Abstract
Objective
Successful stepping reactions, led by either the paretic or nonparetic leg, in response to a loss of balance are critical to safe mobility poststroke. The purpose of this study was to measure sagittal plane hip, knee, ankle, and trunk kinematics during 2-step stepping reactions initiated by paretic and nonparetic legs of people who had stroke and members of a control group.
Methods
Principal component analysis (PCA) was used to reduce the data into movement patterns explaining interlimb coordination of the stepping and stance legs. Correlations among principal components loading scores and clinical measures of balance ability (as measured on the Community Balance and Mobility scale), motor impairment (as measured on the foot and leg sections of the Chedoke-McMaster Stroke Assessment), and step characteristics (length and velocity) were used to examine the effect of stroke on stepping reaction movement patterns.
Results
The first 5 principal components explained 95.9% of the movement pattern of stepping reactions and differentiated between stepping reactions initiated by paretic legs, nonparetic legs, or the legs of controls. Moderate-strong associations (ρ/r > 0.50) between specific principal component loading scores and clinical measures and step characteristics were dependent on the initiating leg. Lower levels of motor impairment, higher levels of balance ability, and faster and longer steps were associated with stepping reactions initiated by the paretic leg that comprised paretic leg flexion and nonparetic leg extension. Step initiation with the nonparetic leg showed associations between higher scores on clinical measures and movement patterns of flexion in both paretic and nonparetic legs.
Conclusions
Movement patterns of stepping reactions poststroke were influenced by the initiating leg. After stroke, specific movement patterns showed associations with clinical measures depending on the initiating leg, suggesting that these movement patterns are important to retraining of stepping reactions. Specifically, use of flexion patterning and assessment of between-leg pattern differentiation may be important aspects to consider during retraining of stepping reactions poststroke.
Impact
Evidence-based interventions targeting balance reactions are still in their infancy. This investigation of stepping reactions poststroke addresses a major gap in research.
Keywords: Balance, Interlimb Coordination, Movement Patterns, Stepping Reactions, Stroke
Introduction
When standing balance is challenged, a successful stepping reaction is commonly required to prevent a fall. Fall occurrence has been reported to be as high as 73% of all people living in the community who recover the ability to walk poststroke, and falls commonly occur within the first few months of returning home from rehabilitation.1,2 The occurrence of falls may be a result of the impaired ability of patients poststroke to perform an effective stepping reaction in response to a loss of balance.3–5 Specifically, people poststroke have been shown to have stepping reactions with decreased step length and speed, regardless of the stepping leg, in comparison to age-matched controls,6 demonstrating impact on the motor control of both the paretic and nonparetic leg (inclusive of compensatory movement). Importantly, the inability to initiate stepping reactions with the paretic leg has been identified as a contributing factor to impaired balance reactions.5 The optimal clinical approach to retraining stepping reactions poststroke remains unclear.
Rehabilitation following stroke involves the retraining of task-specific movement patterns known to be associated with optimal functional outcomes. There is a paucity of research that explains the impact of stroke on the interlimb movement patterns of stepping reactions. The most recent international Stroke Recovery and Rehabilitation Roundtable prioritized a need to add kinematic and kinetic movement quantification to its core recommendations for standardized measurements of sensorimotor recovery in stroke trials.7 Currently standards are outlined for the upper limb and the need for similar kinematic recommendations in the lower limb has been identified.7 Principal components (PC) analysis, applied to kinematic measurement of movement, identifies patterns that exist within the diverse range of individual movement patterns observable among a clinical population3,8 and can be used to identify the most common movement patterns explaining interlimb coordination of the stepping and stance legs during a stepping reaction. Identifying the movement patterns that are associated with clinical measures of functional balance ability and motor impairment following stroke provides clinically meaningful interpretation of these movement patterns to focus future rehabilitation development. Furthermore, examining relationships between movement patterns with step length and speed, 2 key measures known to affect the success of stepping reactions poststroke, will add to an improved understanding of stepping reactions kinematics.
Various methods of evoking stepping reactions are noted in the current literature. For this reason, it is important to understand the relationship between the stimuli provoking the stepping reaction and the resultant movement pattern.9 Balance during walking is defined by the ability to maintain the center of mass (COM) within the base of support (BOS). Therefore, exploring the relationship between stepping reaction movement patterns and the distance of the COM from the BOS and the velocity that the COM is traveling, at the time when a stepping reaction is initiated, may inform our understanding of the influence of the stimuli evoking a stepping reaction.
This study aimed to test for differences between movement patterns describing stepping reactions of people poststroke compared to age-matched controls. Specifically, we aimed to test whether components of the movement patterns of the legs and trunk that differentiated between people poststroke and controls were associated with 1) motor impairment and walking balance, 2) step length and step velocity, and 3) the movement of the COM characterizing the destabilizing event.
