- Email: [email protected]
Attenuation of the evoked responses with repeated exposure to proprioceptive disturbances is muscle specific
Gait & Posture 32 (2010) 161–168
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Attenuation of the evoked responses with repeated exposure to proprioceptive disturbances is muscle specific Se´bastien Caudron a, Lucas Langlois b, Vincent Nougier c, Michel Guerraz a,* a
Laboratoire de Psychologie et de NeuroCognition, UMR 5105 CNRS–Universite´ de Savoie, BP 1104, 73011 Chambe´ry Cedex, France Laboratoire de Physiologie de l’exercice. Universite´ de Savoie, 73376 Le Bourget du lac, France c Laboratoire TIMC-IMAG - Equipe Sante´, Plasticite´, Motricite´, UMR 5525 CNRS–Universite´ Joseph Fourier, Faculte´ de Me´decine, 38706 La Tronche Cedex, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 July 2009 Received in revised form 7 April 2010 Accepted 13 April 2010
In response to repetitive proprioceptive disturbances (vibration) applied to postural muscles, the evoked response has been shown to decrease in amplitude within the first few trials. The present experiment investigated whether this attenuation of the response to vibration stimulation (90 Hz, 5 s) was muscle specific or would be transferred to the antagonist muscles. Sixteen participants stood upright with eyes closed. One half of the participants practiced 15 tibialis vibrations followed by 15 calf vibrations (TIBCALF order), while the other half practiced the opposite order (CALF-TIB order). Antero-posterior trunk displacements were measured at the level of C7 and centre of foot pressure (COP). EMG activity of the tibialis anterior (TA) and gastrocnemius lateralis (GL) was also measured. Results showed that evoked postural responses as well as EMG activity decreased with practice when vibration was applied to either calf or tibialis muscles. However, such attenuation of the response appeared muscle specific since it did not generalise when the same vibration stimulus was later applied onto the antagonist muscles. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Posture Sensory attenuation Muscle vibration
1. Introduction In order to limit the likelihood of balance loss, the postural control system has to continuously adapt to biomechanical or sensory varying conditions encountered in everyday life. Such behaviour can be experimentally objectivised when subjects are submitted to sensorial disturbances [1–8]. Whatever the channel manipulated, healthy subjects respond usually quite intensively to the first unexpected sensory disturbance. Responses appear sensibly reduced on subsequent stimulation [3,6,9,10]. However, despite their apparent similarity, attenuation of postural response to repeated disturbances are restricted to the sensory modality of the disturbance and do not transfer automatically to other sensory conditions [8]. Whatever the sensory stimulation (visual, vestibular, proprioceptive) postural responses appear directionally specific [3,11–16]. For instance, vibration applied to either the tendon or the belly of any muscle involved in postural adjustments gives rise to compensatory postural responses, which direction, size or speed depend on the location of the vibrators [17–19]. When vibration is applied to the Achilles tendons, muscle spindles of the calf-muscles (gastrocnemius and soleus) respond as if stretched, that is, as if a
* Corresponding author. Tel.: +33 4 79 75 91 86; fax: +33 4 76 82 78 34. E-mail address: [email protected] (M. Guerraz). 0966-6362/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2010.04.003
standing person is swaying forward. In response, subjects sway backward to balance the simulated forward sway [20]. Conversely, tibialis vibration causes subjects to sway forward. An important question is whether response attenuation to sensory disturbances can be transferred to similar sensory disturbances but inducing oppositely directed postural responses. Tendon-muscle vibration technique was used in the present experiment to assess this intra-modality transferability. Subjects were submitted to repeated vibration trials applied to either tibialis or calf muscles. Afterwards, they were vibrated on the opposite side, the calf muscles if previously vibrated at the level of the tibialis or vice versa. The purpose was to investigate whether adaptation to stimulation applied to one particular muscle, either the tibialis or the calf muscles could be transferred to its antagonist. 2. Method 2.1. Participants Sixteen participants (5 females and 11 males, 21.4 years 3.2; weight m = 64.5 kg and height m = 172 cm) took part in the experiment and gave their informed consent prior to the study. They had no history of balance or neuromuscular disorders. None of them had ever experienced muscle vibration. This experiment was approved by the local ethics committee and was performed in accordance with the Helsinki Declaration of 1975.
162
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
2.2. Vibratory stimulation Two pairs of inertial vibrators (VB115, Techno Concept, France) were attached to the ankle with elastic straps both on the tibialis (over the bellies of both tibialis muscles) and Achilles tendons (Fig. 1). Vibration was applied bilaterally to one of these two pairs of muscles (tibialis vs. calf muscles) for 5 s (90 Hz, 0.85 mm). 2.3. Procedure Before the experiment, participants were stimulated in a seated position to avoid startle reaction to the vibratory stimuli. Then, participants were instructed to stand still and relaxed, hands at their side, feet in a natural position (10 cm apart) and with eyes
closed. The experiment was divided in two sequences of stimulation differing by the stimulus location. One half of the participants practiced 15 tibialis followed by 15 calf vibrations (TIB-CALF), while the other half practiced the opposite order (CALF-TIB). Each vibratory stimulation was preceded and followed by a short and variable period of quiet stance (between 5 and 15 s). No additional time separated the two phases as the first trial of phase 2 directly followed the last trial of phase 1. The experiment lasted 1 h. 2.4. Data recording Trunk displacements were recorded with an electromagnetic motion capture system (Polhemus FastrakTM, USA). The sensor was
Fig. 1. The photo represents vibrators position. Traces are representative sample antero-posterior postural displacements measured at the level of C7 and COP in response to the first trial of tibialis (left panel) and calf (right panel) vibration. Upward deflection indicates a forward displacement while downward deflection indicates a backward displacement. On the lower panels are represented EMG activities. EMG traces represent the subtraction of gastrocnemius EMG recordings from tibialis recordings. One can notice an early limited agonist (early response) activity followed by a much larger antagonist activity (late response) in both tibialis and calf vibration conditions.
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
163
positioned at the level of C7. COP displacements were recorded in 12 subjects with a triangular force-plate (PF01TM; Equi+, France). Data were collected with a sampling frequency of 40 Hz (C7 and COP). EMG activities of the tibialis anterior (TA) and gastrocnemius lateralis (GL) muscles of the right leg were recorded via surface electrodes placed with a 2 cm inter-electrode distance longitudinally over the bellies of the muscles. A reference electrode was attached to a body area remote from the studied muscles. EMG data were sampled at 1000 Hz with a ME3000P8 device (MegaTM Electronics, Finland). The measured EMG was band pass filtered (8–500 Hz) and amplified 375 times close to the recording site. Stimulus artifacts due to movement of the electrodes were excluded from the EMG data using a digital filter (band stop 80–100 Hz [20]).
over the first second of vibration stimulation and that measured over a 1 s period preceding vibration onset. This parameter characterised muscle activity acting to move the body towards the vibrated side (2) the iEMG of the late (corrective) motor response calculated as the difference between the iEMG measured over a 4 s period from 1 to 5 s after the onset of the stimulus and that measured over a 1 s-period preceding vibration onset. This response (normalised for a 1 s period) characterises muscle activity acting to either limit body excursion towards the vibrated side or bring it back towards a baseline position. In supplementary analysis, we measured iEMG activity of each individual muscle during (1) a baseline period of 3 s preceding vibration onset and (2) the vibration period calculated as the difference between iEMG over a 4 s period from 1 to 5 s (normalised for a 1 s period) after the onset of the stimulus and that measured over the baseline period.
