In this study, two force platforms were synchronized to investigate the coordination of the right and left foot center of pressure (COPR and COPL) and its relation ...
Motor Control, 2012, 16, 425-443 © 2012 Human Kinetics, Inc.
Coordination Patterns of Foot Dynamics in the Control of Upright Standing Zheng Wang, Kimberlee Jordan, and Karl M. Newell In this study, two force platforms were synchronized to investigate the coordination of the right and left foot center of pressure (COPR and COPL) and its relation to the COPNET in the control of 5 upright postures with and without visual information. The results revealed that the standard deviation (SD) of COPL, COPR, and COPNET progressively increased in the more challenging staggered and tandem stances, respectively, and to a lesser degree with the absence of visual information. Circular analysis of the relative phase of COPL and COPR revealed that the coupling pattern and variability were dependent on postural stances and the availability of vision. A negative correlation between the variability of the relative phase of the two feet COPs and the SD of the COPNET in the anterior-posterior (AP) direction was evident most strongly in the no vision conditions. Thus, the asymmetry of the mechanical constraints on the feet as a function of stance organize the coordination patterns of the feet COPs while the degree of adaptive variation between the feet COPs is dependent on both the mechanical constraints and the availability of vision. Keywords: Postural control, coordination, center of pressure, limb load
To sustain standing posture the neuromuscular control system must support the body against gravity, stabilize the supporting elements of the body when other elements are moved, and ensure that the body is balanced through the vertical projection of the center of gravity lying within the base of support (Massion, 1994; Rothwell, 1993; Winter, 1995). The primary variable that has been analyzed to investigate the control of human upright standing is that of the net body center of pressure (COPNET) recorded from a single force platform (e.g., Goldie, Bach, & Evans, 1989; Murray, Seirewg & Sepic, 1975). From this experimental set-up the control strategies of human upright standing posture have been inferred as ankle and hip control in inverted pendulum models (McCollum & Leen, 1989; Mergner, Maurer, & Peterka, 2003; Winter, 1995). The motion of the COPNET is a controlled variable determined by the collective average of the center of pressure of each foot (COPL and COPR), the body weight loading of each foot and mediated by the availability of the visual, vestibular and somatosensory feedback (Winter, Prince, Stergiou, & Powell, 1993; Riccio, The authors are with the Department of Kinesiology, The Pennsylvania State University, University Park, PA. 425
426 Wang, Jordan and Newell
1993; Horak, Nashner, & Diener, 1990). Mizrahi, Groswasser, Susak and ReiderGroswasser (1989) recognized the limitation of using the single force platform for the measurement and modeling of postural motion. They conducted a series of experiments on the bilateral reactive force (i.e., resultant force vector of Fx and Fy) on healthy adults, stroke and craniocerebral injured patients and found that even the healthy subjects displayed different reactive force patterns during side-by-side stance indicating multiple levels of stabilizing activities on each foot. Winter, Prince, Frank, Powell, and Zabjek (1996) required the participants to stand with different erect postures by loading their body weight evenly on the two feet to isolate the contribution of the ankle and hip COP motions to the COPNET and found that different combinations of the ankle and hip mechanisms were evident with different postural positions. The availability of visual information typically reduces the variability of postural COPNET motion as measured by the traditional linear techniques such as the traveling distance, area and the SD of the COPNET trajectory (Paulus, Straube, & Brandt, 1984; Newell, Slobounov, Slobounova, & Molenaar, 1997). Recent experiments using diffusion analysis and other nonlinear methods have shown that the stochastic properties of the COPNET are reduced with eyes open (Collins & De Luca, 1995; Newell, 1998). In a two-force platform protocol, Blaszczyk, Prince, Raiche, and Hébert (2000) found that limb load asymmetry was stronger in older adults than younger adults, and this effect was magnified in the absence of visual information. Similarly, Kinsella-Shaw, Harrison, and Turvey (2011) have shown that illumination of the visual environment influences (as reflected in cross-recurrence quantification analysis) the interleg dynamics of older adults more than young adults (see also Kinsella-Shaw, Harrison, Colon-Semenza, & Turvey, 2006). However, the effect of vision on foot coordination dynamics with different postural stances has not been fully investigated. The first aim of the current experiment was to investigate the coordination patterns of the foot center of pressures in the control of different upright postural stances and the availability of vision. We identify the coupling dynamics between the two feet COPs (COPL and COPR) by analyzing relative phase. It was hypothesized that the different mechanical constraints of the postural stances (Kirby, Price, & MacLeod, 1987) and visual information (eyes open—EO/ eyes closed— EC—Blaszczrk et al., 2000) would lead to asymmetrical weight loading on each foot, which would induce different coupling patterns (as indexed by the mean of the relative phase) and variability (as indexed by the circular standard deviation of the relative phase) of the coordination of the COPL and COPR. A second aim was to investigate the relation between the coupling dynamics of the two feet COPs to the COPNET. COPL and COPR are two peripheral degrees of freedom that have been traditionally summarized as COPNET. It was anticipated that the adaptive compensation of the individual COPL and COPR would induce stabilization of the performance variable—COPNET (Latash, Scholz, & Schöner, 2007). From this perspective, we tested the hypothesis that the increased variability in the relative phase of COPL and COPR is directly related to the decreased amount
Coordination of Foot Dynamics 427
of variability of COPNET and this relation would be strengthened without vision (Ranganathan & Newell, 2008; Hu, Loncharich, & Newell, 2011) and the increased challenge of less stable postural stances.
