturbation trial, subjects oscillated more and exhibited a greater variability.

Figure 4 : Mean velocity of the CP along the A-P and M-L axes recorded during the quiet standing trials as a function of the four blocs analyzed.

Vertical bars denote the 0.95 confidence intervals.

The total activity of the ankle postural muscles (TA and GM) involved in the regulation of the upright stance has been analyzed to examine if subjects adopted an ankle stiffening strategy to prevent upcoming perturbations. No difference was observed for the mean integrated values of the TA activity across the four blocs of trials (p > 0.05). For the GM (Fig.5), the ANOVA showed a main effect of Bloc (F(3,30) = 3.68, p < 0.05). A decomposition of the main effect showed a significant decrease of the total activity of the GM for the middle and the last bloc compared to that observed for the control bloc (F(1,10) = 6.17, p < 0.05). This suggests that subjects did not adopt a stiffening ankle strategy when they apprehended an external perturbation. Rather, we observed no change of activity for the TA and a decrease of activity for the GM after the first-perturbation trial.

Figure 5 : Total integrated EMG activity of the GM involved in the regulation of the quiet standing trials as a function of the four blocs.

Vertical bars denote the 0.95 confidence intervals.

3.2. Responses to the perturbed trials

For the first-perturbation trial, 7 of the 12 subjects (58%) made a step. For all other perturbed trials, five subjects showed less than 4 trials with a step (out of 15 perturbations) and for these subjects the step occurred randomly throughout the session. Seven subjects stepped at least 8 times. Overall, no change in the frequency of the stepping response was observed through the perturbed trials.

Although subjects’ initial position could vary slightly during testing due to normal A-P postural oscillations, the CM position and velocity just prior to the beginning of the external perturbation were not statistically different for all blocs of perturbed trials (ps > 0.05). Similar observations were made for the hip angular position and velocity (ps > 0.05). This simply suggests that, across trials, subjects did not modify their initial CM and hip angular position to counteract the destabilizing effect of the upcoming perturbation.

The perturbation created a forward displacement of the CM. After the impact (on average, 320 ms after the perturbation onset) the CM position was 3.9 cm forward of the initial position (velocity of 30.2 cm/s). Across trials, there were clear modifications in the response strategy. Figure 6 shows hip angular position and velocity for the four blocs of trials at the perturbation onset and 160, 320 and 480 ms after the perturbation onset. Across blocs of trials, subjects adopted a response strategy characterized by a more important hip flexion and a greater rate of change of hip angular position. At the perturbation onset and at 160 and 320 ms after the perturbation onset, the hip angular position did not vary across blocs (ps > 0.05). At 480 ms after the perturbation onset, however, the main effect of bloc was significant (F(3,33) = 13.52, p < 0.001). A decomposition of this main effect showed a significant and systematic linear increased hip flexion (F(1,10) = 39.29, p < 0.001) from the first-perturbation trial to the last bloc of perturbed trials (Fig. 6a). For hip angular velocity (Fig. 6b), the main effects of bloc were significant at 320 ms and 480 ms after the perturbation onset (F(3,33) = 3.14, p < 0.05 and F(3,33) = 4.70, p < 0.01, respectively). A decomposition of these main effects also showed significant and systematic linear increases of the hip angular velocity from the first-perturbation trial to the last bloc of perturbed trials (F(1,11) = 7.52, p < 0.05 and F(1,10) = 6.26, p < 0.05, respectively). Overall, these observations suggest that, after being perturbed once, subjects gradually learned to respond to the perturbation with a greater and faster hip flexion. This yielded a smaller forward displacement of the CM. This effect was most notable 480 ms after the perturbation onset. The main effect of Bloc, however, did not reach the level of significance (p = 0.08).

Figure 6 : Hip angular position and velocity of the perturbed trials.

Hip angular position (left panel) and velocity (right panel) for the four blocs of the perturbed trials at 0 (perturbation onset), 160, 320 (end of the contact between the pendulum and the subject’s upper back) and 480 ms. Vertical bars denote the 0.95 confidence intervals. Asterisk denotes a significant and systematic linear increase from the first-perturbation trial to the last bloc of perturbed trials (p < 0.05).