2.5. Data analysis
2.6. Statistical analyses
Postural displacement in response to the vibrated muscle was analyzed from C7 and COP movements by: (1) The velocity of the response measured from the slope based on a least-squares linear regression fitted over a period covering the displacement of C7 and COP from the first to the third second (2) the peak amplitude of the response calculated as the difference between the baseline level (mean position of C7 or COP over the 3 s before stimulus onset) and the maximal C7 or COP displacement measured during the 5 s vibration period. In addition, the off-peak amplitude of COP displacement (see [18]) was computed. It corresponds to the difference between the maximum amplitude of COP displacement during a 1 s period after cessation of vibration and the baseline level. Data are expressed as positive for displacements in the direction of the vibrated muscle. Potential body lean strategy in response to repeated vibration periods was also analysed by subtracting the baseline body position (C7 and COP) measured over the 3 s before stimulus onset for each trial with the baseline body position measured over the 3 s before stimulus onset of the first trial of the experiment (i.e., first trial in tibialis and calf vibration conditions for the TIB-CALF and the CALF-TIB group respectively). EMG activity was measured as follow: since TA and GL muscles act as antagonists in response to both tibialis and calf vibration conditions, we used the method initially proposed by Fitzpatrick et al. [21] (see also [22]) consisting in subtracting the GL EMG activity of each trial from that of the TA EMG activity. This method allows analyzing the orientation and variation of motor activity changes in pairs of antagonist muscles. To compare responses between subjects, all individual EMG data signals were first rectified and normalized to the maximal voluntary forces recorded for the tibialis and gastrocnemius muscles. iEMG activity was calculated as the percentage of maximal forces. From EMG activity, two main parameters were computed (1) the iEMG of the initial motor response calculated as the difference between the iEMG
Parameters were submitted to analyses of variance (ANOVA) with one between-subjects factor (‘‘Order’’ [TIB-CALF vs. CALFTIB]) and two within-subjects factors (‘‘Muscle’’ [Tibialis vs. Calf vibration] ‘‘Trial’’ [1st to 15th]). Normality was evaluated with the Kolmogorov Smirnov test. P level of significance was 0.05. 3. Results 3.1. Trunk (C7) displacements (16 participants) Vibration induced trunk (C7) displacements in the direction of the vibrated muscle with backward and forward displacements in condition of calf and tibialis vibration, respectively (Fig. 1). Trunk displacements were slightly larger in response to calf than tibialis vibration but this difference reached significance only for the velocity parameter (Table 1). The evoked response in condition of either calf or tibialis vibration decreased with practice (Fig. 2). Statistical analysis confirmed this attenuation when both velocity and peak amplitude were concerned (factor ‘‘Trial’’: p < .001). As can be seen in Fig. 2, the evoked response decreased markedly within the first trials in condition of calf vibration but decreased more gradually throughout the 15 trials in condition of tibialis vibration. The significant interaction between ‘‘Muscle’’ and ‘‘Trial’’ factors confirmed this differential decrease (p < .01). Newman–Keuls post-hoc comparisons revealed that whatever the vibrated muscle and order, the evoked response in the 15th trial was significantly smaller than that measured in the first trial (p < .001 for velocity and peak amplitude). The important issue was to know whether the decreased response to vibration of either calf or tibialis muscles would be transferred, at least partly, to vibration of their antagonists. As can be seen in Fig. 2, the benefit of experiencing repeated vibrations applied to one particular muscle was not transferred to its
Table 1 Statistical analyses: F-values for main effects and interactions obtained from the ANOVA for C7 and, COP analysis. df indicated degrees of freedom. p-values are mentioned in superscript (*: p < 05, **: p < 01, ***: p < .001, NS: p > .05). Trunk displacements (C7)
Order Muscle Trial Order Muscle Order Trial Muscle Trial Order Muscle Trial
COP displacements
df
Velocity
Peak amplitude
Baseline position
df
Velocity
Peak amplitude
Peak-off amplitude
Baseline position
1, 14 1, 14 14, 196 1, 14 14, 196 14, 196 14, 196
1.5NS 6.9* 11.4*** 1.4NS 0.7NS 2.9*** 0.7NS
0.3NS 4.5p = .05 4.4*** 1.5NS 0.5NS 2.6** 0.6NS
0.4NS 1.8NS 3.0*** 0.5NS 0.6NS 1.9** 1.2NS
1, 10 1, 10 14, 140 1, 10 14, 140 14, 140 14, 140
0.3NS 8.8* 4.4*** 0.7NS 1.9* 1.4NS 1.2NS
1.0NS 2.1NS 6.4*** 0.2NS 1.3* 2.3** 1.4NS
0.3NS 4.1 p=.07 0.5NS 1.2NS 0.9NS 0.9NS 0.8NS
0.1NS 3.9NS 1.6NS 2.5NS 1.2NS 2.7** 0.5NS
164
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
Fig. 2. (A) Mean velocity and (B) mean peak amplitude of C7 antero-posterior displacements depending on the trial number and the vibrated muscle for the Tib-Calf (dashed lines) and Calf-Tib (solid lines) groups. Positive values indicate displacements in the direction of the vibrated muscle. (C) Pre-trial body tilts (baseline C7 position). For clarity purposes, positive and negative values of baseline C7 position indicate forward and backward tilt respectively. Error bars indicate standard error of the mean.
antagonist: whatever the parameter concerned. Indeed, the factor ‘‘order’’ did not reach significance either as a main effect or in interaction with the other factors (Table 1). Baseline body position analysis (prior to vibration onset) showed that, after the first trial, participants tended to orient their body slightly in the direction opposite to the vibrated muscle (factor ‘‘Trial’’: p < .001, see Table 1 and Fig. 2C). The significant interaction ‘‘Muscle’’ ‘‘trial’’ (p < .05) and T-Student tests comparisons to the 0-norm (normalized position at the beginning of the experiment) showed that the initial body position changed significantly only through tibialis vibration sessions.
transferred to its antagonist. Analysis of the off-peak response of the COP within one second after cessation of vibration revealed a trend to a significant effect of the ‘‘muscle’’ factor (p = .07) with a slightly larger response in calf vibration condition. No other significant effect was noticed. Baseline COP position moved slightly in the opposite direction to the vibrated muscle after the first trial, but only in tibialis vibration condition (significant interaction ‘‘Muscle’’ ‘‘Trial’’: p < .01, Fig. 3C). No other significant result emerged (Table 1).
3.2. COP displacements (12 participants)
As can be seen in Fig. 1 muscle activity in response to vibration was characterised by a modest initial EMG response occurring within the first second after the onset of vibration, likely acting to move the body in the direction of the vibrated muscle. This early response was of a magnitude much smaller than the following corrective response (whatever the vibrated muscle considered) and was not systematically detectable on individual EMG traces.