Methods
Participants
Ten people with chronic stroke (>1 year poststroke) and 10 age-matched controls provided written informed consent to participate in this study. Participants poststroke were recruited from local community stroke groups, and controls were recruited from the university community. Individuals poststroke were included if they were ambulatory, with or without a walking aid, and could stand independently. Individuals were excluded if they had any additional health conditions that negatively affected balance. Controls were included if they were free from neurological or musculoskeletal conditions that impaired balance. The study conformed to the standards set by the latest revision of the Declaration of Helsinki and was approved by the University of British Columbia Clinical Research Ethics Board.
Clinical Measures of Balance and Motor Impairment
Ambulatory Balance
All participants were assessed with the Community Balance and Mobility Scale (CB&M, total score/96).10,11 Higher scores reflect higher level walking balance.
Motor Recovery Poststroke
Severity of motor impairment following stroke was measured using the foot and leg sections of the Chedoke-McMaster Stroke Assessment (CMSA).12The CMSA describes 7 stages of motor recovery; 0 of 7 refers to flaccid paralysis, and 7 of 7 refers to movement equated to a “normal” sensory-perceptual-motor system.12,13 All clinical measures were performed by a licensed physiotherapist and lasted 45 minutes per participant.
Experimental Protocol
Stepping reactions were performed in the forward direction with the use of an overhead safety harness. Participants were instructed on the performance of the forward-leaning task, including demonstration of the task. Participants were given the verbal cue of letting themselves “fall like a tree” to represent the action of a forward loss of balance. The visual example of fall like a tree was emphasized as participants were instructed to maintain a straight trunk position while initiating the lean at the ankles rather than bending forward at the hips or through the trunk. Participants were provided with signaling of “ready, set, go” to initiate the perturbation. Instructions with respect to the stepping reactions were to simply stop themselves from an actual fall. Twenty-five trials were performed consecutively with the right and left legs (paretic or nonparetic) as the initial stepping leg (50 steps per participant). The order of each block of stepping reactions was randomly determined to ensure stepping reactions were initiated with each leg equally (eg, half of participants poststroke performed stepping reactions initiated by the paretic leg first). Participants were instructed as to which leg to initiate the stepping reaction with; limbs were not physically restrained to direct the stepping reaction. Any errors in the stepping leg initiating were noted and participants were reminded as to the stepping leg during each block. No walking aids were used during data collection. Rests were provided as needed by the participants with a minimum of rest after completion of the first 10 steps and after each block of 25 stepping reactions. The duration of measurement of these stepping reactions, inclusive of rest breaks, was no longer than 30 minutes.
Kinematic Data
Twenty-two passive reflective markers were affixed to participants according to a modified Helen Hayes marker set.14 Eight high-speed digital cameras (Raptor-E; Motion Analysis Corp, Santa Rosa, CA, USA) sampled the movement of the reflective markers at 100 Hz. Joint angles (hip, knee, ankle) and trunk angle coordinates were calculated from marker coordinate data using commercially available software (OrthoTrak; Motion Analysis Corp, Santa Rosa, CA, USA). This software was also used for calculation of COM from marker coordinate data and published anthropometric values.15 Anterior displacement of the COM in relation to the BOS was used as a measure of the forward-leaning position that participants obtained at the point of the initial toe-off of the stepping reaction. The COM velocity (derivative of COM displacement) at the point of the initial toe-off further characterized the stimuli that the participant needed to respond to. Analysis of kinematics was limited to the sagittal plane. Step length, specific to the initial position of each limb, was calculated from the forward displacement of the ankle marker from toe-off to foot contact. Step velocity was calculated from displacement of the ankle marker divided by the time for the step.
Data Analysis
Kinematic data were analyzed with a custom-written program in Mathematica (Wolfram Research Inc, Champaign, IL, USA). Kinematic data were time normalized into 30 equally spaced time points for each step. Stepping reactions were included in the analysis if the participant did not experience a loss of balance that required the assistance of the harness or the therapist. Each step cycle (referred to as step 1 and step 2 or the first step and second step of the stepping reaction) were defined as the initial step (toe-off, defined as upward movement of the marker affixed to the great-toe, of the first stepping leg to toe-off of the second stepping leg) and the second step (toe-off of the second stepping leg to heel strike of the same leg). The kinematic pattern of the two steps taken, therefore, contained the interlimb coordination pattern of the stepping and stance limbs. For each stepping reaction, kinematic patterns were defined by 210 joint angle windows: ([3 lower extremity joints × 2 legs] plus trunk movement)—that is, ([3 × 2] + 1)— × 30 time points. PC analysis was used to reduce this large multivariate data set to a number of components that describe the most important patterns of each step, and to distinguish the effect of stroke on the interlimb coordination pattern3.
