| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1School of Arts, Sciences and Humanities, University of Sao Paulo, Sao Paulo, Brazil; and 2Department of Kinesiology, University of Waterloo, Waterloo, Canada
Submitted 13 January 2006; accepted in final form 3 August 2007
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
; Patla et al. 1999). Vision and proprioceptive inputs provide information about undesirable landing spots in the pathway and the amount of foot displacement in each direction relative to the normal landing position necessary to avoid them. However, when more than one option exists for minimizing foot displacement, vision and proprioceptive inputs alone are not enough to help determine where to step. It has been shown that the selection of alternate foot placement is based on three determinants: minimum foot displacement (i.e., economy), stability, and maintenance of forward progression (Moraes and Patla 2006; Moraes et al. 2004; Patla et al. 1999). When only one option offers the minimum foot displacement from its normal landing spot, it is the preferred choice. However, often the amount of foot displacement is similar for more than one option. In this case, two additional rules apply. If the minimum foot displacement from its normal landing spot is the same for one adjustment in the frontal plane and for another in the plane of progression, the preference is to make the adjustment in the plane of progression (Patla et al. 1999). It has been hypothesized that preference for changes in the plane of progression would minimize deviations from the endpoint goal (i.e., maintenance of forward progression). If the minimum foot displacement is the same within the same plane of movement, the preferred choice is long (i.e., lengthening the step) in the plane of progression and medial (i.e., narrowing the step) in the frontal plane (Patla et al. 1999). Moraes et al. (2004) suggested that these preferences are stability related.
Moraes et al. (2004) recently validated the primacy of minimum foot displacement in determining the choice of alternate foot placement. In this study, the predicted minimum foot displacement for each combination of options (i.e., long, short, lateral, and medial) and obstacle position was calculated based on the average foot placement during normal walking. The dominant choice minimized foot displacement from its normal landing position for the majority of cases, therefore validating minimum foot displacement as one of the determinants used for alternate foot placement selection. Weerdesteyn et al. (2005) have also shown that young females minimized foot displacement when avoiding a planar obstacle when walking on a treadmill. More recently, Moraes and Patla (2006) have shown that the minimization of foot displacement is economy related. The other two determinants (i.e., stability and maintenance of forward progression) have not yet been addressed and, consequently, validated. This is the focus of this study.
During walking, stability maintenance is a dynamic process caused by continuous changes in the base of support (BOS) and center of mass (COM) location. Walking is a cyclical sequence of falling forward and recovering balance by properly placing the swing foot, which determines the future location of the center of pressure (Patla 2003; Redfern and Schumann 1994; Winter 1995). This suggests that COM location relative to the BOS is a good indicator of body balance during walking. This measure of stability has been extensively used in studies involving insects (see Ting et al. 1994) where it is appropriate to consider a static measure of stability because they usually have more than two legs on the ground. However, distance parameters alone may not be adequate to assess dynamic stability. During walking, the dynamics of the body play an important role, and this can be captured by combining COM position and velocity (Iqbal and Pai 2000; Pai and Patton 1997). More recently, Hof et al. (2005) have proposed a simple and elegant calculation that captures the contribution of COM velocity to dynamic stability, which generates similar predictions as proposed by Pai and Patton (1997). They used the current location of COM and its velocity to extrapolate COM position. This extrapolated COM position can be compared against the maximum reach of center of pressure to identify whether or not the system is stable. Therefore the measure of stability is evaluated by computing the distance between the extrapolated COM position and the limit of the BOS, which is called margin of dynamic stability (MDS). During single support, COM projection is outside the BOS in both anteroposterior (AP) and mediolateral (ML) directions. Because, at each heel contact, balance is re-established by proper foot placement, it is appropriate to measure dynamic stability during the double support phase. Thus in this study, COM projection relative to the BOS during the beginning of the double support phase is used to assess stability. Furthermore, Patla (2003) has shown that more than one step is needed to completely overcome the instability generated by different sources of perturbation. During the alternate foot placement task, balance may be compromised because there are substantial changes in foot placement location. Therefore to unequivocally validate the stability determinant, it is necessary to assess the MDS not only in the adaptive step, but also in the subsequent step.
The preference for changes in the plane of progression over changes in the frontal plane has been attributed to maintenance of forward progression (Patla et al. 1999). More recently, Moraes et al. (2004) proposed that when making changes in the frontal plane, people try to move the foot not only medially, but also forward to minimize deviation from the forward goal. Although this foot displacement partially supports the maintenance of the forward progression determinant, more global measures of body trajectory are needed. Maintenance of forward progression is best assessed by examining COM trajectory.
Experimentally, this study introduces a new manipulation through a forced condition that allows analysis of how movement in the nonpreferred direction is achieved. In the forced condition, participants are visually cued to select one foot placement option. Because there is a dominant response for each obstacle when allowed free choice, the forced condition fills the gaps with nondominant choices. Thus the purpose of this study is to validate both stability and forward progression determinants when participants select an alternate foot placement and when foot placement is cued. It is hypothesized that these two determinants will have a major influence on selection and execution of alternate foot placement.
|
|
METHODS |
|---|
|
Eight participants volunteered for this study (4 women and 4 men; age, 25.4 ± 4.7 yr; height, 1.75 ± 0.10 m; mass, 69.5 ± 6.7 kg). Participants did not report any neurological, muscular, or joint disorders that could affect their performance in this study. The Office of Research Ethics at the University of Waterloo approved the procedures used in this study.