As can be seen in Figs. 1–3, upper trunk (C7) and COP followed a similar trend in response to both calf and tibialis vibration indicating that participants largely adopted an ankle strategy (like inverted pendulums [23]). Therefore, as reported for C7, vibration induced COP displacements in the direction of the vibrated muscle which decreased with practice (p < .001 for velocity and peak amplitude). The absence of significant effect of the factor ‘‘order’’ (Table 1) confirmed that the attenuation of the evoked response with repeated exposure to vibration to one particular muscle is not
3.3. EMG analysis (16 participants)
3.3.1. Analysis of the initial EMG motor response The initial EMG response was significantly larger in response to calf than tibialis vibration (Muscle effect: p < .01, see Table 2) and
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
165
Fig. 3. (A) Mean velocity and (B) mean peak amplitude of COP antero-posterior displacements depending on the trial number and the vibrated muscle for the Tib-Calf (dashed lines) and Calf-Tib (solid lines) group orders. Positive values indicated displacements in the direction of the vibrated muscle. (C) Pre-trial position (baseline COP position) with positive and negative values indicating forward and backward COP displacements respectively. Error bars indicate standard error of the mean.
decreased with practice (Trial effect: p < .001). This attenuation can be seen notably in response to tibialis vibration from the first to the second trial and to a lesser extent in response to calf vibration (Fig. 4, top panels). In contrast, there was neither ‘‘order’’ effect nor any significant interaction between the three manipulated factors (p > .05).
3.3.2. Analysis of the late EMG motor response Results revealed that the late EMG response was not significantly different for tibialis and calf vibration conditions (p > .05). As can be seen in Fig. 4, this EMG activity decreased with practice (p < .05). No other significant effect was observed (Table 2).
Table 2 Statistical analyses: F-values for main effects and interactions obtained from the ANOVA for iEMG analyses. Agonist/Antagonist iEMG analysis consisted in subtracting the gastrocnemius EMG recordings of each subject from those of the tibialis muscle. df indicated degrees of freedom. p-values are mentioned in superscript (*: p < 05, **: p < 01, ***: p < .001, NS: p > .05). Agonist/Antagonist iEMG response
Order Muscle Trial Order Muscle Order Trial Muscle Trial Order Muscle Trial
Separate TA/GL iEMG analysis
df
Early response
Late response
1, 14 1, 14 14, 196 1, 14 14, 196 14, 196 14, 196
0.1NS 21.6*** 1.8** 1.5NS 1.0NS 1.2NS 0.7NS
0.9NS 2.3NS 2.0* 3.8p = .07 0.4NS 1.4NS 1.4NS
Baseline activity
Activity during vibration
TA
GL
TA
GL
0.5NS 0.03NS 1.3NS 0.3NS 1.1NS 0.7NS 0.9NS
0.5NS 7.5* 1.7NS 1.8NS 1.1NS 3.2*** 0.9NS
1.3NS 10.4** 5.8*** 0.7NS 0.4NS 6.0*** 0.4NS
0.4NS 24.0*** 2.6** 1.2NS 0.9NS 1.3NS 0.8NS
166
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
Fig. 4. EMG activity: Analysis consisted in subtracting the gastrocnemius EMG recordings of each subject from those of the tibialis muscle. This method allows measuring the orientation and variation of motor activity changes in pairs of antagonist muscles (see Fitzpatrick [19]) Upper figures (A) represent the mean early iEMG activity (calculated over the first second of vibration) and medium figures (B) represent the mean late iEMG activity (measured over a 4 s period from 1 to 5 s after stimulus onset. Negative values indicate EMG activity acting to move the body towards the vibrated side while positive values reflect EMG activity acting to limit body excursion toward the vibrated side and or bring back the body towards baseline position. In lower figures is represented the mean iEMG activity of TA and GL muscles measured during the 5 s period of stimulation. EMG is expressed in percentage of maximal voluntary forces and is reported according to trial number, vibrated muscle and order. Error bars indicate standard error of the mean.
3.3.3. Baseline TA and GL iEMG activity Baseline GL iEMG slightly increased and decreased by maximum 6–7% in calf and tibialis vibration conditions respectively [‘‘Muscle’’: p < .05; ‘‘Muscle’’ ‘‘Trial’’: p < .001]. In contrast, baseline TA iEMG was constant throughout the experiment, indicating that there was no anticipatory EMG co-activation (Table 2). 3.3.4. TA and GL iEMG activity during the 5 s period of vibration In response to calf vibration, a large TA EMG activity appeared, attesting of the key role of TA muscles in counteracting the evoked backward body sway. This activity decreased with practice (Fig. 4C, Table 2). In contrast, GL EMG activity in response to Calf vibration was either equal to baseline level or even decreased (negative values) attesting the absence of any co-contraction strategy to counteract postural disturbance. In response to tibialis vibration, a
large GL activity which counteracted the evoked forward body sway appeared (Fig. 4C, Table 2) but decreased with practice. Similarly, but to a much lesser extent, tibialis activity increased compared to baseline activity as expressed by positive values. Statistical analysis confirmed the significant effect of practice (p < .001) but also the absence of any order effect either as a main effect or in interaction with the other factors (p > .05). 4. Discussion In the present experiment, was observed an attenuation of the vibration evoked responses measured at the level of the COP, C7 but also in EMG activities when standing persons were submitted to repeated tendon-muscle vibration applied to either the tibialis or calf muscles. The main result was that such attenuation did not
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
generalise when the same vibration stimulus was later applied onto the antagonist muscles. In the present experiment, the evoked responses were larger with calf than with tibialis vibration and decreased markedly within 1–2 trials in the former condition but more gradually in the latter condition. These muscles specificities are reminiscent of the work of Gomez et al. [19] who observed differences when vibration was applied either to the calf or neck muscles. In agreement, proprioceptive perturbations have clearly different effects in terms of nature, degree and adaptative response depending on the site of vibratory stimulation. Attenuation of postural responses to repeated sensory stimulation is considered as reflecting the ability of the central nervous system (CNS) to disregard unreliable sensory afferents. It could alternatively be the consequence of changes in other parameters of the postural control system. In order to overcome postural destabilisation, participants could change their body alignment [7] to biomechanically counteract the disturbance. Such postural strategy was observed in the present experiment but was significant only in response to tibialis vibration. The leaning strategy was therefore not used to adapt to disturbances induced by calf vibration which appeared more destabilising than tibialis vibration. Not efficient on the long term, co-contraction of the antagonist muscles (tibialis and calf) may also be transiently appropriate to counteract sensory perturbations [24–26]. Analysis of baseline iEMG activity of TA and GL did not support this cocontraction hypothesis since baseline TA iEMG did not change throughout the experiment. In contrast, a slight increase of baseline GL iEMG was observed under calf vibration. Such an increased iEMG GL activity in concert with a forward body lean (gravitational forces would then be used as a counter-acting force) could theoretically act as a co-contraction effect. However, as reported above such a forward body lean did not appear to be significant in the present experiment. This increased GL iEMG was therefore more likely the indication of a vibration post-effect [27,28]. Similarly, analysis of iEMG activity of TA and GL muscles during the 5 s vibration period revealed that such co-contraction did not occur. Overall, these results indicated that modifying some parameters of the postural control such as changes in body alignment may transiently help, but cannot embrace the overall postural attenuation. Such analysis indirectly confirmed the central origin of this attenuation but also the ability of the CNS to disregard, at least partly, unreliable sensory afferents. The attenuation of postural responses to repeated sensory stimulation is often considered as reflecting changes in sensory weight [4,6–8,29–32]. The key idea of this hypothesis is that the postural control system is able to shift priorities from one sensory modality which appears unreliable to alternative more reliable sensory modalities. Evidence of such a shift was provided by Oie et al. [32]: subjects were submitted simultaneously to sinusoidal sensory stimulations (visual and somatosensory inputs) at different frequencies and amplitudes. Results revealed decreased reliance on one sensory input systematically benefited to the other sensory input. Within that context, the reduced weight allocated to the unreliable sensory channel, the proprioceptive channel in the present experiment should be compensated for by an increased weight to alternative more reliable sensory modalities. The absence of transfer from one muscle to its antagonist in the present results rather suggested that response attenuation is specific to the site of stimulation (restricted to the muscle from which the unreliable signal was originating) and therefore does unlikely result from a proprioceptive sensory-reweighting process. Does response attenuation to repetitive trials reflect pure adaptive mechanisms or habituation? Decreased responses from a rather intense response to the first unexpected sensory disturbance and a stabilized response after several identical disturbances
167
is defined as habituation [33]. Adaptation refers to the integration of novel sensorial cues through learning processes. Adaptation processes can be easily evaluated in motor tasks such as mirror drawing for instance. In the field of posture, these two processes are quite difficult to disentangle and the terms habituation and adaptation are often interchangeably used. Our experiment cannot ascertain as to whether changes in behavioural responses are the consequences of either integration of the novel proprioceptive input or the mark of the absence of reliance on this input. The rapid decrement of postural responses, particularly in condition of calf vibration, but also the absence of attenuation transfer from one muscle to the other argues however in favour of habituation instead of adaptation processes. In summary, the present results confirmed the ability of the postural control system not to rely, at least partly, on disturbing sensory afferents after repeated exposure. However, this appeared to be muscle specific, and therefore not transferred when the same disturbance was later applied onto the antagonist muscles. This absence of transfer should be considered for rehabilitation programs in which multiplying experiences in different conditions might be facilitating. Acknowledgment We express our gratitude to Anne Hillairet de Boisferon, Benjamin Fredembach and Julien Ochs for their technical advices during the pre-tests. Conflict of interest statement All authors disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work References [1] Nashner LM. Adapting reflexes controlling the human posture. Exp Brain Res 1976;26:59–72. [2] Mummel P, Timman D, Krause UWH, Boering D, Thilmann AF, Diener HC, et al. Postural responses to changing task conditions in patients with cerebellar lesions. J Neurol Neurosurg Psychiatry 1998;65:734–42. [3] Bronstein A. Suppression of visually evoked postural responses. Exp Brain Res 1986;63:655–8. [4] Loughlin PJ, Redfern MS. Spectral characteristics of visually induced postural sway in healthy elderly and healthy young subjects. IEEE Trans Neural Syst Rehabil Eng 2001;9:24–30. [5] Johansson R, Magnusson M, Fransson PA. Galvanic vestibular stimulation for analysis of postural adaptation and stability. IEEE Trans Biomed Eng 1995;42:282–92. [6] Smiley-Oyen AL, Cheng HY, Latt LD, Redfern MS. Adaptation of vibrationinduced postural sway in individuals with Parkinson’s disease. Gait Posture 2002;16:188–97. [7] Fransson PA, Tjernstro¨m F, Hafstro¨m A, Magnusson M, Johansson R. Analysis of short- and long-term effects of adaptation in human postural control. Biol Cybern 2002;86:355–65. [8] Fransson PA, Analysis of adaptation in human postural control. Ph.D. thesis. Lund University, 2005. [9] Bronstein AM, Hood JD, Gresty MA, Panagi C. Visual control of balance in cerebellar and parkinsonian syndromes. Brain 1990;113:767–79. [10] Caudron S, Nougier V, Guerraz M. Postural challenge and adaptation to proprioceptive perturbation. Exp Brain Res 2010;202:935–41. [11] Lee DN, Lishman JR. Visual proprioceptive control of stance. J Hum Mov Stud 1975;1:87–95. [12] Guerraz M, Gianna C, Burchill P, Gresty MA, Bronstein AM. Effect of visual surrounding motion on body sway in a 3D environment. Percept Psychophys 2001;63:47–58. [13] Day BL, Se´verac Cauquil A, Bartolomei L, Pastor MA, Lyon IN. Human bodysegment tilts induced by galvanic stimulation: a vestibularly driven balance protection mechanism. J Physiol (Lond) 1997;500:661–72. [14] Day BL, Guerraz M. Feedforward versus feedback modulation of vestibularevoked balance responses by visual self-motion information. J Physiol 2007;582:153–61. [15] Guerraz M, Day BL. Expectation and the vestibular control of balance. J Cog Neurosci 2005;17:463–9. [16] Eklund G. General features of vibration-induced effects on balance. Upsala J Med Sci 1972;77(2):112–24.
168
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
[17] Roll JP, Roll R. From eye to foot. A proprioceptive chain involved in postural control. In: Amblard B, Berthoz A, Clarac F, editors. Posture and gait. Amsterdam: Elsevier; 1988. p. 155–64. [18] Magnusson M, Andersson G, Gomez S, Johansson R, Martensson A, Karlberg M, et al. Cervical muscle afferents play a dominant role over vestibular afferents during bilateral vibration of neck muscles. J Vestib Res 2006;16:127–36. [19] Gomez S, Patel M, Magnusson M, Johansson L, Einarsson EJ, Fransson PA. Differences between body movement adaptation to calf and neck muscle vibratory proprioceptive stimulation. Gait Post 2009;30:93–9. [20] Ceyte H, Cian C, Zory R, Barraud PA, Roux A, Guerraz M. Effect of tendon vibration on postural orientation. Neurosci Lett 2007;416:71–5. [21] Fitzpatrick R, Burke D, Gandevia SC. Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol 1994;478:363–72. [22] Kavounoudias A, Roll R, Roll JP. Foot sole and ankle muscle inputs contribute jointly to human erect posture regulation. J Physiol 2001;532:869–78. [23] Smetanin BN, Popov KE, Kozhina GV. Human postural responses to vibratory stimulation of calf muscles under conditions of visual inversion. Hum Physiol 2002;28:53–8. [24] Carpenter MG, Frank JS, Silcher CP. Surface height effects on postural control: A hypothesis for a stiffness strategy for stance. J Vest Res 1999;9:277–86.
