- Email: [email protected]
Manipulating post-stroke gait: Exploiting aberrant kinematics
Accepted Manuscript Manipulating Post-Stroke Gait: Exploiting Aberrant Kinematics Megan E Reissman, Keith E. Gordon, Yasin Y. Dhaher PII: DOI: Reference:
S0021-9290(17)30686-3 https://doi.org/10.1016/j.jbiomech.2017.11.031 BM 8479
To appear in:
Journal of Biomechanics
Accepted Date:
28 November 2017
Please cite this article as: M.E. Reissman, K.E. Gordon, Y.Y. Dhaher, Manipulating Post-Stroke Gait: Exploiting Aberrant Kinematics, Journal of Biomechanics (2017), doi: https://doi.org/10.1016/j.jbiomech.2017.11.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title: Manipulating Post-Stroke Gait: Exploiting Aberrant Kinematics Authors: Megan E Reissman PhD1, Keith E Gordon PhD2,3, Yasin Y Dhaher PhD4,5 1. Department of Mechanical Engineering, University of Dayton, Dayton, Ohio 2. Department of Physical Therapy and Human Movement Sciences, Northwestern University Feinberg School of Medicine, Chicago, Illinois 3. Edward Hines Jr. VA Hospital, Hines, Illinois 4. Department of Biomedical Engineering, Northwestern University, Evanston, Illinois 5. Rehabilitation Institute of Chicago, Chicago, Illinois Address correspondence to: Megan E. Reissman 300 College Park 365E Kettering Labs Dayton, OH 45469 Email: [email protected] Phone: 937-229-5332 Fax: 937-229-2835 Word Count: 3998 Figure Count: 4 Table Count: 2 Keywords: gait, stroke, rehabilitation, kinematics
1
Abstract Post-stroke individuals often exhibit abnormal kinematics, including increased pelvic obliquity and hip abduction coupled with reduced knee flexion. Prior examinations suggest these behaviors are expressions of abnormal cross-planar coupling of muscle activity. However, few studies have detailed the impact of gait-retraining paradigms on three-dimensional joint kinematics. In this study, a cross-tilt walking surface was examined as a novel gait-retraining construct. We hypothesized that relative to baseline walking kinematics, exposure to cross-tilt would generate significant changes in subsequent flat-walking joint kinematics during affected limb swing. Twelve post-stroke participants walked on a motorized treadmill platform during a flat-walking condition and during a 10-degree cross-tilt with affected limb up-slope, increasing toe clearance demand. Individuals completed 15 minutes of cross-tilt walking with intermittent flat-walking catch trials and a final washout period (5 minutes). For flat-walking conditions, we examined changes in pelvic obliquity, hip abduction/adduction and knee flexion kinematics at the spatiotemporal events of swing initiation and toe-off, and the kinematic event of maximum angle during swing. Pelvic obliquity significantly reduced at swing initiation and maximum obliquity in the final catch trial and late washout. Knee flexion significantly increased at swing initiation, toe-off, and maximum flexion across catch trials and late washout. Hip abduction/adduction was not significantly influenced following cross-tilt walking. Significant decrease in the rectus femoris and medial hamstrings muscle activity across catch trials and late washout was observed. Exploiting the abnormal features of post-stroke gait during retraining yielded desirable changes in muscular and kinematic patterns post-training.
2
Introduction Individuals recovering from stroke often exhibit abnormal kinematics, including frontal plane pelvic obliquity coupled with reduced sagittal plane knee flexion. Prior examinations report reduced knee flexion in the affected limb, compared to healthy controls, at toe-off (~14°) and maximum flexion (~20°) (Chen et al., 2005), as well as exaggerations in maximum pelvic obliquity angle of 2° for community ambulators and 5° for limited community ambulators (Stanhope et al., 2014; Cruz and Dhaher, 2009). These abnormal frontal plane kinematics can lead to a circumduction movement pattern (Kerrigan et al., 2000), are energetically costly (Chen et al., 2005), and may increase incidence of sideways falls during gait (Mackintosh et al., 2005). Unfortunetly, few studies have detailed the impact of gait training paradigms on frontal plane joint behaviors (Sulzer et al., 2010; Tyrell et al., 2011; Massaad et al., 2009) and the factors that contribute to these pathologic behaviors remain unclear. Prior work suggested that this specific gait pattern is a compensatory strategy to ensure toe clearance (Perry and Burnfield, 1992). However, recent evidence suggests the origins are more complex, involving abnormal neuromechanical interactions between the pelvis and lower limb. Our recent post stroke isometric studies provide evidence of lack of independent control expressed as abnormal across-joint torque coupling at the hip and knee in a swing initiation limb configuration (Cruz and Dhaher, 2008; Tan and Dhaher, 2014). In particular, we observed hip abduction coupled with knee flexion, and hip adduction coupled with knee extension. The increased expression of these kinetic couplings strongly correlates with abnormal frontal plane pelvic kinematics observed during gait (Cruz et al., 2009). Additionally, an external assistance to knee flexion torque resulted in an increased expression of hip hiking and circumduction in chronic stroke survivors at swing initiation during walking (Sulzer et al., 2010). This finding is
3
inconsistent with the hypothesis that frontal-plane kinematic deviations are a compensation for reduced knee flexion. We argue that a gait-retraining paradigm that simultaneously engages the coupled frontal and sagittal impairments may be critical for modifying these abnormal behaviors and any underlying muscle activation patterns. In the context of the circumduction behavior, we sought to explore if a training paradigm that increases the demand for toe clearance during swing would enhance ability to perform selective control between frontal and sagittal plane degrees-offreedom, expressed as kinematic changes post-exposure. In this paradigm, subjects walked on a custom, motorized treadmill during flat and 10° cross-tilt (frontal plane slope) conditions with their affected limb on the up-slope side. In this context, the cross-tilt increases the mechanical demand for toe clearance of the affected (upslope) limb. The paradigm followed established motor learning constructs including periods of cross-tilt exposure, and flat-walking catch trials and washout phases. During catch trials and washout, we explored changes relative to baseline in frontal (pelvic obliquity and hip abduction/adduction) and sagittal plane (knee flexion) kinematics, particularly at swing initiation and during the swing phase of the gait cycle. We hypothesized that relative to baseline walking kinematics, a significant change in joint kinematics would be expressed during catch and washout periods. Based on our earlier report on the interplay between motor impairment and function (Cruz et al., 2009), we hypothesize that these kinematic changes will be primarily expressed in the knee and pelvic degrees of freedom. We additionally anticipated that significant changes in muscle activation, relative to baseline, would underlie these kinematic changes.
4
Methods Study participants The Northwestern University Institutional Review Board approved the protocol. Informed, written consent was obtained from 12 subjects with hemiparesis due to chronic stroke, aged 55.3±8.7 years (mean±standard deviation), from a single unilateral stroke occurring 74.9±37.5 months prior to testing (Table 1). Exclusion criteria included: stroke less than 3 months prior, stroke with cerebellar or brain stem involvement, history of lower limb surgery or injury, severe cognitive deficits, or concurrent severe medical illness. Study criteria were confirmed via inspection of each participant’s medical record. Subjects with a range of walking speeds and impairment levels were recruited (Table 1). During the study, ankle foot orthosis (AFO) users walked while wearing their device. Any participants who used canes for outdoor walking were capable of independent walking and did not use their cane during the study.
Experimental procedures A licensed physical therapist performed clinical assessments of all participants (Table 1). Gait was assessed at self-selected velocity using the Functional Gait Assessment, Two Minute Walk Test, and Ten Meter Walk Test. Balance was assessed using the Berg Balance Scale, and Timed Up-and-Go protocols. Lower limb motor ability was assessed using the Lower Extremity Fugl-Meyer. Cognitive impairment and ability to provide informed consent were assessed using the Mini-Mental examination. Kinematic data during treadmill walking was recorded at 100-Hz using an 8-camera video system (Motion Analysis Corporation, Santa Rosa, CA) with reflective markers (n=38) placed on the lower limbs, pelvis, torso, and head. Bi-polar surface EMG sensors (Motion Lab Systems, Baton Rouge, LA) recorded bilateral muscle activity data at 1000-Hz from the gluteus
5
medius, adductor magnus/gracilius (adductor), rectus femoris, vastus medialis, and semitendinosus (medial hamstring). EMG ground was placed on the dorsal unaffected hand. Subjects did not receive body weight support but wore a safety harness for fall protection. Footswitch pressure sensors on each foot determined heel-strike and toe-off events. Handrails were not present; however, research staff closely monitored subject position and comfort. Selfselected treadmill walking speed was determined by starting with the subject’s overground speed and making adjustments as requested by the subject. All subjects requested a treadmill speed slower than their overground walking speed, consistent with prior reports of self-selected speed differences between overground and treadmill walking (Bayat et al., 2005). Each subject walked on a single-belt treadmill (Figure 1.A) connected to a motorized platform that could provide a 10° cross-tilt (frontal plane slope). During cross-tilt walking the treadmill was inclined to a 10° position with the affected side of the body on the up-slope side (Figure 1.B). Prior examination of mean toe clearance in chronic stroke gait suggested that a cross-tilt of 2° would generate a zero toe clearance, assuming no adaptations (Cruz and Dhaher, 2009). To insure a clear challenge to toe clearance, a fivefold increase in this theoretical minimum was used. Additionally, in an amputee population, a 10° cross-tilt has been previously shown to generate gait pattern changes during exposure (Villa et al., 2015). Following acclimatization to flat treadmill walking at their self-selected treadmill speed, subjects completed the following protocol (Figure 1.C). Subjects walked for two minutes in a flat (no cross-tilt) treadmill condition to establish their baseline walking behavior. Subjects then completed 15 minutes of a cross-tilt walking exposure period with one minute flat-walking catch trials after 5, 10, and 15 minutes. Subjects were not given any specific guidance nor encouraged to react to the cross-tilt in any particular way. Following Catch 1 and Catch 2 subjects rested for 6
two minutes while standing on the treadmill platform, with longer breaks given as requested. Following Catch 3 subjects continued walking on the flat treadmill for four additional minutes (washout). Transitions between flat and cross-tilt conditions lasted 15 seconds and were automated using closed loop control of a linear actuator. Subjects walked continuously during the transition period.