For the PC analysis, the kinematic patterns were compiled into p × N data matrix A, where p represents time windows and N represents steps. There were 2000 steps (20 participants, 25 stepping reactions per leg [2 legs], 2 steps during each stepping reaction) and 210 time windows (3 joints per leg: paretic, nonparetic, or control right and left legs and trunk × 30 time points per step). The principal components of the data were determined16 from covariance matrix B of data A without prior subtraction of the mean.17 The PC weightings of data A are given by the unit eigenvectors ξ of covariance matrix B. The relative proportion of the data explained by each component is given by ξ′Bξ, and the loading scores for each of the PC for a given trial are given by ξ′A. The PC 1 (PC1) weightings represent the common angular joint pattern, and subsequent PCs represent variations in the joint angle patterns of the leg from PC1. PC weightings are presented as the mean joint angle for each of the joints during the stepping reaction. Presentation of the weightings in this way provides a summative representation of the emerging patterns of variance at each of the joints of the stepping or the stance leg. The loading scores describe the contribution that each PC makes to the movement pattern of a stepping reaction and are grouped as steps initiated by the paretic, nonparetic, or control legs. The joint angular patterns can be reconstructed from the product of the weightings and loading scores for each of the major PCs. PC were included for further analysis if they contributed to explaining 95% of the variability of the angular joint movement pattern.
Statistical Analysis
Analysis of the stepping reactions initiated by the right leg and the left leg of controls was performed separately as there may have been effects of leg dominance among controls. Review of the results revealed no significant differences between the loading scores of steps initiated by the right and left legs of controls (PC1–PC5; P > .05). Therefore, for simplicity, control data are presented, and were analyzed, as stepping reactions initiated with the right leg only.
A one-way analysis of variance compared the performance variables and PC loading scores among the 3 initiating limbs (control, paretic, and nonparetic legs). The level of significance was set at P less than or equal to .05. When significant between-limb differences were found, a post hoc analysis was performed with the Tukey honestly significant difference test. Spearman correlations (ρ) were used to determine relationships between the PC loading scores of each PC and the clinical measures of CB&M and CMSA. Pearson correlations (r) were used to explore the relationships between the PC loading scores of each PC and step characteristics (step length and speed) and perturbation stimuli (anterior displacement and velocity of the COM). Associations of ρ/r equal to 0.25 to 0.50 were considered fair, those of ρ/r equal to 0.50 to 0.75 were considered moderate, and those of ρ/r greater than 0.75 were considered strong.18 A Bonferroni correction was used for multiple comparisons resulting in a significance level of P less than or equal to .001 for all correlation coefficients. Data are reported as means (SDs).
Results
Participants
Table 1 presents participant characteristics. Participants poststroke demonstrated significantly lower balance abilities compared to controls (P < .01). The scores on the CMSA of the foot for participants poststroke reflect motor control impairment described as marked spasticity present, some voluntary movement, and synergistic patterns with inability to move quickly between plantar flexion and dorsiflexion. The scores on the CMSA of the leg reflect motor control impairment described as waning spasticity, increased range of voluntary movement, and synergistic patterns less evident.13
Table 1.
Participant Characteristicsa
| Group | Age, y | Sex (No. of Men/No. of Women) | Years After Onset | Paretic Side (No. Right/No. Left) | Foot Section of CMSA b (0–7) | Leg Section of CMSA b (0–7) | CB&M (0–96) |
|---|---|---|---|---|---|---|---|
| Stroke | 63.6 (7.0) | 7/3 | 7.1 (3.6) | 6/4 | 3.0 (3–5) | 4.5 (3–7) | 35.0 (19.2) |
| Control | 68.4 (6.3) | 6/4 | NA | NA | NA | NA | 85.2 (5.0) |
a Data are reported as mean (SD) unless otherwise indicated. CB&M = Community Balance and Mobility Scale; NA = not applicable.
b Chedoke-McMaster Stroke Assessment (CMSA) data are reported as median (interquartile range).
Stimuli and Step Characteristics
There was a statistically significant main effect of leg initiating the stepping reaction in each of the measures of COM movement and step characteristics (P < .01). Stepping reactions initiated by controls demonstrated greater anterior COM displacement (mean [SD] = 36.68 cm [5.83 cm]) and velocity (85.40 cm/s [15.87 cm/s]) at initial toe-off compared to both stepping reactions initiated by the paretic (COM displacement = 24.58 cm [5.65 cm]; velocity = 42.53 cm/s [18.90 cm/s]; P < .05) and nonparetic (COM displacement = 26.27 cm [5.40 cm]; velocity = 50.58 cm/s [16.43 cm/s]; P < .05) legs of participants poststroke. Control participants demonstrated longer and faster steps than participants poststroke during both the first step and the second step of the stepping reactions (P < .01) (Fig. 1A and 1B). Participants poststroke took a longer and faster first step when initiated by the nonparetic leg compared to the paretic leg (P < .01) (see Fig. 1A).
Figure 1.
Characteristics of the stepping reaction during steps 1 and 2. (A) Step length. (B) Step velocity. Controls (green circles) demonstrated longer and faster steps than both the nonparetic (red square) and the paretic (blue diamond) legs of participants poststroke during both the first and the second steps of the stepping reactions. **P less than .01. There were no significant differences between right and left legs of controls.
