CHAPITRE III PERTURBATION OF THE POSTURAL CONTROL SYSTEM INDUCED BY MUSCULAR FATIGUE

(Accepté dans Gait and Posture)

Philippe Corbeil1-2, Jean-Sébastien Blouin1, François Bégin1,

Vincent Nougier2 & Normand Teasdale1

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

2 Université Joseph-Fourier, UFR-APS, Grenoble, France.

Correspondence to:

Normand Teasdale

Division de kinésiologie, PEPS, Université Laval, Québec, Canada G1K 7P4

Fax: 418-656-2441, Phone: 418-656-2147

Email: Normand.Teasdale@kin.msp.ulaval.ca

RÉSUMÉ

Dans cette étude, onze sujets masculins ont participé à un protocole expérimental visant à comprendre les mécanismes adaptatifs du système de contrôle postural lorsque les principaux muscles participant au contrôle de l’équilibre orthostatique sont fatigués. Chaque sujet a effectué dix essais par condition (vision et non vision avec et sans fatigue musculaire) et chaque essai avait une durée de 60 secondes. Deux techniques d’analyse ont été utilisées afin d’étudier le comportement global et stochastique des déplacements du centre de pression. Les oscillations posturales des sujets sont plus importantes dans les essais impliquant une fatigue musculaire (augmentation de la vitesse moyenne, de la vitesse instantanée maximale et des fréquences moyenne et médiane). De plus, la composante caractérisant la stochasticité des oscillations posturales sur une longue durée tend à décroître avec la fatigue. Aucune différence n’a été observée pour l’étendue et la variabilité des oscillations posturales. Les effets de la fatigue musculaire sont semblables autant dans les essais sollicitant la vision que ceux réalisés avec les yeux fermés. Les résultats de cette étude suggèrent que le système nerveux compense les effets de la fatigue en augmentant la fréquence des corrections posturales. Le système de contrôle opérerait d’une façon moins stochastique en agissant avec un degré d’anti-persistence plus élevé lorsque les principaux actionneurs du maintien de l’équilibre sont fatigués.

ABSTRACT

In the present experiment, we induced muscular fatigue of ankle plantar-flexors to examine how it deteriorates the regulation of bipedal quiet upright standing. Postural stability was assessed in conditions with and without vision over 60-s periods to examine not only classical postural variables (time- and frequency-domain analyses) but also structural variables (stabilogram-diffusion analysis). Muscular fatigue was induced with repeated plantar-flexion of both legs. With muscular fatigue, subjects exhibited an increased postural sway [faster center of pressure (CP) velocity, greater CP mean and median frequency] and a decreased long term scaling exponent compared to the control conditions. The fatigue conditions, however, did not modify the range of oscillations and the variability of the postural oscillations around the mean position of the CP. The effects of muscular fatigue were similar with eyes open and eyes closed. These results suggest that fatigue did induce some changes in the control mode of postural stability but that the detection/action capabilities of the sensorimotor system remained partly efficient when the ankle plantar-flexors were fatigued. Furthermore, the decreased long-term scaling exponent observed with fatigue suggests that the control of upright stance operates in a less stochastic and more antipersistent manner when fatigue is present (that is past and future behaviors were more negatively correlated and thus more tightly regulated). Altogether, the present results suggest that, compared to the no fatigue conditions, fatigue places higher demands on the postural control system by increasing the frequency of actions needed to regulate the upright stance.

Keywords: Fatigue, Postural control, Center of pressure

1. Introduction

Multiple sensory systems and motor components of the nervous system are involved during the control of quiet standing [1]. For quasi-static (unperturbed) conditions, it is believed that ankle proprioception is critical for the establishment of an internal representation essential to the organisation and planification of motor acts. When standing upright, low-level of muscular forces are necessary for stabilizing the center of mass over the base of support. This has led several authors to explore the general idea that humans sometimes behave as an inverted pendulum [2]. When the sensory or motor components are altered or defective, body sway generally increases and muscle activity increases concurrently, in order to maintain postural equilibrium [3]. Minor perturbations occurring during normal stance can be counteracted by the regulation of ankle muscles [4,5].