Experimental analysis In this study we examined the kinematic changes in key lower limb joint angles during the catch and washout periods. Particularly, we focused on changes at key events in the preswing and swing phase of the affected limb. We define pre-swing phase as the period from opposite limb heel-strike to swing limb toe-off. We additionally define the start of the pre-swing phase (opposite limb heel-strike) as a key event, which we call swing initiation. We define swing phase as the period from swing limb toe-off to swing limb heel-strike. Body segment coordinates (head, torso, pelvis, thigh, shank, foot) and relative orientation were defined with commercial software (Visual 3D, C-Motion Inc, Germantown, MD) from a static standing pose and key anatomical markers. Body segment motion and joint angles during walking were tracked using 4-6 markers per segment. As the primary focus of this work, kinematic data was accessed for all flat-walking conditions. Kinematic data from the cross-tilt exposure period is presented for key intervals (minutes 1-5, 10, and 15) to aid in the interpretation of subsequent flat-walking responses. Because motion of the platform may have an additional effect on subject response, walking during transition periods was excluded from the analysis. The primary outcome measures are pelvic obliquity angle, hip abduction/adduction angle, and knee flexion angle. Pelvic obliquity is measured relative to the laboratory coordinate system 7
and defined as positive when the ipsilateral swing side of the pelvis is elevated (Figure 2). Hip angle is measured between the pelvis and femur segments. Knee angle is measured between the femur and tibia/fibula segments. Our examination of these joint angles focuses on changes relative to baseline behavior at the spatiotemporal events of swing initiation and toe-off, and the kinematic event of maximum angle. Maximum angle is defined for each joint as: maximum pelvic obliquity (abduction) (Figure 2), maximum hip abduction (Figure 2), and maximum knee flexion (Figure 3) occurring any time during pre-swing or swing phase (Kerrigan et al., 2000). Being a kinematic event, the timing associated with the maximum angle varied depending on the joint of interest. For each subject and event, the mean angle value during baseline is subtracted from the angles measured in the catch and washout periods. Thus, the baseline mean is zero by definition and values for subsequent conditions represent changes from the baseline behavior (Figure 2, Figure 3). Muscle activation signals were high-pass filtered (20-Hz) with a zero-lag fourth-order Butterworth filter, rectified, and low-pass filtered (12-Hz) with a zero-lag fourth-order Butterworth filter to attain an overall muscle activation envelope. For each muscle, maximum signal value was determined for each step during the baseline period and the maximum values were averaged. Only muscle activations at the spatiotemporal event of toe-off were examined. Prior examinations of stiff knee gait, across pathologies, have been synonymous with muscle activity at toe-off (Reinbolt et. al. 2008) and guided our choice of this event for analysis. Muscle activity at toe-off was normalized to the average gait cycle maximum during baseline walking, a normalization procedure that minimizes sensitivity to outliers. All statistical analysis of significance was performed using NCSS software (v10, Kaysville, Utah). We assessed differences in flat-walking kinematic joint angles (catch and 8
washout) in response to the cross-tilt exposure period. Data for the flat-walking conditions included each step during Baseline, Catch 1, Catch 2, Catch 3, and the final minute of Washout (Late Washout). Kolmogorov-Smirnov test was used to examine normality in the 540 data sets (specific to subject, joint, and event). The test rejected normality for 10% of the data sets, with non-normal sets appearing equally across joints and events. Analysis of variance (ANOVA) was utilized based on good sample size of individual data sets (39.2±7.3) and low occurrence of normality deviations. We compared joint angles using ANOVA with a factor of condition (Baseline, Catch1, Catch2, Catch3, Late Washout). Power values (P) and F-Ratios (F) are given for the main factor of condition. Tukey-Kramer post-hoc testing was used to assess significance (set at p<0.05). All significance values (p) given are for the Tukey-Kramer pairwise comparisons.
Results Kinematic changes At swing initiation, pelvic obliquity (Figure 2) was significantly reduced during Catch 3 and Late Washout compared to Baseline (p<0.05, P=0.87, F=3.93) (Table 2). At toe-off, pelvic obliquity was not significantly different across conditions. Maximum pelvic obliquity was significantly reduced during Catch 3 and Late Washout compared to Baseline (p<0.05, P=0.87, F=4.17). Frontal hip angle (Figure 2) was not significantly different across conditions at swing initiation, toe-off, or maximum abduction (Table 2). At swing initiation, the sagittal knee angle (Figure 3) was significantly increased during Catch 1 and Catch 3 compared to Baseline (p<0.02, P=0.69, F=4.82) (Table 2). The change in sagittal knee angle during Catch 2 and Late Washout approached but was not significant (p<0.07). At toe-off, the sagittal knee angle was significantly increased during Catch 2, Catch 3, 9
and Late Washout compared to Baseline (p<0.02, P=0.96, F=5.53). Maximum knee flexion angle was significantly increased during Catch 2, Catch 3, and Late Washout compared to Baseline (p<0.02, P=0.94, F=5.89). Equivalent analysis of unaffected limb hip and knee angles showed no significant differences.
Muscle activity changes At toe-off, gluteus medius, adductor, and vastus lateralis muscle activity (Figure 4) were not significantly different across conditions compared to Baseline activation values. Significant muscle activity reductions in the rectus femoris (p<0.05, P=0.91, F=4.50) and medial hamstrings (p<0.05, P=0.94, F=5.06) were observed during Catch 2, Catch 3, and Late Washout compared to Baseline (Figure 4).
Discussion Our findings indicate that following short exposures to a cross-tilt walking condition (affected limb high), subjects demonstrated significant kinematic changes in subsequent flatwalking catch trials. Additionally, these changes were sustained after five minutes of flatwalking (Late Washout) following the last cross-tilt exposure period. Relative to Baseline values, pelvic obliquity was significantly reduced at swing initiation and maximum obliquity during Catch 3 and Late Washout. This change in pelvic behavior was accompanied by increased knee flexion at swing initiation, toe-off, and maximum flexion during Catch 3 and Late Washout. No significant changes in hip abduction/adduction angle were observed during Catch trials or Late Washout compared to Baseline values. An examination of the underlying muscle activations at the Catch 3 and Late Washout events indicates that the kinematic changes are associated with reductions in rectus femoris and medial hamstrings activity at toe-off. Increased pelvic obliquity and decreased knee flexion during swing are common kinematic gait deviations in the post stroke
10
population (Chen et al., 2005; Stanhope et al., 2014; Cruz and Dhaher 2009; Kerrigan 2000). Thus, an intervention able to generate reductions in pelvic obliquity, combined with increases in knee flexion, may be beneficial to post stroke individuals demonstrating these abnormal kinematic patterns. The examination of the three-dimensional response to an intervention is critical in our understanding of abnormal kinematics post-stroke. Specifically, interventions that generate sagittal plane improvements may have concomitant undesired changes in frontal plane behaviors (Sulzer et al., 2010). Further investigation of how gait-retraining impacts segmental and joint angle changes at multiple key swing events will shape our understanding of which changes generate the greatest functional impact and should be targeted. Worth noting is that cross-tilt exposure does not appear to negatively impact the swing behavior of the unaffected limb, equivalent analysis of unaffected limb hip and knee angles showed no significant differences. However, significant changes in pelvic obliquity were noted at Catch 3 and Late Washout consistent with those observed during affected swing for the same segment. We argue that the cross-tilt construct generates clinically relevant short-term improvements in both the pelvic obliquity and knee flexion behavior during swing, without any undesired changes in hip abduction behavior.
Clinical significance of joint angle changes The significance of the observed changes in knee flexion and pelvic obliquity are interpreted in the context of the reported training-mediated changes in these metrics in individuals post-stroke. Knee flexion differences between post-stroke and non-impaired subjects at specific swing events have been previously reported (Chen et al., 2005), including reduced knee flexion in the affected limb at toe-off (~14°) and maximum swing phase flexion (~20°). In 11
the current study, we observed at Catch 3 an increase in knee flexion of 6° at swing initiation and 3° at maxmimum flexion. Accompanying increased knee flexion was a 1° reduction in pelvic obliquity at swing initiation and maximum obliquity. Prior examinations report mean maximum pelvic obliquity angles of 2° for community ambulators and 5° for limited community ambulators (Stanhope et al., 2014; Cruz and Dhaher 2009). In this context, the 1° change in the current training construct represents up to 50% change in the maximum pelvic obliquity. Unfortunately, few studies report on changes in pelvic behavior due to a given gait-retraining paradigm; thus meaningful comparison to other retraining approaches is difficult. However, in the context of post-stroke circumduction, inducing a reduction in the pelvic obliquity combined with an increase in knee flexion angle may represent a clinically important change. Use of intramuscular functional electrical stimulation (FES) based intervention poststroke has reported significant kinematic changes (Daly et al., 2004). A mean 10° improvement in maximum knee flexion during swing was observed after 12 weeks of treatment that included FES, strengthening exercises, body-weight supported treadmill training, and overground gait training. Prior pharmacological interventions, particularly botulinum toxin injections to the rectus femoris muscle of the affected limb, have reported enhancements of maximum knee flexion up to 5° at two months post injection (Stoquart et al., 2008). In addition, injections in multiple muscles (rectus femoris, lateral hamstrings, and gastrocnemius) have improved maximum knee flexion up to 5° (Caty et al., 2008). In the current study, we observed up to 6° of knee flexion improvement following cross-tilt exposure. This highlights the potential clinical viability and robustness of our current approach compared with the more complicated aforementioned methodologies. For example, while promising, the invasive nature of the intramuscular FES paradigm likely limits its clinical translation. While statistically significant, 12
the functional impact of observed, “favorable” kinematic joint changes remain to be seen. However, prior studies report improved functional and clinical gait metrics associated with a 10° improvement in knee flexion (Daly et al., 2006), suggesting an association between kinematic changes at the joint/segmental level and functional improvements.