Principal Components
Kinematic Movement Patterns
PC1 through PC5 explained 95.9% of the variability of the movement pattern of the legs and trunk during the 2-step stepping reactions. Figure 2 shows the mean flexion and extension angles occurring at each joint that contributed to the overall movement pattern identified by each PC. To assist with interpretation and discussion, each PC is given a descriptive name that aims to highlight the distinguishing feature of each movement pattern (see Fig. 2). PC1 explains 77.8% (the largest percentage) of the overall movement pattern with a greater amount of flexion in the stepping leg compared to the stance leg. PC1 can be referred to as the main movement pattern of generalized flexion in each leg. Of the 4 remaining PC, PC2 accounts for the largest amount of variability in the movement pattern of the stepping reaction (PC2 describes 11.7% of the pattern). PC2 (pattern of between-leg flexion-extension differentiation) differentiates the flexion movement of the stepping leg from the extension of the stance leg, particularly at the hip and knee joints.
Figure 2.
Mean of principal components (PC) weightings of the joint angle across a step cycle for each joint segment. The mean angular movement of the hip, knee, and ankle for the stepping (black bars) and stance (gray bars) legs and trunk (white) during stepping reactions is shown. The percentage of the data set explained for each PC1 to PC5 is noted, and a total of 95.9% of the pattern is explained. PC1 (main movement pattern of generalized flexion in each leg) demonstrates the most common aspect of the movement pattern with greater amount of flexion of the swing compared to the stance legs. PC2 (pattern of between-leg flexion-extension differentiation) increases flexion in the hip and knee of the stepping leg while increasing extension of the hip and knee in the stance leg. PC3 (symmetrical influence of movement pattern of both legs) increases hip flexion and extension in the knees and ankles of both the stepping and the stance legs and increases trunk flexion. PC4 (trunk extension) most notably decreases the use of trunk flexion. PC5 (stepping leg ankle flexion) shows a small increase in extension of the hip together with flexion of the knee and considerable increase of flexion of the ankle in the stepping leg and flexion of the hip with extension of knee and ankle in the stance leg.
Figure 3 shows the loading scores of all PCs for steps 1 and 2. There was a statistically significant main effect of leg initiating the stepping reaction in the loading scores of PC1, PC2, PC4, and PC5 (P < .01). The loading scores for all PCs except PC3 (symmetrical influence of movement pattern of both legs), which increases bilateral hip flexion and knee extension together with trunk flexion, in step 1 were significantly different between steps initiated by the paretic, nonparetic, and control legs (P < .01). Therefore, how each of these components of movement patterns come together to describe the overall movement pattern is dependent on whether the stepping reaction was led by the paretic, nonparetic, or control leg.
Figure 3.
Mean principal components (PC) loading scores for 2-step stepping reactions initiated by nonparetic, paretic, and control legs for PC1 to PC5 for (A) step 1 and (B) step 2. Values are expressed as mean and SD. †, Differences between nonparetic and paretic legs were significant at P less than .01. *, Differences between control and nonparetic legs and between control and paretic legs were significant at P less than .01.
Figure 4 presents the reconstructed movement patterns based on the sum of the products of each PC weighting (see Fig. 2) and loading scores (see Fig. 3) specific to the leg leading the stepping reaction. Stepping reactions initiated by the paretic leg see (Fig. 2A, step 1, blue) demonstrate less flexion at the hip, knee, and ankle than stepping reactions initiated by either the nonparetic see (Fig. 2A, red) or control see (Fig. 2A, green) legs. The control and nonparetic legs move quite similarly to the initial stepping leg. However, in step 2, when the nonparetic leg steps after an initial paretic step, there is less flexion of the nonparetic hip, knee, and ankle compared to controls.
Figure 4.
Mean kinematic reconstructions of joint angle movement patterns in the sagittal plane (product of PC1–PC5 weighting and loading scores) and the SEM for the hip, knee, and ankle of the stepping and stance legs and the trunk during (A) step 1 and (B) step 2 of a 2-step stepping reaction. Line colors are consistent with the leg that first initiated the stepping reaction. The legends corresponding to step 1 and step 2 provide details for the stepping and stance legs during each step.
Association Between Principal Components and Clinical Measures and Step Characteristics
Only correlation values that reached significance (P ≤ .001) are reported, and only moderate to strong relationships between variables are highlighted in Table 2 and discussed in results. In stepping reactions initiated by the paretic leg, the loading scores of PC2 (pattern of between-leg flexion-extension differentiation) during steps 1 and 2 and the loading scores of PC5 (stepping leg ankle flexion) during step 2 demonstrated only moderate to strong (r > 0.50) relationships with most clinical measures, step characteristics, and descriptors of the perturbation stimuli (see Tab. 2).
Table 2.