[25] De Serres SJ, Milner TE. Wrist muscle activation patterns and stiffness associated with stable and unstable mechanical loads. Exp Brain Res 1991;86:451–8. [26] Milner TE. Adaptation to destabilizing dynamics by means of muscle cocontraction. Exp Brain Res 2002;143:406–16. [27] Gilhodes JC, Gurfinkel VS, Roll JP. Role of Ia muscle spindle afferents in postcontraction and post-vibration motor effect genesis. Neurosci Lett 1992;135: 247–51. [28] Duclos C, Roll R, Kavounadias A, Roll JP, Forget R. Vibration-induced posteffects: a means to improve postural asymmetry in lower leg amputees. Gait Posture 2007;26:595–602. [29] Carver S, Kiemel T, Jeka JJ. Modeling the dynamics of sensory reweighting. Biol Cyber 2006;95:123–34. [30] Peterka RJ. Sensorimotor integration in human postural control. J Neurophysiol 2002;88:1097–118. [31] O’Connor P, Loughlin M, Redfern P, Sparto P. Postural adaptations to repeated optic flow stimulation in older adults. Gait Posture 2008;28:385–91. [32] Oie KS, Kiemel T, Jeka JJ. Multisensory fusion: simultaneous re-weighting of vision and touch for the control of human posture. Cog Brain Res 2002;14: 164–76. [33] Harris JD. Habituatory response decrement in the intact organism. Psychol Bull 1943;40:385–422.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Attenuation of the evoked responses with repeated exposure to proprioceptive disturbances is muscle specific Se´bastien Caudron a, Lucas Langlois b, Vincent Nougier c, Michel Guerraz a,* a
Laboratoire de Psychologie et de NeuroCognition, UMR 5105 CNRS–Universite´ de Savoie, BP 1104, 73011 Chambe´ry Cedex, France Laboratoire de Physiologie de l’exercice. Universite´ de Savoie, 73376 Le Bourget du lac, France c Laboratoire TIMC-IMAG - Equipe Sante´, Plasticite´, Motricite´, UMR 5525 CNRS–Universite´ Joseph Fourier, Faculte´ de Me´decine, 38706 La Tronche Cedex, France b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 July 2009 Received in revised form 7 April 2010 Accepted 13 April 2010
In response to repetitive proprioceptive disturbances (vibration) applied to postural muscles, the evoked response has been shown to decrease in amplitude within the first few trials. The present experiment investigated whether this attenuation of the response to vibration stimulation (90 Hz, 5 s) was muscle specific or would be transferred to the antagonist muscles. Sixteen participants stood upright with eyes closed. One half of the participants practiced 15 tibialis vibrations followed by 15 calf vibrations (TIBCALF order), while the other half practiced the opposite order (CALF-TIB order). Antero-posterior trunk displacements were measured at the level of C7 and centre of foot pressure (COP). EMG activity of the tibialis anterior (TA) and gastrocnemius lateralis (GL) was also measured. Results showed that evoked postural responses as well as EMG activity decreased with practice when vibration was applied to either calf or tibialis muscles. However, such attenuation of the response appeared muscle specific since it did not generalise when the same vibration stimulus was later applied onto the antagonist muscles. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Posture Sensory attenuation Muscle vibration
1. Introduction In order to limit the likelihood of balance loss, the postural control system has to continuously adapt to biomechanical or sensory varying conditions encountered in everyday life. Such behaviour can be experimentally objectivised when subjects are submitted to sensorial disturbances [1–8]. Whatever the channel manipulated, healthy subjects respond usually quite intensively to the first unexpected sensory disturbance. Responses appear sensibly reduced on subsequent stimulation [3,6,9,10]. However, despite their apparent similarity, attenuation of postural response to repeated disturbances are restricted to the sensory modality of the disturbance and do not transfer automatically to other sensory conditions [8]. Whatever the sensory stimulation (visual, vestibular, proprioceptive) postural responses appear directionally specific [3,11–16]. For instance, vibration applied to either the tendon or the belly of any muscle involved in postural adjustments gives rise to compensatory postural responses, which direction, size or speed depend on the location of the vibrators [17–19]. When vibration is applied to the Achilles tendons, muscle spindles of the calf-muscles (gastrocnemius and soleus) respond as if stretched, that is, as if a
* Corresponding author. Tel.: +33 4 79 75 91 86; fax: +33 4 76 82 78 34. E-mail address: [email protected] (M. Guerraz). 0966-6362/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2010.04.003
standing person is swaying forward. In response, subjects sway backward to balance the simulated forward sway [20]. Conversely, tibialis vibration causes subjects to sway forward. An important question is whether response attenuation to sensory disturbances can be transferred to similar sensory disturbances but inducing oppositely directed postural responses. Tendon-muscle vibration technique was used in the present experiment to assess this intra-modality transferability. Subjects were submitted to repeated vibration trials applied to either tibialis or calf muscles. Afterwards, they were vibrated on the opposite side, the calf muscles if previously vibrated at the level of the tibialis or vice versa. The purpose was to investigate whether adaptation to stimulation applied to one particular muscle, either the tibialis or the calf muscles could be transferred to its antagonist. 2. Method 2.1. Participants Sixteen participants (5 females and 11 males, 21.4 years 3.2; weight m = 64.5 kg and height m = 172 cm) took part in the experiment and gave their informed consent prior to the study. They had no history of balance or neuromuscular disorders. None of them had ever experienced muscle vibration. This experiment was approved by the local ethics committee and was performed in accordance with the Helsinki Declaration of 1975.
162
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
2.2. Vibratory stimulation Two pairs of inertial vibrators (VB115, Techno Concept, France) were attached to the ankle with elastic straps both on the tibialis (over the bellies of both tibialis muscles) and Achilles tendons (Fig. 1). Vibration was applied bilaterally to one of these two pairs of muscles (tibialis vs. calf muscles) for 5 s (90 Hz, 0.85 mm). 2.3. Procedure Before the experiment, participants were stimulated in a seated position to avoid startle reaction to the vibratory stimuli. Then, participants were instructed to stand still and relaxed, hands at their side, feet in a natural position (10 cm apart) and with eyes
closed. The experiment was divided in two sequences of stimulation differing by the stimulus location. One half of the participants practiced 15 tibialis followed by 15 calf vibrations (TIB-CALF), while the other half practiced the opposite order (CALF-TIB). Each vibratory stimulation was preceded and followed by a short and variable period of quiet stance (between 5 and 15 s). No additional time separated the two phases as the first trial of phase 2 directly followed the last trial of phase 1. The experiment lasted 1 h. 2.4. Data recording Trunk displacements were recorded with an electromagnetic motion capture system (Polhemus FastrakTM, USA). The sensor was
Fig. 1. The photo represents vibrators position. Traces are representative sample antero-posterior postural displacements measured at the level of C7 and COP in response to the first trial of tibialis (left panel) and calf (right panel) vibration. Upward deflection indicates a forward displacement while downward deflection indicates a backward displacement. On the lower panels are represented EMG activities. EMG traces represent the subtraction of gastrocnemius EMG recordings from tibialis recordings. One can notice an early limited agonist (early response) activity followed by a much larger antagonist activity (late response) in both tibialis and calf vibration conditions.
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
163
positioned at the level of C7. COP displacements were recorded in 12 subjects with a triangular force-plate (PF01TM; Equi+, France). Data were collected with a sampling frequency of 40 Hz (C7 and COP). EMG activities of the tibialis anterior (TA) and gastrocnemius lateralis (GL) muscles of the right leg were recorded via surface electrodes placed with a 2 cm inter-electrode distance longitudinally over the bellies of the muscles. A reference electrode was attached to a body area remote from the studied muscles. EMG data were sampled at 1000 Hz with a ME3000P8 device (MegaTM Electronics, Finland). The measured EMG was band pass filtered (8–500 Hz) and amplified 375 times close to the recording site. Stimulus artifacts due to movement of the electrodes were excluded from the EMG data using a digital filter (band stop 80–100 Hz [20]).
over the first second of vibration stimulation and that measured over a 1 s period preceding vibration onset. This parameter characterised muscle activity acting to move the body towards the vibrated side (2) the iEMG of the late (corrective) motor response calculated as the difference between the iEMG measured over a 4 s period from 1 to 5 s after the onset of the stimulus and that measured over a 1 s-period preceding vibration onset. This response (normalised for a 1 s period) characterises muscle activity acting to either limit body excursion towards the vibrated side or bring it back towards a baseline position. In supplementary analysis, we measured iEMG activity of each individual muscle during (1) a baseline period of 3 s preceding vibration onset and (2) the vibration period calculated as the difference between iEMG over a 4 s period from 1 to 5 s (normalised for a 1 s period) after the onset of the stimulus and that measured over the baseline period.
