Effect of increasing difficulty in standing balance tasks with visual feedback on postural sway and EMG: Complexity and performance

Human Movement Science 31 (2012) 1224–1237

Contents lists available at SciVerse ScienceDirect

Human Movement Science journal homepage: www.elsevier.com/locate/humov

Effect of increasing difficulty in standing balance tasks with visual feedback on postural sway and EMG: Complexity and performance David Barbado Murillo a, Rafael Sabido Solana a,⇑, Francisco J. Vera-Garcia a, Narcis Gusi Fuertes b, Francisco J. Moreno a a b

Sport Research Centre, Miguel Hernández University of Elche, Elche, Spain Sport Science Faculty, University of Extremadura, Cáceres, Spain

a r t i c l e

i n f o

Article history: Available online 2 June 2012 PsycINFO classification: 2330 Keywords: Postural sway EMG Fuzzy Entropy Task difficulty Adaptability Visual feedback

a b s t r a c t Studies about the relationship between complexity and performance in upright standing balance have yielded mixed results and interpretations. The aim of the present study was to assess how the increasing difficulty in standing balance task affects performance and the complexity of postural sway and neuromuscular activation. Thirty-two young healthy participants were asked to stand still on a stability platform with visual feedback in three levels of difficulty. EMG signals from gastrocnemius medialis, tibialis anterior, rectus femoris and biceps femoris were measured with surface electromyography. As task difficulty increased, the amplitude of postural sway also increased. In the antero-posterior axis, Fuzzy Entropy (complexity) of postural sway decreased from the stable condition to the medium instability condition, and increased again at the highest instability condition. Fuzzy Entropy in the medio-lateral axis was higher in the stable condition; however, no differences were observed between the two instability conditions. Lower values of Fuzzy Entropy in postural sway during stable condition correlated with greater percent increases in postural sway in medio-lateral and antero-posterior axis from the standing still condition to the highest instability condition. In addition, mean and coefficient of variation of EMG increased and Fuzzy Entropy of EMG decreased when the difficulty in standing balance tasks increased. These results suggest that the higher postural sway complexity in stable condition, the greater capacity of the

⇑ Corresponding author. Address: Sport Research Centre of Elche, Miguel Hernández University, Avda. de la Universidad s/n, 03202 Elche, Spain. Tel.: +34 965 22 24 37; fax: +34 965 22 24 56. E-mail address: [email protected] (R. Sabido Solana). 0167-9457/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.humov.2012.01.002

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

1225

postural control system to adapt to the platform instability increases. In addition, changes in the complexity of EMG modulated by task difficulty do not necessarily reflect similar changes on postural sway. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction The maintenance of upright standing posture is a fundamental motor act that provides the basis for locomotion and most other movement tasks (Vaillancourt & Newell, 2002). The postural control system regulates the body’s postural sway in upright standing through the complex interaction of somatosensory, visual and vestibular sensory feedback networks, numerous brain regions, and the musculoskeletal system (Manor et al., 2010; Palmieri et al., 2003; Winter, Patla, & Frank, 1990). So ‘‘complexity’’ is defined as the number of system components and coupling functions (interaction) among them (Newell & Vaillancourt, 2001). This can be observed in the upright standing posture through fluctuations of postural sway (Lipsitz, 2002; Manor et al., 2010; Thurner, Mittermaier, & Ehrenberger, 2002). Numerous authors have suggested that complexity is related to the capacity of the system to generate adaptive responses to stressors (Goldberger, 1996; Goldberger et al., 2002; Lipsitz, 2002). In this sense higher system complexity is connected to a better performance. Therefore, a loss of complexity is thought to be linked to a reduced ability to adapt (Goldberger, 1996; Manor et al., 2010). This takes place when the difficulty of the task increases or when there is a motor control deficit (Seigle, Ramdani, & Bernard, 2009). These conditions caused a reduction in the number of individual structural components making up the system or an alteration in the coupling between components (Goldberger et al., 2002; Newell, 1998; Vaillancourt & Newell, 2002). These results have led to the general hypothesis that a higher complexity of the biological signals is associated to healthy systems, while there is a loss of complexity with aging and disease (Borg & Laxåback, 2010; Goldberger, Peng, & Lipsitz, 2002; Goldberger, Rigney, & West, 1990). Nevertheless, there are studies which do not support this hypothesis (Borg & Laxåback, 2010; Duarte & Sternad, 2008; Vaillancourt & Newell, 2002). So, for example, in some studies that have examined the relationship between complexity and aging in stability tests, they have found that aged people in addition to having a worse performance in these tests. That is to say, they showed a more complex response than young people in the postural stability control (Borg & Laxåback, 2010; Duarte & Sternad, 2008). Based on these results, and in contrast to the initial hypothesis, high complexity levels may be taken as signs that the system is becoming less sustainable. This is close to the traditional interpretation of entropy as a measure of disorder and noise (Borg & Laxåback, 2010). Vaillancourt and Newell (2002) suggest that there is not a universal increase or decrease in complexity with age and disease. Rather, the directional change in complexity is dependent on the dimension of the intrinsic dynamic of the behavioral or physiological system. These authors suggest a bidirectional complexity hypothesis that postulates that tasks with a stable equilibrium point show an increase in complexity are highly correlated with enhanced performance, in contrast to tasks with limit cycle stability that show a decrease. In this way, Newell and Vaillancourt (2001) have shown that more degrees of freedom are employed by subjects when trying to maintain a constant isometric force output than during the production of a force sine-wave. On the other hand, Morrison, Hong, and Newell (2007) suggest the possibility that the introduction of more degrees of freedom generally opposes the requirements of the task, finding a higher complexity displacement of center of pressure (COP) when trying to carry out tasks that require the use of a higher number of muscular groups. Most of the studies that have analyzed the relationship between complexity and performance in balance standing tasks have limited their analysis mainly to postural sway. Therefore, many of them focus on observing the complexity of COP displacement, while only a few have analyzed the complexity of variables such as the activation of muscles involved in postural sway. It is important to examine whether the properties of the EMG for individual muscles will be altered in a similar fashion to that of

