Elsevier

Journal of Biomechanics

Volume 53, 28 February 2017, Pages 136-143
Journal of Biomechanics

Two biomechanical strategies for locomotor adaptation to split-belt treadmill walking in subjects with and without transtibial amputation

https://doi.org/10.1016/j.jbiomech.2017.01.012Get rights and content

Abstract

Locomotor adaptation is commonly studied using split-belt treadmill walking, in which each foot is placed on a belt moving at a different speed. As subjects adapt to split-belt walking, they reduce metabolic power, but the biomechanical mechanism behind this improved efficiency is unknown. Analyzing mechanical work performed by the legs and joints during split-belt adaptation could reveal this mechanism. Because ankle work in the step-to-step transition is more efficient than hip work, we hypothesized that control subjects would reduce hip work on the fast belt and increase ankle work during the step-to-step transition as they adapted. We further hypothesized that subjects with unilateral, trans-tibial amputation would instead increase propulsive work from their intact leg on the slow belt. Control subjects reduced hip work and shifted more ankle work to the step-to-step transition, supporting our hypothesis. Contrary to our second hypothesis, intact leg work, ankle work and hip work in amputees were unchanged during adaptation. Furthermore, all subjects increased collisional energy loss on the fast belt, but did not increase propulsive work. This was possible because subjects moved further backward during fast leg single support in late adaptation than in early adaptation, compensating by reducing backward movement in slow leg single support. In summary, subjects used two strategies to improve mechanical efficiency in split-belt walking adaptation: a CoM displacement strategy that allows for less forward propulsion on the fast belt; and, an ankle timing strategy that allows efficient ankle work in the step-to-step transition to increase while reducing inefficient hip work.

Introduction

When walking under novel conditions, the nervous system gradually modifies locomotor control in response to sensory feedback. Adaptation is characterized biomechanically by a movement parameter changing gradually in response to an altered environment (Martin et al., 1996) and by aftereffects, which linger when the environmental perturbation is removed. Because aftereffects do not immediately dissipate when the perturbation is removed, they indicate predictive control of movement (Ogawa et al., 2014, Kagerer et al., 2007).

Locomotor adaptation can be studied during split-belt treadmill walking, in which each belt moves at a different speed (Prokop et al., 1995). Over several minutes of split-belt walking, interlimb parameters such as step length symmetry and double support time adapt (Reisman et al., 2005). Step length symmetry adaptation is a robust finding replicated in stroke survivors (Reisman et al., 2007, Reisman et al., 2013), Parkinson׳s disease patients (Roemmich et al., 2014a, Roemmich et al., 2014b), and many conditions in healthy subjects (Malone and Bastian, 2010, Torres-Oviedo and Bastian, 2010, Torres-Oviedo and Bastian, 2012, Finley et al., 2014). During split-belt walking, ground reaction forces (GRF) exhibit predictive changes at initial contact, although vertical GRF in single support and propulsive GRF changes immediately, indicating reactive control (Mawase et al., 2013, Ogawa et al., 2014). Over the course of split-belt adaptation, subjects reduce muscle activity and metabolic power (Finley et al., 2013).

Metabolic power decreases during split-belt adaptation, but changes in joint mechanical work that may cause this improved efficiency are unknown. During steady walking, metabolic cost is determined largely by trailing leg work in the step-to-step transition (STS), which increases with metabolic rate (Donelan et al., 2002a). Trailing leg work in STS is more efficient than single support work (Kuo, 2002) and comes predominantly from the ankle while single support work comes more from the hip (Farris and Sawicki, 2012, Zelik et al., 2015). This result explains increased metabolic cost in trans-tibial amputees (Houdijk et al., 2009a, Houdijk et al., 2009b, Adamczyk and Kuo, 2015) and healthy subjects with immobilized ankles (Wutzke et al., 2012). However, changes in joint work during split-belt adaptation are largely unknown. Only ankle work in late stance and hip work in swing have been quantified in split-belt adaptation (Roemmich et al., 2014a, Hinkel-Lipsker and Hahn, 2016). Furthermore, split-belt walking has not been tested on trans-tibial amputees. Because trans-tibial amputees have limited propulsion from their prosthesis (Silverman et al., 2008, Fey et al., 2011), split-belt walking with the prosthesis on the fast belt is a useful model of compensation during locomotor adaptation.