Role of error in gait-retraining paradigms A retained ability following stroke to adapt the gait pattern to perturbation has been observed in other gait-retraining paradigms, and changes are particularly noted for paradigms that exaggerate asymmetry or motor deficts (Reisman et al., 2007; 2009; Savin et al., 2014). The cross-tilt paradigm shares the feature of accentuating asymmetry. Although the current design relies on “error” to drive change, there are interesting differences between the current design and the expected outcomes from classical error augmentation or “error” based constructs. Error augmentation typically yields a decrease in performance errors, a desirable outcome but one that decays over time. Motor learning due to this “error” based training has been attributed to many structures along the neuronal axis, including the cerebellum (Jayaram et al., 2012). This may have occurred in the case of the pelvic degree-of-freedom, in that the tilted treadmill increased pelvic obliquity during exposure and subsequent flat-walking revealed a decrease in pelvic obliquity. Conversely, knee flexion was increased during cross-tilt exposure, yet did not result in a knee extension response in subsequent flat-walking. On the contrary, subjects maintained a similar directional change in knee flexion between the exposure and post-exposure period. While the split-belt treadmill paradigm (Reisman et al., 2007) and the cross-tilt paradigm are introducing error at the endpoint, the split treadmill design defines the outcomes in terms of the endpoint behaviors (ex: step length). In this examination, post-exposure changes are expressed at the joint level. The response to the cross-tilt constraint requires primarily joint-level 13
kinematic changes and engages both sagittal and frontal plane lower limb mechanics. Arguably, the response to the split-belt constraint requires primarily spatiotemporal changes and can be achieved with sagittal plane limb mechanics alone. The differences between the two constructs suggest that the motor control system differentially interprets the imposed challenge. We argue that endpoint errors which require joint level changes rely on joint specific feedback, leading to more direct adaptation strategies in which subjects must consider the interplay between sagittal and frontal plane lower limb mechanics.
Contributions of muscle activity to kinematic changes Our study found decreased activity in the medial hamstrings and rectus femoris following a cross-tilt exposure period. Prior modeling examinations of the effect of toe-off muscle activity on knee flexion velocity suggest that increased force contributions from vasti and rectus femoris will decrease knee flexion velocity (Goldberg et al., 2004). Botulinum toxin injections frequently target the rectus femoris and have been effective in generating long-term increases in peak knee flexion (Stoquart et al., 2008). Our finding of decreased rectus femoris activity following crosstilt exposure supports the idea that interventions which reduce rectus femoris activity may improve post-stroke knee flexion during swing. However, subject response to cross-tilt exposure is inconsistent with the overactive knee extensor hypothesis for the knee extension deviation post-stroke. Based on these observations, we argue that stroke subjects retain a rich repertoire of motor sequences that can be re-activated under the appropriate conditions. It is likely that the rectus femoris activity changes had a cross-joint and cross-planar effect. Electrical stimulation of rectus femoris, biceps femoris, and vastus lateralis have been shown to generate significant frontal plane contributions in a post-stroke population (Hunter et al., 2009). Rectus femoris activity at toe-off decreased during catch trials, a change sustained 14
across washout. Hunter et al. reported a correlation between increased rectus femoris activity at toe-off and increased hip abduction, whereas biceps femoris and vastus lateralis activity correlated with adduction. Please see also the online supplement for further discussion.
Implications for clinical use Although the cross-tilt treadmill system in this study was motorized to allow continuous walking, a standard treadmill raised on one side would generate identical cross-tilt exposure for clinical implementation. It is currently unknown if changes in the kinematic gait pattern following cross-tilt walking exposure would transfer to overground walking. Other treadmillbased gait-retraining paradigms have demonstrated short-term transfer of step length and double stance time changes to overground walking (Savin et al., 2014; Reissman et al., 2009). Though studies suggest that stroke subjects adapt to challenge paradigms more slowly than control subjects, they also appear to de-adapt more slowly, which may increase the feasibility of aftereffect durations in rehabilitation (Savin et al., 2014; Tyrell et al., 2014).
Limitations Our study included both AFO-users and non-users, due to recruitment constraints. Kinematic degrees of freedom in the ankle are arguably altered during AFO use and may impact the response of those subjects during and following the cross-tilt exposure period. In this group, AFO-users had reduced maximum knee flexion at baseline and made larger improvements than non-users on average. This suggests that AFO use does not limit the effectiveness of the intervention. For further characterization, please see the supplementary online material. Overall, further examination with larger, equal sized AFO/non-user populations is warranted to fully determine if AFO use differentially influences cross-tilt response.
15
The preferred treadmill walking speed for the participants was on average 0.40±0.19 m/s slower than their overground speed, consistent with prior studies post-stroke in which preferred treadmill speed was slower than preferred overground walking speed (Bayat et al., 2005). This may have influenced kinematic behaviors in both flat and cross-tilt walking. However, a prior examination found no significant differences in peak frontal plane pelvis and hip angles as a function of increasing treadmill speeds (Tyrell et al., 2011). While slower walking speeds were not examined, this suggests that frontal plane gait behaviors are not sensitive for moderate deviations away from the preferred overground speed. Following stroke, decreased strength in the affected limb is likely to impact compensatory changes in gait kinematics. The potential impact of different levels of joint torque production ability in the affected limb was not assessed. However, this study focused on swing response in which strength differences could be expected to have a smaller impact compared to stance response changes.
Conclusions Gait-retraining paradigms that exaggerate pathological behaviors appear effective in activating more normative motor patterns. Using a cross-tilt walking condition in which subjects’ affected limb was on the up-slope side, we accentuated the required toe clearance during swing. In the post-exposure response, subjects significantly reduced pelvic obliquity and increased knee flexion at key events in pre-swing and swing, without concomitant increases in hip abduction. Appropriate muscle activation pattern changes appear to underlie these kinematic changes, specifically reduced rectus femoris and medial hamstrings activity at toe-off. These changes followed a short 15-minute training period and were maintained over a 5-minute washout; a longer duration relative to other gait-retraining paradigms. Future examinations should explore 16
post-stroke response to repeated cross-tilt walking exposure over several weeks, with longer washout periods to characterize motor retention, and transfer to over ground walking. Overall, clinical implementation of the cross-tilt gait-retraining construct is highly accessible given the simplistic modification of standard treadmill equipment.
Acknowledgments The authors acknowledge Heidi Roth for her assistance in performing clinical assessments and Kathryn Berry for her assistance with data collection and initial processing. This work has been supported by the National Institute of Neurological Disorders and Stroke (R01NS064084-02). Supporting institutions had no role in study design, data collection or analysis, or manuscript preparation.
Conflict of Interest Statement All authors declare that they have no financial or personal conflicts of interest regarding this work.
References 1. Bayat, R., Barbeau, H., Lamontagne, A., 2005. Speed and Temporal-Distance Adaptations during Treadmill and Overground Walking Following Stroke. Neurorehabilitation & Neural Repair. 19, 115-124. 2. Caty, G.D., Detrembleur, C., Bleyenheuft, C., Deltombe, T., Lejeune, T.M., 2008. Effect of Simultaneous Botulinum Toxin Injections Into Several Muscles. Stroke. 39, 28032808. 3. Chen, G., Patten, C., Kothari, D.H., Zajac, F.E., 2005. Gait differences between individuals with post-stroke hemiparesis and non-disabled controls at matched speeds. Gait & Posture. 22, 51-56. 17
4. Cruz, T.H., Dhaher, Y.Y., 2008. Evidence of Abnormal Lower-Limb Torque Coupling After Stroke: An Isometric Study. Stroke. 39, 139-147. 5. Cruz, T.H., Dhaher, Y.Y., 2009. Impact of ankle-foot-orthosis on frontal plane behaviors post-stroke. Gait & Posture. 30, 312-316. 6. Cruz, T.H., Lewek, M.D., Dhaher, Y.Y., 2009. Biomechanical impairments and gait adaptations post-stroke: Multi-factorial associations. J Biomechanics. 42, 1673-1677. 7. Daly, J.J., Roenigk, K.L., Butler, K.M., Gansen, J.L., 2004. Response of sagittal plane gait kinematics to weight-supported treadmill training and functional neuromuscular stimulation following stroke. J Rehabilitation Research Development. 41, 807. 8. Daly, J.J., Roenigk, K.L., Holcomb, P.J., Rogers, J.M., Butler, K.M., Gansen, J.L., Ruff, R.L., 2006. A Randomized Controlled Trial of Functional Neuromuscular Stimulation in Chronic Stroke Subjects. Stroke. 37, 172-178. 9. Goldberg, S.R., Anderson, F.C., Pandy, M.G., Delp, S.L., 2004. Muscles that influence knee flexion velocity in double support: implications for stiff-knee gait. J Biomechanics. 37, 1189-1196. 10. Hunter, B.V., Thelen, D.G., Dhaher, Y.Y., 2009. A Three-Dimensional Biomechanical Evaluation of Quadriceps and Hamstrings Function Using Electrical Stimulation. IEEE Neural Systems Rehabilitation Engineering. 17, 167-175. 11. Jayaram, G., Tang, B., Pallegadda, R., Vasudevan, E.V., Celnik, P., Bastian, A., 2012. Modulating locomotor adaptation with cerebellar stimulation. J Neurophysiology. 107, 2950-2957.