1226

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

the COP (Morrison et al., 2007). Although only a limited number of papers have characterized surface EMG with measures of complexity, these measures provided useful information about neuromuscular activation (e.g., index of fatigue) and motor control (Chen, Wang, Xie, & Yu, 2007; Chen, Zhuang, Yu, & Wang, 2009; Farina, Fattorini, Felici, & Filligoi, 2002; Morrison et al., 2007; Rodrick & Karwowski, 2006; Xie, Guo, & Zheng, 2010). For example, Morrison et al. (2007) have demonstrated that timedependent structure of the EMG signal changes significantly as a function of the postural task being performed. The authors suggest that the more a task requires the intervention of more muscular groups and more joints, the complexity of the system at a macroscopic level increases (COP analysis), while at a microscopic level (electromyographic analysis) the complexity of the system decreases. In this situation, increases in complexity at one level of the system will be compensated for by decreases in complexity at the other (Morrison et al., 2007). In view of the controversy of results from the complexity and performance, the objectives of this study were: (1) to assess how the level of difficulty affects the complexity and performance; and (2) to explore the relationship between the complexity and the capacity of the system to adapt to stress (task difficulty). In addition, due to the lack of studies that analyze the complexity of the EMG signal in response to changes in the difficulty of balance task, we examined whether the neuromuscular activation complexity of the muscle groups involved in the control of upright stance posture is altered or not in the same way as the postural sway. For these reasons, we have carried out a study to investigate the modifications in postural sway and electromyographic signal features depending on the changes in the difficulty of the balance standing tasks on a stability platform (Biodex Balance System): stable, medium instability and high instability conditions. 2. Methods 2.1. Participants Thirty-two healthy volunteers took part in this study (age: 23.72 ± 2.80 years; height: 1.68 ± 0.09 m; mass: 64.66 ± 8.70 kg; M ± SD), sixteen women (age: 22.37 ± 1.71 years; height: 1.61 ± 0.04 m; mass: 57.92 ± 4.80 kg) and sixteen men (age: 25.06 ± 3.06 years; height: 1.74 ± 0.05 m; mass: 71.39 ± 6.04 kg). They had no history of neurological or visual disorders. All participants were right leg dominant. Written informed consent was obtained from each participant prior to testing. The experimental procedures used in this study were in accordance with the Declaration of Helsinki and were approved by the ethics standards of the committee on Human Experimentation of the University of Extremadura. 2.2. Instrumentation and data collection 2.2.1. The Biodex Balance System (BBS) To assess balance measures in static and dynamic conditions, this study used the BBS (Biodex, USA), which consists of a movable balance platform that provides up to 20° of surface tilt in a 360° range of motion. The platform is free to move on the antero-posterior (AP) and medio-lateral (ML) axes simultaneously. The degree of surface instability is controlled by the system’s microprocessor-based actuator by adjusting the level of resistance of eight springs located at the perimeter of the balance platform. Each spring, uncompressed, is 13.97 cm long, the outside diameter is 3.11 cm, and the wire diameter is 0.24 cm. The spring, made of music wire, is compressed to 7.52 cm and has a spring rate of 13.81 N/cm. When the spring is compressed, it creates 88.9 N of force (Arnold & Schmitz, 1998). To assess balance in static condition, the platform can be locked in the horizontal plane. Then, a series of strain gauges allows the BBS to determine variations in the participants’ resultant COP. The BBS has a support rail around the platform for safety (Biodex Medical Systems, 1999). The BBS has a display to provide visual feedback in real time through a cursor that should be kept as close as possible to the center of a screen. During dynamic assessments this cursor represents the degree of platform tilt on the AP and ML axes. The participant’s ability to control the platform angle is quantified as a variance from the locked (level) position, as well as degrees of deflection over time.

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

1227

During static assessments, the cursor represents the angular displacement of the participant’s center of gravity (COG) with respect to the center of the platform (zero point). The BBS software calculates COG angular excursion around the center of the platform applying a 0.4 s moving-average filter to the COP displacements, taking body height into consideration (55% of the participant’s height). So for static measures, the height of the participant must be previously input into the system (Biodex, 1999). Data from BBS, in static or dynamic tests, were collected at 20 Hz.

2.2.2. EMG recording Surface electromyograms were bilaterally recorded from gastrocnemius medialis (electrodes placed on the most prominent bulge of the muscle), tibialis anterior (electrodes placed at 1/3 from the fibula on the line between the tip of the fibula and the tip of the medial malleolus), rectus femoris (electrodes placed at the center of the line from the anterior superior iliac spine to the superior part of the patella) and biceps femoris (electrodes placed at the center of the line between the ischial tuberosity and the lateral epicondyle of the tibia). The A/D converter System (MP100, BIOPAC, USA) was used for collecting the EMG data, with a high level transducer (HLT100, BIOPAC, USA) and dedicated software (AcqKnowledge 3.7., BIOPAC, USA) on a personal computer. The myoelectric signals were amplified (gain: 1000; band pass filter: 20–450 Hz; common mode rejection ratio: 95 dB), A/D converter (16-bit resolution) at 1000 Hz. EMG activity was recorded using bipolar surface electrodes (TSD150B, BIOPAC, USA) with a diameter of 11.4 mm, interelectrode distance of 20 mm, built-in pre-amplification (350) and 3 dB band pass 12–500 Hz, and input impedance of 100 MX. The SENIAM guidelines (Hermens, Freriks, Disselhorst-Klug, & Rau, 2000) for surface EMG electrode placement were used to direct placement of the sensors. The area where the sensor was to be placed was shaved with a disposable razor, and vigorously abraded with an alcohol swab to remove dead skin cells and oils from the surface of the skin. The EMG sensor was attached with double sided sticky tape (BSN Medical Strappal). A conductive gel was also used as it was found this inhibited the effects of excessive sweat during long wearing periods. The reference electrode was placed on the lateral malleolus of the ankle. The influence of peripheral factors on EMG signal, such as the distance between the muscle and recording electrode, were controlled by the within subject design.