Methods Subjects Twelve subjects (6 female, 6 male), between the age of 25 and 35 years with no diagnosed musculoskeletal and balance pathology, participated in this study. All subjects provided informed consent to this project that was approved by the Institutional Review Board (IRB) of Pennsylvania State University.
Apparatus The foot kinetic data were collected by two adjacent AMTI force platforms (Advanced Mechanical Technology Inc., OR6–5-1000) at a sample rate of 100 Hz. The platforms were calibrated and synchronized for data collection. In addition, the origin was set in between the two force platforms using a global coordination reference system. System noise was also estimated by measurement of platform variability from a rigid mannequin before actual data collection.
Tasks and Procedures Five upright postures were investigated in this study (Figure 1). They were: feet side-by-side standing about hip width apart, staggered position (with either right or left foot forward) and tandem position (with either right or left foot forward and one foot directly in front of the other). Each posture was tested twice under both EO and EC conditions. Under the EO condition, participants were instructed to focus on a target, about human eye height level, 2 m in front of the force platforms. In the EC condition they were asked to keep their eyes closed and their head oriented straight ahead. There were 2 trials at each of the 5 postures × 2 visual conditions. Each trial was 60 s in duration. The conditions were randomly assigned to subjects and the 2 trials at each condition were blocked. A piece of paper was spread flat on each force platform (one piece for each platform) and an outline of the subject’s feet was traced for each stance allowing foot position to be accurately repeated for each trial of a stance. For each postural stance the participant was asked to stand with their feet at a comfortable distance apart in the stance required. The mean ± SD of foot distance in both directions (AP and ML) for each stance are reported in the caption of Figure 1. The participants were instructed to stand with their arms relaxed by their sides and keep their knees straight during the testing. Data recording was initiated 5 s after the participants could balance themselves in the respective posture on the force platforms. A 15s break was given between each trial of a condition and there was a 1 min break between conditions.
428 Wang, Jordan and Newell
Figure 1 — Schematic of the feet positions for the 5 postural stances: SS—side by side, SR -staggered right foot forward, SL—staggered left foot forward, TR—tandem right foot forward, TL—tandem left foot forward. Feet distances are measured by the length between the forefoot heel and the rear foot’s big toe in the AP direction and the first metatarsophalangeal articulations between left and right foot in the ML direction (landmarks: dash line; distances: solid line). These distances were: SS: ML 19.04 ± 5.54(cm); SR: ML 10.88 ± 5.41(cm), AP 8.20 ± 4.31(cm); SL: ML 10.35 ± 4.51 (cm), AP 7.62 ± 2.97(cm); TR: AP 10.49 ± 5.75(cm); TL: AP 9.69 ± 5.08(cm).
Data Analysis Center of Pressure
The raw force (N) and moment (N.m) data were low-pass filtered by a 4th order double pass Butterworth filter with 10 Hz cut-off frequency. The COPs in the AP and ML directions were calculated independently for each limb (COPL and COPR). The COPNET was derived from the COPL, COPR and ground reactions forces (Fz) under each limb based on the equation by Winter et al. (1993):
COPNET (t ) = COPL (t )
Fz L (t ) Fz R (t ) (1) + COPR (t ) Fz L (t ) + Fz R (t ) Fz L (t ) + Fz R (t )
where COPL(t) and COPR (t) are the center of pressure under the left and right foot, respectively, FzL(t) is the vertical ground reaction force under the left foot and FzR(t) is the vertical ground reaction force under the right foot. To make the data suitable for analysis of the relative phase the mean of the COPs was removed from the respective data.
COP Variability This analysis focused on the SD of the COPL, COPR and COPNET trajectories as a function of postural stance and the availability of vision.
Coordination of Foot Dynamics 429
Ground Reaction Force The ground reaction force ratio of the more loaded foot to the less loaded foot was calculated. According to this definition, the ratio will have a value greater or equal to one (when the body weight is evenly loaded on the two feet, the ratio will be one), which does not satisfy a normal distribution.
Coupling Between COPL and COPR Hilbert transformed relative phase. The coupling effects between COPL and COPR, in both ML and AP directions, were quantified by the relative phase. There are problems in deriving relative phase from x vs. x (position vs. velocity) state space for nonharmonic oscillations (Fuchs, Jirsa, Haken, & Kelso, 1996; Rosenblum & Kurths, 1998), thus the relative phase of COPL and COPR was obtained via the Hilbert transform as follows:
(2)
Where SL (t) and SR (t) are the real parts of Hilbert transformed COPL and COPR (which are the original time series data), respectively, S L (t ) and S R (t ) are the imaginary parts of COPL and COPR after Hilbert transform (which are the actual Hilbert transformed data), while correspondingly, φL (t) – φR (t) represents the relative phase of the two COPs. Since the relative phase is expressed in angle, each relative phase φL (t) – φR (t) time series has been plotted on a circular histogram with 120 bins (3° per bin) for the purpose of showing specific patterns of the phase angular distribution (Figure 2). Test for Directedness. To further process the circular data, the directedness of the angular distribution was measured first by Rao’s (1976) spacing test which is a sensitive test for circular data distributed by unimodal, nonunimodal or multimodal patterns. The null hypothesis of Rao’s test was that the sampled circular data are
Figure 2 — Representative relative phase circular histograms of the directed (or onesidedness) and directionless distribution patterns with 120 bins (3° in each bin). Left panel: directed distribution; Right panel: directionless distribution. Directionless distribution is identified by the Rao’s spacing test.