Protocol
Participants were asked to walk on level ground at a self-selected pace on a pathway containing a force plate (AMTI, Boston, MA) and an embedded liquid crystal display (LCD) monitor (Samsung SyncMaster TFT 181T Black) (Fig. 1). A piece of Plexiglas was placed over the LCD monitor so that participants could step normally on it. The starting point was adjusted for each participant to ensure that the entire left foot landed on the force plate and the subsequent step by the right foot landed on the center of the screen. Participants were required to avoid stepping with the right foot on a virtual white planar obstacle that would be displayed in the LCD screen. The obstacle appeared at left heel contact (HC) on the force plate (vertical component >5 N), which provided one step for implementing the alternate foot placement. The pathway was covered with a black rubber carpet that had specific cuts to accommodate the force plate and the LCD monitor. The force plate was also covered with the same black rubber carpet. The monitor edge was also black. For the trials with no obstacle, the screen background was kept completely black. One obstacle labeled ML was designed to facilitate adjustments in the frontal plane; the other obstacle labeled AP produced foot placement in the sagittal plane. Obstacles were displayed in the middle of the screen. Participants performed the task under two conditions: free and forced. In the free condition, participants chose the alternate foot placement most appropriate for avoiding the planar obstacle. In the forced condition, a green arrow projected over the white planar obstacle indicated the direction in which the alternate foot placement was to be performed (Fig. 1). The white obstacle and the green arrow were displayed simultaneously at left HC on the force plate. For the AP obstacle, two forced conditions were used: long and short. For the ML obstacle, two forced conditions were used: medial and lateral. Therefore a total of six conditions was collected: AP free, ML free, long forced, short forced, lateral forced, and medial forced. Six trials were collected per condition (36 trials in all). To keep a probability of obstacle appearance equal to 20%, 144 walk-through (WT) trials were also collected. Trials were completely randomized.
|
Data analysis
Videotapes were used to identify the successful trials for all the adjustments in the forced condition. In addition, in the unsuccessful trials, the errors were classified as follows: wrong adjustment, stepped on the obstacle, and others. Wrong adjustments are the trials where one adjustment was requested and the participant made some other adjustment. Others are trials in which participants missed the force plate, terminated gait just before or after the obstacle, or simply did not see the obstacle at all. The percentage for each of these parameters was calculated based on the total number of trials across participants.
Marker coordinates were filtered using a fourth-order zero lag low-pass digital Butterworth filter with a cut-off frequency of 6 Hz. A model with 16 segments was used to calculate whole body COM position: feet, legs, thighs, arms, forearms, head, pelvis, and trunk modeled with four components (Winter 2005). Anthropometric parameters were obtained from Winter (2005). COM velocity was calculated as the first derivative of COM position (central difference procedure). HC was determined by visual inspection of the foot stick figure using the OptoFix software (Mishac Kinetics, Waterloo, Ontario, Canada). This software allowed moving the stick figure back and forth on a frame-by-frame basis. HC was assumed as the frame number just before the initiation of the foot roll-over movement. Estimation of HC based on the vertical component of the force plate (Fy > 5 N) was used as the gold standard to validate the visual inspection. Differences between these two methods were always within the range of two frames (i.e., 0.03 s).
Ankle markers (i.e., lateral malleolus) were defined as the limb endpoint and used to calculate the foot placement modification vector (Fig. 2). For each experimental trial, relative coordinates (RCs) were computed as the subtraction of the ankle coordinates of the trial (x or AP and z or ML) from the average coordinates of the WT trials at HC. The average values were obtained from 20 randomly selected WT trials. The RCs were used to calculate the foot placement modification vector magnitude and foot placement modification vector orientation (i.e., angle: 
). Vector orientation was used to define the adjustment made in the free condition. Classification included four directions of adjustment: lateral (0–45° and >315–360°), long (>45–135°), medial (>135–225°), and short (>225–315°). Percentage of adjustment in each direction was calculated relative to the total number of trials successfully performed by all the participants. The same vector analysis performed for the lower limb endpoint was performed for the COM at HC. The cosine of the difference between foot and COM foot placement modification vector angle [cos(Foot – COM)] was used to identify whether or not foot and COM movement vectors were oriented in the same direction.
|
|
).
Extrapolated COM location was calculated based on the work done by Hof et al. (2005). According to these authors, COM location can be extrapolated based on its actual velocity as follows
 | (1) |
M is the velocity of the center of mass, and 
0 is defined by Eq. 2
 | (2) |
Based on the XcoM, margin of dynamic stability (MDS) was calculated at HC as follows
 | (3) |
Figure 11 shows the MDS calculation. For step N (i.e., adaptive step), calculations were made relative to right foot, whereas for step N + 1, calculations were made relative to left foot. Because velocity polarity changes from step-to-step in the ML direction, Eq. 3 was rearranged for MDS calculation for the left foot (MDS = XcoM – Foot Edge). A positive value for MDS indicates that XcoM is located before the edge of the foot and therefore the system is dynamically stable and vice versa for negative values.
|
|
|
|
|
|
|
|
The following step parameters were calculated: step length (SL) and step width (SW). SL was defined as the difference of the x-coordinate between two consecutive heel contacts. SW was defined as the difference of the z-coordinate between two consecutive heel contacts. For the SW, negative values indicate the presence of a cross-over between right and left limbs. Percentage of cross-over was calculated as a proportion to the total number of trials for each condition.