2.3. Procedure Participants performed one trial of each of the three different sway tasks in the BBS: (1) standing still on a locked platform of BBS (stable condition), (2) standing on an unstable surface at level 6 of BBS (medium instability condition), and (3) standing on an unstable surface at level 1 of BBS (high instability condition). The three tasks were counterbalanced. The length of each trial was 20 s and the rest period between trials was 1 min. The length of 20 s was sufficient because the postural sway was not non-stationary due to visual feedback task. In addition, data collection was initiated after participants achieved a steady state. For the standing still condition, participants were asked to stand still in their preferred standing position with arms crossed over the chest. This position was maintained throughout all the trials. In the instability conditions, participants were able to keep the standing posture without grasping the support rail or stepping in any direction. In order to normalize the EMG signal, the participants performed two maximal voluntary isometric contractions (MVC). The maximal exertions were performed by reaching maximal force and maintaining it for 3–4 s. For rectus femoris, a single knee extension in a 90° knee flexion position was performed. For biceps femoris, a single knee flexion in a 45° knee flexion position was performed. Finally, for gastrocnemius medialis and tibialis anterior a unilateral plantar flexion and a unilateral dorsal flexion at 90° of ankle position were performed, respectively. The participants were verbally encouraged during the execution. To prevent fatigue, the subject rested for 2 min after each maximal contraction.

1228

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

2.4. Data analysis and reduction 2.4.1. Measures of stability AP stability index (APSI) and ML stability index (MLSI) were calculated (Arnold & Schmitz, 1998) to assess participant’s balance. These indexes are standard deviations of the cursor which is assessing fluctuations around the center of the screen (zero point) in dynamic (degrees of platform tilt) or static conditions (degrees of COG angular excursion). The MLSI and the APSI represent the fluctuations along the AP and ML axes around the center of the platform, respectively:

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð0  ½Y2 Þ APSI ¼ n

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð0  ½X2 Þ MLSI ¼ n

Y means antero-posterior displacements; X means medio-lateral displacements; and n means number of samples. Higher score of these indexes indicates poorer balance (Arnold & Schmitz, 1998). 2.4.2. Measures of EMG amplitude The root mean square (RMS) EMG amplitude values were calculated for 30 ms time frame (Hatton, Dixon, Martin, & Rome, 2009; Hatton, Dixon, Rome, & Martin, 2011) and then normalized to MVC amplitudes. We selected 30 ms time frame because short time windows have been recommended for fast movements (Sinkjaer & Arendt-Nielsen, 1991) and here, the upright stance posture is controlled through intermittent stabilization bursts (fast adjustments) to produce ballistic torque impulses (Loram & Lakie, 2002; Loram, Maganaris, & Lakie, 2005; Morasso & Sanguineti, 2002). The mean and coefficient of variation of the normalized EMG amplitude were calculated across participants for each muscle and balance condition. 2.4.3. Measures of complexity A method of nonlinear time series analysis was applied to displacement of COG, platform tilt and raw EMG signal. Complexity was calculated by the Fuzzy Entropy (FuzzyEn) method (Chen et al., 2007, 2009; Xie et al., 2010)1. Higher values of FuzzyEn thus represent lower repeatability of vectors Xi of length m to that of m + 1, marking lower predictability of future data points, and greater irregularity within the time-series. Lower values represent a greater repeatability of vectors of length m + 1, and thus are a marker of higher predictability in the time series. The parameters chosen for FuzzyEn were the vector length m = 2, the gradient n = 2 and the tolerance window r = 0.2. Even though we have carried out the analyses of other related complexity measures (Approximate Entropy and Sample Entropy), we have finally chosen FuzzyEn. FuzzyEn is an evolution of the Approximate Entropy (Pincus, 1991) and Sample Entropy (Richman & Moorman, 2000), which are broadly used as complexity measures. Even though they have a similar meaning, the precision of the FuzzyEn is higher than either of them (Chen et al., 2007, 2009; Xie et al., 2010). Briefly, the main difference is that, unlike Approximate Entropy and Sample Entropy, which assess the similarity of two vectors by the Heaviside function, FuzzyEn employs an exponential function to bind two vectors’ similarity. In a Heaviside function, the boundary is rigid: the contributions of all the data points inside it are treated equally, whereas the data points outside it are left out. The hard boundary causes discontinuity, which may lead to abrupt changes of entropy values when the tolerance (r) changes slightly and even to the failure in Sample Entropy definition if no template-match can be found for small tolerance. In an exponential function, on the contrary, there is no rigid boundary. The exponential function value a around certain vector Xi can be viewed as the fuzzy membership to indicate the similarity between it and its neighbor Xj. The closer the neighboring vector Xj is, the more similar Xj is to Xi, and the similarity between Xi and Xj is almost zero when Xj is far away from Xi. As all the data points are considered as members of fuzzily exponential function, entropy values 1 In order to measure the complexity of EMG signal by FuzzyEn method, we have used raw data according to Chen et al. (2007, 2009). Nevertheless, to assess the robustness of FuzzyEn method, we have applied it based on three types of signal treatment: (1) Raw EMG signal (used by Chen et al., 2007, 2009; Xie et al., 2010); (2) The root mean square (RMS) EMG amplitude with 30 ms time frame (used in our study for EMG amplitude analysis); and (3) Full-wave rectification and bandpass filter at 2–400 Hz (used by Morrison et al., 2007). Filters and smoothing entail a reduction of complexity. Nevertheless its influence seems similar in all instability conditions and does not affect result interpretation.

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

1229

of FuzzyEn will change continuously and gracefully, and there is no parameter definition limitation. For an extensive description of Fuzzy Entropy see Chen et al. (2007, 2009). 2.4.4. Statistical analysis Data normality was examined using the Kolmogorov-Smirnov statistic with a Lilliefors correction. One-way repeated-measures ANOVA and task difficulty being within-subjects factor was performed in order to investigate the effects of increasing difficulty in standing balance tasks on postural sway and EMG variables. Post hoc analysis with Bonferroni adjustment was used for multiple comparisons. Mass and height were used as co-variables. With the aim of valuating sphericity, Mauchly’s test was carried out. Where the assumption of sphericity was violated (Epsilon < 1), Greenhouse-Geisser corrections were used. Pearson’s correlations were calculated to determine the following relationships: (a) between the complexity of postural sway in ML and AP axis during stable condition and the change (absolute and percentage values) in the index of stability (MLSI and APSI) from stable condition to medium and high instability condition (Manor et al., 2010); and (b) between the change in complexity of postural sway in ML and AP axis from stable condition to medium and high instability condition and the change (absolute and percentage values) in the index of stability (MLSI and APSI) from stable condition to medium and high instability condition. In the same way, Pearson’s correlations were calculated to determine the following relationship: (a) between the EMG complexity during stable condition and the change in amplitude and coefficient of variation of the EMG signal from stable condition to medium and high instability condition; and (b) between the change in amplitude, coefficient of variation and complexity of the EMG signal from stable condition to medium and high instability condition.