430 Wang, Jordan and Newell
uniformly distributed (directionless). The null hypothesis could be rejected when p value less than α = 0.05 and, a conclusion can be drawn as the relative phase of COPL and COPR φL (t) – φR (t) is distributed directed or one-sided. Figure 2 depicts the representative relative phase φL (t) – φR (t) circular histograms of the directed (left panel) and directionless (right panel) angular distribution patterns. Circular Statistics. Circular mean vector and variability were calculated as
follows:
m=
1 (e1 + e2 + ... + en ) (3) n (4)
(5)
where e1, e2, . . . en are the relative phase circular data, m is the mean vector of the angular samples (with direction and value), r is a scalar which only represents the value of m, s0 is the standard deviation of the relative phase (Mardia,1975). Following Equations 3–5, if r approaches to zero, which could happen on directionless distributed data, the mean vector m tends to be unreliable both in direction and value. The standard deviation s0 could also have a wide range of unreliable values (when r leans to zero, s0 diverges to ∞). Thus, the directionless distributed relative phase time series (11 trials out of the total 480 trials) were excluded from the circular statistic analysis. In addition, note that the circular mean r can also approach to zero under the axial (the circular distribution with two equal and opposite modes), trimodal (the case that modes are 120° apart from each other) and quadri-modal (the modes are well 90° apart from each other) distributions. However, our datasets did not contain these particular patterns. As a result, the circular mean r and standard deviation s0 calculated by the circular statistics are reliable for these directed angular data. Oriana 2.02e (Kovach Computing Services, Anglesey, Wales) software was used for the circular statistic analyses. Coupling Effects. It was interpreted that the mean of a circular relative phase φL
(t) – φR (t) distribution around 0° represents an in phase coupling between COPL and COPR whereas a distribution around 180° represents an antiphase coordination. A small s0 indicates a tight or strengthened coupling whereas a large value reveals a more flexible coupling dynamics.
Statistical Analysis COP Variability. The independent factors were postural stance and vision. The dependent variables were the SDs of COPL, COPR and COPNET in the AP and ML directions. For each dependent variable the average across the two trials was calculated. A 5 (posture) × 2 (vision) fixed-effect repeated-measures ANOVA for each COP was performed in these two directions. The Bonferroni post hoc test was used to determine the differences on all pairs of levels of independent variables. Only effects that were statistically significant at p < .05 are reported. Where Mauchly’s test indicated violation of sphericity, the Greenhouse-Geisser estimate was used to provide a conservative test of ANOVA main and interaction effects.
Coordination of Foot Dynamics 431
Ground Reaction Force. The independent factors were postural stance and the availability of vision and the dependent variable was the ground reaction force ratio of the two feet. Due to the nonnormally distributed property of the data, Friedman’s nonparametric test (also called the repeated-measures ANOVA for ranks, comparing three or more paired groups or levels) was conducted on the posture main effect. Post hoc comparison was tested by Wilcoxon single-ranked test on the different posture stance combinations. Bonferroni-adjusted significance level was 0.005 due to the 10 comparisons between postures. Thus, only effects that were statistically significant at p < .005 are reported. The vision main effect was analyzed by Wilcoxon single-ranked test because this factor has only two levels. Coupling Between COPL and COPR. The independent factors were posture and vision. The dependent variables were the circular mean r and standard deviation s0 of φL (t) – φR (t) in both AP and ML directions. The circular mean r of the φL (t) – φR (t) reveals different directional coupling patterns by which the participants used to maintain their balance whereas the circular s0 reveals the strength of the coupling. Due to the circular property of r, Harrison-Kanji test (equivalent to the two-way ANOVA for the normally distributed linear data) was conducted to test the posture and vision main effects. The dependent variable circular standard deviation s0 is linearly distributed, so a 5 (posture) × 2(vision) fixed-effects repeated-measures ANOVA was conducted. Correlation Between the Circular s0 of Relative Phase and the SD of the COPNET. The intraparticipant Pearson correlation coefficients were calculated
as a function of posture and vision.