Force plate data were used to calculate braking and propulsive impulses in three directions (AP, ML, and vertical). The transition between braking and propulsive impulse was defined by identifying the zero crossing point in the AP component of the ground reaction force (Fig. 6). Braking impulse was obtained by computing the area under the curve from HC to zero-crossing, whereas propulsive impulse was defined as the area from zero-crossing to toe-off (vertical component <5 N). Braking impulse corresponds to the deceleration period of the support phase, whereas the propulsive impulse corresponds to the acceleration phase of the support phase.
Statistical analyses
For the onset time of limb trajectory change, impulse, and maintenance of forward progression, one-way ANOVAs (condition) with repeated measures were carried out. For the predicted minimum foot displacement (PMFD), a one-way ANOVA (option) with repeated measures was performed for each obstacle position. For the remaining dependent variables, two-way ANOVAs (condition x step) with repeated measures in both factors were carried out. For each dependent variable, the mean value of the successful trials was calculated per participant and used in the statistical analyses. The medial free condition was not used for any statistical analyses because only four participants selected this adjustment. Means and SD of these four participants are shown in the graphs for illustrative purposes only. Because comparisons between long/short and lateral/medial were not of interest, analyses were divided into two groups: medial/lateral and long/short conditions. For the long/short conditions, analyses were split because one participant never freely chose to shorten the step for the AP obstacle. Therefore four ANOVAs were carried out: 1) conditions (long forced and short forced) x step; 2) conditions (long free and short free) x step; 3) conditions (long forced and long free) x step; and 4) conditions (short forced and short free) x step. 
value was set to 0.05. When main or interaction effects were found, least squares means post hoc was used to identify which treatments differed from one another.
|
|
RESULTS |
|---|
|
Short and medial/lateral options would result in the minimization of foot displacement for the AP and ML obstacles, respectively
For both AP and ML obstacles, main effects of option were observed for the predicted minimum foot displacement (AP: F3,21 = 4.38, P = 0.0152; ML: F3,21 = 222.97, P < 0.0001). Least squares means post hoc analyses are presented in Table 1.
|
Both long and lateral choices were preferred in >50% of the trials for the free condition (Fig. 3). For the AP obstacle, five participants showed a long preference, one participant exhibited a short preference, and two participants had no preference (i.e., number of long and short choices were the same). For the ML obstacle, seven participants exhibited a lateral preference and one participant showed a medial preference.
Participants were successful in performing the forced condition
For the forced condition, long (89.1%) and lateral (79.2%) adjustments showed the highest success rates for the AP and ML obstacles, respectively (Fig. 3). The success rates for short and medial adjustments in the forced condition were equal to 47.9 and 52.1%, respectively. Short forced adjustment resulted in the highest rate of wrong adjustments (41.7%), followed by the medial forced adjustment (29.2%). These two adjustments were not the preferred ones in the free condition. The wrong adjustment was dominantly long for the short forced condition (95.0%). The wrong adjustment in the medial forced condition was more distributed (lateral: 50.0%; long: 28.6%; short: 21.4%).
In addition, for 26.1% of the successful trials involving short forced adjustments, participants initially performed a forward foot movement followed by a backward foot movement. This indicated that the original planning was for a long adjustment (Fig. 4, aborted long adjustment), but because of the forced nature of the task participants needed to reverse foot trajectory to perform the task properly. For the participants who exhibited aborted long adjustments in the forced condition, 50.0% of them showed a long preference, 25.0% had a short preference, and 25.0% had no preference in the free choice condition (AP obstacle). The same aborted behavior was observed in the ML direction for the medial forced adjustment. In this case, 24.0% of the successful trials initially showed a lateral foot movement, indicating that the original planning involved lateral adjustment. All the participants who exhibited aborted lateral adjustments in the forced condition had a lateral preference in the free choice condition (ML obstacle).
Adjustment location had no effect on limb trajectory change latency
One-way ANOVAs (condition) with repeated measures for the onset time of limb trajectory change identified a main effect of condition only for the comparison involving long forced and free adjustments (F1,7 = 8.39, P = 0.0231). As Fig. 5 shows, long free adjustments started earlier than long forced adjustments.
Initial response to obstacle appearance was both specific to the final foot placement and generic
For the braking impulse, main effects were identified for all four one-way ANOVAs combining long and short adjustments (see Statistical analysis) in both AP and vertical directions (P values ranging from <0.0001 to 0.0419). Least squares means post hoc analyses identified that braking impulse increased for the long and short adjustments compared with WT in both AP and vertical directions (long: 11.6 and 8.4%; short: 19.6 and 27.3%, respectively; Fig. 6). For the propulsive impulse, main effects were identified for all four one-way ANOVAs in all directions (P values ranging from <0.0001 to 0.0023). The propulsive impulses increased in all three directions for long adjustments (AP: 82.2%; vertical: 25.7%; ML: 55.0%) and decreased for short adjustments (AP: –56.1%; vertical: –38.9%; ML: –71.4%) in comparison to WT (Fig. 6).