4. Discussion

In the present study, we first submitted subjects to five quiet standing trials without perturbation. Then, a postural perturbation was applied unexpectedly. This procedure allowed us to specifically compare the postural control for quiet standing trials without and with a postural threat. After the first-perturbation trial, subjects modified their postural behavior during quiet standing trials.

Overall, about 51.6% of the large perturbation trials yielded a postural response characterized by the use of at least one-step. The perturbation required an appropriate postural response to maintain stability; it also created a psychological context in which the postural control of subjects was affected by apprehension factors.

Several authors have suggested that postural threat yields subjects to adopt a stiffening-like behavior [1,4,5]. This was clearly not the case in the present experiment. After the first-perturbation trial, subjects were clearly submitted to a postural threat; when no perturbation was given, they showed an increased amplitude and variability of the CP displacement. Mean velocity of the CP increased. Following the first-perturbation trial, we also observed no change of the total EMG activity for the TA and a decrease of the total EMG activity for the GM. Altogether, these results do not support the suggestion that subjects adopted a stiffening ankle strategy to face the postural threat of a forward destabilisation.

A major difference between the present experiment and that of others is the perturbation per se. For example, for some experiments, the postural threat (or fear of falling) was created by elevating subjects from the floor. In these experiments, subjects never experienced a perturbation. For instance, in Adkin et al.’s experiment [1], the amplitude of the CP oscillations decreased and the mean power frequency increased with an increasing height of the standing surface. A similar height protocol was used for a walking task [4]. In this latter experiment, anxiety was measured through galvanic skin response. A greater skin conductivity was observed when subjects had to walk on an elevated and more narrow platform and gait was characterized by less variability in the CM medio-lateral movement. In both of these experiments, subjects presumably adopted a stiffer postural and gait behavior when a postural threat was present. It is possible that the mere presence of a postural threat without experiencing any perturbation yields a behavior different than when subjects actually experience a real perturbation. Also, a stiffer behavior could be adapted to specific conditions (e.g., reducing forward sway prior to going downstairs) but such a stiffer behavior may not be universal nor optimal for other environmental conditions.

There are some suggestions that psychological stress could lead to an increase of the postural sway area [15]. In particular, Maki et al. [11] observed that, in blindfolded spontaneous-sway tests, elderly fallers with a fear of falling swayed more and were more variable than elderly fallers without a fear of falling. This hypothesis of a more variable postural behavior when a psychological stress is present also was put forward recently in a literature review [16]. Yardley and Redfern [16] suggested that psychological factors may aggravate dizziness and retard recovery from balance disorders.

In the present experiment, subjects clearly modified their postural responses across perturbations. The response to the first perturbation was feedback based. For all subsequent trials, a mixture of feedforward and feedback control strategies could have been adopted: with repeated exposure to the perturbation, subjects adopted a postural strategy characterized by a more important hip flexion and a greater rate of change of hip angular position. This progressive change in the postural set presumably led to a more stable and effective response to perturbations. This progressive change toward a hip response is reminiscent of the hip strategy observed in several surface translation perturbation experiments [8,9]. In these experiments, subjects most often switched from an initial ankle response to a hip response suggesting that they gradually learned to use more hip motion. In the present experiment, the subjects did not show any reduction in the number of trials for which a step was taken but they were not specifically instructed to avoid using a stepping response. When a step clearly cannot be taken during perturbed trials, Brown and Frank [3] also showed a behavior characterized by a more relaxed strategy. In their experiment, with the experience gained with repeated exposure to the perturbation, their subjects allowed: 1) the CM to travel closer to the base of support and 2) for three out of eight subjects, a more important hip angular displacement.

Anticipatory adaptations have been reported in previous experiments. For instance, in Brown and Frank’s experiment [3], subjects adopted a more backward lean position prior to perturbation onset (push to the upper back). In another experiment, Maki et al. [12] observed a greater forward lean when prior information was provided regarding the perturbation magnitude (forward and backward translation of the support surface). We did not observe similar anticipatory adaptations in the present experiment. The increased range of oscillations observed when no perturbation was given, however, could be the direct consequence of an anticipatory behavior. This would be the case if subjects, after the first unexpected perturbation, allocated more attention to sensory signals involved in the detection of the perturbation and to the selection and programming of the motor response necessary for maintaining stability when a perturbation was given. Such a change in the allocation policy of the cognitive ressources (i.e. shift from the active control of upright standing to the detection of a perturbation) is supported by recent findings of Redfern et al. [14]. These authors, showed that, when subjects were perturbed, the RTs to auditory stimuli were increased during the pre-perturbation time and during the initiation of the postural response to the perturbation, suggesting that the preparation and programming of the postural response required additional cognitive ressources.