Results COP Variability Figure 3 shows three examples of one representative subject’s trajectory variability for COPL, COPNET and COPR for a single trial in three different postures (side-byside, staggered and tandem with the left foot forward) for the EO condition. The data show different patterns for the COPs as a function of posture and vision that are now analyzed in more detail. ML Direction. Figure 4 shows the SD of COPL, COPR and COPNET in the ML direction as a function of posture and vision. There were significant main effects for posture (F4,44 =109.109, p< .001), vision (F1,11=56.641, p< .001) and COP (F2, 22 =89.838, p< .001) and interaction effects for posture × vision (F1.913, 21.040 =35.397, p< .001), posture × COP (F2.486, 27.349=16.381, p< .001) and posture × vision × COP (F3.193, 35.128=4.349, p= .009). Tandem stances (TR M= .732 cm/s; TL M= .714 cm/s) displayed the largest COP variability followed by the staggered stances (SR M= .339 cm/s; SL M=. 352 cm/s) and the side-by-side position (M= .128 cm/s). The unavailability of vision (M= .535 cm/s) induced larger COP variation as compared with the eyes open conditions (M= .370 cm/s). The COPNET SD (M= .575 cm/s) was significantly larger than that of the COPL (M= .385 cm/s) and COPR (M= .399 cm/s). Post-hoc analysis of the COP trajectory revealed that vision effect was prominent only for staggered left foot forward and tandem positions. COPNET
432 Wang, Jordan and Newell
Figure 3 — COP trajectories of one representative subject (No.6) under individual feet (COPL and COPR) and COPNET (as could be recorded from a single force platform) in three stances with eyes open. (A) Side by side: COPL and COPR move in phase in the AP direction, antiphase in the ML direction; (B) Staggered with left foot forward: COPL and COPR trajectories move along the diagonal line from upper left to lower right; (C) Tandem with left foot forward: COPL and COPR trajectories have a “half moon” shape.
variation was larger than that of the COPL and COPR in side-by-side and staggered stances both with and without the availability of vision. In staggered stances, the front foot displayed the smallest COP SD followed by the rear foot and then the COPNET in both visual conditions. AP Direction. Figure 4 illustrates that the main effect of all three factors was significant (posture F1.982, 21.800=9.525, p< .001; vision F1, 11=23.676, p< .001; COP F2, 22= 15.625, p< .001). There was a significant interaction for posture × vision (F4, 44= 7.203, p< .001) and posture × COP (F3.932, 43.255= 6.246, p< .001). The COP variation in tandem stances (TR M= .910 cm/s; TL M=. 875 cm/s) was significantly larger than that of the side-by-side (M= .546 cm/s) and staggered right foot forward positions (M= .658 cm/s). The eyes closed conditions (M= .836 cm/s) induced larger COP variation as compared with the vision conditions (M= .640 cm/s). Opposite to the ML direction, the COPnet SD (M= .588 cm/s) was significantly
Figure 4 — (A) Mean COP (COPL, COPR and COPNET) SD of the ML motion as a function of postural stance and vision (EO/ EC). (B) Mean COP (COPL, COPR and COPNET) SD of the AP motion as a function of postural stance and vision (EO/ EC). Data shown are in the mean and standard error format. 433
434 Wang, Jordan and Newell
smaller than that of the COPL (M= .799 cm/s) and COPR (M= .828 cm/s) in the AP direction. The vision effect was prominent in all positions except the staggered right foot forward stance. Post-hoc analysis showed that the rear foot COP variation displayed the largest value followed by the front foot and COPnet in the staggered stances. The SD of the COPnet was significantly smaller than that of the front and rear foot. In the side-by-side stance, COPnet variation was smaller in comparison with that of the COPR.
Ground Reaction Force Figure 5 shows the ground reaction force ratio as a function of postural stance and vision. There was a significant difference in the ground reaction force ratio depending on the postural stance, χ2 (4) = 43.433, p = .000. The median ground reaction force ratio for the five postures were, side-by-side: 1.0245 (.9533–1.0992); staggered right foot forward: 2.3906 (1.4584–2.6400); staggered left foot forward: 2.0470 (1.8029–2.8177); tandem right foot forward: 2.3002 (1.8891–3.0953) and tandem left foot forward: 2.1313 (1.6439–3.2557). There were significant reductions in the ratio between the side-by-side stance to the other postures (side-by-side vs. staggered right foot forward: Z = -4.257, p = .000; side-by-side vs. staggered left foot forward: Z = -4.114, p = .000; side-by-side vs. tandem right foot forward: Z = -4.286, p = .000 and side-by-side vs. tandem left foot forward: Z= -4.286, p =
Figure 5 — Mean Fz ratio of the more loaded and less loaded feet as a function of postural stance and vision (Black bar: EO/ White bar: EC). Data shown are in the mean and standard error format.
Coordination of Foot Dynamics 435
.000). The vision effect was conducted by Wilcoxon Single-Rank Test indicating no significant difference between EO and EC conditions (Z = -1.185, p = .236).
Coupling Between COPL and COPR Test of Directedness. The distribution pattern of the relative phase φL (t) – φR
(t) for each trial (480 trials in total: 5 posture × 2 vision × 2 direction × 2 attempts/ subject × 12 subjects) was classified using Rao’s (1976) spacing test as either directed or directionless. Overall, 11 trials were detected as directionless. Circular Analysis. Table 1 and Figure 6 show the mean and the 95% confidence
interval of the COP L and COPR coupling across different postural × vision conditions in both AP and ML directions. The mean angles were consistently distributed around 0° in the AP direction and around 180° in the ML direction in side-by-side standing. However, in the staggered stances, the distribution of these mean angles switched to 0° in the ML direction and stayed the same as that in the AP direction. In the tandem postures, the mean angles were in-phase in the ML direction, but they were even more spread out in the AP direction revealing a wider range of coupling strategies between COPL and COPR in this postural stance. Vision did not show a significant main effect on the circular mean r (AP: χ2 (8)= 4.460, p = .813; ML: χ2 (2)= 2.854, p = .240).