For the lateral and medial adjustments, main effects of condition were identified in all three directions for the braking and propulsive impulses (P values ranging from <0.0001 to 0.0146). Least squares means post hoc analyses identified that AP and vertical braking impulses increased for the lateral (13.5 and 15.1%, respectively) and medial (11.2 and 19.4%, respectively) adjustments compared with WT. Braking impulse also increased in the ML direction for the lateral adjustment (57.0%). For the propulsive impulse, there was an increase of 28.4% for the lateral adjustment in the AP direction and a decrease of 11.2% for the medial adjustment in the vertical direction in comparison to WT. ML propulsive impulse increased 97.8% for the lateral adjustment and decreased 149.9% for the medial adjustment.
Adjustment location had no effect on foot placement modification vector magnitude in step N
The two-way ANOVA including long and short forced conditions identified an interaction effect (F1,14 = 37.27, P < 0.0001). Least squares means post hoc analysis showed that the foot placement modification vector magnitude was the same in step N (long forced: 27.8 ± 4.8 cm; short forced: 24.3 ± 7.6 cm), but it was larger for the long forced adjustment (31.6 ± 6.9 cm) than for the short forced adjustment (18.8 ± 7.0 cm) in step N + 1. For the ANOVA including medial and lateral adjustments, an interaction effect (F2,21 = 23.15, P < 0.0001) was also found. Foot placement modification vector magnitude was equal among conditions in step N (lateral forced: 15.7 ± 5.1 cm; lateral free: 13.9 ± 3.3 cm; medial forced: 14.1 ± 6.8 cm), but it substantially increased in step N + 1 for the medial forced adjustment (lateral forced: 14.1 ± 3.6 cm; lateral free: 10.2 ± 1.9 cm; medial forced: 28.0 ± 11.1 cm).
Foot and COM point to the same direction for adjustments in the AP direction, except for short forced adjustment
For the long (forced and free) and short free adjustments, there was a good coupling between foot and COM in steps N and N + 1 (Fig. 7). This is not the case for the short forced adjustment in step N. In some of the trials, the foot moves backward to achieve the goal of the task, whereas the COM moves forward, creating a very unstable gait as shown in Fig. 8. ANOVAs 1 (long forced/short forced) and 4 (short forced/short free) for the cosine of the angle difference presented interaction effects (1: F1,14 = 7.12, P = 0.0184; 4: F1,12 = 6.44, P = 0.0261). The interaction effects for both analyses resulted from the increased angle difference for the short forced adjustment in step N (Fig. 10).
Foot and COM do not point to the same direction for adjustments in the ML direction
For the lateral (free and forced) and medial adjustments, the decoupling between foot and COM in step N is quite clear (Fig. 9). For the medial free condition, this is not clear. The two-way ANOVA (condition x step) with repeated measures revealed an interaction effect (F2,21 = 8.67, P = 0.0018). Least squares means post hoc analysis showed no significant difference between lateral forced and medial forced adjustments in step N, but there was a significant difference in step N + 1, where a tightening coupling was observed for the medial forced adjustment (Fig. 10).
Long and lateral adjustments are more stable than short and medial adjustments, respectively
MDS: LONG-SHORT ADJUSTMENTS. ANOVAs 1 and 4 identified interactions effects (1: F2,21 = 6.05, P = 0.0084; 4: F2,18 = 14.91, P = 0.0002) for MDS in the AP direction. Least squares means post hoc analysis for ANOVA 1 showed that MDS was more negative for the short forced adjustment than for the long forced adjustment and WT in step N (Fig. 11) . In step N + 1, no difference was found between long and short forced adjustments, but MDS was more negative for the short forced adjustment than for the WT. For ANOVA 4, post hoc analysis found that both short forced and free adjustments presented a more negative MDS than WT in step N. In addition, MDS was more negative for the short forced adjustment than for the short free adjustment. In step N + 1, no difference was observed between short forced and free adjustments, but both adjustments exhibited a more negative MDS compared with WT. ANOVAs 2 and 3 identified main effects of condition (1: F2,12 = 11.16, P = 0.0018; 4: F2,14 = 10.00, P = 0.0020) for MDS in the AP direction. For ANOVA 2, post hoc analysis found that both long and short free adjustments exhibited a MDS more negative than WT. The difference between long and short free adjustments just failed to achieve statistical significance (P = 0.0552). For ANOVA 3, post hoc analysis showed that both long forced and free adjustments exhibited a more negative MDS than WT.
MDS: LATERAL-MEDIAL ADJUSTMENTS. The two-way ANOVA (condition x step) with repeated measures identified an interaction effect (F3,28 = 35.05, P < 0.0001) in the ML direction. Post hoc analysis indicated that MDS was larger for the lateral adjustments (forced and free) than for the WT and medial forced adjustment in step N (Fig. 11). In addition, WT exhibited a larger MDS than medial forced adjustment in the same step. In step N + 1, no difference was observed between lateral adjustments (forced and free) and WT, but the medial forced adjustment exhibited a smaller MDS than all other three conditions.