Correlation of PC1 to PC5 Loading Scores Indicative of Movement Patterns of First and Second Steps of a 2-Step Stepping Reactiona
| PC | Balance Ability | Motor Recovery | Step Characteristics | Stimuli | |||
|---|---|---|---|---|---|---|---|
| CB&M | CMSA Foot | CMSA Leg | Step Length | Step Velocity | COM Velocity | COM Dist BOS | |
| Step 1 | |||||||
| Paretic leg initiating stepping reaction | |||||||
| PC1 | 0.42 | 0.22 | 0.50 b | 0.36 | 0.24 | ||
| PC2 | 0.89 b | 0.65 b | 0.78 b | 0.82 b | 0.80 b | 0.78 b | 0.57 b |
| PC3 | −0.22 | 0.22 | 0.21 | 0.38 | |||
| PC4 | 0.38 | 0.34 | 0.39 | 0.50 b | 0.34 | ||
| PC5 | 0.42 | 0.28 | 0.26 | 0.39 | 0.45 | 0.36 | |
| Nonparetic leg initiating stepping reaction | |||||||
| PC1 | −0.40 | −0.40 | −0.43 | −0.47 | −0.49 | −0.58 | −0.38 |
| PC2 | −0.32 | −0.56 b | -0.23 | ||||
| PC3 | 0.21 | 0.28 | 0.45 | 0.35 | 0.56 b | ||
| PC4 | −0.21 | −0.28 | −0.28 | ||||
| PC5 | −0.61 | −0.59 | –0.45 | −0.60 | −0.60 | −0.60 | −0.54 |
| Control (right) leg initiating step | |||||||
| PC1 | 0.28 | 0.23 | 0.27 | 0.30 | |||
| PC2 | 0.56 b | 0.72 b | 0.34 | ||||
| PC3 | −0.48 | ||||||
| PC4 | −0.33 | −0.39 | −0.65 | ||||
| PC5 | 0.22 | ||||||
| Step 2 | |||||||
| Paretic leg initiating stepping reaction | |||||||
| PC1 | 0.24 | 0.31 | 0.27 | 0.44 | |||
| PC2 | −0.84 | −0.62 | −0.60 | −0.63 | −0.43 | −0.73 | |
| PC3 | 0.22 | 0.24 | |||||
| PC4 | −0.22 | −0.26 | −0.22 | ||||
| PC5 | −0.85 | −0.49 | −0.57 | −0.71 | −0.63 | −0.77 | |
| Nonparetic leg initiating stepping reaction | |||||||
| PC1 | 0.84 b | 0.71 b | 0.71 b | 0.66 b | 0.69 b | 0.80 b | |
| PC2 | 0.38 | 0.23 | 0.40 | 0.37 | 0.29 | 0.44 | |
| PC3 | −0.23 | −0.21 | −0.28 | ||||
| PC4 | 0.54 b | 0.61 b | 0.36 | 0.37 | 0.38 | 0.49 | |
| PC5 | 0.50 b | 0.24 | 0.39 | 0.34 | |||
| Control (right) leg initiating step | |||||||
| PC1 | 0.31 | 0.35 | 0.33 | 0.38 | |||
| PC2 | −0.49 | −0.43 | −0.24 | −0.52 | |||
| PC3 | −0.36 | ||||||
| PC4 | −0.29 | ||||||
| PC5 | −0.75 | −0.71 | −0.38 | −0.64 | |||
a Specific to initiation by paretic, nonparetic, and control legs, with indicators of balance ability (Community Balance and Mobility Scale [CB&M]), motor impairment following stroke (Chedoke-McMaster Stroke Assessment [CMSA]), and discrete variables describing step characteristics and center-of-mass (COM) movement as stimuli of stepping reactions. Dist BOS = distance from base of support; PC1 = principal component (PC) 1. bCorrelations ρ and r greater than 0.50, indicative of moderate to strong associations. Only correlations significant at a level of P less than or equal to .001 are reported.
Stepping reactions initiated by the nonparetic leg showed PC5 (stepping leg ankle flexion) during step 1 to have a moderate (r = 0.50–0.75) relationship with most clinical measures, step characteristics, and descriptors of the perturbation stimuli. In step 2 of this reaction (paretic leg stepping), PC1 (the main movement pattern of generalized flexion in each leg) demonstrated moderate to strong relationships with clinical measures characterizing participants, step characteristics, and descriptors of the perturbation stimuli see (Tab. 2).
Stepping reactions performed by controls also demonstrated moderate (r = 0.50–0.75) relationships between both PC2 (pattern of between-leg flexion-extension differentiation) and PC5 (stepping leg ankle flexion) and balance ability see (Tab. 2).
Role of the Funding Source
The funders played no role in the design, conduct, or reporting of this study.
Discussion
Five movement patterns explained 95% of the variability of performance of stepping reactions in people with stroke and age-matched controls. The extent to which each of these movement patterns contributed to the makeup of the overall (reconstructed) stepping reaction pattern was specific to initiation of the reaction by the paretic, nonparetic, and control legs. Three patterns (PC1, PC2, and PC5) emerged that showed moderate to strong correlations between their respective PC loading scores and clinical measures. Moderate to strong correlations with clinical measures suggest these movement patterns may be clinically important in the retraining of stepping reactions poststroke. Specifically, the pattern of between-leg flexion-extension differentiation (PC2), which explained 11.7% of the overall pattern, showed moderate to strong associations with most clinical measures when the paretic leg led the reaction. In contrast, the main movement pattern of generalized flexion in each leg (PC1), which explained 77.8% of the overall pattern, showed moderate to strong associations with most clinical measures during the second step specifically when the nonparetic leg led the reaction.