2.5. Data analysis
2.6. Statistical analyses
Postural displacement in response to the vibrated muscle was analyzed from C7 and COP movements by: (1) The velocity of the response measured from the slope based on a least-squares linear regression fitted over a period covering the displacement of C7 and COP from the first to the third second (2) the peak amplitude of the response calculated as the difference between the baseline level (mean position of C7 or COP over the 3 s before stimulus onset) and the maximal C7 or COP displacement measured during the 5 s vibration period. In addition, the off-peak amplitude of COP displacement (see [18]) was computed. It corresponds to the difference between the maximum amplitude of COP displacement during a 1 s period after cessation of vibration and the baseline level. Data are expressed as positive for displacements in the direction of the vibrated muscle. Potential body lean strategy in response to repeated vibration periods was also analysed by subtracting the baseline body position (C7 and COP) measured over the 3 s before stimulus onset for each trial with the baseline body position measured over the 3 s before stimulus onset of the first trial of the experiment (i.e., first trial in tibialis and calf vibration conditions for the TIB-CALF and the CALF-TIB group respectively). EMG activity was measured as follow: since TA and GL muscles act as antagonists in response to both tibialis and calf vibration conditions, we used the method initially proposed by Fitzpatrick et al. [21] (see also [22]) consisting in subtracting the GL EMG activity of each trial from that of the TA EMG activity. This method allows analyzing the orientation and variation of motor activity changes in pairs of antagonist muscles. To compare responses between subjects, all individual EMG data signals were first rectified and normalized to the maximal voluntary forces recorded for the tibialis and gastrocnemius muscles. iEMG activity was calculated as the percentage of maximal forces. From EMG activity, two main parameters were computed (1) the iEMG of the initial motor response calculated as the difference between the iEMG
Parameters were submitted to analyses of variance (ANOVA) with one between-subjects factor (‘‘Order’’ [TIB-CALF vs. CALFTIB]) and two within-subjects factors (‘‘Muscle’’ [Tibialis vs. Calf vibration] ‘‘Trial’’ [1st to 15th]). Normality was evaluated with the Kolmogorov Smirnov test. P level of significance was 0.05. 3. Results 3.1. Trunk (C7) displacements (16 participants) Vibration induced trunk (C7) displacements in the direction of the vibrated muscle with backward and forward displacements in condition of calf and tibialis vibration, respectively (Fig. 1). Trunk displacements were slightly larger in response to calf than tibialis vibration but this difference reached significance only for the velocity parameter (Table 1). The evoked response in condition of either calf or tibialis vibration decreased with practice (Fig. 2). Statistical analysis confirmed this attenuation when both velocity and peak amplitude were concerned (factor ‘‘Trial’’: p < .001). As can be seen in Fig. 2, the evoked response decreased markedly within the first trials in condition of calf vibration but decreased more gradually throughout the 15 trials in condition of tibialis vibration. The significant interaction between ‘‘Muscle’’ and ‘‘Trial’’ factors confirmed this differential decrease (p < .01). Newman–Keuls post-hoc comparisons revealed that whatever the vibrated muscle and order, the evoked response in the 15th trial was significantly smaller than that measured in the first trial (p < .001 for velocity and peak amplitude). The important issue was to know whether the decreased response to vibration of either calf or tibialis muscles would be transferred, at least partly, to vibration of their antagonists. As can be seen in Fig. 2, the benefit of experiencing repeated vibrations applied to one particular muscle was not transferred to its
Table 1 Statistical analyses: F-values for main effects and interactions obtained from the ANOVA for C7 and, COP analysis. df indicated degrees of freedom. p-values are mentioned in superscript (*: p < 05, **: p < 01, ***: p < .001, NS: p > .05). Trunk displacements (C7)
Order Muscle Trial Order Muscle Order Trial Muscle Trial Order Muscle Trial
COP displacements
df
Velocity
Peak amplitude
Baseline position
df
Velocity
Peak amplitude
Peak-off amplitude
Baseline position
1, 14 1, 14 14, 196 1, 14 14, 196 14, 196 14, 196
1.5NS 6.9* 11.4*** 1.4NS 0.7NS 2.9*** 0.7NS
0.3NS 4.5p = .05 4.4*** 1.5NS 0.5NS 2.6** 0.6NS
0.4NS 1.8NS 3.0*** 0.5NS 0.6NS 1.9** 1.2NS
1, 10 1, 10 14, 140 1, 10 14, 140 14, 140 14, 140
0.3NS 8.8* 4.4*** 0.7NS 1.9* 1.4NS 1.2NS
1.0NS 2.1NS 6.4*** 0.2NS 1.3* 2.3** 1.4NS
0.3NS 4.1 p=.07 0.5NS 1.2NS 0.9NS 0.9NS 0.8NS
0.1NS 3.9NS 1.6NS 2.5NS 1.2NS 2.7** 0.5NS
164
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
Fig. 2. (A) Mean velocity and (B) mean peak amplitude of C7 antero-posterior displacements depending on the trial number and the vibrated muscle for the Tib-Calf (dashed lines) and Calf-Tib (solid lines) groups. Positive values indicate displacements in the direction of the vibrated muscle. (C) Pre-trial body tilts (baseline C7 position). For clarity purposes, positive and negative values of baseline C7 position indicate forward and backward tilt respectively. Error bars indicate standard error of the mean.
antagonist: whatever the parameter concerned. Indeed, the factor ‘‘order’’ did not reach significance either as a main effect or in interaction with the other factors (Table 1). Baseline body position analysis (prior to vibration onset) showed that, after the first trial, participants tended to orient their body slightly in the direction opposite to the vibrated muscle (factor ‘‘Trial’’: p < .001, see Table 1 and Fig. 2C). The significant interaction ‘‘Muscle’’ ‘‘trial’’ (p < .05) and T-Student tests comparisons to the 0-norm (normalized position at the beginning of the experiment) showed that the initial body position changed significantly only through tibialis vibration sessions.
transferred to its antagonist. Analysis of the off-peak response of the COP within one second after cessation of vibration revealed a trend to a significant effect of the ‘‘muscle’’ factor (p = .07) with a slightly larger response in calf vibration condition. No other significant effect was noticed. Baseline COP position moved slightly in the opposite direction to the vibrated muscle after the first trial, but only in tibialis vibration condition (significant interaction ‘‘Muscle’’ ‘‘Trial’’: p < .01, Fig. 3C). No other significant result emerged (Table 1).
3.2. COP displacements (12 participants)
As can be seen in Fig. 1 muscle activity in response to vibration was characterised by a modest initial EMG response occurring within the first second after the onset of vibration, likely acting to move the body in the direction of the vibrated muscle. This early response was of a magnitude much smaller than the following corrective response (whatever the vibrated muscle considered) and was not systematically detectable on individual EMG traces.