18
12. Kerrigan, D.C., Frates, E.P., Rogan, S., Riley, P.O., 2000. Hip Hiking and Circumduction: Quantitative Definitions. Amer. J Physical Medicine Rehabilitation. 79, 247-252. 13. Lennon, D., Baxter, A., Ashburn, S., 2001. Physiotherapy based on the Bobath concept in stroke rehabilitation: a survey within the UK. Disability and Rehabilitation. 23, 254-262. 14. Mackintosh, S.F., Hill, K., Dodd, K.J., Goldie, P., Culham, E., 2005. Falls and injury prevention should be part of every stroke rehabilitation plan. Clinical Rehabilitation. 19, 441-451. 15. Massaad, F., Lejeune, T.M., Detrembleur, C., 2009. Reducing the Energy Cost Of Hemiparetic Gait Using Center of Mass Feedback. Neurorehabilitation & Neural Repair. 16. Mehrholz, J., Elsner, B., Werner, C., Kugler, J., Pohl, M., 2013. Electromechanical‐assisted training for walking after stroke. Cochrane Library. 17. Perry, J., Burnfield, J.M., 1992. Gait analysis: normal and pathological function. Slack Incorporated, New Jersey. 18. Reinbolt, J.A., Fox, M.D., Arnold, A.S., Õunpuu, S., Delp, S.L., 2008. Importance of preswing rectus femoris activity in stiff-knee gait. J Biomechanics, 41, 2362-2369. 19. Reisman, D.S., Wityk, R., Silver, K., Bastian, A.J., 2007. Locomotor adaptation on a split-belt treadmill can improve walking symmetry post-stroke. Brain. 130, 1861-1872. 20. Reisman, D.S., Wityk, R., Silver, K., Bastian, A.J., 2009. Split-Belt Treadmill Adaptation Transfers to Overground Walking in Persons Poststroke. Neurorehabilitation & Neural Repair. 23, 735-744.
19
21. Savin, D.N., Morton, S.M., Whitall, J., 2014. Generalization of improved step length symmetry from treadmill to overground walking in persons with stroke and hemiparesis. Clinical Neurophysiology. 125, 1012-1020. 22. Stanhope, V.A., Knarr, B.A., Reisman, D.S., Higginson, J.S., 2014. Frontal plane compensatory strategies associated with self-selected walking speed in individuals poststroke. Clinical Biomechanics. 29, 518-522. 23. Stoquart, G.G., Detrembleur, C., Palumbo, S., Deltombe, T., Lejeune, T.M., 2008. Effect of botulinum toxin injection in the rectus femoris on stiff-knee gait in people with stroke. Arch. Physical Medicine Rehabilitation. 89, 56-61. 24. Sulzer, J.S., Gordon, K.E., Dhaher, Y.Y., Peshkin, M.A., Patton, J.L., 2010. Preswing Knee Flexion Assistance Is Coupled With Hip Abduction in People With Stiff-Knee Gait After Stroke. Stroke. 41, 1709-1714. 25. Tan, A.Q., Dhaher, Y.Y., 2014. Evaluation of lower limb cross planar kinetic connectivity signatures post-stroke. J Biomechanics. 47, 949-956. 26. Tyrell, C.M., Helm, E., Reisman, D.S., 2014. Learning the spatial features of a locomotor task is slowed after stroke. J Neurophysiology. 112:480. 27. Tyrell, C.M., Roos, M.A., Rudolph, K.S., Reisman, D.S., 2011. Influence of Systematic Increases in Treadmill Walking Speed on Gait Kinematics After Stroke. Physical Therapy. 91, 392-403. 28. Villa, C., Drevelle, X., Bonnet, X., Lavaste, F., Loiret, I., Fodé, P., Pillet, H., 2015. Evolution of vaulting strategy during locomotion of individuals with transfemoral amputation on slopes and cross-slopes compared to level walking. Clinical Biomechanics. 30, 623-628.
20
Figure Captions Figure 1: Cross-tilt protocol. A) Subjects completed flat treadmill walking during Baseline, Catch, and Washout periods. B) Subjects walked at a 10° cross-tilt with their affected limb on the upslope side during Cross-tilt periods, 15 total minutes. C) Subjects walked continuously on the treadmill platform during transitions between flat and crosstilt walking (15 seconds) which are represented as vertical dashed lines. Transition data is not included in the analysis. Breaks were given following Catch 1 and Catch 2. Figure 2: Changes in frontal pelvis and hip angles. Frontal pelvis angles (pelvic obliquity) reported are relative to the laboratory coordinate system. Change in frontal pelvis angle is positive when the elevation of the ipsilateral swing side of the pelvis is increased. Change in frontal hip angle (abduction/adduction) is positive when abduction is increased. Figure 3: Changes in sagittal knee angles. Change in sagittal knee angle is positive when knee flexion is increased. Figure 4: Affected lower limb muscle activity at toe-off, normalized to the mean of the muscle activity maximums for each Baseline walking step. The group mean at Baseline is extended across each plot for reference as a horizontal gray line.
21
Figure 1
Figure 2
Figure 3
Figure 4
Table 1
Table 1. Subject characteristics and clinical measures Age
Months
Sex
Side
Height
since
Assistive
Belt
Mini
Fugl-
Berg
Device
Speed
Mental
Meyer
Balance
(m)
FGA
TUG
2MWT
10mWT
SSV
SSV
(m/s)
(m/s)
(years)
stroke
S01
59.4
58
M
R
1.73
SC
0.72
28
27
53
21
10.6
0.99
0.89
S02
54.1
62
F
R
1.63
SC/AFO
0.67
29
25
49
19
14.1
0.83
0.88
S03
50.0
53
M
R
1.83
0.89
30
23
56
24
12.1
1.23
1.31
S04
60.7
118
M
R
1.68
SC
0.27
29
19
45
14
21.4
0.52
0.48
S05
67.0
141
F
R
1.60
SC/AFO
0.49
30
22
47
17
15.5
0.83
1.08
S06
44.3
104
M
R
1.75
0.85
29
31
55
22
8.8
1.28
1.19
S07
43.3
69
M
R
1.83
0.49
20
26
51
18
11.7
1.17
1.35
S08
60.6
33
F
L
1.52
SC/AFO
0.40
25
15
48
12
14.5
0.83
0.88
S09
61.0
86
M
L
1.80
SC/AFO
0.54
30
22
50
20
17.4
0.91
0.90
S10
68.4
37
M
R
1.68
0.58
23
23
52
18
14.4
0.88
0.96
S11
46.8
114
F
L
1.68
0.40
28
28
54
16
11.5
0.78
0.88
S12
48.4
24
M
R
1.75
0.36
22
22
53
24
9.6
1.22
1.45
Mean
55.3
74.9
1.71
0.56
26.9
23.6
51.1
18.8
13.5
0.96
1.02
SD
8.7
37.5
0.09
0.19
3.5
4.2
3.4
3.7
3.5
0.23
0.27
(m/s)
AFO
Lower
M – Male, F – Female, R – Right, L – Left, SC – Straight Cane, AFO – Ankle Foot Orthosis, FGA – Functional Gait Assessment, TUG – Timed Up-And-Go, 2MWT – Two Minute Walk Test, 10mWT – Ten Meter Walk Test, SSV – Self Selected Velocity, SD – Standard Deviation
Table 1. Subject characteristics and clinical measures Age
Months
Sex
Side
Height
since
Assistive
Belt
Mini
Fugl-
Berg
Device
Speed
Mental
Meyer
Balance
(m)
FGA
TUG
2MWT
10mWT
SSV
SSV
(m/s)
(m/s)
(years)
stroke
S01
59.4
58
M
R
1.73
SC
0.72
28
27
53
21
10.6
0.99
0.89
S02
54.1
62
F
R
1.63
SC/AFO
0.67
29
25
49
19
14.1
0.83
0.88
S03
50.0
53
M
R
1.83
0.89
30
23
56
24
12.1
1.23
1.31
S04
60.7
118
M
R
1.68
SC
0.27
29
19
45
14
21.4
0.52
0.48
S05
67.0
141
F
R
1.60
SC/AFO
0.49
30
22
47
17
15.5
0.83
1.08
S06
44.3
104
M
R
1.75
0.85
29
31
55
22
8.8
1.28
1.19
S07
43.3
69
M
R
1.83
0.49
20
26
51
18
11.7
1.17
1.35
S08
60.6
33
F
L
1.52
SC/AFO
0.40
25
15
48
12
14.5
0.83
0.88
S09
61.0
86
M
L
1.80
SC/AFO
0.54
30
22
50
20
17.4
0.91
0.90
S10
68.4
37
M
R
1.68
0.58
23
23
52
18
14.4
0.88
0.96
S11
46.8
114
F
L
1.68
0.40
28
28
54
16
11.5
0.78
0.88
S12
48.4
24
M
R
1.75
0.36
22
22
53
24
9.6
1.22
1.45
Mean
55.3
74.9
1.71
0.56
26.9
23.6
51.1
18.8
13.5
0.96
1.02
SD
8.7
37.5
0.09
0.19
3.5
4.2
3.4
3.7
3.5
0.23
0.27
(m/s)
AFO
Lower
M – Male, F – Female, R – Right, L – Left, SC – Straight Cane, AFO – Ankle Foot Orthosis, FGA – Functional Gait Assessment, TUG – Timed Up-And-Go, 2MWT – Two Minute Walk Test, 10mWT – Ten Meter Walk Test, SSV – Self Selected Velocity, SD – Standard Deviation
22
Table 2. Mean group kinematic angle changes Change from Baseline at
Group
Group
Mean at
SD at
Baseline
Baseline
Catch 1
Catch 2
Catch 3
Washout
(degree)
(degree)
(degree)
(degree)
(degree)
(degree)
Late
Pelvic Obliquity (+ increased obliquity) Swing Initiation
1.30
1.77
-0.89
-0.55
-0.92
-1.03
Toe Off
1.45
2.88
-0.28
-0.21
-0.52
-0.68
Maximum Obliquity
3.55
1.82
-0.65
-0.58
-0.80
-0.94
Hip Abduction/Adduction (+ increased abduction) Swing Initiation
-3.75
2.51
+0.85
+0.12
+0.60
+0.27
Toe Off
0.43
3.80
+0.17
-0.08
+0.43
+0.64
Maximum Abduction
2.61
3.83
+0.07
-0.06
+0.29
+0.45
Knee Flexion (+ increased flexion) Swing Initiation
7.2
10.0
+6.4
+5.3
+6.6
+5.3
Toe Off
40.2
7.6
+2.7
+3.5
+4.0
+3.6
Maximum Flexion
48.4
13.0
+1.7
+2.5
+2.8
+2.3
23
S0021-9290(17)30686-3 https://doi.org/10.1016/j.jbiomech.2017.11.031 BM 8479
To appear in:
Journal of Biomechanics
Accepted Date:
28 November 2017
Please cite this article as: M.E. Reissman, K.E. Gordon, Y.Y. Dhaher, Manipulating Post-Stroke Gait: Exploiting Aberrant Kinematics, Journal of Biomechanics (2017), doi: https://doi.org/10.1016/j.jbiomech.2017.11.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title: Manipulating Post-Stroke Gait: Exploiting Aberrant Kinematics Authors: Megan E Reissman PhD1, Keith E Gordon PhD2,3, Yasin Y Dhaher PhD4,5 1. Department of Mechanical Engineering, University of Dayton, Dayton, Ohio 2. Department of Physical Therapy and Human Movement Sciences, Northwestern University Feinberg School of Medicine, Chicago, Illinois 3. Edward Hines Jr. VA Hospital, Hines, Illinois 4. Department of Biomedical Engineering, Northwestern University, Evanston, Illinois 5. Rehabilitation Institute of Chicago, Chicago, Illinois Address correspondence to: Megan E. Reissman 300 College Park 365E Kettering Labs Dayton, OH 45469 Email: [email protected] Phone: 937-229-5332 Fax: 937-229-2835 Word Count: 3998 Figure Count: 4 Table Count: 2 Keywords: gait, stroke, rehabilitation, kinematics
1