Overall, the present experiment shows that, following a sudden and unexpected postural perturbation, subjects did not adopt a stiffening-like strategy. When no perturbation was given, subjects showed a more variable postural behavior. With repeated exposure to the perturbation, they adopted a postural strategy characterized by a more important hip flexion and a greater rate of change of hip angular position. This progressive change in the postural set presumably led to a more stable and effective response to perturbations. This suggests that environmental and cognitive factors may modulate the postural response in multiple ways.

Acknowledgements

The authors gratefully acknowledge Marcel Kaszap for the technical support in software developement. This study was supported by NSERC Canada, Ministère de l’Éducation du Québec and Égide France.

References

[1] A.L. Adkin, J.S. Frank, M.G. Carpenter, G.W. Peysar, Postural control is scaled to level of postural threat, Gait Posture 12 (2000) 87-93.

[2] D.J. Beckley, B.R. Bloem, M.P. Remler, R.A. Roos, J.G. Van Dijk, Long latency postural responses are functionally modified by cognitive set, Electroencephalogr. Clin. Neurophysiol. 81 (1991) 353-358.

[3] L.A. Brown, J.S. Frank, Postural compensations to the potential consequences of instability: kinematics, Gait Posture 6 (1997) 89-97.

[4] L.A. Brown, W.H. Gage, M.A. Polych, R.J. Sleik, T.R. Winder, Central set influences on gait. Age-dependent effects of postural threat, Exp. Brain Res. 145 (2002) 286-296.

[5] M.G. Carpenter, J.S. Frank, C.P. Silcher, G.W. Peysar, The influence of postural threat on the control of upright stance, Exp. Brain Res. 138 (2001) 210-218.

[6] W.T. Dempster, Space requirements of the seated operator. Aerospace medical research laboratories, Dayton, OH, 1955.

[7] M. Guerraz, K.V. Thilo, A.M. Bronstein, M.A. Gresty, Influence of action and expectation on visual control of posture, Cognitive Brain Res. 11 (2001) 259-266.

[8] F.B. Horak, H.C. Diener, L.M. Nashner, Influence of central set on human postural responses, J. Neurophysiol. 62 (1989) 841-853.

[9] F.B. Horak, L.M. Nashner, Central programming of postural movements: adaptation to altered support-surface configurations, J. Neurophysiol. 55 (1986) 1369-1381.

[10] B.E. Maki, P.J. Holliday, G.R. Fernie, Aging and postural control. A comparison of spontaneous- and induced-sway balance tests, J. Am. Geriatr. Soc. 38 (1990) 1-9.

[11] B.E. Maki, P.J. Holliday, A.K. Topper, Fear of falling and postural performance in the elderly, J. Gerontol. 46 (1991) M123-131.

[12] B.E. Maki, R.S. Whitelaw, Influence of expectation and arousal on center-of-pressure responses to transient postural perturbations, J. Vestib. Res. 3 (1993) 25-39.

[13] O. Martin, N. Teasdale, M. Simoneau, P. Corbeil, C. Bourdin, Pointing to a target from an upright position in human: tuning of postural responses when there is target uncertainty, Neurosci. Lett. 281 (2000) 53-56.

[14] M.S. Redfern, M.L. Muller, J.R. Jennings, J.M. Furman, Attentional dynamics in postural control during perturbations in young and older adults, J. Gerontol. A Biol. Sci. Med. Sci. 57 (2002) B298-303.

[15] Y. Tanaka, T. Takeya, Y. Ohno, K. Ueno, H. Sugano, The influence of psychological stress on fine body movement, Agressologie 20 (1979) 165-166.

[16] L. Yardley, M.S. Redfern, Psychological factors influencing recovery from balance disorders, J. Anxiety Disord. 15 (2001) 107-119.