The purposes of this study were to: (1) determine how joint work changes as subjects adapt to split-belt walking; and, (2) explore biomechanical compensation mechanisms trans-tibial amputees use during split-belt walking. Our primary hypothesis is that controls would reduce hip work and increase trailing leg STS ankle work on the fast belt over adaptation. We propose this switch from hip work to trailing ankle work as a joint-level biomechanical mechanism for reducing metabolic power during split-belt adaptation. Our primary focus was work from the fast leg joints, which have much higher work than slow leg joints and should thus contribute more to metabolic power. Given that amputees lack a biological ankle on the fast belt, have less adaptability on the amputated side (Houdijk et al., 2012), and do more work with the intact than the amputated leg (Houdijk et al., 2009a, Houdijk et al., 2009b, Adamczyk and Kuo, 2015), we further hypothesized that amputees would be unable to increase fast (prosthetic) leg ankle work and would instead increase slow (intact) leg work as they adapted. Lastly, based on decreasing slow leg braking force during able-bodied split-belt adaptation (Ogawa et al., 2014), we hypothesize that, for both amputees and controls, slow leg collisional energy loss would decrease as subjects adapted.

Section snippets

Subjects

Eight trans-tibial amputees (6 male, 1 congenital, 7 traumatic, BW:80.4±16.9 kg, intact leg length:92.0±6.4 cm) and eight matched controls (6 male, BW:81.5±14.1 kg, leg length:91.8±4.7 cm) gave informed consent prior to participating in this study approved by the Georgia Tech Institutional Review Board. Amputees wore their own prostheses with energy storage and return feet, passive elastic feet that provide more propulsion than a solid-ankle cushioned heel foot but less than a powered ankle (

Results

All subjects displayed stereotypical responses in step kinematics. Compared to baseline, both controls and amputees exhibited negative step length asymmetry in early adaptation (controls: ES=2.74; p=0.034; amputees: ES=1.51; p=0.013) and positive asymmetry in early post-adaptation (controls: p=0.031; amputees: p=0.031). Although they appear to overshoot zero asymmetry as they adapt, neither group significantly differed from baseline step length symmetry in late adaptation (controls: ES=1.17; p

Step length symmetry has typical adaptation pattern in both controls and amputees

Changes in step length symmetry were consistent with previous studies (Reisman et al., 2005, Malone and Bastian, 2010, Vazquez et al., 2015). Both controls and amputees initially responded to split-belt walking by taking longer steps with the slow leg leading, but returned to baseline step length symmetry by late adaptation. Both groups had the opposite aftereffect in early post-adaptation, indicating predictive control. Therefore, the lack of sensory feedback in the amputees’ distal limb did

Conclusion

Control subjects adapt to split-belt walking using two strategies: they change ankle work timing, increasing fast ankle work during STS while reducing inefficient hip work, and they allow their CoM to move farther backward during fast single support while limiting backward motion on the slow belt. Because amputees had their prostheses on the fast belt, they quickly relied on the CoM displacement strategy throughout adaptation, but they still adapted step length symmetry without using the ankle

Conflicts of interest

The authors have no conflicts of interest to disclose.

Acknowledgments

This research was funded through NIH Grants NICHD 5T32-HD-055180 and NINDS 5R01NS069655.

References (41)

  • J. Yeom et al.

    A gravitational impulse model predicts collision impulse and mechanical work during a step-to-step transition

    J. Biomech.

    (2011)
  • P.G. Adamczyk et al.

    Mechanisms of gait asymmetry due to push-off deficiency in unilateral amputees

    IEEE Trans. Neural Syst. Rehab Eng.

    (2015)
  • P.G. Adamczyk et al.

    Redirection of center-of-mass velocity during the step-to-step transition of human walking

    J. Exp. Biol.

    (2009)
  • J.M. Donelan et al.

    Mechanical work for step-to-step transitions is a major determinant of the metabolic cost of human walking

    J. Exp. Biol.

    (2002)
  • D.J. Farris et al.

    The mechanics and energetics of human walking and running: a joint level perspective

    J. R. Soc. Interface

    (2012)
  • J.M. Finley et al.

    Learning to be economical: the energy cost of walking tracks motor adaptation

    J. Physiol.

    (2013)
  • J.M. Finley et al.

    A novel optic flow pattern speeds split-belt locomotor adaptation

    J. Neurophysiol.

    (2014)
  • H.M. Herr et al.

    Bionic ankle-foot prosthesis normalizes walking gait for persons with leg amputation

    Proc. R. Soc. B

    (2012)
  • J.W. Hinkel-Lipsker et al.

    Novel kinetic strategies adopted in asymmetric split-belt treadmill walking

    J. Mot. Behav.

    (2016)
  • H. Houdijk et al.

    Assessing gait adaptability in people with a unilateral amputation on an instrumented treadmill with projected visual context

    Phys. Ther.

    (2009)
  • Cited by (0)

    View full text