Abstract Post-stroke individuals often exhibit abnormal kinematics, including increased pelvic obliquity and hip abduction coupled with reduced knee flexion. Prior examinations suggest these behaviors are expressions of abnormal cross-planar coupling of muscle activity. However, few studies have detailed the impact of gait-retraining paradigms on three-dimensional joint kinematics. In this study, a cross-tilt walking surface was examined as a novel gait-retraining construct. We hypothesized that relative to baseline walking kinematics, exposure to cross-tilt would generate significant changes in subsequent flat-walking joint kinematics during affected limb swing. Twelve post-stroke participants walked on a motorized treadmill platform during a flat-walking condition and during a 10-degree cross-tilt with affected limb up-slope, increasing toe clearance demand. Individuals completed 15 minutes of cross-tilt walking with intermittent flat-walking catch trials and a final washout period (5 minutes). For flat-walking conditions, we examined changes in pelvic obliquity, hip abduction/adduction and knee flexion kinematics at the spatiotemporal events of swing initiation and toe-off, and the kinematic event of maximum angle during swing. Pelvic obliquity significantly reduced at swing initiation and maximum obliquity in the final catch trial and late washout. Knee flexion significantly increased at swing initiation, toe-off, and maximum flexion across catch trials and late washout. Hip abduction/adduction was not significantly influenced following cross-tilt walking. Significant decrease in the rectus femoris and medial hamstrings muscle activity across catch trials and late washout was observed. Exploiting the abnormal features of post-stroke gait during retraining yielded desirable changes in muscular and kinematic patterns post-training.
2
Introduction Individuals recovering from stroke often exhibit abnormal kinematics, including frontal plane pelvic obliquity coupled with reduced sagittal plane knee flexion. Prior examinations report reduced knee flexion in the affected limb, compared to healthy controls, at toe-off (~14°) and maximum flexion (~20°) (Chen et al., 2005), as well as exaggerations in maximum pelvic obliquity angle of 2° for community ambulators and 5° for limited community ambulators (Stanhope et al., 2014; Cruz and Dhaher, 2009). These abnormal frontal plane kinematics can lead to a circumduction movement pattern (Kerrigan et al., 2000), are energetically costly (Chen et al., 2005), and may increase incidence of sideways falls during gait (Mackintosh et al., 2005). Unfortunetly, few studies have detailed the impact of gait training paradigms on frontal plane joint behaviors (Sulzer et al., 2010; Tyrell et al., 2011; Massaad et al., 2009) and the factors that contribute to these pathologic behaviors remain unclear. Prior work suggested that this specific gait pattern is a compensatory strategy to ensure toe clearance (Perry and Burnfield, 1992). However, recent evidence suggests the origins are more complex, involving abnormal neuromechanical interactions between the pelvis and lower limb. Our recent post stroke isometric studies provide evidence of lack of independent control expressed as abnormal across-joint torque coupling at the hip and knee in a swing initiation limb configuration (Cruz and Dhaher, 2008; Tan and Dhaher, 2014). In particular, we observed hip abduction coupled with knee flexion, and hip adduction coupled with knee extension. The increased expression of these kinetic couplings strongly correlates with abnormal frontal plane pelvic kinematics observed during gait (Cruz et al., 2009). Additionally, an external assistance to knee flexion torque resulted in an increased expression of hip hiking and circumduction in chronic stroke survivors at swing initiation during walking (Sulzer et al., 2010). This finding is
3
inconsistent with the hypothesis that frontal-plane kinematic deviations are a compensation for reduced knee flexion. We argue that a gait-retraining paradigm that simultaneously engages the coupled frontal and sagittal impairments may be critical for modifying these abnormal behaviors and any underlying muscle activation patterns. In the context of the circumduction behavior, we sought to explore if a training paradigm that increases the demand for toe clearance during swing would enhance ability to perform selective control between frontal and sagittal plane degrees-offreedom, expressed as kinematic changes post-exposure. In this paradigm, subjects walked on a custom, motorized treadmill during flat and 10° cross-tilt (frontal plane slope) conditions with their affected limb on the up-slope side. In this context, the cross-tilt increases the mechanical demand for toe clearance of the affected (upslope) limb. The paradigm followed established motor learning constructs including periods of cross-tilt exposure, and flat-walking catch trials and washout phases. During catch trials and washout, we explored changes relative to baseline in frontal (pelvic obliquity and hip abduction/adduction) and sagittal plane (knee flexion) kinematics, particularly at swing initiation and during the swing phase of the gait cycle. We hypothesized that relative to baseline walking kinematics, a significant change in joint kinematics would be expressed during catch and washout periods. Based on our earlier report on the interplay between motor impairment and function (Cruz et al., 2009), we hypothesize that these kinematic changes will be primarily expressed in the knee and pelvic degrees of freedom. We additionally anticipated that significant changes in muscle activation, relative to baseline, would underlie these kinematic changes.
4
Methods Study participants The Northwestern University Institutional Review Board approved the protocol. Informed, written consent was obtained from 12 subjects with hemiparesis due to chronic stroke, aged 55.3±8.7 years (mean±standard deviation), from a single unilateral stroke occurring 74.9±37.5 months prior to testing (Table 1). Exclusion criteria included: stroke less than 3 months prior, stroke with cerebellar or brain stem involvement, history of lower limb surgery or injury, severe cognitive deficits, or concurrent severe medical illness. Study criteria were confirmed via inspection of each participant’s medical record. Subjects with a range of walking speeds and impairment levels were recruited (Table 1). During the study, ankle foot orthosis (AFO) users walked while wearing their device. Any participants who used canes for outdoor walking were capable of independent walking and did not use their cane during the study.
Experimental procedures A licensed physical therapist performed clinical assessments of all participants (Table 1). Gait was assessed at self-selected velocity using the Functional Gait Assessment, Two Minute Walk Test, and Ten Meter Walk Test. Balance was assessed using the Berg Balance Scale, and Timed Up-and-Go protocols. Lower limb motor ability was assessed using the Lower Extremity Fugl-Meyer. Cognitive impairment and ability to provide informed consent were assessed using the Mini-Mental examination. Kinematic data during treadmill walking was recorded at 100-Hz using an 8-camera video system (Motion Analysis Corporation, Santa Rosa, CA) with reflective markers (n=38) placed on the lower limbs, pelvis, torso, and head. Bi-polar surface EMG sensors (Motion Lab Systems, Baton Rouge, LA) recorded bilateral muscle activity data at 1000-Hz from the gluteus
5
medius, adductor magnus/gracilius (adductor), rectus femoris, vastus medialis, and semitendinosus (medial hamstring). EMG ground was placed on the dorsal unaffected hand. Subjects did not receive body weight support but wore a safety harness for fall protection. Footswitch pressure sensors on each foot determined heel-strike and toe-off events. Handrails were not present; however, research staff closely monitored subject position and comfort. Selfselected treadmill walking speed was determined by starting with the subject’s overground speed and making adjustments as requested by the subject. All subjects requested a treadmill speed slower than their overground walking speed, consistent with prior reports of self-selected speed differences between overground and treadmill walking (Bayat et al., 2005). Each subject walked on a single-belt treadmill (Figure 1.A) connected to a motorized platform that could provide a 10° cross-tilt (frontal plane slope). During cross-tilt walking the treadmill was inclined to a 10° position with the affected side of the body on the up-slope side (Figure 1.B). Prior examination of mean toe clearance in chronic stroke gait suggested that a cross-tilt of 2° would generate a zero toe clearance, assuming no adaptations (Cruz and Dhaher, 2009). To insure a clear challenge to toe clearance, a fivefold increase in this theoretical minimum was used. Additionally, in an amputee population, a 10° cross-tilt has been previously shown to generate gait pattern changes during exposure (Villa et al., 2015). Following acclimatization to flat treadmill walking at their self-selected treadmill speed, subjects completed the following protocol (Figure 1.C). Subjects walked for two minutes in a flat (no cross-tilt) treadmill condition to establish their baseline walking behavior. Subjects then completed 15 minutes of a cross-tilt walking exposure period with one minute flat-walking catch trials after 5, 10, and 15 minutes. Subjects were not given any specific guidance nor encouraged to react to the cross-tilt in any particular way. Following Catch 1 and Catch 2 subjects rested for 6
two minutes while standing on the treadmill platform, with longer breaks given as requested. Following Catch 3 subjects continued walking on the flat treadmill for four additional minutes (washout). Transitions between flat and cross-tilt conditions lasted 15 seconds and were automated using closed loop control of a linear actuator. Subjects walked continuously during the transition period.