3. Results 3.1. Measures of stability and complexity of angular displacement of COG and displacement of platform tilt Examples of angular displacement of COG and platform tilt are shown in Fig. 1. Repeated-measures ANOVA showed significant main effects for task difficulty on the stability indexes (MLSI: F(1.071, 32.211) = 30.635, p < .001; APSI, F(1.061, 32.884) = 32.652, p < .001). Pairwise comparisons showed that APSI and MLSI increased when the difficulty of the balance task also increased from stable to high instability (Fig. 2). Repeated-measures ANOVA showed significant main effects on FuzzyEn of ML axis, F(1.569, 48631) = 22.237, p < .001 (Fig. 2) and AP axis, F(2, 62) = 10.642, p < .001. Pairwise comparisons showed that FuzzyEn of ML axis was greater in the stable condition compared to the two instability conditions, but there were no statistical differences between the medium and high instability conditions (Fig. 2). In the AP axis, the highest complexity of sway was also found in the stable condition (Fig. 2). However, unlike the ML axis, in the AP axis there were statistical differences between the two instability conditions; specifically, the complexity of postural sway was greater in the high instability condition (level 1 of BBS) compared to the medium instability condition (levels 6 of BBS). In this sense, for the AP axis, the relationship between the difficulty of the task (stable, medium instability and high instability) and the changes in the complexity of postural sway showed a quadratic function, F(1, 31) = 36.853, p < .001, while for the ML axis, this relationship showed a more linear function, F(1, 31) = 14.312, p < .001. Correlational analysis showed an inverse relationship between FuzzyEn in stable condition (Fig. 3) and the percentage of increase in postural sway from stable condition to medium instability condition in ML axis and from stable condition to high instability condition in ML and AP axis. 3.2. Measures of EMG amplitude and complexity: mean, coefficient of variation and Fuzzy Entropy Examples of normalized RMS EMG signals from the muscles of the right leg for each of the three conditions are shown in Fig. 4.

1230

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

A – Static

4 2

4

ML

0

2

-2 -4 4 2

0

AP -2

0 -2

-4

B – Level 6 of BBS 4 2

ML

0 -2 -4 4 2

AP

0 -2 -4

AP Angular Displacement (degrees)

Angular Displacement (degrees)

-4

-4

-2

0

2

4

-4

-2

0

2

4

-4

-2

0

2

4

4

2

0

-2

-4

C – Level 1 of BBS 4 2

ML

0

4

2

-2 -4 4 2

0

AP -2

0 -2 -4

-4

ML Angular Displacement (degrees) Fig. 1. Representative examples of antero-posterior (AP) and medio-lateral (ML) angular displacements of the center of gravity and platform tilt during the three balance conditions: standing on a static surface (stable), and standing on a labile surface at level 6 and level 1 of BBS (medium and high instability, respectively). Examples of the angular displacement in each dimension (AP and ML) over the time course of 20 s is presented in the left column. Representative AP–ML plots are shown in the right column. All traces were obtained from a single subject.

Repeated-measures ANOVA showed significant main effects on mean EMG amplitude of the leg muscles (RRF: F(2, 62) = 13.839, p < .001; LRF: F(2, 62) = 52.960, p < .001; RBF: F(2, 62) = 23.189, p < .001; LBF: F(2, 62) = 19.544, p < .001; RTA: F(2, 62) = 30.423, p < .001; LTA: F(2, 62) = 20.112, p < .001; RGM: F(2, 62) = 13.267, p < .001; LGM: F(2, 62) = 20.992, p < .001; Fig. 5). Pairwise comparisons showed that the mean EMG amplitude increased when the difficulty of the balance task also increased from stable to high instability condition.

1231

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

5

MLSI (º) FuzzyEn MLSI

4.5 4

Degrees

0.9 0.8

3.5 3

AB

2.5

0.7 0.6

A A

2

0.5

A

1.5

0.4

1

0.3

0.5 0

0.2 Stable

Medium instability

High instability

APSI (º) FuzzyEn APSI

5 4.5

AB

4

1 0.9 0.8

3.5

Degrees

1

AB

0.7

3 A

2.5

0.6

2

0.5

1.5

0.4

A

1

0.3

0.5

0.2

0

Stable

Medium instability

High instability

Fig. 2. Mean and standard deviation (error bars) of the sway variables [anterior–posterior stability index (APSI), and medio– lateral stability index (MLSI)] and Fuzzy Entropy (FuzzyEn) of medio-lateral axis (ML) and antero-posterior axis (AP), for stable (static surface of BBS) and medium and high instability conditions (level 6 and 1of the Biodex Balance System respectively). Post hoc analysis with Bonferroni adjustment was used for multiple comparisons: A Significantly different from Stable with p < .01; B Significantly different from Medium instability (level 6 of BBS) with p < .01. Differences in gray are referred to the postural sway (MLSI and APSI). Differences in black are referred to Fuzzy Entropy.

In addition, repeated-measures ANOVA showed significant main effects on the coefficient of variation of the EMG amplitude of the leg muscles (RRF: F(2, 62) = 36.516, p < .001; LRF: F(2, 62) = 11.589, p < .001; RBF: F(2, 62) = 18.317, p < .001; LBF: F(2, 62) = 23.142, p < .001; RTA: F(2, 62) = 72.155, p < .001; LTA: F(2, 62) = 59.398, p < .001; RGM: F(2, 62) = 33.542, p < .001; LGM: F(2, 62) = 24.703, p < .001). Pairwise comparisons showed that coefficient of variation of the EMG amplitude also increased when the difficulty of the balance task increased from stable to high instability condition (Fig. 5). As happened before, although most of the differences between levels of difficulty were statistically significant, there were some exceptions for the left leg (rectus femoris, biceps femoris and gastrocnemius medialis). Repeated-measures ANOVA showed significant main effects on the FuzzyEn of raw EMG signal of the leg muscles (RRF: F(2, 62) = 55.505, p < .001; LRF: F(2, 62) = 45.560, p < .001; RBF: F(2, 62) = 16.838, p < .001; LBF: F(2, 62) = 14.218, p < .001; RTA: F(2, 62) = 56.423, p < .001; LTA, F(2, 62) = 57.896, p < .001; RGM, F(2, 62) = 25.790, p < .001; LGM, F(2, 62) = 25.533, p < .001; Fig. 5). Pairwise comparisons showed that FuzzyEn of EMG signal decreased when the difficulty of the balance task increased from stable to high instability condition, indicating that there are more regular signals in the unstable surface conditions. Correlational analysis showed that the decrease in FuzzyEn of EMG signal (all muscles) from stable condition to medium and high instability condition correlated inversely with the increase and the percent increase of coefficient of variation of EMG amplitude from stable condition to medium and high instability condition (.5 < r < .8). However, the correlation analysis did not show relationship between FuzzyEn of the EMG signal and the mean EMG amplitude.