As can be seen in Figs. 1–3, upper trunk (C7) and COP followed a similar trend in response to both calf and tibialis vibration indicating that participants largely adopted an ankle strategy (like inverted pendulums [23]). Therefore, as reported for C7, vibration induced COP displacements in the direction of the vibrated muscle which decreased with practice (p < .001 for velocity and peak amplitude). The absence of significant effect of the factor ‘‘order’’ (Table 1) confirmed that the attenuation of the evoked response with repeated exposure to vibration to one particular muscle is not
3.3. EMG analysis (16 participants)
3.3.1. Analysis of the initial EMG motor response The initial EMG response was significantly larger in response to calf than tibialis vibration (Muscle effect: p < .01, see Table 2) and
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
165
Fig. 3. (A) Mean velocity and (B) mean peak amplitude of COP antero-posterior displacements depending on the trial number and the vibrated muscle for the Tib-Calf (dashed lines) and Calf-Tib (solid lines) group orders. Positive values indicated displacements in the direction of the vibrated muscle. (C) Pre-trial position (baseline COP position) with positive and negative values indicating forward and backward COP displacements respectively. Error bars indicate standard error of the mean.
decreased with practice (Trial effect: p < .001). This attenuation can be seen notably in response to tibialis vibration from the first to the second trial and to a lesser extent in response to calf vibration (Fig. 4, top panels). In contrast, there was neither ‘‘order’’ effect nor any significant interaction between the three manipulated factors (p > .05).
3.3.2. Analysis of the late EMG motor response Results revealed that the late EMG response was not significantly different for tibialis and calf vibration conditions (p > .05). As can be seen in Fig. 4, this EMG activity decreased with practice (p < .05). No other significant effect was observed (Table 2).
Table 2 Statistical analyses: F-values for main effects and interactions obtained from the ANOVA for iEMG analyses. Agonist/Antagonist iEMG analysis consisted in subtracting the gastrocnemius EMG recordings of each subject from those of the tibialis muscle. df indicated degrees of freedom. p-values are mentioned in superscript (*: p < 05, **: p < 01, ***: p < .001, NS: p > .05). Agonist/Antagonist iEMG response
Order Muscle Trial Order Muscle Order Trial Muscle Trial Order Muscle Trial
Separate TA/GL iEMG analysis
df
Early response
Late response
1, 14 1, 14 14, 196 1, 14 14, 196 14, 196 14, 196
0.1NS 21.6*** 1.8** 1.5NS 1.0NS 1.2NS 0.7NS
0.9NS 2.3NS 2.0* 3.8p = .07 0.4NS 1.4NS 1.4NS
Baseline activity
Activity during vibration
TA
GL
TA
GL
0.5NS 0.03NS 1.3NS 0.3NS 1.1NS 0.7NS 0.9NS
0.5NS 7.5* 1.7NS 1.8NS 1.1NS 3.2*** 0.9NS
1.3NS 10.4** 5.8*** 0.7NS 0.4NS 6.0*** 0.4NS
0.4NS 24.0*** 2.6** 1.2NS 0.9NS 1.3NS 0.8NS
166
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
Fig. 4. EMG activity: Analysis consisted in subtracting the gastrocnemius EMG recordings of each subject from those of the tibialis muscle. This method allows measuring the orientation and variation of motor activity changes in pairs of antagonist muscles (see Fitzpatrick [19]) Upper figures (A) represent the mean early iEMG activity (calculated over the first second of vibration) and medium figures (B) represent the mean late iEMG activity (measured over a 4 s period from 1 to 5 s after stimulus onset. Negative values indicate EMG activity acting to move the body towards the vibrated side while positive values reflect EMG activity acting to limit body excursion toward the vibrated side and or bring back the body towards baseline position. In lower figures is represented the mean iEMG activity of TA and GL muscles measured during the 5 s period of stimulation. EMG is expressed in percentage of maximal voluntary forces and is reported according to trial number, vibrated muscle and order. Error bars indicate standard error of the mean.
3.3.3. Baseline TA and GL iEMG activity Baseline GL iEMG slightly increased and decreased by maximum 6–7% in calf and tibialis vibration conditions respectively [‘‘Muscle’’: p < .05; ‘‘Muscle’’ ‘‘Trial’’: p < .001]. In contrast, baseline TA iEMG was constant throughout the experiment, indicating that there was no anticipatory EMG co-activation (Table 2). 3.3.4. TA and GL iEMG activity during the 5 s period of vibration In response to calf vibration, a large TA EMG activity appeared, attesting of the key role of TA muscles in counteracting the evoked backward body sway. This activity decreased with practice (Fig. 4C, Table 2). In contrast, GL EMG activity in response to Calf vibration was either equal to baseline level or even decreased (negative values) attesting the absence of any co-contraction strategy to counteract postural disturbance. In response to tibialis vibration, a
large GL activity which counteracted the evoked forward body sway appeared (Fig. 4C, Table 2) but decreased with practice. Similarly, but to a much lesser extent, tibialis activity increased compared to baseline activity as expressed by positive values. Statistical analysis confirmed the significant effect of practice (p < .001) but also the absence of any order effect either as a main effect or in interaction with the other factors (p > .05). 4. Discussion In the present experiment, was observed an attenuation of the vibration evoked responses measured at the level of the COP, C7 but also in EMG activities when standing persons were submitted to repeated tendon-muscle vibration applied to either the tibialis or calf muscles. The main result was that such attenuation did not
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
generalise when the same vibration stimulus was later applied onto the antagonist muscles. In the present experiment, the evoked responses were larger with calf than with tibialis vibration and decreased markedly within 1–2 trials in the former condition but more gradually in the latter condition. These muscles specificities are reminiscent of the work of Gomez et al. [19] who observed differences when vibration was applied either to the calf or neck muscles. In agreement, proprioceptive perturbations have clearly different effects in terms of nature, degree and adaptative response depending on the site of vibratory stimulation. Attenuation of postural responses to repeated sensory stimulation is considered as reflecting the ability of the central nervous system (CNS) to disregard unreliable sensory afferents. It could alternatively be the consequence of changes in other parameters of the postural control system. In order to overcome postural destabilisation, participants could change their body alignment [7] to biomechanically counteract the disturbance. Such postural strategy was observed in the present experiment but was significant only in response to tibialis vibration. The leaning strategy was therefore not used to adapt to disturbances induced by calf vibration which appeared more destabilising than tibialis vibration. Not efficient on the long term, co-contraction of the antagonist muscles (tibialis and calf) may also be transiently appropriate to counteract sensory perturbations [24–26]. Analysis of baseline iEMG activity of TA and GL did not support this cocontraction hypothesis since baseline TA iEMG did not change throughout the experiment. In contrast, a slight increase of baseline GL iEMG was observed under calf vibration. Such an increased iEMG GL activity in concert with a forward body lean (gravitational forces would then be used as a counter-acting force) could theoretically act as a co-contraction effect. However, as reported above such a forward body lean did not appear to be significant in the present experiment. This increased GL iEMG was therefore more likely the indication of a vibration post-effect [27,28]. Similarly, analysis of iEMG activity of TA and GL muscles during the 5 s vibration period revealed that such co-contraction did not occur. Overall, these results indicated that modifying some parameters of the postural control such as changes in body alignment may transiently help, but cannot