STEP PARAMETERS. All four ANOVAs revealed interaction effects for the dependent variable step length (1: F2,21 = 117.29, P < 0.0001; 2: F2,18 = 40.43, P < 0.0001; 3: F2,21 = 23.70, P < 0.0001; 4: F2,18 = 32.84, P < 0.0001). Step length increased for the long adjustments (forced and free) and decreased for the short adjustments (forced and free) in relation to WT in step N (Fig. 12). In step N + 1, step length was not different among the conditions, except for the short forced adjustment, where it increased. For adjustments in the ML direction, the ANOVA revealed an interaction effect (F3,28 = 57.67, P < 0.0001) for step width. Step width increased for the lateral adjustments (forced and free) and decreased for the medial forced adjustment in comparison to WT in step N. Interestingly, it increased even more in step N + 1 for the lateral adjustments. For the medial adjustment, it was negative in step N + 1, indicating that participants crossed their steps.
|
The only ANOVA that indicated a main effect of condition was the one including long forced and short forced adjustments (F2,13 = 7.58, P = 0.0066). The deviation was larger for the long forced adjustment (1.6 ± 1.1°) than for the short forced adjustment (0.3 ± 1.3°). For the ML obstacle, the one-way ANOVA with repeated measures showed a main effect of condition (F3,21 = 39.05, P < 0.0001). Post hoc analysis indicated that there was a difference between lateral adjustments (forced: –3.4 ± 2.4°; free: –1.6 ± 2.2°) and the medial forced adjustment (10.2 ± 3.8°). Lateral forced and medial forced adjustments increased the goal deviation compared with WT (0.6 ± 0.6°).
|
|
DISCUSSION |
|---|
|
and expanded by Moraes et al. (2004). In the free condition, participants decided where to place their foot to avoid an obstacle, allowing us to study the preferred choice in each plane of movement (i.e., frontal and sagittal) when the amount of foot displacement for the alternate foot placement was similar within each plane of motion. The forced condition allowed analyzing how movement in the nonpreferred direction is achieved and how stability is compromised in these extreme cases. In the forced condition, a green arrow was displayed on top of the obstacle, indicating the direction of alternate foot placement. An arrow as a cue is a common signal for indicating directions present in our environment. In fact, Kingstone et al. (2003) noted that arrows are a good cue for shifting attention and helping reduce reaction time. Therefore the arrow was an appropriate trigger to cue the location of alternate foot placement. There was no difference between forced and free adjustments, except for the short adjustment where the forced adjustment generated a more unstable movement, indicating the validity of using such paradigm. Initial slowing response to all adjustments provides more time for planning and decision-making
An overall increase in braking impulse was observed for all adjustments implemented. This consistency suggests that such changes may not be directly related to step change implementation, but rather to the gathering of more time for planning alternate foot placement. Patla et al. (1989) have shown that braking impulse responses are specific for shortening and lengthening the step. Previous studies (Patla et al. 1989, 1991b) used a visual cue to indicate what change to implement, and the decision-making was simplified. Because, in this study, participants had to decide what modification to make to avoid the obstacle, the increased braking impulse in the AP and vertical directions could be seen as a strategy for slowing down the whole body movement to get extra time for planning alternate foot placement. However, the amount of change is not similar across conditions. For instance, the mean increase in braking impulse in the vertical direction was equal to 8.4 and 27.3% for the long and short adjustments, respectively. For the short adjustment, it is clearly higher than for the long adjustment, which may also suggest that part of the changes in braking impulse in the AP and vertical directions are directly related to the alternate foot placement implementation, because increasing the braking impulse helps in shortening the step (Patla et al. 1989). Therefore the changes in the braking impulse are likely related to both explanations: 1) to get more time for decision-making and 2) to implement effective changes related to the alternate foot placement.
Long and lateral adjustments are dominant over short and medial adjustments, respectively
As expected, long adjustment was preferred over short adjustment in the free condition (Patla et al. 1999), even though long adjustment had a greater predicted minimum foot displacement than short adjustment. However, Moraes et al. (2004) proposed that minimum foot displacement is not the major determinant guiding the selection of alternate foot placement. When more steps are available for planning and implementing the alternate foot placement, maintenance of forward progression seems to be more important than minimizing foot displacement. These finding suggests the existence of a threshold, where the switch to short adjustments occurs only when long adjustments represent a substantial increase in step length that may be uneconomical and/or unstable. The switch from long to short seems to occur when the amount of foot displacement for the long adjustment exceeds the short one by >7 cm (the difference was 
4 cm in this study) (Greig et al. 2004). It is also possible that long adjustments give participants more time to perform the next foot placement.
In addition, lateral adjustment was preferred over medial adjustment, which contradicts the medial preference reported by Patla et al. (1999). Predicted minimum foot displacement does not bias the lateral choice because no difference was found between medial and lateral adjustments for this variable. This lateral preference is in accordance with the results from Reynolds and Day (2005), who studied the visual control of foot trajectory from a standing position. In their study, participants were instructed to step on a target that could shift sideways at foot-off. They found that participants were able to make appropriate directional changes, although the magnitude of such changes was dramatically reduced for the medial (7.4 cm) compared with the lateral (16.9 cm) target location. In addition, recent work of Moraes and Patla (2006) has also shown a lateral adjustment prevalence over medial adjustment when the amount of foot displacement was the same. The reason(s) for this inconsistency between this study and the study of Patla et al. is not yet clear, but some methodological differences can be considered. Patla et al. (1999) used a mechanical apparatus with a piece of black cardboard with cuts and lights underneath it, indicating the planar obstacle. With such an apparatus, participants could get visual cues before the obstacle light was turned on, even though the room light was dim. Therefore they could have used such visual cues to plan and implement the alternate foot placement in advance. This contrasts with this experiment, where participants had no clue in advance because of the use of the LCD monitor. Moraes et al. (2004) found a medial preference when participants had three steps to plan and implement the alternate foot placement and they could see it from the beginning of the trial. Another issue may be related to stimulus identification (i.e., planar obstacle identification). Schmidt and Lee (2005) have suggested that stimulus intensity, i.e., the brightness of a light stimulus, has an effect on reaction time. In the original work, a mechanical apparatus with lights underneath was used and it created a brighter stimulus than the use of a LCD monitor. Further research is needed to identify the reasons for the bias in the frontal plane toward medial or lateral.