Best fit line (r = -0.764; p = 0.001)

2000 1500 1000 500 0 0

0.2

-500

0.4

0.6

0.8

1

Fuzzy Entropy of ML axis in Stable Condition

1000

Best fit line (r = -0.102; p = 0.538)

800 600 400 200 0 0.4

0.5

0.6

0.7

0.8

0.9

1

Fuzzy Entropy of AP axis in Stable Condition

-200

% change in APSI from Stable Condition to High Instability Condition

Medium Instability Condition

2500

% change in MLSI from Stable Condition to High Instability Condition

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

Medium Instability Condition

% change in APSI from Stable Condition to

% change in MLSI from Stable Condition to

1232

Best fit line (r = -0.493; p = 0.004)

12000 10000 8000 6000 4000 2000 0 0 -2000

0.2

0.4

0.6

0.8

1

Fuzzy Entropy of ML axis in Stable Condition

Best fit line (r = -0.381; p = 0.031)

4000 3500 3000 2500 2000 1500 1000 500 0 0.4

0.5

0.6

0.7

0.8

0.9

1

Fuzzy Entropy of AP axis in Stable Condition

Fig. 3. Relationship between postural sway complexity (Fuzzy Entropy) of Medial Lateral and Antero Posterior axis and the percent of increase in the Medial Lateral (MLSI) and Antero Posterior (APSI) Stability Index from Stable Condition to High Instability Condition.

Rectus Femoris

Level 6 of BBS Level 1 of BBS

EMG Amplitude (% MVC)

Static

50

Biceps Femoris

50

Tibialis Anterior

50

40

40

40

30

30

30

30

20

20

20

20

10

10

10

0

0

0

1

5001

10001

15001

1

5001

10001

10 0 1

15001

5001

10001

15001

50

50

50

50

40

40

40

40

30

30

30

30

20

20

20

20

10

10

10

0

0

0

1

5001

10001

1

15001

50

5001

10001

15001

50

5001

10001

15001

50

40

40

40

30

30

30

20

20

20

20

10

10

10

0

0

0

10001

15001

1

5001

10001

15001

5001

10001

15001

1

5001

10001

15001

1

5001

10001

15001

0

50

30

5001

1

10 1

40

1

Gastrocnemius Medialis

50

40

10 0 1

5001

10001

15001

Fig. 4. Representative examples of normalized RMS EMG of rectus femoris, biceps femoris, tibialis anterior and gastrocnemius medialis recorded during the three balance conditions: standing on a static surface (stable), and standing on a labile surface at level 6 and level 1 of BBS (medium and high instability, respectively). All traces were obtained from the right leg of a single subject.

4. Discussion The main objectives of this study were to assess the effect of the increase in the difficulty of balance task on postural sway and motor complexity, and to explore the relationship between the complexity and the capacity of the system to adapt to task difficulty. As a secondary objective, the neuromuscular activation complexity of the muscle groups involved in the control of upright stance posture was analyzed.

1233

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

Static (stable)

Level 6 of BBS (medium instability)

Level 1 of BBS (high instability)

30

AB AB

Amplitude (%MVC)

25

AB

AB

20

A

A

A

R-TA

L-TA

R-GM

A

15 10

AB A A

AB

AB A

5

A

Ab

0 R-RF

L-RF

R-BF

L-BF

L-GM

90

A

Coefficient of variation (%)

80

AB

A

AB

AB

70 A

60

AB

A

50

A

A A

A

AB

40 30

A

A A

20 10 0 R-RF

L-RF

R-BF

L-BF

R-TA

L-TA

A

A

R-GM

L-GM

2.5

Fuzzy Entropy

2 A

A AB

AB

1.5

A

AB

A

Ab AB

A

A Ab

AB

AB

1 0.5 0 R-RF

L-RF

R-BF

L-BF

R-TA

L-TA

R-GM

L-GM

Fig. 5. Mean and standard deviation (error bars) of the average and coefficient of variation of normalized EMG amplitudes and of Fuzzy Entropy of raw EMG signals of right (R) and left (L) rectus femoris (RF), biceps femoris (BF), tibialis anterior (TA) and gastrocnemius medialis (GM) for static surface and labile surface conditions of the Biodex Balance System (BBS). Post hoc analysis with Bonferroni adjustment was used for multiple comparisons: A Significantly different from static (stable) with p < .01; B significantly different from level 6 of BBS (medium instability) with p < .01; b Significantly different from level 6 of BBS (medium instability) with p < .05.

4.1. Stability and complexity of angular displacement of COG and platform tilt displacement Our results are consistent with other studies which found that an increase in the difficulty of the stability task entails an increase of the postural sway amplitude (Lamoth, van Lummel, & Beek, 2009; Negahban et al., 2010; Stins, Ledebt, Emck, van Dokkum, & Beek, 2009; Stins, Michielsen, Roerdink, & Beek, 2009). Observing the complexity of the postural sway, we find that in both the ML axis and the AP axis, the complexity of the balance is lower in the unstable conditions than in the stable condition. These