Experimental analysis In this study we examined the kinematic changes in key lower limb joint angles during the catch and washout periods. Particularly, we focused on changes at key events in the preswing and swing phase of the affected limb. We define pre-swing phase as the period from opposite limb heel-strike to swing limb toe-off. We additionally define the start of the pre-swing phase (opposite limb heel-strike) as a key event, which we call swing initiation. We define swing phase as the period from swing limb toe-off to swing limb heel-strike. Body segment coordinates (head, torso, pelvis, thigh, shank, foot) and relative orientation were defined with commercial software (Visual 3D, C-Motion Inc, Germantown, MD) from a static standing pose and key anatomical markers. Body segment motion and joint angles during walking were tracked using 4-6 markers per segment. As the primary focus of this work, kinematic data was accessed for all flat-walking conditions. Kinematic data from the cross-tilt exposure period is presented for key intervals (minutes 1-5, 10, and 15) to aid in the interpretation of subsequent flat-walking responses. Because motion of the platform may have an additional effect on subject response, walking during transition periods was excluded from the analysis. The primary outcome measures are pelvic obliquity angle, hip abduction/adduction angle, and knee flexion angle. Pelvic obliquity is measured relative to the laboratory coordinate system 7
and defined as positive when the ipsilateral swing side of the pelvis is elevated (Figure 2). Hip angle is measured between the pelvis and femur segments. Knee angle is measured between the femur and tibia/fibula segments. Our examination of these joint angles focuses on changes relative to baseline behavior at the spatiotemporal events of swing initiation and toe-off, and the kinematic event of maximum angle. Maximum angle is defined for each joint as: maximum pelvic obliquity (abduction) (Figure 2), maximum hip abduction (Figure 2), and maximum knee flexion (Figure 3) occurring any time during pre-swing or swing phase (Kerrigan et al., 2000). Being a kinematic event, the timing associated with the maximum angle varied depending on the joint of interest. For each subject and event, the mean angle value during baseline is subtracted from the angles measured in the catch and washout periods. Thus, the baseline mean is zero by definition and values for subsequent conditions represent changes from the baseline behavior (Figure 2, Figure 3). Muscle activation signals were high-pass filtered (20-Hz) with a zero-lag fourth-order Butterworth filter, rectified, and low-pass filtered (12-Hz) with a zero-lag fourth-order Butterworth filter to attain an overall muscle activation envelope. For each muscle, maximum signal value was determined for each step during the baseline period and the maximum values were averaged. Only muscle activations at the spatiotemporal event of toe-off were examined. Prior examinations of stiff knee gait, across pathologies, have been synonymous with muscle activity at toe-off (Reinbolt et. al. 2008) and guided our choice of this event for analysis. Muscle activity at toe-off was normalized to the average gait cycle maximum during baseline walking, a normalization procedure that minimizes sensitivity to outliers. All statistical analysis of significance was performed using NCSS software (v10, Kaysville, Utah). We assessed differences in flat-walking kinematic joint angles (catch and 8
washout) in response to the cross-tilt exposure period. Data for the flat-walking conditions included each step during Baseline, Catch 1, Catch 2, Catch 3, and the final minute of Washout (Late Washout). Kolmogorov-Smirnov test was used to examine normality in the 540 data sets (specific to subject, joint, and event). The test rejected normality for 10% of the data sets, with non-normal sets appearing equally across joints and events. Analysis of variance (ANOVA) was utilized based on good sample size of individual data sets (39.2±7.3) and low occurrence of normality deviations. We compared joint angles using ANOVA with a factor of condition (Baseline, Catch1, Catch2, Catch3, Late Washout). Power values (P) and F-Ratios (F) are given for the main factor of condition. Tukey-Kramer post-hoc testing was used to assess significance (set at p<0.05). All significance values (p) given are for the Tukey-Kramer pairwise comparisons.
Results Kinematic changes At swing initiation, pelvic obliquity (Figure 2) was significantly reduced during Catch 3 and Late Washout compared to Baseline (p<0.05, P=0.87, F=3.93) (Table 2). At toe-off, pelvic obliquity was not significantly different across conditions. Maximum pelvic obliquity was significantly reduced during Catch 3 and Late Washout compared to Baseline (p<0.05, P=0.87, F=4.17). Frontal hip angle (Figure 2) was not significantly different across conditions at swing initiation, toe-off, or maximum abduction (Table 2). At swing initiation, the sagittal knee angle (Figure 3) was significantly increased during Catch 1 and Catch 3 compared to Baseline (p<0.02, P=0.69, F=4.82) (Table 2). The change in sagittal knee angle during Catch 2 and Late Washout approached but was not significant (p<0.07). At toe-off, the sagittal knee angle was significantly increased during Catch 2, Catch 3, 9
and Late Washout compared to Baseline (p<0.02, P=0.96, F=5.53). Maximum knee flexion angle was significantly increased during Catch 2, Catch 3, and Late Washout compared to Baseline (p<0.02, P=0.94, F=5.89). Equivalent analysis of unaffected limb hip and knee angles showed no significant differences.
Muscle activity changes At toe-off, gluteus medius, adductor, and vastus lateralis muscle activity (Figure 4) were not significantly different across conditions compared to Baseline activation values. Significant muscle activity reductions in the rectus femoris (p<0.05, P=0.91, F=4.50) and medial hamstrings (p<0.05, P=0.94, F=5.06) were observed during Catch 2, Catch 3, and Late Washout compared to Baseline (Figure 4).
Discussion Our findings indicate that following short exposures to a cross-tilt walking condition (affected limb high), subjects demonstrated significant kinematic changes in subsequent flatwalking catch trials. Additionally, these changes were sustained after five minutes of flatwalking (Late Washout) following the last cross-tilt exposure period. Relative to Baseline values, pelvic obliquity was significantly reduced at swing initiation and maximum obliquity during Catch 3 and Late Washout. This change in pelvic behavior was accompanied by increased knee flexion at swing initiation, toe-off, and maximum flexion during Catch 3 and Late Washout. No significant changes in hip abduction/adduction angle were observed during Catch trials or Late Washout compared to Baseline values. An examination of the underlying muscle activations at the Catch 3 and Late Washout events indicates that the kinematic changes are associated with reductions in rectus femoris and medial hamstrings activity at toe-off. Increased pelvic obliquity and decreased knee flexion during swing are common kinematic gait deviations in the post stroke
10
population (Chen et al., 2005; Stanhope et al., 2014; Cruz and Dhaher 2009; Kerrigan 2000). Thus, an intervention able to generate reductions in pelvic obliquity, combined with increases in knee flexion, may be beneficial to post stroke individuals demonstrating these abnormal kinematic patterns. The examination of the three-dimensional response to an intervention is critical in our understanding of abnormal kinematics post-stroke. Specifically, interventions that generate sagittal plane improvements may have concomitant undesired changes in frontal plane behaviors (Sulzer et al., 2010). Further investigation of how gait-retraining impacts segmental and joint angle changes at multiple key swing events will shape our understanding of which changes generate the greatest functional impact and should be targeted. Worth noting is that cross-tilt exposure does not appear to negatively impact the swing behavior of the unaffected limb, equivalent analysis of unaffected limb hip and knee angles showed no significant differences. However, significant changes in pelvic obliquity were noted at Catch 3 and Late Washout consistent with those observed during affected swing for the same segment. We argue that the cross-tilt construct generates clinically relevant short-term improvements in both the pelvic obliquity and knee flexion behavior during swing, without any undesired changes in hip abduction behavior.
Clinical significance of joint angle changes The significance of the observed changes in knee flexion and pelvic obliquity are interpreted in the context of the reported training-mediated changes in these metrics in individuals post-stroke. Knee flexion differences between post-stroke and non-impaired subjects at specific swing events have been previously reported (Chen et al., 2005), including reduced knee flexion in the affected limb at toe-off (~14°) and maximum swing phase flexion (~20°). In 11
the current study, we observed at Catch 3 an increase in knee flexion of 6° at swing initiation and 3° at maxmimum flexion. Accompanying increased knee flexion was a 1° reduction in pelvic obliquity at swing initiation and maximum obliquity. Prior examinations report mean maximum pelvic obliquity angles of 2° for community ambulators and 5° for limited community ambulators (Stanhope et al., 2014; Cruz and Dhaher 2009). In this context, the 1° change in the current training construct represents up to 50% change in the maximum pelvic obliquity. Unfortunately, few studies report on changes in pelvic behavior due to a given gait-retraining paradigm; thus meaningful comparison to other retraining approaches is difficult. However, in the context of post-stroke circumduction, inducing a reduction in the pelvic obliquity combined with an increase in knee flexion angle may represent a clinically important change. Use of intramuscular functional electrical stimulation (FES) based intervention poststroke has reported significant kinematic changes (Daly et al., 2004). A mean 10° improvement in maximum knee flexion during swing was observed after 12 weeks of treatment that included FES, strengthening exercises, body-weight supported treadmill training, and overground gait training. Prior pharmacological interventions, particularly botulinum toxin injections to the rectus femoris muscle of the affected limb, have reported enhancements of maximum knee flexion up to 5° at two months post injection (Stoquart et al., 2008). In addition, injections in multiple muscles (rectus femoris, lateral hamstrings, and gastrocnemius) have improved maximum knee flexion up to 5° (Caty et al., 2008). In the current study, we observed up to 6° of knee flexion improvement following cross-tilt exposure. This highlights the potential clinical viability and robustness of our current approach compared with the more complicated aforementioned methodologies. For example, while promising, the invasive nature of the intramuscular FES paradigm likely limits its clinical translation. While statistically significant, 12
the functional impact of observed, “favorable” kinematic joint changes remain to be seen. However, prior studies report improved functional and clinical gait metrics associated with a 10° improvement in knee flexion (Daly et al., 2006), suggesting an association between kinematic changes at the joint/segmental level and functional improvements.