How localized muscle fatigue may affect the control of posture is not clear. Muscle fatigue is a complex phenomenon, that has been defined as a reduction in the force-generating capacity, regardless of the task performed [6]. There are several fatigue-related mechanisms involved at different levels of the nervous system that could affect the regulation of these small forces. At the peripheral level, pre- and postsynaptic mechanisms and sites are potentially implicated, including a failure in the transmission of the neural signal or a failure of the muscle to respond to neural excitation [6]. At the central level, fatigue may induce a failure of excitation of the motoneurons caused by changes in the nervous system (supraspinal, segmental, sensory feedback) [7,8]. The origin of the changes in motoneuron firing has been attributed to the intrinsic properties of the motoneurons, recurrent inhibition (Renshaw cell), reflex inhibition or disfacilitation and changes in descending drive to the motoneuron pool [9-15].

In order to examine how fatigue affects postural stability, some authors have induced a generalized fatigue through a strenuous aerobic physical exercise. Generally, these authors reported a mild effect when vision is available [16,17]. In fact, in one experiment the magnitude of the effect of fatigue was smaller than that of visual withdrawal [18]. This effect was also short-lasting (transient increase in body sway for about 5 min with a plateau that did not exceed 15 minutes) [17]. When vision, ankle proprioception or both sources of sensory information were perturbed using a dynamic posturography test (or Sensory Organisation Test-Equitest Neurocom Int, Clackamas, Oregon, USA), however, the magnitude of the effect of fatigue was exacerbated [16].

When muscles of the lower limb muscular chain were fatigued with an isokinetic dynamometer (50% of the maximal strength), subjects showed a loss of stability when attempting to maintain their equilibrium on a balancing device [19]. More recently, some authors showed that, when performing a one-leg stance with vision, subjects were able to limit the destabilizing effect induced by localized fatigue of the ankle muscles [20]. The absence of vision yielded an increased body sway but only after subjects had been in absence of vision for at least 6 s. Overall, the above experiments suggest that, with vision, the magnitude of the fatigue effect on postural stability is somewhat mitigate. It is difficult, however, to make specific comparisons across experiments due to the nature of the fatigue protocol (global vs localized) and balance measurement (composite scores, one- vs two-leg stance, duration of the data analyzed).

In the present experiment, we induced muscular fatigue of ankle plantar-flexors to examine how it deteriorates the regulation of bipedal quiet standing. Postural stability was measured over 60-s periods in order to examine not only classical and more global variables (time- and frequency-domain analyses) but also structural variables (stabilogram-diffusion analysis). It is believed that the former set of variables provide insights into the size of the sway pattern (time or frequency domain) while the latter may identify sub-units underlying motor control processes [21,22].

Fatigue was induced in conditions with and without vision to examine if the absence of vision would exacerbate the effect of fatigue because subjects would have to rely more on ankle proprioceptive inputs to regulate their posture. We expected that fatigue would yield a less stochastic and more antipersistent behavior for mechanisms of the closed-loop control (decreased slope for the long-term region of the stabilogram-diffusion analysis) and a more persistent behavior for the open-loop control (increased slope for the short-term region of the stabilogram-diffusion analysis).

2. Methods

2.1 Subjects and apparatus

Eleven healthy male subjects participated in the study (age: 27 ± 7 years, height: 177.1 ± 6.6 cm, body weight: 76.0 ± 9.1 kg). The subjects were recruited at Laval University and had no evidence of gait, postural or musculo-skeletal abnormalities. Informed consent was obtained from each subject according to the research ethics committee. Postural stability was evaluated with the help of a force platform (AMTI OR6-5-1 model). Force and moment components were amplified (Ectron 563H) prior to being fed to a computer (200 Hz sampling rate; 12 bit A/D conversion). Data were digitally filtered with a second-order Butterworth filter (7 Hz 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 center of pressure (CP) were derived from filtered data.

2.2 Procedure

We evaluated postural sway for four different conditions: Standing eyes open and eyes closed with and without bilateral muscular fatigue of the soleus and gastrocnemius. The subjects were instructed to stand upright using a standardized stance on the force platform: they had their feet touching each other (modified Romberg protocol) in the center of the force platform. They stood barefoot on the force platform with their arm comfortably lying on each side and were instructed to fixate a point located 4 meters in front of them. Each trial lasted 60 seconds. A series of ten trials was performed with eyes open. Another series of ten trials was conducted with eyes closed. The order of testing (eyes open versus eyes closed) was randomized between subjects. Rest periods of 30 seconds were provided between each trial and five minutes between conditions. Two identical series of ten trials with fatigue were then performed. Trials without fatigue were always performed before trials with fatigue.