1234

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

results are similar to those obtained in previous studies in the ML axis (Negahban et al., 2010; Stins, Michielsen et al., 2009) and in the AP axis (Lamoth et al., 2009; Negahban et al., 2010; Stins, Michielsen et al., 2009), in which participants showed a higher complexity postural sway in quiet standing on rigid support as opposed to foam support. Especially significant were the results in the correlation analyses. They showed that higher postural sway complexity in stable condition correlates with less decrement in performance from stable to instability conditions. These findings agree with Manor et al. (2010), and may suggest that the higher postural sway complexity in the stable condition, the greater capacity of the postural control system to adapt to stressful constraints (caused here by instability of the platform). However, comparing changes in complexity between the moderate and the high instability conditions of our study, we observe that while in the ML axis no differences were appreciated, in the AP axis (axis where the higher amplitude oscillations are seen) participants showed a higher complexity sway in the highest instability situation. These results question the statement that the system’s predictability increases as the task difficulty increases (Seigle et al., 2009). Specifically, as we have observed in the results of the AP axis (Fig. 2) the data distribution across the three levels of difficulty resembles a quadratic function. It is impossible to compare these results with those of previous studies, as to our knowledge no previous studies have examined the differences in postural sway complexity comparing several levels of instability. However, we have found a similar result from a single study which analyzed the effects of speed variation on gait stability and complexity (Arif, Ohtaki, Nagatomi, & Inooka, 2004). In this study the authors found, for an elderly sample, that stability (evaluated via mediolateral acceleration) reduced as the speed increased (that is, as the difficulty of the task increased). Nevertheless, similarly to what was observed in our study, the relationship between complexity and speed (constraint that modulates task difficulty) described a U-shape curve. One explanation could be that the highest instability condition (level 1 of BBS) is a novel task for most participants, who are not used to such a high level of instability in their daily lives, and show an exploratory behavior to adapt to it (Ko, Challis, & Newell, 2003). This exploratory behavior is characterized by an increase of the body’s degrees of freedom (reflected in higher levels of complexity) and relatively large variability (higher postural sway). According to Ko et al. (2003), this high degree of biomechanical redundancy leads to a search for the most appropriate mode from among several alternative solutions. An alternative interpretation could be based upon the participants’ difficulties to control the standing posture in the most difficult condition of BBS. This may result in near loss of stability and in irregular behavior and hence a reduction in the predictability or an increase in entropy (Borg & Laxåback, 2010). Nevertheless, each trial was visually controlled by a researcher and no participant showed difficulties to maintain the standing posture in the high instability condition (e.g., grasping the support rail and/or making a step). In addition, participants did not show a great postural sway in this condition (normally it was not higher than 5-6° of platform tilt). In order to provide more clarity to the results obtained here, it is necessary that future studies on balance tasks are designed with a higher number of difficulty levels. 4.2. Amplitude and complexity of EMG signal Although we have not compared between muscles, the EMG amplitudes of the gastrocnemius medialis and tibialis anterior seemed higher in the three balance tasks than those of rectus and biceps femoris (Fig. 2). In our opinion, the higher activation of the ankle muscles could be due to the fact that the ankle strategy is the main mechanism to counteract the gravitational torque in upright standing posture (Winter, Patla, Prince, Ishac, & Gielo-Perczak, 1998; Winter, Patla, Rietdyk, & Ishac, 2001). On the other hand, the higher values in rectus and biceps femoris EMG activity under the most challenging conditions (instability level 6 and 1 of BBS) could suggest an increase in the role played by the knee and hip joints. Our results are consistent with other studies which found that an increase in the difficulty of the stability task entails an increase of the muscle activity (Fransson, Gomez, Patel, & Johansson, 2007). The higher extensor/flexor co-activation (gastrocnemius medialis/tibialis anterior and rectus femoris/biceps femoris) could increase the stiffness of the leg joints, and as a result increase joint stability (Winter et al., 1998, 2001). Additionally, the higher coefficient of variations of EMG amplitude when

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

1235

task difficulty increased may indicate that many bursts of muscle activity were required to control postural sway and prevent the body from falling (Bottaro, Casadio, Morasso, & Sanguineti, 2005; Sasagawa, Ushiyama, Masani, Kouzaki, & Kanehisa, 2009). In addition to the increases in amplitude and coefficient of variation, we observed a reduction in complexity, that is to say, the EMG complexity was higher in the stable condition than in the unstable conditions. These results may reflect a change in the strategy used to control the posture when difficulty increased. In standing still, participants may carry out postural corrections at irregular intervals with the aim of altering the COG angular displacement (Morrison et al., 2007), but when task difficulty increased, postural corrections might have been made at more regular intervals and with more similar amplitudes. Previous studies using linear analysis have shown intermittent and irregular activity, with very quick variations, in the lower leg muscles while standing still (Masani, Vette, & Popovic, 2006; Mochizuki, Ivanova, & Garland, 2005). Nevertheless, the changes in EMG complexity could be also related to modifications in the influence of lower-level physiological variables. There are several factors that contribute to the surface EMG amplitude such as recruitment in motor fiber, changes in firing rate, modification in muscle fiber conduction velocity, motor unit synchronization and even peripheral factors such as the distance between the muscle and recording electrode (Fuglevand, Winter, & Patla, 1993; Zhou & Rymer, 2004). It is possible that changes in EMG complexity could be modulated in some way by the increase in the intensity of muscular activation. At low levels of activation, variations in applied force are mainly caused by changes in the recruitment of muscle fibers, while at higher levels of force (50–80% of MVC), variations in intensity depend more on the changes in the firing rate (Moritani & Muro, 1987). So changes in complexity could be related to changes in the muscular contraction intensity used by the system. However, due to the low level of EMG amplitude (<25% of MVC) in this study, we may exclude this option. On the other hand, decreased EMG complexity could also reflect a reduction of muscle fiber conduction velocity and/or an increase of motor unit synchronization due to muscle fatigue (Farina et al., 2002; Xie et al., 2010). To avoid the fatigue effects in our study, the 3 balance tasks were counterbalanced. Minimization of sway size is caused in standing by an improvement in the accuracy of the ballistic torque impulses (Loram & Lakie, 2002; Loram et al., 2005). As balance task difficulty increases, synchronization of motor unit may facilitate higher ballistic torque impulses (Fling, Christie, & Kamen, 2008; Gabriel & Kamen, 2009). Just as Farina et al. (2002) have showed, an increase in synchronization of EMG caused a higher regularity of the signal, and in consequence, a lower EMG complexity. The correlational analyses showed that there is not a relationship between EMG complexity and amplitude of the EMG. The results of this study suggest that the complexity of the EMG signal is more related to postural corrections than lower-level physiological variables. However, future studies should be carried out to explore it in depth. 4.3. Comparison of task difficulty effect between postural sway and EMG complexity Although the complexity of both, postural sway and EMG signal, decreased from stable condition to medium instability condition, different EMG and postural sway complexity modifications were observed from medium to high instability condition (Figs. 2 and 5). Specifically, while EMG complexity decreased from medium to high instability condition, postural sway complexity increased in the AP axis and did not show significant differences in the ML axis. These unexpected results also do not support the findings by Morrison et al. (2007). These authors observed an inverse relationship between the postural sway and the electromyographic signal complexity. According to the aforementioned authors, increases in complexity at one level of the system would be compensated for by decreases in complexity at the other (Morrison et al., 2007). The results found in our study can complement those found by Morrison et al. (2007), as they provide information on motor control in balance tasks in function of the degree of difficulty. In our opinion, changes in complexity are an indicator of how the system components’ relationship modifies to adapt to the characteristics of the task. So, these changes are dependent of the intrinsic dynamic of the system (Seigle, Ramdani, & Bernard, 2009). In addition, according to Morrison et al. modifications in the complexity of EMG, modulated by task difficulty, do not necessarily reflect similar changes on postural sway. The differences in the relationship of the postural sway complexity and the EMG in the results obtained in our study and those