Role of error in gait-retraining paradigms A retained ability following stroke to adapt the gait pattern to perturbation has been observed in other gait-retraining paradigms, and changes are particularly noted for paradigms that exaggerate asymmetry or motor deficts (Reisman et al., 2007; 2009; Savin et al., 2014). The cross-tilt paradigm shares the feature of accentuating asymmetry. Although the current design relies on “error” to drive change, there are interesting differences between the current design and the expected outcomes from classical error augmentation or “error” based constructs. Error augmentation typically yields a decrease in performance errors, a desirable outcome but one that decays over time. Motor learning due to this “error” based training has been attributed to many structures along the neuronal axis, including the cerebellum (Jayaram et al., 2012). This may have occurred in the case of the pelvic degree-of-freedom, in that the tilted treadmill increased pelvic obliquity during exposure and subsequent flat-walking revealed a decrease in pelvic obliquity. Conversely, knee flexion was increased during cross-tilt exposure, yet did not result in a knee extension response in subsequent flat-walking. On the contrary, subjects maintained a similar directional change in knee flexion between the exposure and post-exposure period. While the split-belt treadmill paradigm (Reisman et al., 2007) and the cross-tilt paradigm are introducing error at the endpoint, the split treadmill design defines the outcomes in terms of the endpoint behaviors (ex: step length). In this examination, post-exposure changes are expressed at the joint level. The response to the cross-tilt constraint requires primarily joint-level 13
kinematic changes and engages both sagittal and frontal plane lower limb mechanics. Arguably, the response to the split-belt constraint requires primarily spatiotemporal changes and can be achieved with sagittal plane limb mechanics alone. The differences between the two constructs suggest that the motor control system differentially interprets the imposed challenge. We argue that endpoint errors which require joint level changes rely on joint specific feedback, leading to more direct adaptation strategies in which subjects must consider the interplay between sagittal and frontal plane lower limb mechanics.
Contributions of muscle activity to kinematic changes Our study found decreased activity in the medial hamstrings and rectus femoris following a cross-tilt exposure period. Prior modeling examinations of the effect of toe-off muscle activity on knee flexion velocity suggest that increased force contributions from vasti and rectus femoris will decrease knee flexion velocity (Goldberg et al., 2004). Botulinum toxin injections frequently target the rectus femoris and have been effective in generating long-term increases in peak knee flexion (Stoquart et al., 2008). Our finding of decreased rectus femoris activity following crosstilt exposure supports the idea that interventions which reduce rectus femoris activity may improve post-stroke knee flexion during swing. However, subject response to cross-tilt exposure is inconsistent with the overactive knee extensor hypothesis for the knee extension deviation post-stroke. Based on these observations, we argue that stroke subjects retain a rich repertoire of motor sequences that can be re-activated under the appropriate conditions. It is likely that the rectus femoris activity changes had a cross-joint and cross-planar effect. Electrical stimulation of rectus femoris, biceps femoris, and vastus lateralis have been shown to generate significant frontal plane contributions in a post-stroke population (Hunter et al., 2009). Rectus femoris activity at toe-off decreased during catch trials, a change sustained 14
across washout. Hunter et al. reported a correlation between increased rectus femoris activity at toe-off and increased hip abduction, whereas biceps femoris and vastus lateralis activity correlated with adduction. Please see also the online supplement for further discussion.
Implications for clinical use Although the cross-tilt treadmill system in this study was motorized to allow continuous walking, a standard treadmill raised on one side would generate identical cross-tilt exposure for clinical implementation. It is currently unknown if changes in the kinematic gait pattern following cross-tilt walking exposure would transfer to overground walking. Other treadmillbased gait-retraining paradigms have demonstrated short-term transfer of step length and double stance time changes to overground walking (Savin et al., 2014; Reissman et al., 2009). Though studies suggest that stroke subjects adapt to challenge paradigms more slowly than control subjects, they also appear to de-adapt more slowly, which may increase the feasibility of aftereffect durations in rehabilitation (Savin et al., 2014; Tyrell et al., 2014).
Limitations Our study included both AFO-users and non-users, due to recruitment constraints. Kinematic degrees of freedom in the ankle are arguably altered during AFO use and may impact the response of those subjects during and following the cross-tilt exposure period. In this group, AFO-users had reduced maximum knee flexion at baseline and made larger improvements than non-users on average. This suggests that AFO use does not limit the effectiveness of the intervention. For further characterization, please see the supplementary online material. Overall, further examination with larger, equal sized AFO/non-user populations is warranted to fully determine if AFO use differentially influences cross-tilt response.
15
The preferred treadmill walking speed for the participants was on average 0.40±0.19 m/s slower than their overground speed, consistent with prior studies post-stroke in which preferred treadmill speed was slower than preferred overground walking speed (Bayat et al., 2005). This may have influenced kinematic behaviors in both flat and cross-tilt walking. However, a prior examination found no significant differences in peak frontal plane pelvis and hip angles as a function of increasing treadmill speeds (Tyrell et al., 2011). While slower walking speeds were not examined, this suggests that frontal plane gait behaviors are not sensitive for moderate deviations away from the preferred overground speed. Following stroke, decreased strength in the affected limb is likely to impact compensatory changes in gait kinematics. The potential impact of different levels of joint torque production ability in the affected limb was not assessed. However, this study focused on swing response in which strength differences could be expected to have a smaller impact compared to stance response changes.
Conclusions Gait-retraining paradigms that exaggerate pathological behaviors appear effective in activating more normative motor patterns. Using a cross-tilt walking condition in which subjects’ affected limb was on the up-slope side, we accentuated the required toe clearance during swing. In the post-exposure response, subjects significantly reduced pelvic obliquity and increased knee flexion at key events in pre-swing and swing, without concomitant increases in hip abduction. Appropriate muscle activation pattern changes appear to underlie these kinematic changes, specifically reduced rectus femoris and medial hamstrings activity at toe-off. These changes followed a short 15-minute training period and were maintained over a 5-minute washout; a longer duration relative to other gait-retraining paradigms. Future examinations should explore 16
post-stroke response to repeated cross-tilt walking exposure over several weeks, with longer washout periods to characterize motor retention, and transfer to over ground walking. Overall, clinical implementation of the cross-tilt gait-retraining construct is highly accessible given the simplistic modification of standard treadmill equipment.
Acknowledgments The authors acknowledge Heidi Roth for her assistance in performing clinical assessments and Kathryn Berry for her assistance with data collection and initial processing. This work has been supported by the National Institute of Neurological Disorders and Stroke (R01NS064084-02). Supporting institutions had no role in study design, data collection or analysis, or manuscript preparation.
Conflict of Interest Statement All authors declare that they have no financial or personal conflicts of interest regarding this work.
References 1. Bayat, R., Barbeau, H., Lamontagne, A., 2005. Speed and Temporal-Distance Adaptations during Treadmill and Overground Walking Following Stroke. Neurorehabilitation & Neural Repair. 19, 115-124. 2. Caty, G.D., Detrembleur, C., Bleyenheuft, C., Deltombe, T., Lejeune, T.M., 2008. Effect of Simultaneous Botulinum Toxin Injections Into Several Muscles. Stroke. 39, 28032808. 3. Chen, G., Patten, C., Kothari, D.H., Zajac, F.E., 2005. Gait differences between individuals with post-stroke hemiparesis and non-disabled controls at matched speeds. Gait & Posture. 22, 51-56. 17
4. Cruz, T.H., Dhaher, Y.Y., 2008. Evidence of Abnormal Lower-Limb Torque Coupling After Stroke: An Isometric Study. Stroke. 39, 139-147. 5. Cruz, T.H., Dhaher, Y.Y., 2009. Impact of ankle-foot-orthosis on frontal plane behaviors post-stroke. Gait & Posture. 30, 312-316. 6. Cruz, T.H., Lewek, M.D., Dhaher, Y.Y., 2009. Biomechanical impairments and gait adaptations post-stroke: Multi-factorial associations. J Biomechanics. 42, 1673-1677. 7. Daly, J.J., Roenigk, K.L., Butler, K.M., Gansen, J.L., 2004. Response of sagittal plane gait kinematics to weight-supported treadmill training and functional neuromuscular stimulation following stroke. J Rehabilitation Research Development. 41, 807. 8. Daly, J.J., Roenigk, K.L., Holcomb, P.J., Rogers, J.M., Butler, K.M., Gansen, J.L., Ruff, R.L., 2006. A Randomized Controlled Trial of Functional Neuromuscular Stimulation in Chronic Stroke Subjects. Stroke. 37, 172-178. 9. Goldberg, S.R., Anderson, F.C., Pandy, M.G., Delp, S.L., 2004. Muscles that influence knee flexion velocity in double support: implications for stiff-knee gait. J Biomechanics. 37, 1189-1196. 10. Hunter, B.V., Thelen, D.G., Dhaher, Y.Y., 2009. A Three-Dimensional Biomechanical Evaluation of Quadriceps and Hamstrings Function Using Electrical Stimulation. IEEE Neural Systems Rehabilitation Engineering. 17, 167-175. 11. Jayaram, G., Tang, B., Pallegadda, R., Vasudevan, E.V., Celnik, P., Bastian, A., 2012. Modulating locomotor adaptation with cerebellar stimulation. J Neurophysiology. 107, 2950-2957.