Table 1 Mean Angle and 95% Confidence Interval of the Directed Data Trials AP EO
EC
SS
0.327° (2.981°-357.753°)
0.919° (357.212°-4.627°)
SR
347.797° (342.1°-353.495°)
354.544° (350.627°-358.46°)
SL
12.926° (358.554°-27.298°)
10.763° (4.192°-17.334°)
TR
305.124° (257.376°-352.871°)*
320.484° (285.334°-355.633°)
TL
162.201° (87.495°-236.906°)*
40.919° (350.177°-91.66°)* ML
EO
EC
SS
182.462°(175.214°-189.71°)
175.737° (166.346°-185.129°)
SR
358.49° (341.475°-15.506°)
352.544° (338.009°-7.099°)
SL
1.743° (357.65°-5.836°)
5.02° (354.906°-15.134°)
TR
356.75° (353.891°-359.61°)
356.133° (354.019°-358.248°)
TL
358.255°(354.427°-2.082°)
359.062° (354.232°-3.892°)
*Confidence interval may not be reliable due to the spread out of the mean vectors in tandem stances.
Figure 6 — Circular mean angle of all subjects’ directed data within different postural stances. Each triangular marker (Black dot: EO/ White dot: EC) represents the mean angle of one relative phase time series. Top panel: side-by-side stance; Middle panel: staggered left foot forward; Bottom panel: tandem left foot forward. 436
Coordination of Foot Dynamics 437
Figure 7 shows the circular standard deviation s0 of the COPL and COPR relative phase in both AP and ML directions as a function of posture and vision. The main effect of posture was significant for both directions (AP: F (4, 40) = 21.674, p = .000 and ML: F (2.266, 22.662) = 3.826, p < .05). Post hoc analyses showed that, in general, the variability of the COPL and COPR coupling for the tandem stances was significantly higher than that for the side-by-side and staggered stances in the AP direction. However, the variability of the coupling for the tandem right foot forward stance was significantly lower than that of the side-by-side stance in the ML direction. The main effect of vision was significant only in the AP direction, F (1, 10) = 5.792, p < .05, that the circular s0 of the φL (t) – φR (t) in EO condition was significantly higher in comparison with that in EC indicating a more flexible or less coupled strength with the availability of the visual information. Table 2 shows the Pearson correlation coefficients between the circular standard deviation s0 of the COPL and COPR relative phase and the SD of the COPNET as a function of posture, direction of motion (AP-AP and ML-ML) and the availability of vision. This relation was negative in all but one condition with the effect stronger in the AP-AP no vision postural conditions.
Discussion The experiment was set-up to investigate the coordination patterns of the COP dynamics of the two feet in the control of several erect postural stances under the availability or nonavailability of vision. The two-force platform protocol not only affords a measure of the center of pressure of each foot in addition to that of the COPNET but also provides the basis to examine the relation between the two feet coordination dynamics and the variability of the COPNET. The motion of the COPNET is a controlled variable determined by the center of pressure and the body weight loading of each foot according to Equation 1 (Winter et al., 1993). Our results show that the coupling of the two feet COPs and the SD of the COPNET are mediated by different postural stances and the availability of visual information. There were different foot center of pressure coordination patterns for different postural stances (Table 1 and Figure 6). According to Kinsella-Shaw et al. (2011), interleg coordination was prevalent between the time series that were in the same direction. In our study, analysis of the foot COP coordination dynamics was only focused on the AP and ML directions. The findings showed that the side-by-side stance was dominated by a coordination pattern of the foot dynamics that had inphase and antiphase motion in the AP and ML directions, respectively. This finding parallels that of Winter (1995) who showed that COPNET in the AP direction is roughly the average of COPL and COPR as the body weight carried by each foot is about 50% while the COPs of the two individual feet in the ML direction oscillate in opposite directions. In contrast to the side-by-side stance, the staggered stances led to a greater range of the mean relative phase in the AP direction while the ML pattern was in-phase. The tandem stances showed multiple coupling phase relations in the AP direction while the ML direction showed a consistent in-phase pattern. There are several factors that could contribute to the different foot coordination patterns in the different postural stances.
Figure 7 — S0 of the relative phase time series as a function of postural stance and vision (Black bar: EO/ White bar: EC) in both AP and ML directions. Error bar represents one standard error.