embrace the overall postural attenuation. Such analysis indirectly confirmed the central origin of this attenuation but also the ability of the CNS to disregard, at least partly, unreliable sensory afferents. The attenuation of postural responses to repeated sensory stimulation is often considered as reflecting changes in sensory weight [4,6–8,29–32]. The key idea of this hypothesis is that the postural control system is able to shift priorities from one sensory modality which appears unreliable to alternative more reliable sensory modalities. Evidence of such a shift was provided by Oie et al. [32]: subjects were submitted simultaneously to sinusoidal sensory stimulations (visual and somatosensory inputs) at different frequencies and amplitudes. Results revealed decreased reliance on one sensory input systematically benefited to the other sensory input. Within that context, the reduced weight allocated to the unreliable sensory channel, the proprioceptive channel in the present experiment should be compensated for by an increased weight to alternative more reliable sensory modalities. The absence of transfer from one muscle to its antagonist in the present results rather suggested that response attenuation is specific to the site of stimulation (restricted to the muscle from which the unreliable signal was originating) and therefore does unlikely result from a proprioceptive sensory-reweighting process. Does response attenuation to repetitive trials reflect pure adaptive mechanisms or habituation? Decreased responses from a rather intense response to the first unexpected sensory disturbance and a stabilized response after several identical disturbances
167
is defined as habituation [33]. Adaptation refers to the integration of novel sensorial cues through learning processes. Adaptation processes can be easily evaluated in motor tasks such as mirror drawing for instance. In the field of posture, these two processes are quite difficult to disentangle and the terms habituation and adaptation are often interchangeably used. Our experiment cannot ascertain as to whether changes in behavioural responses are the consequences of either integration of the novel proprioceptive input or the mark of the absence of reliance on this input. The rapid decrement of postural responses, particularly in condition of calf vibration, but also the absence of attenuation transfer from one muscle to the other argues however in favour of habituation instead of adaptation processes. In summary, the present results confirmed the ability of the postural control system not to rely, at least partly, on disturbing sensory afferents after repeated exposure. However, this appeared to be muscle specific, and therefore not transferred when the same disturbance was later applied onto the antagonist muscles. This absence of transfer should be considered for rehabilitation programs in which multiplying experiences in different conditions might be facilitating. Acknowledgment We express our gratitude to Anne Hillairet de Boisferon, Benjamin Fredembach and Julien Ochs for their technical advices during the pre-tests. Conflict of interest statement All authors disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work References [1] Nashner LM. Adapting reflexes controlling the human posture. Exp Brain Res 1976;26:59–72. [2] Mummel P, Timman D, Krause UWH, Boering D, Thilmann AF, Diener HC, et al. Postural responses to changing task conditions in patients with cerebellar lesions. J Neurol Neurosurg Psychiatry 1998;65:734–42. [3] Bronstein A. Suppression of visually evoked postural responses. Exp Brain Res 1986;63:655–8. [4] Loughlin PJ, Redfern MS. Spectral characteristics of visually induced postural sway in healthy elderly and healthy young subjects. IEEE Trans Neural Syst Rehabil Eng 2001;9:24–30. [5] Johansson R, Magnusson M, Fransson PA. Galvanic vestibular stimulation for analysis of postural adaptation and stability. IEEE Trans Biomed Eng 1995;42:282–92. [6] Smiley-Oyen AL, Cheng HY, Latt LD, Redfern MS. Adaptation of vibrationinduced postural sway in individuals with Parkinson’s disease. Gait Posture 2002;16:188–97. [7] Fransson PA, Tjernstro¨m F, Hafstro¨m A, Magnusson M, Johansson R. Analysis of short- and long-term effects of adaptation in human postural control. Biol Cybern 2002;86:355–65. [8] Fransson PA, Analysis of adaptation in human postural control. Ph.D. thesis. Lund University, 2005. [9] Bronstein AM, Hood JD, Gresty MA, Panagi C. Visual control of balance in cerebellar and parkinsonian syndromes. Brain 1990;113:767–79. [10] Caudron S, Nougier V, Guerraz M. Postural challenge and adaptation to proprioceptive perturbation. Exp Brain Res 2010;202:935–41. [11] Lee DN, Lishman JR. Visual proprioceptive control of stance. J Hum Mov Stud 1975;1:87–95. [12] Guerraz M, Gianna C, Burchill P, Gresty MA, Bronstein AM. Effect of visual surrounding motion on body sway in a 3D environment. Percept Psychophys 2001;63:47–58. [13] Day BL, Se´verac Cauquil A, Bartolomei L, Pastor MA, Lyon IN. Human bodysegment tilts induced by galvanic stimulation: a vestibularly driven balance protection mechanism. J Physiol (Lond) 1997;500:661–72. [14] Day BL, Guerraz M. Feedforward versus feedback modulation of vestibularevoked balance responses by visual self-motion information. J Physiol 2007;582:153–61. [15] Guerraz M, Day BL. Expectation and the vestibular control of balance. J Cog Neurosci 2005;17:463–9. [16] Eklund G. General features of vibration-induced effects on balance. Upsala J Med Sci 1972;77(2):112–24.
168
S. Caudron et al. / Gait & Posture 32 (2010) 161–168
[17] Roll JP, Roll R. From eye to foot. A proprioceptive chain involved in postural control. In: Amblard B, Berthoz A, Clarac F, editors. Posture and gait. Amsterdam: Elsevier; 1988. p. 155–64. [18] Magnusson M, Andersson G, Gomez S, Johansson R, Martensson A, Karlberg M, et al. Cervical muscle afferents play a dominant role over vestibular afferents during bilateral vibration of neck muscles. J Vestib Res 2006;16:127–36. [19] Gomez S, Patel M, Magnusson M, Johansson L, Einarsson EJ, Fransson PA. Differences between body movement adaptation to calf and neck muscle vibratory proprioceptive stimulation. Gait Post 2009;30:93–9. [20] Ceyte H, Cian C, Zory R, Barraud PA, Roux A, Guerraz M. Effect of tendon vibration on postural orientation. Neurosci Lett 2007;416:71–5. [21] Fitzpatrick R, Burke D, Gandevia SC. Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol 1994;478:363–72. [22] Kavounoudias A, Roll R, Roll JP. Foot sole and ankle muscle inputs contribute jointly to human erect posture regulation. J Physiol 2001;532:869–78. [23] Smetanin BN, Popov KE, Kozhina GV. Human postural responses to vibratory stimulation of calf muscles under conditions of visual inversion. Hum Physiol 2002;28:53–8. [24] Carpenter MG, Frank JS, Silcher CP. Surface height effects on postural control: A hypothesis for a stiffness strategy for stance. J Vest Res 1999;9:277–86.
[25] De Serres SJ, Milner TE. Wrist muscle activation patterns and stiffness associated with stable and unstable mechanical loads. Exp Brain Res 1991;86:451–8. [26] Milner TE. Adaptation to destabilizing dynamics by means of muscle cocontraction. Exp Brain Res 2002;143:406–16. [27] Gilhodes JC, Gurfinkel VS, Roll JP. Role of Ia muscle spindle afferents in postcontraction and post-vibration motor effect genesis. Neurosci Lett 1992;135: 247–51. [28] Duclos C, Roll R, Kavounadias A, Roll JP, Forget R. Vibration-induced posteffects: a means to improve postural asymmetry in lower leg amputees. Gait Posture 2007;26:595–602. [29] Carver S, Kiemel T, Jeka JJ. Modeling the dynamics of sensory reweighting. Biol Cyber 2006;95:123–34. [30] Peterka RJ. Sensorimotor integration in human postural control. J Neurophysiol 2002;88:1097–118. [31] O’Connor P, Loughlin M, Redfern P, Sparto P. Postural adaptations to repeated optic flow stimulation in older adults. Gait Posture 2008;28:385–91. [32] Oie KS, Kiemel T, Jeka JJ. Multisensory fusion: simultaneous re-weighting of vision and touch for the control of human posture. Cog Brain Res 2002;14: 164–76. [33] Harris JD. Habituatory response decrement in the intact organism. Psychol Bull 1943;40:385–422.