2.3 Fatigue protocol

For the fatigue conditions, we used a bloc design-training program. Muscular fatigue was induced in the ankle plantar-flexors with repeated plantar-flexion of both legs. Subjects were sitting in a standard ankle flexors training device. They had to lift upward a bar by rising the heels. The bar was loaded with weight and placed on the distant extremity of their thighs. A maximal load was first determined by adding weight until subjects were able to perform a single repetition. Then, subjects were instructed to perform 100 repetitions starting at 75% of their maximal workload with a reverse pyramidal technique consisting in diminishing gradually the load whenever subjects were unable to perform plantar-flexion. Immediately after the 100 repetitions, subjects were tested for postural stability. Nardone et al. [18] observed that, following fatiguing treadmill exercise, a significant increase in body sway with respect to pre-exercise values lasted until about 15 min from the end of the exercise. Therefore, after the fifth trial, subjects performed a maximal number of ankle plantar-flexions at 50% of their maximal workload followed by isometric contractions of the ankle plantar-flexors for as long as they could keep their heels from making contact with the ground. Postural stability of the subjects was then assessed for another five trials. Because fatigue was already induced, subjects were instructed to perform 40 repetitions at 75% of their maximal workload before the first trial of the second fatigue condition (no vision or vision). After five trials, subjects again performed a maximal number of ankle flexions at 50% of their maximal workload followed by isometric contractions of ankle flexors for as long as they could hold on.

2.4 Data reduction

2.4.1 Postural analyses

Range of the CP trajectory, mean velocity, standard deviation and maximal instantaneous velocity along A-P and M-L axes were calculated for all conditions. The range of the CP displacement indicates the maximal deviation of the CP displacement along A-P and M-L axes. The calculation of the standard deviation of the signal provides a measure of amplitude variability of the CP around the mean position. 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. The derivative of CP displacement was calculated using a finite difference technique using a 110-ms weighted window. The maximal instantaneous velocity represents the worst case scenario of the anterior CP velocity occuring during a trial.

Some authors suggested that periodicity exhibited by a physiological system may be an important marker of its functional ability [23]. An approach to studying the periodicity of a system is the frequency domain analysis. Power spectra were calculated from smoothed detrended data of the A-P and M-L displacement of the CP using a no-overlapping Fast Fourier Transform (FFT) window of 8192 points, permitting a frequency resolution of 0.0244 Hz. Power spectra were averaged for individual subjects for each condition resulting in a spectral signature characterizing postural control [5]. Median and mean frequencies were then calculated. Median frequency separates the power spectrum into two equal energy areas and mean frequency represents the centroid of the spectrum. Both dependant variables characterised the frequency components of the CP fluctuations.

2.4.2 Stabilogram-diffusion analysis

CP trajectories for M-L and A-P postural oscillations also were studied using the stabilogram-diffusion analysis developed by Collins and De Luca [21]. Squared distances between all pairs of points separated by a given time interval (Δt) were calculated and averaged, yielding a stabilogram-diffusion curves for each trial. An exponential model was used to extract parameters from the mean stabilogram-diffusion curves obtained per condition for each subject. The linear fit is performed on the resultant double-log-natural stabilogram-diffusion plot. Two distinct regions are identified from the stabilogram-diffusion curves: a short-term region and a long-term region (subscripts s and l will be used hereafter to refer to these regions, respectively). These two regions are determined by fitting two slopes through the resulting plot while minimizing the sum of squared deviations. The critical point coordinates approximate the temporal (Δtc) and the spatial (Yc) characteristics of the transition region that separates the short-term and long-term regions of the resultant double-log-natural stabilogram-diffusion plot. The half values of the slopes of the short-term and long term regions are labeled scaling exponents (Hs and Hl, respectively). Scaling exponent (H) gives the correlation between the step increments making up the trajectory of a random walker. For H equal to 0.5, the increments in displacement are statistically independent (uncorrelated), i.e. the system has no memory. For H superior to 0.5, the stochastic process is positively correlated: a random walker moving in a particular direction for some t0 will tend to continue in the same direction for t > t0 (this type of behavior is known as persistence). For H inferior to 0.5, past and future increments are negatively correlated: a random walker moving in a particular direction for some t0 will tend to continue in the other direction for t > t0 (antipersistence). Persistence and antipersistence behaviors in the CP trajectories have been interpreted as implicating mechanisms of the open-loop and closed-loop control, respectively [21]. Other authors hypothesized that persistence is relied to an exploratory mode (detection) and antipersistence is asssociated with a performatory mode (action) [24]. With respect to the effect of fatigue, we hypothesized that both the short and the long term regions, would be affected. More specifically, a more persistent behavior was expected for the short term region (increased slope) and more antipersistent behavior was expected for the long term region (decreased slope). The decreased slope for the long term region would be indicative of a less stochastic behavior. Similar software for performing stabilogram-diffusion analysis is available on the International Society of Biomechanics website (http://isb.ri.ccf.org/software/stamp/).