18
12. Kerrigan, D.C., Frates, E.P., Rogan, S., Riley, P.O., 2000. Hip Hiking and Circumduction: Quantitative Definitions. Amer. J Physical Medicine Rehabilitation. 79, 247-252. 13. Lennon, D., Baxter, A., Ashburn, S., 2001. Physiotherapy based on the Bobath concept in stroke rehabilitation: a survey within the UK. Disability and Rehabilitation. 23, 254-262. 14. Mackintosh, S.F., Hill, K., Dodd, K.J., Goldie, P., Culham, E., 2005. Falls and injury prevention should be part of every stroke rehabilitation plan. Clinical Rehabilitation. 19, 441-451. 15. Massaad, F., Lejeune, T.M., Detrembleur, C., 2009. Reducing the Energy Cost Of Hemiparetic Gait Using Center of Mass Feedback. Neurorehabilitation & Neural Repair. 16. Mehrholz, J., Elsner, B., Werner, C., Kugler, J., Pohl, M., 2013. Electromechanical‐assisted training for walking after stroke. Cochrane Library. 17. Perry, J., Burnfield, J.M., 1992. Gait analysis: normal and pathological function. Slack Incorporated, New Jersey. 18. Reinbolt, J.A., Fox, M.D., Arnold, A.S., Õunpuu, S., Delp, S.L., 2008. Importance of preswing rectus femoris activity in stiff-knee gait. J Biomechanics, 41, 2362-2369. 19. Reisman, D.S., Wityk, R., Silver, K., Bastian, A.J., 2007. Locomotor adaptation on a split-belt treadmill can improve walking symmetry post-stroke. Brain. 130, 1861-1872. 20. Reisman, D.S., Wityk, R., Silver, K., Bastian, A.J., 2009. Split-Belt Treadmill Adaptation Transfers to Overground Walking in Persons Poststroke. Neurorehabilitation & Neural Repair. 23, 735-744.
19
21. Savin, D.N., Morton, S.M., Whitall, J., 2014. Generalization of improved step length symmetry from treadmill to overground walking in persons with stroke and hemiparesis. Clinical Neurophysiology. 125, 1012-1020. 22. Stanhope, V.A., Knarr, B.A., Reisman, D.S., Higginson, J.S., 2014. Frontal plane compensatory strategies associated with self-selected walking speed in individuals poststroke. Clinical Biomechanics. 29, 518-522. 23. Stoquart, G.G., Detrembleur, C., Palumbo, S., Deltombe, T., Lejeune, T.M., 2008. Effect of botulinum toxin injection in the rectus femoris on stiff-knee gait in people with stroke. Arch. Physical Medicine Rehabilitation. 89, 56-61. 24. Sulzer, J.S., Gordon, K.E., Dhaher, Y.Y., Peshkin, M.A., Patton, J.L., 2010. Preswing Knee Flexion Assistance Is Coupled With Hip Abduction in People With Stiff-Knee Gait After Stroke. Stroke. 41, 1709-1714. 25. Tan, A.Q., Dhaher, Y.Y., 2014. Evaluation of lower limb cross planar kinetic connectivity signatures post-stroke. J Biomechanics. 47, 949-956. 26. Tyrell, C.M., Helm, E., Reisman, D.S., 2014. Learning the spatial features of a locomotor task is slowed after stroke. J Neurophysiology. 112:480. 27. Tyrell, C.M., Roos, M.A., Rudolph, K.S., Reisman, D.S., 2011. Influence of Systematic Increases in Treadmill Walking Speed on Gait Kinematics After Stroke. Physical Therapy. 91, 392-403. 28. Villa, C., Drevelle, X., Bonnet, X., Lavaste, F., Loiret, I., Fodé, P., Pillet, H., 2015. Evolution of vaulting strategy during locomotion of individuals with transfemoral amputation on slopes and cross-slopes compared to level walking. Clinical Biomechanics. 30, 623-628.
20
Figure Captions Figure 1: Cross-tilt protocol. A) Subjects completed flat treadmill walking during Baseline, Catch, and Washout periods. B) Subjects walked at a 10° cross-tilt with their affected limb on the upslope side during Cross-tilt periods, 15 total minutes. C) Subjects walked continuously on the treadmill platform during transitions between flat and crosstilt walking (15 seconds) which are represented as vertical dashed lines. Transition data is not included in the analysis. Breaks were given following Catch 1 and Catch 2. Figure 2: Changes in frontal pelvis and hip angles. Frontal pelvis angles (pelvic obliquity) reported are relative to the laboratory coordinate system. Change in frontal pelvis angle is positive when the elevation of the ipsilateral swing side of the pelvis is increased. Change in frontal hip angle (abduction/adduction) is positive when abduction is increased. Figure 3: Changes in sagittal knee angles. Change in sagittal knee angle is positive when knee flexion is increased. Figure 4: Affected lower limb muscle activity at toe-off, normalized to the mean of the muscle activity maximums for each Baseline walking step. The group mean at Baseline is extended across each plot for reference as a horizontal gray line.
21
Figure 1
Figure 2
Figure 3
Figure 4
Table 1
Table 1. Subject characteristics and clinical measures Age
Months
Sex
Side
Height
since
Assistive
Belt
Mini
Fugl-
Berg
Device
Speed
Mental
Meyer
Balance
(m)
FGA
TUG
2MWT
10mWT
SSV
SSV
(m/s)
(m/s)
(years)
stroke
S01
59.4
58
M
R
1.73
SC
0.72
28
27
53
21
10.6
0.99
0.89
S02
54.1
62
F
R
1.63
SC/AFO
0.67
29
25
49
19
14.1
0.83
0.88
S03
50.0
53
M
R
1.83
0.89
30
23
56
24
12.1
1.23
1.31
S04
60.7
118
M
R
1.68
SC
0.27
29
19
45
14
21.4
0.52
0.48
S05
67.0
141
F
R
1.60
SC/AFO
0.49
30
22
47
17
15.5
0.83
1.08
S06
44.3
104
M
R
1.75
0.85
29
31
55
22
8.8
1.28
1.19
S07
43.3
69
M
R
1.83
0.49
20
26
51
18
11.7
1.17
1.35
S08
60.6
33
F
L
1.52
SC/AFO
0.40
25
15
48
12
14.5
0.83
0.88
S09
61.0
86
M
L
1.80
SC/AFO
0.54
30
22
50
20
17.4
0.91
0.90
S10
68.4
37
M
R
1.68
0.58
23
23
52
18
14.4
0.88
0.96
S11
46.8
114
F
L
1.68
0.40
28
28
54
16
11.5
0.78
0.88
S12
48.4
24
M
R
1.75
0.36
22
22
53
24
9.6
1.22
1.45
Mean
55.3
74.9
1.71
0.56
26.9
23.6
51.1
18.8
13.5
0.96
1.02
SD
8.7
37.5
0.09
0.19
3.5
4.2
3.4
3.7
3.5
0.23
0.27
(m/s)
AFO
Lower
M – Male, F – Female, R – Right, L – Left, SC – Straight Cane, AFO – Ankle Foot Orthosis, FGA – Functional Gait Assessment, TUG – Timed Up-And-Go, 2MWT – Two Minute Walk Test, 10mWT – Ten Meter Walk Test, SSV – Self Selected Velocity, SD – Standard Deviation
Table 1. Subject characteristics and clinical measures Age
Months
Sex
Side
Height
since
Assistive
Belt
Mini
Fugl-
Berg
Device
Speed
Mental
Meyer
Balance
(m)
FGA
TUG
2MWT
10mWT
SSV
SSV
(m/s)
(m/s)
(years)
stroke
S01
59.4
58
M
R
1.73
SC
0.72
28
27
53
21
10.6
0.99
0.89
S02
54.1
62
F
R
1.63
SC/AFO
0.67
29
25
49
19
14.1
0.83
0.88
S03
50.0
53
M
R
1.83
0.89
30
23
56
24
12.1
1.23
1.31
S04
60.7
118
M
R
1.68
SC
0.27
29
19
45
14
21.4
0.52
0.48
S05
67.0
141
F
R
1.60
SC/AFO
0.49
30
22
47
17
15.5
0.83
1.08
S06
44.3
104
M
R
1.75
0.85
29
31
55
22
8.8
1.28
1.19
S07
43.3
69
M
R
1.83
0.49
20
26
51
18
11.7
1.17
1.35
S08
60.6
33
F
L
1.52
SC/AFO
0.40
25
15
48
12
14.5
0.83
0.88
S09
61.0
86
M
L
1.80
SC/AFO
0.54
30
22
50
20
17.4
0.91
0.90
S10
68.4
37
M
R
1.68
0.58
23
23
52
18
14.4
0.88
0.96
S11
46.8
114
F
L
1.68
0.40
28
28
54
16
11.5
0.78
0.88
S12
48.4
24
M
R
1.75
0.36
22
22
53
24
9.6
1.22
1.45
Mean
55.3
74.9
1.71
0.56
26.9
23.6
51.1
18.8
13.5
0.96
1.02
SD
8.7
37.5
0.09
0.19
3.5
4.2
3.4
3.7
3.5
0.23
0.27
(m/s)
AFO
Lower
M – Male, F – Female, R – Right, L – Left, SC – Straight Cane, AFO – Ankle Foot Orthosis, FGA – Functional Gait Assessment, TUG – Timed Up-And-Go, 2MWT – Two Minute Walk Test, 10mWT – Ten Meter Walk Test, SSV – Self Selected Velocity, SD – Standard Deviation
22
Table 2. Mean group kinematic angle changes Change from Baseline at
Group
Group
Mean at
SD at
Baseline
Baseline
Catch 1
Catch 2
Catch 3
Washout
(degree)
(degree)
(degree)
(degree)
(degree)
(degree)
Late
Pelvic Obliquity (+ increased obliquity) Swing Initiation
1.30
1.77
-0.89
-0.55
-0.92
-1.03
Toe Off
1.45
2.88
-0.28
-0.21
-0.52
-0.68
Maximum Obliquity
3.55
1.82
-0.65
-0.58
-0.80
-0.94
Hip Abduction/Adduction (+ increased abduction) Swing Initiation
-3.75
2.51
+0.85
+0.12
+0.60
+0.27
Toe Off
0.43
3.80
+0.17
-0.08
+0.43
+0.64
Maximum Abduction
2.61
3.83
+0.07
-0.06
+0.29
+0.45
Knee Flexion (+ increased flexion) Swing Initiation
7.2
10.0
+6.4
+5.3
+6.6
+5.3
Toe Off
40.2
7.6
+2.7
+3.5
+4.0
+3.6
Maximum Flexion
48.4
13.0
+1.7
+2.5
+2.8
+2.3
23