1236

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

obtained by Morrison et al., could indicate different postural control strategies between balance tasks (analyzed in our study) and voluntarily random and regular sway motions (analyzed in the Morrison’s study). Therefore, the relationship between the complexity of neuromuscular activation and postural sway seems task dependent. 5. Summary and conclusions In the present study we analyzed changes in postural sway and EMG signals in three different balance tasks with visual feedback. We found greater postural sway complexity in the upright standing in the stable condition than in the instability conditions. This may be a sign of an increased capacity of the postural control system to adapt to stresses caused by increasing instability on a stability platform. These results agree with the hypothesis that there is a loss of complexity as task difficulty increases. On the other hand, we did not observe the same changes in postural sway complexity in both axes when we compared between the two instability conditions; that is, while there were no differences in the ML axis, the postural sway complexity was higher in the high instability condition than in the moderate instability condition for the AP axis. These results could suggest that the higher instability situation would be a novel task and the participants showed an exploratory behavior to find the most appropriate solution from among several possible alternatives. EMG amplitude (mean and coefficient of variation) increased and EMG complexity decreased as the difficulty of the stability task increased. The neuromuscular activation complexity of the muscle groups involved in the control of upright standing posture was not altered in the same way as the postural sway complexity. Based on this and previous studies, the relationship between EMG and postural sway complexity seems task dependent. Overall, these results seem to indicate that the relationship between complexity and system performance is not linear. References Arif, M., Ohtaki, Y., Nagatomi, R., & Inooka, H. (2004). Estimation of the effect of cadence on gait stability in young and elderly people using approximate entropy technique. Measurement Science Review, 4, 29–40. Arnold, B. L., & Schmitz, R. J. (1998). Examination of balance measures produced by the Biodex Stability System. Journal of Athletic Training, 33, 323–327. Biodex Medical Systems, Inc. (1999). Balance system operations and service manual, #945–300. Shirley, NY: Author. Borg, F. G., & Laxåback, G. (2010). Entropy of balance – Some recent results. Journal of Neuroengineering and Rehabilitation, 7, 38. Bottaro, A., Casadio, M., Morasso, P. G., & Sanguineti, V. (2005). Body sway during quiet standing: Is it the residual chattering of an intermittent stabilization process? Human Movement Science, 24, 588–615. Chen, W., Wang, Z., Xie, H., & Yu, W. (2007). Characterization of surface EMG signal based on fuzzy entropy. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 15, 266–272. Chen, W., Zhuang, J., Yu, W., & Wang, Z. (2009). Measuring complexity using FuzzyEn, ApEn, and SampEn. Medical Engineering and Physics, 31, 61–68. Duarte, M., & Sternad, D. (2008). Complexity of human postural control in young and older adults during prolonged standing. Experimental Brain Research, 191, 265–276. Farina, D., Fattorini, L., Felici, F., & Filligoi, G. (2002). Nonlinear surface EMG analysis to detect changes of motor unit conduction velocity and synchronization. Journal of Applied Physiology, 93, 1753–1763. Fling, B. W., Christie, A., & Kamen, G. (2008). Motor unit synchronization in FDI and biceps brachii muscles of strength-trained males. Journal of Electromyography and Kinesiology, 5, 800–809. Fransson, P. A., Gomez, S., Patel, M., & Johansson, L. (2007). Changes in multi-segmented body movements and EMG activity while standing on firm and foam support surfaces. European Journal of Applied Physiology, 101, 81–89. Fuglevand, A. J., Winter, D. A., & Patla, A. E. (1993). Models of recruitment and rate coding organization in motor-unit pools. Journal of Neurophysiology, 70, 2470–2488. Gabriel, D. A., & Kamen, G. (2009). Experimental and modeling investigation of spectral compression of biceps brachii SEMG activity with increasing force levels. Journal of Electromyography Kinesiology, 19, 437–448. Goldberger, A. L. (1996). Nonlinear dynamics for clinicians: Chaos theory, fractals, and complexity at the bedside. Lancet, 347, 1312–1314. Goldberger, A. L., Amaral, L. A., Hausdorff, J. M., Ivanov, P. C., Peng, C. K., & Stanley, H. E. (2002). Fractal dynamics in physiology: Alterations with disease and aging. Proceedings of the National Academy of Science of United State of America, 99, 2466–2472. Goldberger, A. L., Peng, C. K., & Lipsitz, L. A. (2002). What is physiologic complexity and how does it change with aging and disease? Neurobiology of Aging, 23, 23–26. Goldberger, A. L., Rigney, D. R., & West, B. J. (1990). Chaos and fractals in human physiology. Scientific American, 262, 42–49. Hatton, A. L., Dixon, J., Martin, D., & Rome, K. (2009). The effects of textured surfaces on postural stability and lower limb muscle activity. Journal of Electromyography and Kinesiology, 19, 957–964. Hatton, A. L., Dixon, J., Rome, K., & Martin, D. (2011). Standing on textured surfaces: Effects on standing balance in healthy older adults. Age and Ageing, 40, 363–368.