In all adjustments, participants tried to minimize foot displacement because the difference between predicted minimum foot displacement and foot placement modification vector magnitude was quite small (3.0 cm on average). Therefore because the difference between predicted minimum foot displacement and foot placement modification vector magnitude is the same for long and short adjustments and for lateral and medial adjustments, and long adjustment is preferred even though it results in a greater displacement of the foot, the amount of foot displacement from its normal landing position is not the sole predictor of the preferred alternate foot placement choice.
Maintenance of forward progression is consistent even for foot placement changes in the frontal plane
For all adjustments, deviation from forward progression was minimal, except for the medial forced adjustment. In the plane of progression, there was no deviation from the goal, although there is a difference between long and short forced. In addition, Fig. 7 shows that COM vectors always lie in the plane of progression. In the frontal plane, as shown in Fig. 9, COM continues to move forward, whereas the foot moves sideways in step N. This is shown by the decoupling between foot placement modification vector orientation and COM modification vector orientation (Fig. 10). For the medial forced condition, the high degree of forward progression deviation was related to the instability in step N + 1. Medial free adjustment resulted in a similar deviation, but in the opposite direction, as observed for the lateral adjustment. Thus maintenance of forward progression is one of the major determinants guiding the selection of alternate foot placement.
Stability guides the alternate foot placement choice under time pressure and similar minimum foot displacement
Margin of dynamic stability is more negative for the short adjustments compared with long adjustments and WT, which is a clear indication of increased instability for short adjustments. In addition, the fact that margin of dynamic stability for short adjustments does not return to baseline values in step N + 1 suggests that they are not preferred because it takes more time to recover from the changes in the locomotor behavior. Thus the CNS has the ability to predict the consequences of shortening or lengthening a step to body stability and uses that information not only to choose what to do, but also to anticipate the disturbances, as proposed by Patla (2003). Patla et al. (1999) have proposed that long adjustment is preferred over short adjustment because the latter could lead to a substantial increase in angular momentum if not properly controlled. The results for the short forced adjustment clearly exemplify this case (Fig. 8). The reduced coupling between foot and COM vectors is an indication of such instability because the COM continued to move forward while the foot moved backward and stopped at HC. This creates a tendency for the body to fall forward because of an increase in angular momentum. Therefore shortening the step is not preferred because it takes longer to return to baseline value and, under time pressure, may lead to a substantial increase in angular momentum and body instability, which may result in a fall if recovery is unsuccessful.
In the ML direction, lateral adjustments are more stable than medial adjustments, and this is shown by the margin of dynamic stability and by the cross-over in step N + 1 (Fig. 12). Cross-over was very high in both conditions for medial adjustments (free: 45.5%; forced: 88%). Lateral adjustment is not only stable in step N, but also in step N + 1, where it shows the same margin of dynamic stability in the ML direction as normal walking. For the medial adjustments, the margin of dynamic stability is reduced compared with WT in both steps. Thus narrowing the step is not preferred because it takes longer to return to baseline values. The increase in positive margin of dynamic stability for the lateral adjustments was mainly caused by increase in step width, increasing the distance between COM projection and BOS (i.e., foot). Although the increased distance of the foot relative to the COM may increase the destabilizing moment of force at the hip joint because of the large upper body mass (MacKinnon and Winter 1993), appropriate anticipatory control is needed to prevent this (Patla 2003). Therefore the CNS again uses its predictive capacity to determine the more stable adjustment in the ML direction.
It is important to mention that the values of the margin of dynamic stability are slightly higher because we estimated center of pressure from heel and fifth metatarsal markers. In their original work, Hof et al. (2005, 2006) measured the limits of center of pressure excursion directly and used that as the limit of the BOS. Although the margin of dynamic stability variable was quite revealing of the stability requirements involved with alternate foot placement, there are some limitations that should be addressed. The margin of dynamic stability in the AP direction was always negative, including the WT condition, which indicates that the forward movement is unstable. Despite this, participants never fell over. Margin of dynamic stability was calculated at HC and, therefore the impact of corrective joint torques was not taken into account. During the course of the stance phase, changes in joint torque would reflect on the current velocity of the COM and, consequently, would affect the COM extrapolation. For instance, Winter (1991) has shown that increase or decrease in walking speed is achieved, in part, by a synergistic change in the ankle torques. Increase in walking speed is characterized by a decrease in torque generation during the energy absorption phase and an increase in torque generation during the push-off phase and vice versa when decreasing walking speed.
Long latencies are indicative of cortical involvement in the selection of alternate foot placement
The mean latency (292 ms) obtained in this study was 2.4 times greater than latencies reported in previous studies. Weerdesteyn et al. (2004) reported latencies of 122 ms for a similar task involving planar obstacle avoidance while walking on a treadmill. Also, Reynolds and Day (2005) found latencies of 122 ms for a task involving stepping on shifting targets. These latencies are below visual simple reaction time (190 ms; Schmidt and Lee 2005), suggesting the existence of a short-latency visuomotor pathway for the leg (Reynolds and Day 2005; Weerdesteyn et al. 2004), similar to the one involved in correcting for unexpected change in target location during reaching movements (Day and Brown 2001). These authors proposed that this visuomotor pathway is subcortical in origin (Day and Brown 2001; Weerdesteyn et al. 2004).