2.5 Statistical analysis

Dependent variables (range, standard deviation, mean velocity, instantaneous maximal velocity, mean and median frequency, Δtc, Yc, Hs and Hl) were all submitted to a 2-way ANOVA (2 vision x 2 fatigue conditions) with repeated measures on both factors for M-L and A-P oscillations. The level of significance was set at p <0.05.

3. Results

The main objective of the analyses was to determine the effects of muscular fatigue of the ankle plantar-flexor muscles on the postural control system. It was expected that the absence of vision would exacerbate the effect of fatigue because subjects would have to rely more on ankle proprioceptive inputs to regulate their posture.

3.1 Postural analyses

For the range of CP oscillations, the ANOVAs showed no main effects of Fatigue (ps > 0.05), but main effects of Vision for both M-L and A-P axes (F = 36.12 , p < 0.001 and F = 17.72, p<0.01, respectively). On average, the range of CP oscillations along the M-L axis were 2.90 and 3.58 cm for the vision and no-vision conditions, respectively. The range along the A-P axis were 3.02 and 3.56 cm, respectively. For both axes, the interaction of Fatigue x Vision was not significant (ps > 0.05). For the standard deviation, for both M-L and A-P axes, the ANOVAs showed no main effects of Fatigue (ps > 0.05), main effects of Vision (F = 20.19, p < 0.01 and F = 8.25, p < 0.05, respectively) and no interaction (ps < 0.05). On average, the standard deviations of CP oscillations along the M-L axis were 0.53 and 0.62 cm for the vision and no-vision conditions, respectively. The standard deviations along the A-P axis were 0.57 and 0.63 cm, respectively. Hence, the fatigue conditions did not modify the range of oscillations and the variability of the postural oscillations around the mean position of the CP.

Results obtained for the mean velocity and maximal instantaneous velocity of the CP along the M-L and A-P axes are presented in Fig. 1 for all conditions. Upper panels present data for the mean velocity of the CP and lower panels present data for the maximal instantaneous velocity of the CP. On average, the mean velocity along M-L and A-P axes increased after fatigue of ankle plantar-flexor muscles. For both axes, the ANOVAs showed main effects of Fatigue (F = 8.55, p < 0.05 and F = 11.70, p < 0.05, respectively) and Vision (F = 25.51, p < 0.001 and F = 62.63, p < 0.001, respectively), but no interaction (ps > 0.05). On average, fatigue yielded an increased velocity of 0.22 and 0.26 cm/s for the A-P and M-L axes, respectively.

For the maximal instantaneous velocity of the CP, fatigue of the ankle plantar-flexor muscles yielded a significant increase for the A-P direction only (F = 8.40 , p < 0.05). On average, fatigue yielded an increased velocity of 1.86 cm/s. The effect was not significant for the M-L direction (p > 0.05). The main effect of Vision was significant for both axes (F = 14.58, p < 0.01 and F = 24.79, p < 0.01 for the M-L axis and the A-P axis, respectively). The interactions of Fatigue x Vision again were not significant (ps > 0.05).

Figure 1 : Mean velocity and maximal instantaneous velocity of the CP displacements.