The importance of understanding the differences between stepping reactions initiated by the paretic or nonparetic leg of people poststroke has been recently highlighted.4,19 People poststroke have been shown to be reluctant to initiate a stepping reaction with the paretic leg.4,5 Therefore, retraining stepping reactions without a focus on retraining both strategies will limit the flexibility of the situation-specific stepping reaction strategy that a person poststroke can employ in response to a loss of balance. In other words, the ability to avoid a fall will be influenced by whether the nonparetic leg is able to initiate the stepping reaction (eg, not physically blocked or in a disadvantaged position). Understanding the movement patterns specific to initiation with the paretic and nonparetic legs in the context of clinical measures of balance and motor recovery serves to inform future development of rehabilitation strategies to retrain stepping reactions poststroke.
The overall movement patterns, specific to the leg leading the stepping reaction, are shown in the reconstructed movement patterns (Fig. 4). These patterns highlight the similarity between the stepping leg movement of the control and nonparetic leg in contrast to the decreased flexion across the joints of the paretic leg as the initial stepping leg. However, when transitioning from the stepping leg to stance leg, both the paretic and the nonparetic legs of participants poststroke demonstrated notably less flexion at the hip, knee, and ankle than controls. This stance leg is critical to the control of forward and downward velocity of the COM and contributes significantly to controlling momentum and regaining balance.20–22 The decreased flexion of both the paretic and the nonparetic legs is suggestive of difficulties in the ability to dissipate COM downward velocity both in the paretic and the nonparetic legs of people poststroke.19,21 Following stroke, extensor patterning predominates in the paretic leg,12,23 and an extensor patterning is commonly heightened with increased postural challenges.12,23 This present finding is suggestive of extensor patterning poststroke somewhat affecting both the paretic and the nonparetic legs during response to a movement challenge.
Examining the movement patterns identified by specific components (PC) that showed moderate to strong associations with clinical measures provides more detailed understanding of clinically relevant patterns. In line with the aims of this study, we can identify the movement patterns that are associated with higher levels of recovery of motor impairment of the leg and foot as measured with the CMSA, walking balance as measured with the CB&M, and improved step characteristics (length and speed).
When a stepping reaction was initiated with the paretic leg, participants poststroke with higher levels of motor recovery, balance ability, and better step characteristics showed greater PC2 loading scores, which means greater flexion-extension differentiation between legs. This finding suggests that, when stepping reactions are initiated by the paretic leg, the presence of increased flexion across the joints of the paretic stepping leg and extension in the nonparetic stance leg are associated with higher levels of balance ability, motor recovery, and faster and longer steps.
The second step of the stepping reactions initiated by the paretic leg includes transition of the paretic leg to stance leg and the nonparetic leg to stepping leg. This step is also characterized by moderate to strong relationships between PC2 and clinical measures. However, it is important to note that the relationships are now inversely correlated. The negative correlations suggest that extension in the stepping leg (nonparetic) and flexion in the stance leg (paretic leg) is associated with increased balance ability, motor recovery, and step characteristics. Unique to the second step, the role of the stance paretic leg is now increasingly challenged by the need to dissipate the COM downward velocity, maintain balance, and upward support following foot contact of the initial stepping leg. Accordingly, greater flexion of the paretic stance leg, likely reflective of strategies to control the momentum of the COM, is associated with higher levels of balance ability and improved step characteristics.
Increased flexion of the paretic leg as it steps and transitions to the stance limb is associated with improved motor recovery and balance ability. This is suggestive of the extensor pattern dominance and difficulty with moving rapidly from extension (required for stance) to flexion (required for stepping) during recovery from stroke.12,24–26 The reluctance to initiate a stepping reaction with the paretic leg was hypothesized to be due to slower reaction times of the paretic leg.4,5 However, Inness et al4 showed there was no significant difference in the time to toe-off between paretic and nonparetic legs when initiating a stepping reaction. It is possible that the reluctance of people poststroke to initiate stepping reactions with the paretic leg is influenced by the inability to initiate a step rapidly with sufficient flexion. Therefore, rehabilitation strategies directed toward retraining fast flexion patterning of the paretic stepping leg may improve stepping reactions and balance ability poststroke.
The importance of between leg movement patterns that show clear differentiation between the flexion/extension movement of the paretic and nonparetic leg (PC2 and PC5) is further highlighted by moderate to strong positive relationships between these patterns and step characteristics of step length and velocity during stepping reactions initiated by the paretic leg. This means that during stepping reactions led by the paretic leg, movement patterns that differentiate the task-specific pattern of the stepping and stance leg are more likely to result in longer and faster steps. The importance of longer steps has been shown during perturbation trials that use a slip-while-walking stimuli; insufficient step length of the stepping reaction has been shown by people poststroke who experienced a fall during the trial.19 The ability to differentiate the activation of flexion patterning in one leg and while extending the other leg is indicative of advanced motor recovery and higher level of balance poststroke.