Although this study and the study of Weerdesteyn et al. (2004) share a similar task; there are methodological differences that may account for the discrepancy in the latency result. It is well established in the upper limb motor control literature that there is no absolute time to process visual feedback (Schmidt and Lee 2005) and, more importantly, that the time to use visual feedback is task-specific (Carnahan et al. 1993; Elliot and Allard 1985; Paillard 1996). For example, the time to process visual feedback can vary according to the type of visual information available, the predictability of the visual information, and the nature of the task (Schmidt and Lee 2005). Carnahan et al. (1993) found that the time to process visual feedback is longer for pointing than for grasping movements when the distance covered by the arm was the same. The time to process visual feedback during an aiming task can be reduced when the participants are aware of the room light availability (Elliot and Allard 1985; Zelaznik et al. 1983).
Furthermore, processing visual feedback from moving objects involves a mechanism that may differ from processing visual feedback from stationary objects (Schmidt and Lee 2005). Paulignan et al. (1991) found latencies 100 ms for processing visual feedback from moving objects. In the studies of Weerdesteyn et al. (2004) and Reynolds and Day (2005), the obstacle/target moved during the task, whereas in this study, the obstacle was kept static after its appearance. Therefore it is not surprising that the latency was longer in this study than in the previous studies. Carnahan et al. (1993) proposed that the difference between pointing and grasping was related to how visual information was used in both tasks. In the pointing task, participants spent more time preprogramming the movement, whereas in the grasping movement, participants relied more on on-line adjustments while performing the movement. It is possible that because of the more dynamic nature of the task, participants have to preprogram less of the movement and rely more on visual feedback during foot adjustment while walking on the treadmill compared with overground walking. The overall slowing down observed in this study after obstacle appearance (see DISCUSSION) may represent the extra time needed for preprogramming the avoidance response.
Patla et al. (1991a) also found very short latencies (122 ms) to change limb trajectory when a second obstacle appeared unexpectedly on the landing area. However, it is important to point out that changes in foot trajectory occurred in two stages. The first reaction was very fast and generic. The rapid changes in foot trajectory would be enough to clear the highest obstacle, even when the lowest obstacle was triggered. The second reaction occurred at 280 ms after obstacle triggering (i.e., 180 ms after the 1st reaction) and scaled the foot trajectory according to obstacle height. The overall time needed to the second reaction is quite similar to the latencies observed in this study. Patla et al. (1991a) also found that when the participants were certain about the obstacle size appearance, the two-stage correction disappeared. In this study, participants were uncertain about the obstacle orientation that could appear on the pathway, whereas in the study of Weerdesteyn et al. (2004), participants were always aware of which obstacle would appear. This provides an additional explanation for the differences in latencies observed. Because the obstacle was visible all the time suspended on a bridge over the treadmill, attention may have played a very important role. The visual presence of the obstacle may have brought the focus of attention completely to the obstacle, facilitating the fast response observed.
Drew et al. (2004) argued that changes in gait used to avoid obstacles need to be integrated appropriately with the underlying pattern of activity to guarantee smooth adaptations. The same authors have shown the involvement of the motor cortex in cats while stepping over an obstacle seen a few steps before (Drew et al. 1996). These differences between our study and previous studies may indicate that separate pathways may be involved in planning adjustments for the ongoing leg movement. These long latencies may indicate the involvement of cortical pathways. These latencies differences can be seen also as part of a continuum ranging from the most to the least automatic responses (Day and Lyon 2000).
Summary
In general, this study showed that long and lateral adjustments are preferred because they result in a more stable adaptation not only in the adaptive step, but also in subsequent step. Either medial or short adjustments require at least two steps to return the dynamic balance parameters to baseline values. Therefore the CNS plans for the choice that minimizes threats to stability in step N and also guarantees a faster return of the stability parameters to baseline values in subsequent steps (normal walking). Additionally, foot displacement from its normal landing position and maintenance of forward progression were not affected by the alternate foot placement choices. Therefore under time pressure and similar displacement of the foot from its normal landing position, stability is the major determinant driving the selection of the alternate foot placement.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Deceased 29 January 2007. 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Address for reprint requests and other correspondence: R. Moraes, Escola de Artes, Ciências e Humanidades, Univ. de São Paulo, Av. Arlindo Bettio, 1000, Sao Paulo, 03828-000, SP, Brazil (E-mail: renatomoraes{at}usp.br)
|
|
REFERENCES |
|---|
|
Day BL, Brown P. Evidence for subcortical involvement in the visual control of human reaching. Brain 124: 1832–1840, 2001.
Day BL, Lyon IN. Voluntary modification of automatic arm movements evoked by motion of a visual target. Exp Brain Res 130: 159–168, 2000.[CrossRef][Web of Science][Medline]
Drew T, Jiang W, Kably B, Lavoie S. Role of the motor cortex in the control of visually triggered gait modifications. Can J Physiol Pharmacol 74: 426–442, 1996.[CrossRef][Web of Science][Medline]
Drew T, Prentice SD, Schepens B. Cortical and brainstem control of locomotion. Prog Brain Res 143: 251–261, 2004.[Web of Science][Medline]
Elliot D, Allard F. The utilization of visual feedback information during rapid pointing movements. Q J Exp Psychol A 37: 407–425, 1985.[Web of Science][Medline]
Greig MA, Patla AE, Lewis AM. Characterizing step width and step length adjustments for the implementation in robotic locomotion control. Proceedings of the 13th Biennial Conference for the Canadian Society for Biomechanics. Halifax, Nova Scotia, August 4–7, 2004.