Mean velocity (A,B) and maximal instantaneous velocity (C,D) of the CP displacements recorded under both eyes open and closed conditions before and after the fatigue protocol. Each interval averages ten trials for all subjects for each condition. Left panels and right panels represent the M-L and A-P axes, respectively. The error bars represent standard deviations.

Spectral enveloppes generated by averaging individual spectral signatures are presented in Fig. 2. Qualitatively, one can observed that fatigue induced a clear increase in the spectral energy for the A-P axis. The effect was less important for the M-L axis. A subtle shift in the distribution of energy toward higher frequencies can also be observed. This shift can be appreciated with the analyses of the median frequency (50% power frequency) and the mean frequency (centroidal frequency) extracted from the spectra signature of each subject. Results of these analyses are shown in Fig. 3. Under fatigue conditions, subjects exhibited significantly greater median frequency than in the no fatigue conditions for both directions (F = 5.27, p < 0.05 and F = 6.87, p < 0.05 for the M-L and A-P directions, respectively). The ANOVAs also showed main effects of Vison (F = 13.76, p < 0.01 and F = 16.96, p < 0.01 for the M-L and A-P directions). For both axes, the interaction of Fatigue x Vision was not significant (ps > 0.05). For the mean frequency in M-L and A-P directions, the ANOVAs showed main effects of Fatigue (F = 6.64 , p < 0.05 and F = 7.16, p < 0.05, respectively) and Vision (F = 16.34, p < 0.01 and F = 17.25, p < 0.01, respectively), but no interaction (ps > 0.05). Overall, when subjects were fatigued, spectral analyses showed significant increase for both mean and median frequency.The effects were similar both with and without vision.

Figure 2 : Spectral enveloppes calculated by averaging the spectral signatures of each subject for the M-L (left panel) and the A-P (right panel) axes.

Figure 3 : Mean and median frequencies of the CP displacements.

Averaged median frequencies (A,B) (50% power frequency) and mean frequencies (C,D) (centroidal frequency) during the four conditions. Left and right panels represent the M-L and A-P axes, respectively. The error bars represent standard deviations.

3.3 Stabilogram-diffusion analysis

Results of the critical point coordinates obtained from the stabilogram-diffusion analysis with the exponential model are presented in Table 1. Muscular fatigue of the plantar-flexor muscles yielded (1) a significant increase of the spatial (Yc) characteristics of the transition region that separates the short-term and long-term regions of the resultant double-log-natural stabilogram-diffusion plot in the A-P direction (F = 7.71, p < 0.05) and (2) no significant change of the critical time interval (Δtc) along both axes (ps > 0.05). Mean values for the Vision and No-Vision conditions were similar to previous findings [24,25].

The scaling exponents (Hs and Hl) are the slopes of the resultant double-log-natural stabilogram-diffusion plots. For all experimental conditions, the CP trajectories were characterized by a persistent behavior (H > 0.5) over the short-term region and an antipersistent behavior (H < 0.5) over the long-term region. For Hs along the M-L axis (Fig. 4A), the ANOVA showed no effect of Fatigue (p > 0.05), but a main effect of Vision (F = 18.00, p < 0.01) and a significant interaction of Fatigue x Vision (F = 5.67, p < 0.05). In the presence of vision, Hs values were greater without than with fatigue (0.947 vs 0.939); in absence of vision, values did not differ (0.932 vs 0.933). For Hs along the A-P axis (Fig. 4B), ANOVA showed main effects of Fatigue and Vision (F = 5.11, p < 0.05 and F = 9.05, p < 0.05, respectively), but no interaction (p > 0.05). Hs for the A-P direction was smaller with than without fatigue; it was also smaller without vision than when vision was available.

For Hl in the M-L and A-P directions (Fig. 4C et 4D, respectively), the ANOVAs showed main effects of Fatigue (F = 6.36, p < 0.05 and F = 8.31, p < 0.05, respectively), and Vision (F = 38.07, p < 0.001 and F = 26.98, p < 0.001, respectively), but no interaction (ps > 0.05). For both conditions of vision and for both directions, the long-term scaling exponents were smaller with than without fatigue. In other words, the slope of the long term region decreased with fatigue. Overall, this suggests that muscular fatigue decreased the persitent behavior of the short-term region for the A-P direction and increased the antipersistent behavior of the long-term region for both directions. The effects of muscular fatigue, however, were not exacerbated by the absence of vision.