438
Coordination of Foot Dynamics 439
Table 2 Pearson Correlation Coefficients between s0 of Relative Phase and the SD of COPNET as a Function of Posture, Direction of Motion and the Availability of Vision. EO AP-AP
ML-ML
SS
-0.437* (p = 0.033,df = 24)
0.221 (p = 0.322,df = 22)
SR
-0.230 (p = 0.280,df = 24)
-0.473* (p = 0.023,df = 23)
SL
-0.190 (p = 0.374,df = 24)
-0.037 (p = 0.867,df = 23)
TR
-0.347 (p = 0.105,df = 23)
-0.207 (p = 0.332,df = 24)
TL
-0.012 (p = 0.958,df = 21)
-0.182 (p=0.395,df=24) EC
AP-AP
ML-ML
SS
-0.522* (p = 0.009,df = 24)
-0.181 (p = 0.398,df = 24)
SR
-0.631* (p = 0.001,df = 24)
-0.331 (p = 0.114,df = 24)
SL
-0.414* (p = 0.044,df = 24)
-0.020 (p = 0.927,df = 24)
TR
-0.370 (p = 0.090,df = 22)
-0.387 (p = 0.062,df = 24)
TL
-0.479* (p = 0.021,df = 23)
-0.424* (p = 0.039,df = 24)
•AP-AP: s0 of relative phase (AP) vs. SD of COPNET (AP) •ML-ML: s0 of relative phase (ML) vs. SD of COPNET (ML)
Firstly, the more challenged postural stances of the staggered and tandem foot positions led to an asymmetrical weight loading on each foot and a reduction in the stability of the base of support, especially in the frontal plane (Kirby et al., 1987). The Fz ratio of the side-by-side stance was significantly lower than that of the other postures due to the difference of the spatial orientation of the feet in the staggered and tandem stances. The one foot leading the other postural configurations are commonly used in everyday life such as initiating a step into a gait cycle, or preventing a fall, stamping on a bug and so on (Peters, 1988). During these activities, people intentionally load more body weight on their rear foot and less on their front foot so that the rear foot can stabilize the upper body and the front foot could search, explore the base of support boundaries or manipulate the foot motion. Secondly, feet coordination dynamics could also be mediated by the neuromuscular constraints of the lower limb. In the current study, the challenge of the stance was reflected by the gradually increased spontaneous body motion in the ML direction that was especially pronounced in the tandem stances (Figure 4). The AP motion of the COPNET is steadily “locked” by the front foot’s plantarflexion and the rear foot’s dorsiflexion whereas the ML motion of the COPNET was driven by the simultaneous contraction of the ankle invertor/evertor from the side-by-side to the staggered stance. The hip abductor/adductor may not play a main role during the staggered stance because it may lead to a body weight reloading between the two feet and induce an upper body perturbation. The tandem foot position failed to show
440 Wang, Jordan and Newell
a consistent coupling pattern across the participants in the AP direction in contrast to the consistent in-phase pattern in the ML direction. Collectively, these findings show that, it is not just that the staggered and tandem postures lead to higher SD of the COPNET (Goldie et al., 1989; Kirby et al., 1987), but that there are different coordination dynamics associated with different two-footed standing postures. The postural stance showed a significant main effect on the flexibility of the coupling between COPL and COPR (Figure 7). In contrast to the classic absolute phase coordination (Haken, Kelso, & Bunz, 1985), the foot coupling dynamics displayed relative coordination which gives rise to all possible relative phase relations (i.e., a relatively large circular SD) even though a common coupling relationship is still present. In general, in the more challenged stances the coupling variability was decreased in the AP direction whereas it was increased in the ML direction. However, only the difference between the tandem and the side-by-side stances in the AP direction and that between the staggered right foot forward and the sideby-side in the ML direction reached statistical significance. Thus, in addition to the increased coordination strategies in the more challenged stances, the variability of the coupling tended to be less consistent in the AP direction. Nonetheless, the more consistent coupling dynamics was associated with the decreased coupling flexibility of the two feet in ML direction. The effect of visual information on the foot coupling dynamics and the Fz ratio was not as strong as the effect of postural stance but nevertheless there was an influence of visual information on postural control (Collins & De Luca, 1995; Woollacott, Debu, & Mowatt, 1987), in terms of the amount of variability of the respective COP. The effect of visual information on the SD was most prevalent in the staggered and particularly the tandem posture for COPL, COPR and COPNET in the ML direction (Figure 4). This finding is consistent with the hypothesis of different weighting of the multiple sensory feedback based on the orientation of the upper body during quiet stance (Maki, Holliday, & Topper, 1994). More specifically, the somatosensory feedback of the ankle joint, associated with the balance controlling in the AP direction, would be less weighted in the control of the lateral sway due to the increased weighting on the feedback from the plantar surface of the feet and the hip proprioception (Winter et al., 1996). Thus, the role of vision was stronger in the less stable postural coordination modes (Slobounov & Newell, 1994). In addition, the availability of vision