Hof AL, Gazendam MGJ, Sinke WE. The condition for dynamic stability. J Biomech 38: 1–8, 2005.[Web of Science][Medline]
Hof AL, Bockel RM, Schoppen T, Postema K. Control of lateral balance in walking: experimental findings in normal subjects and above-knee amputees. Gait Posture 25: 250–258, 2006.[Web of Science][Medline]
Iqbal K, Pai YC. Predicted region of stability for balance recovery: motion at the knee joint can improve termination of forward movement. J Biomech 33: 1619–1627, 2000.[CrossRef][Web of Science][Medline]
Kingstone A, Smilek D, Ristic J, Friesen CK, Eastwood JD. Attention, researchers! It is time to take a look at the real world. Curr Dir Psychol Sci 12: 176–180, 2003.[CrossRef]
MacKinnon CD, Winter DA. Control of whole body balance in the frontal plane during human walking. J Biomech 26: 633–644, 1993.[CrossRef][Web of Science][Medline]
Moraes R, Lewis MA, Patla AE. Strategies and determinants for selection of alternate foot placement during human locomotion: influence of spatial and temporal constraints. Exp Brain Res 159: 1–13, 2004.[Web of Science][Medline]
Moraes R, Patla AE. Determinants guiding alternate foot placement selection and the behavioral responses are similar when avoiding a real or a virtual obstacle. Exp Brain Res 171: 497–510, 2006.[CrossRef][Web of Science][Medline]
Pai YC, Patton J. Center of mass velocity-position predictions for balance control. J Biomech 30: 347–354, 1997.[CrossRef][Web of Science][Medline]
Paillard J. Fast and slow feedback loops for the visual correction of spatial errors in a pointing task: a reappraisal. Can J Physiol Pharmacol 74: 401–417, 1996.[CrossRef][Web of Science][Medline]
Patla AE. Strategies for dynamic stability during adaptive human locomotion. IEEE Eng Med Biol Mag 22: 48–52, 2003.[CrossRef][Web of Science][Medline]
Patla AE, Beuter A, Prentice S. A two stage correction of limb trajectory to avoid obstacles during stepping. Neurosci Res Commun 8: 153–159, 1991a.[Web of Science]
Patla AE, Prentice SD, Rietdyk S, Allard F, Martin C. What guides the selection of alternate foot placement during human locomotion. Exp Brain Res 128: 441–450, 1999.[CrossRef][Web of Science][Medline]
Patla AE, Prentice SD, Robinson C, Neufeld J. Visual control of locomotion: strategies for changing direction and for going over obstacles. J Exp Psychol Hum Percept Perform 17: 603–634, 1991b.[CrossRef][Web of Science][Medline]
Patla AE, Robinson C, Samways M, Armstrong CJ. Visual control of step length during overground locomotion: task-specific modulation of the locomotor synergy. J Exp Psychol Hum Percept Perform 15: 603–617, 1989.[CrossRef][Web of Science]
Paulignan Y, MacKenzie C, Marteniuk R, Jeannerod M. Selective perturbation of visual input during prehension movements. 1. The effects of changing object position. Exp Brain Res 83: 502–512, 1991.[Web of Science][Medline]
Redfern MS, Schumann T. A model of foot placement during gait. J Biomech 27: 1339–1346, 1994.[CrossRef][Web of Science][Medline]
Reynolds RF, Day BL. Rapid visuo-motor processes drive the leg regardless of balance constraints. Curr Biol 15: R48–R49, 2005.[CrossRef][Web of Science][Medline]
Schmidt RA, Lee TD. Motor Control and Learning: A Behavioral Emphasis. Champaign, IL: Human Kinetics, 2005.
Ting LH, Blickhan R, Full RJ. Dynamic and static stability in hexapedal runners. J Exp Biol 197: 251–269, 1994.[Abstract]
Weerdesteyn V, Nienhuis B, Hampsink B, Duysens J. Gait adjustments in response to an obstacle are faster than voluntary reactions. Hum Mov Sci 23: 351–363, 2004.[CrossRef][Web of Science][Medline]
Weerdesteyn V, Nienhuis B, Mulder T, Duysens J. Older women strongly prefer stride lengthening to shortening in avoiding obstacles. Exp Brain Res 161: 39–46, 2005.[CrossRef][Web of Science][Medline]
Winter DA. The Biomechanics and Motor Control of Human Gait: Normal, Elderly, and Pathological. Waterloo, ON: Waterloo Biomechanics, 1991.
Winter DA. Anatomy, Biomechanics and Control of Balance during Standing and Walking. Waterloo, ON: Waterloo Biomechanics, 1995.
Winter DA. Biomechanics and Motor Control of Human Movement. Hoboken, New Jersey: Wiley, 2005.
Zelaznik HN, Hawkins B, Kisselburgh L. Rapid visual feedback processing in single aiming movements. J Mot Behav 15: 217–236, 1983.[Web of Science][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