Figure 4 : Parameters of the stabilogram-diffusion analysis.

Upper panels (A,B) and lower panels (C,D) present the mean and standard deviation of the scaling coefficients (H) obtained from the exponential model of the stabilogram-diffusion plots along M-L axis and A-P axis, respectively. Left and right panels show respectively the scaling exponents calculated for short-term (Hs) and the long-term (Hl) regions.

Tableau 1 : Mean of the stabilogram diffusion parameters obtained from the exponential model and results of the ANOVAs for M-L and A-P CP directions.

 

Δtc (s)

Yc (mm2)

 

No Fatigue

Fatigue

No Fatigue

Fatigue

M-L

       

Vision

0.405

0.455

13.3

20.3

No Vision

0.472

0.479

32.1

39.2

FATIGUE

NS

NS

VISION

NS

0.001

FATIGUE X VISION

NS

NS

A-P

       

Vision

0.320

0.347

5.23

10.1

No Vision

0.385

0.462

17.4

29.9

FATIGUE

NS

0.05

VISION

0.05

0.001

FATIGUE X VISION

NS

NS

4. Discussion

The postural control system encompasses complex functions for detecting movement using information from the visual, vestibular and somatosensory receptors as well as evoking and controlling coordinated muscular responses. In the present experiment, the range and standard deviation of the CP displacements along both axes were not affected by fatigue. This suggests that the detection/action capabilities of the sensorimotor system remained partly efficient when the ankle plantar-flexors were fatigued. With fatigue, the postural control system was able to maintain the amplitude of the CP oscillations within the same physical limits of the base of support than that observed without fatigue.

Muscular fatigue of the ankle plantar-flexors, however, yielded an increase of the mean velocity along both axes. Increases of body sway were accompanied by a concurrent increase of the frequency of body sway, as indicated by the greater median and mean frequencies observed. This supports and extends previous findings showing that balance was affected by a generalized fatigue induced by a strenuous aerobic physical exercise [16-19]. Faster instantaneous oscillations (maximal instantaneous velocity) in the A-P direction also were observed with fatigue. This suggests that fatigue did induce some changes in the control mode. The faster instantaneous oscillations could be associated with discrete control of the postural oscillations required to compensate the motor and/or sensory deficiencies induced by peripheral muscular fatigue. It is difficult to determine the purpose of these specific events but they could result from specific output behaviors generated to avoid that the CP moved towards more eccentric positions.

The stabilogram-diffusion analysis yields CP parameters which relate to a persistence (short-term region) and antipersistence (long-term region) behaviors underlying the global control of upright stance. The neuromuscular interpretation of diffusion-stabilogram analysis has been subject to some controversy [26]. The objective of this analysis, however, is to find deterministic trends over a fractal Brownian motion by computing the mean values of the CP displacements over different temporal intervals spaced from 0 to 10 seconds. These trends are characterised by the scaling exponent (H). For both directions, the length of the short-term region and the long-term region were similar with and without muscular fatigue. However, significant changes of the spatial characteristic of the transition region were observed in the A-P direction. On average, the CP displacements were greater with than without muscular fatigue. This result could be explained by the increased mean velocity and the concomitant increase of the median and mean frequencies of the CP fluctuations since the distance between two near points increases when the mean velocity of the walker increases. The differences observed for the slope of the short-term region of the stabilogram-diffusion were relatively small. Such small changes, although statistically significant, are somehow difficult to associate with any significant physiological meaning. Under normal condition (vision and no fatigue condition), the long-term region of the stabilogram-diffusion plot showed that the system has less memory (higher Hl), that is the difference between a pair of points of the CP time series spaced by a time interval belonging to the long-term region tend to be intrinsically more stochastic than what was observed when fatigue was introduced (and when vision was withdrawn). In other words, with muscular fatigue, a more antipersistent behavior was observed over the long-term region suggesting that the actions taken by the postural control system were more numerous. There are at least two interpretations that could explain the long-term changes observed with muscular fatigue. One possibility is that a stiffening strategy (by increasing the activation of the antagonist muscles) could have been adopted to prevent the deviations. Results from a modeling study [26] suggests that this is a possibility. Using an inverted pendulum model, Peterka [26] showed that increasing the stiffness component of a PID controller (Kp, which is also the proportional feedback gain in an active control system) yielded stabilogram-diffusion parameters approximating those observed in our experiment when fatigue was present. Another interpretation could reside in the changes of the input-output characteristics of the fatigued effectors. Actions occuring at the peripheral and central level of the nervous system tend to decrease the rate of force development generated by the fatigued skeletal muscles. This altered force production could require an increase in the frequency of the corrections in order to avoid greater displacements of the CP. Both interpretations are not necessarily mutually exclusive, since they suggest that the control of the upright stance operates in a less stochastic manner when fatigue is present.