led to more flexible coupling strategies (the s0 of the relative phase between COPL and COPR) in the AP direction of the foot dynamics (Figure 7A). Collins and De Luca (1995) proposed that visual input induces a muscle stiffness reduction across the joints of the lower limbs leading to lower levels of the stochastic activity of the open-loop control. Kinsella-Shaw et al. (2011) found that older adults exhibited greater interleg dynamical stability than that of the young participants due to the progressive decline of both feedforward and feedback postural control. Our results from the manipulation of visual information support this claim in that without visual information the flexibility of the foot coordination dynamics decreased, especially in the AP direction. However, visual information was not an influential factor in determining the postural foot coordination patterns (the mean of the relative phase between the two feet COPs) and the asymmetrical weight loadings during the quiet stances. When the challenge of the postural stance is increased, visual information might not be as an effective source of feedback as that of the feet cutaneous and somatosensory
Coordination of Foot Dynamics 441
inputs (Sasaki et al., 2002). The asymmetrical loading effect between EO and EC, which positively associates with the amplitude of the postural sway, is more pronounced for the older adults (Blaszczyk et al., 2000). There were negative Pearson correlation coefficients between the strength of dynamical coupling and the SD of the COPNET that varied as a function of posture, direction of the COP motion and the availability of vision. In general, the correlations in AP-AP EC conditions were the most statistically significant in terms of an inverse relation between the variability of the foot coordination dynamics and that of the COPNET. In other words, the coupling strength of the COPL and COPR increases without vision inducing a increase of the variability of the COPNET. A possible reason for the nonsignificant correlations shown in the other conditions is that the COPNET is determined by the center of pressure and the body weight loading of each foot COPL and COPR. The asymmetrical loading could also influence the SD of the COPNET, especially during the staggered and tandem stances. Further research needs to examine how these correlations between the coupling dynamics and the variance of the COPNET are mediated by the influence of different foot positions; or, with evenly distributed body weight and the gradually unevenly loaded body weight on each foot with the same postural stance. In summary, the two-platform protocol reveals features of postural coordination and control that cannot be discerned from the single force platform experimental protocol. Different two-footed postural stances have different qualitative foot coordination patterns of center of pressure dynamics (indexed as the mean and SD of the relative phase) and different quantitative relations between foot coupling variability and the SD of COPNET. This central finding on foot postural coordination patterns leads to different quantitative influences on the standard postural variables such as vision and direction of motion on the COPNET. It appears that the benefits of investigating the relations between the qualitative and quantitative properties of the dynamics in movement tasks (Kelso, 1995) hold also in consideration of postural tasks. The impact of this theoretical approach can be enhanced by the simultaneous consideration of the motions of the torso and limb dynamics (Hsu, Scholz, Schöner, Jeka, & Kiemel, 2007; Krishnamoorthy et al., 2005) with the individual foot center of pressure dynamics. Acknowledgments This project was supported in part by NSF 0848339.
References Blaszczyk, J.W., Prince, F., Raiche, M., & Hébert, R. (2000). Effect of ageing and vision on limb load asymmetry during quiet stance. Journal of Biomechanics, 33, 1243–1248. PubMed doi:10.1016/S0021-9290(00)00097-X Collins, J.J., & De Luca, C.J. (1995). The effects of visual input on open-loop and closed-loop postural control mechanisms. Experimental Brain Research, 103, 151–163. PubMed doi:10.1007/BF00241972 Fuchs, A., Jirsa, V.K., Haken, H., & Kelso, J.A.S. (1996). Extending the HKB model of coordinated movement to oscillators with different eigenfrequencies. Biological Cybernetics, 74, 21–30. PubMed doi:10.1007/BF00199134
442 Wang, Jordan and Newell
Goldie, P.A., Bach, T.M., & Evans, O.M. (1989). Force platform measures for evaluating postural control: Reliability and validity. Archives of Physical Medicine and Rehabilitation, 70, 510–517. PubMed Haken, H., Kelso, J.A., & Bunz, H. (1985). A theoretical model of phase transitions in human hand movements. Biological Cybernetics, 51, 347–356. PubMed doi:10.1007/ BF00336922 Horak, F.B., Nashner, L.M., & Diener, H.C. (1990). Postural strategies associated with somatosensory and vestibular loss. Experimental Brain Research, 82, 167–177. PubMed doi:10.1007/BF00230848 Hsu, W.L., Scholz, J.P., Scho˝ner, G., Jeka, J.J., & Kiemel, T. (2007). Control and estimation of posture during quiet stance depends on multijoint coordination. Journal of Neurophysiology, 97, 3024–3035. PubMed doi:10.1152/jn.01142.2006 Hu, X., Loncharich, M., & Newell, K.M. (2011). Visual information interacts with neuromuscular factors in