The nonparetic and the control leg showed similar patterns at initiation of the stepping reaction. However, the second step of stepping reactions initiated by the nonparetic leg revealed important differences from the control leg. During the second step, the paretic leg was the stepping leg and the nonparetic leg was the stance leg. This step demonstrated moderate to strong associations between the main movement pattern of generalized flexion in each leg (PC1) and most clinical measures. These associations suggest that more flexion of both the stepping and stance legs was associated with higher balance ability, motor recovery, and step length and velocity. While the finding of more flexion of the stepping paretic leg being associated with better clinical outcomes is to be anticipated considering the extensor pattern dominance in the paretic leg following stroke,23,24 the importance of increased flexion of the nonparetic leg as stance leg highlights a novel finding of the impact of stroke on the nonparetic leg during stepping reactions. This is in line with the earlier noted finding of decreased flexion of the nonparetic leg when it is the stepping leg following an initial step with the paretic leg. It has been shown that the sensorimotor impairments following stroke are noted in the nonparetic limb, although to a lesser extent than the paretic limb.27 However, the functional impact of decreased flexion of the nonparetic stance leg, noted here, may also be indicative of a compensation in the nonparetic limb for the lack of flexion across the hip and knee of the paretic leg and possibly decreased foot clearance of the paretic stepping leg. However, this finding also highlights the importance of a clinical focus of flexion patterning of the nonparetic as well as the paretic leg during transition from stepping to stance limb as this is essential to dissipating and controlling momentum during stepping reactions.
Participants poststroke demonstrated less forward displacement and slower movement of their COM than controls, regardless of which leg was initiating the stepping reaction. Our results suggest that the earlier noted kinematic movement patterns that were associated with clinical measures also share similar moderate to strong associations with movement of the COM at the initiation of the stepping reaction. While cause and effect cannot be determined in the present analysis, this relationship suggests that clinical attention to the progressive use of perturbation stimuli, as depicted by the movement of the COM, may be warranted in the retraining of stepping reactions poststroke. The decreased forward displacement and slower movement of the COM of participants poststroke may have contributed to the observed movement patterns of the resultant stepping reactions. These findings also suggest a relationship between the nature of the perturbing stimuli and the movement patterns of the resulting stepping reaction, an important consideration for future intervention trials.
A limitation of this study is that only movement in the sagittal plane was explored. Examination of lateral movements in the mediolateral plane may provide additional understanding of the impairment of forward-stepping reactions poststroke.
Conclusions
The results of this study show that poststroke, interlimb coordination of the movement patterns that describe stepping reactions are dependent on whether the paretic or nonparetic leg is initiating the stepping reaction. In line with the aims of this study, identification of specific aspects of the leg-dependent stepping reaction movement pattern that are associated with clinical measures of balance ability, motor impairment, and step characteristics adds to the applicability of these findings to rehabilitation poststroke. These associations suggest findings to be explored prospectively in future rehabilitation trials: 1) the importance of the ability to employ flexion patterning in the stepping leg and as the stepping leg transitions to stance leg and, 2) the ability to actively differentiate between flexion and extension patterns of the stepping and stance leg may be important in the retraining of stepping reactions initiated both by the paretic and nonparetic leg poststroke.
Author Contributions
Concept/idea/research design: C.L. Pollock, M.A. Hunt, S.J. Garland, J.M. Wakeling
Writing: C.L. Pollock, M.A. Hunt, S.J. Garland, J.M. Wakeling
Data collection: C.L. Pollock, S.J. Garland, T.D. Ivanova
Data analysis: C.L. Pollock, M.A. Hunt, S.J. Garland, T.D. Ivanova, J.M. Wakeling
Project management: C.L. Pollock, M.A. Hunt, T.D. Ivanova
Fund procurement: S.J. Garland, M.A. Hunt
Providing facilities/equipment: S.J. Garland, M.A. Hunt
Ethics Approval
This study was approved by the University of British Columbia Clinical Research Ethics Board and conformed to the standards set by the latest revision of the Declaration of Helsinki.
Funding
This work was supported in part by grants from the Natural Sciences and Engineering Research Council of Canada and Canadian Institutes of Health Research (scholarship to C.L.P.).
Disclosures
The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.