D. Barbado Murillo et al. / Human Movement Science 31 (2012) 1224–1237

1237

Hermens, H. J., Freriks, B., Disselhorst-Klug, C., & Rau, G. (2000). Development of recommendations for SEMG sensors and sensor placement procedures. Journal of Electromyography and Kinesiology, 10, 361–374. Ko, Y. G., Challis, J. H., & Newell, K. M. (2003). Learning to coordinate redundant degrees of freedom in a dynamic balance task. Human Movement Science, 22, 47–66. Lamoth, C. J., van Lummel, R. C., & Beek, P. J. (2009). Athletic skill level is reflected in body sway: A test case for accelometry in combination with stochastic dynamics. Gait and Posture, 29, 546–551. Lipsitz, L. A. (2002). Dynamics of stability: The physiologic basis of functional health and frailty. The Journals of the Gerontology. Series A, Biological Sciences and Medical Sciences, 57, 115–125. Loram, I. D., & Lakie, M. (2002). Human balancing of an inverted pendulum: Position control by small, ballistic-like, throw and catch movements. Journal of Physiology, 540, 1111–1124. Loram, I. D., Maganaris, C. N., & Lakie, M. (2005). Human postural sway results from frequent, ballistic bias impulses by soleus and gastrocnemius. Journal of Physiology, 564, 295–311. Manor, B., Costa, M. D., Hu, K., Newton, E., Starobinets, O., Kang, H. G., et al. (2010). Physiological complexity and system adaptability: Evidence from postural control dynamics of older adults. Journal of Applied Physiology, 109, 1786–1791. Mochizuki, G., Ivanova, T. D., & Garland, S. J. (2005). Synchronization of motor units in human soleus muscle during standing postural tasks. Journal of Neurophysiology, 94, 62–69. Morasso, P., & Sanguineti, V. (2002). Ankle stiffness alone cannot stabilize upright standing. Journal of Neurophysiology, 88, 2157–2162. Moritani, T., & Muro, M. (1987). Motor unit activity and surface electromyogram power spectrum during increasing force of contraction. European Journal of Applied Physiology, 56, 260–265. Morrison, S., Hong, S. L., & Newell, K. M. (2007). Inverse relations in the patterns of muscle and center of pressure dynamics during standing still and movement postures. Experimental Brain Research, 181, 347–358. Negahban, H., Salavati, M., Mazaheri, M., Sanjari, M. H., Hadian, M. R., & Parnianpour, M. (2010). Non-linear dynamical features of center of pressure extracted by recurrence quantification analysis in people with unilateral anterior cruciate injury. Gait and Posture, 31, 450–455. Newell, K. M. (1998). Degrees of freedom and the development of postural center of pressure profiles. In K. M. Newell & P. C. M. Molenaar (Eds.), Applications of nonlinear dynamics to development process modeling. New Jersey: Lawrence Erlbaum Associates. Newell, K. M., & Vaillancourt, D. E. (2001). Dimensional change in motor learning. Human Movement Science, 20, 695–715. Palmieri, R. M., Ingersoll, C. D., Cordova, M. L., Kinzey, S. J., Stone, M. B., & Krause, B. A. (2003). The effect of a simulated knee joint effusion on postural control in healthy subjects. Archives of Physical Medicine and Rehabilitation, 84, 1076–1079. Pincus, S. M. (1991). Approximate entropy as a measure of system complexity. Proceedings of the National Academic of Science of the United States of America, 88, 2297–2301. Richman, J. S., & Moorman, J. R. (2000). Physiological time-series analysis using approximate and sample entropy. American Journal of physiology: Heart and Circulatory Physiology, 278, 2039–2049. Rodrick, D., & Karwowski, W. (2006). Nonlinear dynamical behavior of surface electromyographical signals of biceps muscle under two simulated static work postures. Nonlinear Dynamics, Psychology, and Life Science, 10, 21–35. Sasagawa, S., Ushiyama, J., Masani, K., Kouzaki, M., & Kanehisa, H. (2009). Balance control under different passive contributions of the ankle extensors: Quiet standing on inclined surface. Experimental Brain Research, 196, 537–544. Seigle, B., Ramdani, S., & Bernard, P. L. (2009). Dynamical structure of center of pressure fluctuations in elderly people. Gait and Posture, 30, 106–109. Sinkjaer, T., & Arendt-Nielsen, L. (1991). Knee stability and muscle coordination in patients with anterior cruciate ligament injuries: An electromyographic approach. Journal of Electromyography and Kinesiology, 1, 209–217. Stins, J. F., Ledebt, A., Emck, C., van Dokkum, E. H., & Beek, P. J. (2009). Patterns of postural sway in high anxious children. Behavioral and Brain Function, 5, 42. Stins, J. F., Michielsen, M. E., Roerdink, M., & Beek, P. J. (2009). Sway regularity reflects attentional involvement in postural control: Effects of expertise, vision and cognition. Gait and Posture, 30, 106–109. Thurner, S., Mittermaier, C., & Ehrenberger, K. (2002). Change of complexity patterns in human posture during aging. Audiology and Neurootology, 7, 240–248. Vaillancourt, D. E., & Newell, K. M. (2002). Changing complexity in human behavior and physiology through aging and disease. Neurobiology of Aging, 23, 1–11. Winter, D. A., Patla, A. E., & Frank, J. S. (1990). Assessment of balance control in humans. Medical Progress through Technology, 16, 31–51. Winter, D. A., Patla, A. E., Prince, F., Ishac, M., & Gielo-Perczak, K. (1998). Stiffness control of balance in quiet standing. Journal of Neurophysiology, 80, 1211–1221. Winter, D. A., Patla, A. E., Rietdyk, S., & Ishac, M. G. (2001). Ankle muscle stiffness in the control of balance during quiet standing. Journal of Neurophysiology, 85, 2630–2633. Xie, H. B., Guo, J. Y., & Zheng, Y. P. (2010). Fuzzy approximate entropy analysis of chaotic and natural complex systems: Detecting muscle fatigue using electromyography signals. Annals of Biomedical Engineering, 38, 1483–1496. Zhou, P., & Rymer, W. Z. (2004). Factors governing the form of the relation between muscle force and the EMG: A simulation study. Journal of Neurophysiology, 92, 2878–2886.