the coordination of bimanual isometric force. Experimental Brain Research, 209, 129–138. PubMed doi:10.1007/s00221-010-2528-4 Kelso, J.A.S. (1995). Dynamic patterns: The self-organization of brain and behavior. Cambridge: MIT Press. Kinsella-Shaw, J.M., Harrison, S.J., & Turvey, M.T. (2011). Interleg coordination in quiet standing: influence of age and visual environment on noise and stability. Journal of Motor Behavior, 43, 285–294. PubMed doi:10.1080/00222895.2011.580389 Kinsella-Shaw, J.M., Harrison, S.J., Colon-Semenza, C., & Turvey, M.T. (2006). Effects of visual environment on quiet standing by young and old adults. Journal of Motor Behavior, 38, 251–264. PubMed doi:10.3200/JMBR.38.4.251-264 Krishnamoorthy, V., Yang, J.F., & Scholz, J.P. (2005). Joint coordination during quiet stance: effects of vision. Experimental Brain Research, 164, 1–17. PubMed doi:10.1007/ s00221-004-2205-6 Kirby, R.L., Price, N.A., & MacLeod, D.A. (1987). The influence of foot position on standing posture. Journal of Biomechanics, 20, 423–427. PubMed doi:10.1016/00219290(87)90049-2 Latash, M.L., Scholz, J.P., & Schöner, G. (2007). Toward a new theory of motor synergies. Motor Control, 11, 276–308. PubMed Maki, B.E., Holliday, P.J., & Topper, A.K. (1994). A prospective study of postural balance and risk of falling in an ambulatory and independent elderly population. Journal of Gerontology, 49, M72–M84. PubMed Mardia, K.V. (1975). Statistics of directional data. Journal of the Royal Statistical Society. Series C, Applied Statistics, 37, 349–393. Massion, J. (1994). Postural control system. Current Opinion in Neurobiology, 4, 877–887. PubMed doi:10.1016/0959-4388(94)90137-6 Mizrahi, J., Groswasser, Z., Susak, Z., & Reider-Groswasser, I. (1989). Standing posture of craniocerebral injured patients: bi-lateral reactive force patterns. Clinical Physics and Physiological Measurement, 10, 25–37. PubMed doi:10.1088/0143-0815/10/1/003 McCollum, G., & Leen, T.K. (1989). Form and exploration of mechanical stability in erect stance. Journal of Motor Behavior, 21, 225–244. PubMed Mergner, T., Maurer, C., & Peterka, R.J. (2003). A multisensory posture control model of human upright stance. Progress in Brain Research, 142, 189–201. PubMed doi:10.1016/ S0079-6123(03)42014-1 Murray, M.P., Seirewg, A.A., & Sepic, S.B. (1975). Normal postural stability and steadiness: Quantitative assessment. Journal of Bone and Joint Surgery, 57A, 510–516. PubMed Newell, K.M., Slobounov, S.M., Slobounova, B.S., & Molenaar, P.C.M. (1997). Short-term non-stationary and the development of postural control. Gait & Posture, 6, 56–62. doi:10.1016/S0966-6362(96)01103-4
Coordination of Foot Dynamics 443
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 developmental process modeling (pp. 63–84). Mahwah, NJ: Lawrence Erlbaum Associates. Paulus, W.M., Straube, A., & Brandt, T. (1984). Visual stabilization of posture: physiological stimulus characteristics and clinical aspects. Brain, 107, 1143–1163. PubMed doi:10.1093/brain/107.4.1143 Peters, M. (1988). Footedness: asymmetries in foot preference and skill and neuropsychological assessment of foot movement. Psychological Bulletin, 103, 179–192. PubMed doi:10.1037/0033-2909.103.2.179 Ranganathan, R., & Newell, K.M. (2008). Online feedback and the regulation of degrees of freedom in motor control. Human Movement Science, 27, 577–589. PubMed doi:10.1016/j.humov.2008.01.002 Rao, J.S. (1976). Some tests based on arc-lengths for the circle. Sankhya Series B, 33, 1–10. Riccio, G.E. (1993). Information in movement variability about the qualitative dynamics of posture and orientation. In K.M. Newell & D.M. Corcos (Eds.), Variability and motor control (pp. 317-357). Champaign, IL: Human Kinetics. Rosenblum, M., & Kurths, J. (1998). Analyzing synchronization phenomena from bivariate data by means of the Hilbert transform. In H. Kantz, J. Kurths, & G. Mayer-Kress (Eds.), Nonlinear analysis of physiological data (pp. 91–99). Berlin: Springer. Rothwell, J.C. (1993). Control of human voluntary movement. Verlag, New York: Springer. Sasaki, O., Usami, S-i., Gagey, P-M., Martinerie, J., Le Van Quyen, M., & Arranz, P. (2002). Role of visual input in nonlinear postural control system. Experimental Brain Research, 147, 1–7. PubMed doi:10.1007/s00221-002-1170-1 Slobounov, S., & Newell, K.M. (1994). Postural dynamics as a function of skill level and task constraints. Gait & Posture, 2, 85–93. doi:10.1016/0966-6362(94)90097-3 Winter, D. A., Prince, F., Frank, J. S., Powell, C., & Zabjik, K. F. (1996). Unified theory regarding A/P and M/L balance in quiet standing. Journal of Neurophysiology, 75, 2334-2343. Winter, D.A., Prince, F., Powell, C., & Zabjik, K.F. (1996). Unified theory regarding A/P and M/L balance in quiet standing. Journal of Neurophysiology, 75, 2334–2343. PubMed Winter, D.A. (1995). Human balance and posture control during standing and walking. Gait & Posture, 3, 193–214. doi:10.1016/0966-6362(96)82849-9 Winter, D.A., Prince, F., Stergiou, P., & Powell, C. (1993). Medial-lateral and anteriorposterior motor responses associated with centre of pressure changes in quiet standing. Neuroscience Research Communications, 12, 141–148. Woollacott, M., Debu, B., & Mowatt, M. (1987). Neuromuscular control of posture in the infant and child: is vision dominant? Journal of Motor Behavior, 19, 167–186. PubMed