References
- 1. Mackintosh SF, Hill KD, Dodd KJ, Goldie PA, Culham EG. Balance score and a history of falls in hospital predict recurrent falls in the 6 months following stroke rehabilitation. Arch Phys Med Rehabil. 2006;87:1583–1589. [DOI] [PubMed] [Google Scholar]
- 2. Yates JS, Lai SM, Duncan PW, Studenski S. Falls in community-dwelling stroke survivors: an accumulated impairments model. J Rehabil Res Dev. 2002;39:385–394. [PubMed] [Google Scholar]
- 3. Gray VL, Pollock CL, Wakeling JM, Ivanova TD, Garland SJ. Patterns of muscle coordination during stepping responses post-stroke. J Electromyogr Kinesiol. 2015;25:959–965. [DOI] [PubMed] [Google Scholar]
- 4. Inness EL, Mansfield A, Bayley M, McIlroy WE. Reactive stepping after stroke: determinants of time to foot off in the paretic and nonparetic limb. J Neurol Phys Ther. 2016;40:196–202. [DOI] [PubMed] [Google Scholar]
- 5. Mansfield A, Inness EL, Lakhani B, McIlroy WE. Determinants of limb preference for initiating compensatory stepping poststroke. Arch Phys Med Rehabil. 2012;93:1179–1184. [DOI] [PubMed] [Google Scholar]
- 6. Gray VL, Ivanova TD, Garland SJ. Effects of fast functional exercise on muscle activity after stroke. Neurorehabil Neural Repair. 2012;26:968–975. [DOI] [PubMed] [Google Scholar]
- 7. Kwakkel G, Van Wegen E, Burridge J, et al. Standardized measurement of quality of upper limb movement after stroke: consensus-based core recommendations from the Second Stroke Recovery and Rehabilitation Roundtable. Int J Stroke. 2019;14:783–791. [DOI] [PubMed] [Google Scholar]
- 8. Wakeling JM, Delaney R, Dudkiewicz I. A method for quantifying dynamic muscle dysfunction in children and young adults with cerebral palsy. Gait Posture. 2007;25:580–589. [DOI] [PubMed] [Google Scholar]
- 9. Mansfield A, Maki BE. Are age-related impairments in change-in-support balance reactions dependent on the method of balance perturbation? J Biomech. 2009;42:1023–1031. [DOI] [PubMed] [Google Scholar]
- 10. Howe JA, Inness EL, Venturini A, Williams JI, Verrier MC. The Community Balance and Mobility Scale—a balance measure for individuals with traumatic brain injury. Clin Rehabil. 2006;20:885–895. [DOI] [PubMed] [Google Scholar]
- 11. Knorr S, Brouwer B, Garland SJ. Validity of the community balance and mobility scale in community-dwelling persons after stroke. Arch Phys Med Rehabil. 2010;91:890–896. [DOI] [PubMed] [Google Scholar]
- 12. Gowland C, Stratford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24:58–63. [DOI] [PubMed] [Google Scholar]
- 13. Gowland C, VanHullenaar S, Torresin W, et al. Chedoke-McMaster Stroke Assessment: Development, Validation and Administration Manual. Hamilton, ON, Canada: Chedoke-McMaster Hospitals and McMaster University; 1995. [Google Scholar]
- 14. Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GV. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res. 1989;7:849–860. [DOI] [PubMed] [Google Scholar]
- 15. Dempster WT, Gaughran GRL. Properties of body segments based on size and weight. Am J Anat. 1967;120:33–54. [Google Scholar]
- 16. Morrison D. Multivariate Statistical Methods. 3rd ed. New York, NY, USA: McGraw-Hill Book Company; 1990. [Google Scholar]
- 17. Wakeling JM, Rozitis AI. Spectral properties of myoelectric signals from different motor units in the leg extensor muscles. J Exp Biol. 2004;207:2519–2528. [DOI] [PubMed] [Google Scholar]
- 18. Portney L, Watkins M, eds.. Foundations of Clinical Research: Application to Practice. 3rd ed. Philadelphia, PA, USA: Prentice-Hall; 2007. [Google Scholar]
- 19. Kajrolkar T, Bhatt T. Falls-risk post-stroke: examining contributions from paretic versus non paretic limbs to unexpected forward gait slips. J Biomech. 2016;49:2702–2708. [DOI] [PubMed] [Google Scholar]
- 20. Worthen-Chaudhari L, Bing J, Schmiedeler JP, Basso DM. A new look at an old problem: defining weight acceptance in human walking. Gait Posture. 2014;39:588–592. [DOI] [PubMed] [Google Scholar]
- 21. Pavol MJ, Pai YC. Deficient limb support is a major contributor to age differences in falling. J Biomech. 2007;40:1318–1325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sousa ASP, Tavares JMRS. Interlimb coordination during step-to-step transition and gait performance. J Mot Behav. 2015;47:563–574. [DOI] [PubMed] [Google Scholar]
- 23. Gowland CA. Staging motor impairment after stroke. Stroke. 1990;21:19–21. [PubMed] [Google Scholar]
- 24. Garland SJ, Gray VL, Knorr S. Muscle activation patterns and postural control following stroke. Motor Control. 2009;13:387–411. [DOI] [PubMed] [Google Scholar]
- 25. Gray V, Rice CL, Garland SJ. Factors that influence muscle weakness following stroke and their clinical implications: a critical review. Physiother Can. 2012;64:415–426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Gray VL, Juren LM, Ivanova TD, Garland SJ. Retraining postural responses with exercises emphasizing speed poststroke. Phys Ther. 2012;92:924–934. [DOI] [PubMed] [Google Scholar]
- 27. Raja B, Neptune RR, Kautz SA. Coordination of the non-paretic leg during hemiparetic gait: Expected and novel compensatory patterns. Clin Biomech. 2012;27:1023–1030. [DOI] [PMC free article] [PubMed] [Google Scholar]