Contrary to our initial hypothesis, we observed that the effects of fatigue were similar with and without vision. This was a counter-intuitive result since perturbing the ankle sensory system normally yields an increased postural sway when vision is not available. For example, inserting a foam surface [27], co-vibrating the ankle tendons [28,29] or providing inaccurate ankle sensory information by sway-referencing the support surface [30,31] all yield large increases in body oscillations. Previous authors have shown that vision did not interact with fatigue with the possible exception of a one-leg stance experiment [20] where reinserting vision interacted with localized fatigue of the ankle muscles for a few seconds only. Altogether, these observations suggest that the effect of fatigue may affect more the motor output than the sensory system. Alternatively, it is possible that the sensory system is affected mainly at the peripheral level through a change in the spindle threshold of the fatigued muscles but other receptors (e.g. tactile and pressure sensors) could compensate for this increased threshold. This could explain the absence of a differential effect when vision was withdrawn.

Altogether, these results suggest that localised muscular fatigue of the ankle plantar-flexors seems to affect more the motor output of the postural control system than the sensory system. These results also suggest that, compared to the control no fatigue conditions, fatigue places higher demands on the postural control system by increasing the frequency of actions needed to regulate an upright stance.

Acknowledgements

The technical support of Marcel Kaszap in signal analysis and software developement is gratefully ackowledged. This work was carried out at the Groupe de Recherche en Analyse du Mouvement et en Ergonomie of the Division of Kinesiology at University Laval (Québec). This study was supported in part by NSERC-Canada, FCAR-Québec, Fondation de l’Université Laval, CIHR-FCQ and ÉGIDE France.

References

1. Massion J. Postural control system. Curr Opin Neurobiol 1994; 4: 877-887.

2. Johansson R, Magnusson M, Akesson M. Identification of human postural dynamics. IEEE Trans Biomed Eng 1988; 35: 858-869.

3. Dietz V. Human neuronal control of automatic functional movements: Interaction between central programs and afferent input. Physiol Rev 1992; 72: 33-69.

4. Schieppati M, Hugon M, Grasso M, Nardone A, Galante M. The limits of equilibrium in young and elderly normal subjects and in parkinsonians. Electroencephalogr Clin Neurophysiol 1994; 93: 286-298.

5. McClenaghan BA, Williams HG, Dickerson J, Dowda M, Thombs L, Eleazer P. Spectral characteristics of aging postural control. Gait Posture 1996; 4: 112-121.

6. Bigland-Ritchie B, Woods JJ. Changes in muscle contractile properties and neural control during human muscular fatigue. Muscle Nerve 1984; 7: 691-699.

7. Gandevia SC, Enoka RM, McComas AJ, Stuart DG, Thomas CK. Neurobiology of muscle fatigue. Advances and issues. In: Gandevia SC, Enoka RM, McComas AJ, Stuart DG, Thomas CK, eds. Fatigue. Neural and muscular mechanisms. Advances in experimental medecine and biology. New York: Plenum Press, 1995: 515-525.

8. Gandevia SC, Allen GM, Butler JE, Taylor JL. Supraspinal factors in human muscle fatigue: evidence for suboptimal output from the motor cortex. J Physiol 1996; 490: 529-536.

9. Bigland-Ritchie B. Reflex origin for the slowing of motoneurone firing rates in fatigue of human voluntary contractions. J Physiol 1986; 379: 451-459.

10. Rotto DM, Kaufman MP. Effect of metabolic products of muscular contraction on discharge of group III and IV afferents. J Appl Physiol 1988; 64: 2306-2313.

11. Macefield G, Hagbarth K-E, Gorman R, Gandevia SC, Burke D. Decline in spindle support to alpha motoneurones during sustained voluntary contractions. J Physiol 1991; 440: 497-512.

12. Windhorst U, Boorman G. Overview: Potential role of segmental motor circuitry in muscle fatigue. In: Gandevia SC, Enoka RM, McComas AJ, Stuart DG, Thomas CK, eds. Fatigue. Neural and Muscular Mechanisms. Advances in experimental medecine and biology. New York: Plenum Press, 1995: 241-258.

13. Garland SJ, Kaufman MP. Role of muscle afferents in the inhibition of motoneurons during fatigue. In: Gandevia SC, Enoka RM, McComas AJ, Stuart DG, Thomas CK, eds. Fatigue. Neural and Muscular Mechanisms. Advances in experimental medecine and biology. New York: Plenum Press, 1995: 271-278.

14. Hagbarth K-E, Macefield VG. The fusimotor system. Its role in fatigue. In: Gandevia SC, Enoka RM, McComas AJ, Stuart DG, Thomas CK, eds. Fatigue. Neural and Muscular Mechanisms. Advances in experimental medecine and biology. New York: Plenum Press, 1995: 259-270.

15. Taylor JL, Butler JE, Gandevia SC. Changes in muscle afferents, motoneurons and motor drive during muscle fatigue. Eur J Appl Physiol 2000; 83: 106-113.

16. Lepers R, Bigard AX, Diard J-P, Gouteyron J-F, Guezennec CY. Posture control after prolonged exercise. Eur J Appl Physiol Occup Physiol 1997; 76: 55-61.

17. Nardone A, Tarantola J, Galante M, Schieppati M. Time course of stabilometric changes after a strenuous treadmill exercise. Arch Phys Med Rehabil 1998; 79: 920-924.

18. Nardone A, Tarantola J, Giordano A, Schieppati M. Fatigue effects on body balance. Electroencephalogr Clin Neurophysiol 1997; 105: 309-320.

19. Johnston RB 3rd, Howard ME, Cawley PW, Losse GM. Effect of lower extremity muscular fatigue on motor control performance. Med Sci Sports Exerc 1996; 30: 1703-1707.

20. Vuillerme N, Nougier V, Prieur JM. Can vision compensate for a lower limbs muscular fatigue for controlling posture in humans? Neurosci Lett 2001; 308: 103-106.

21. Collins JJ, De Luca CJ. Open-Loop and Closed-Loop Control of Posture - A Random-Walk Analysis of Center-of-Pressure Trajectories. Exp Brain Res 1993; 95: 308-318.

22. Baratto L, Morasso PG, Re C, Spada G. A new look at posturographic analysis in the clinical context: sway-density versus other parameterization techniques. Motor Control 2002; 6: 246-270.

23. Glass L, Beuter A, Larocque D. Time delays, oscillations and chaos in physiological control systems. Math Biosci 1988; 90: 111-125.

24. Riley MA, Wong S, Miltra S, Turvey MT. Common effects of touch and vision on postural parameters. Exp Brain Res 1997; 117: 165-170.

25. Rougier P. Influence of visual feedback on successive control mechanisms in upright quiet stance in humans assessed by fractional Brownian motion modelling. Neurosci Lett 1999; 266: 157-160.

26. Peterka RJ. Postural control model interpretation of stabilogram diffusion analysis. Biol Cybern 2000; 82: 335-343.

27. Teasdale N, Stelmach GE, Breuing A. Postural sway characteristics of the elderly under normal and altered visual and support surface conditions. J Gerontol 1991; 46: B238-B244.

28. Hay L, Bard C, Fleury M, Teasdale N. Availability of visual and proprioceptive afferent messages and postural control in the elderly. Exp Brain Res 1996; 108: 129-139.

29. Teasdale N, Simoneau M. Attentional demands for postural control: the effects of aging and sensory reintegration. Gait Posture 2001; 14: 203-210.

30. Manchester D, Woollacott M, Zederbauer-Hylton N, Marin O. Visual, vestibular and somatosensory contributions to balance control in the older adult. J Gerontol 1989; 44: M118-M127.

31. Horak FB, Shupert CL, Dietz V, Horstmann G. Vestibular and somatosensory contributions to responses to head and body displacements in stance. Exp Brain Res 1994; 100: 93-106.