Impairments in the mechanical effectiveness of reactive balance control strategies during walking in people post-stroke

Abstract

People post-stroke have an increased risk of falls compared to neurotypical individuals, partly resulting from an inability to generate appropriate reactions to restore balance. However, few studies investigated the effect of paretic deficits on the mechanics of reactive control strategies following forward losses of balance during walking. Here, we characterized the biomechanical consequences of reactive control strategies following perturbations induced by the treadmill belt accelerations. Thirty-eight post-stroke participants and thirteen age-matched and speed-matched neurotypical participants walked on a dual-belt treadmill while receiving perturbations that induced a forward loss of balance. We computed whole-body angular momentum and angular impulse using segment kinematics and reaction forces to quantify the effect of impulse generation by both the leading and trailing limbs in response to perturbations in the sagittal plane. We found that perturbations to the paretic limb led to larger increases in forward angular momentum during the perturbation step than perturbations to the non-paretic limb or to neurotypical individuals. To recover from the forward loss of balance, neurotypical individuals coordinated reaction forces generated by both legs to decrease the forward angular impulse relative to the pre-perturbation step. They first decreased the forward pitch angular impulse during the perturbation step. Then, during the first recovery step, they increased the backward angular impulse by the leading limb and decreased the forward angular impulse by the trailing limb. In contrast to neurotypical participants, people post-stroke did not reduce the forward angular impulse generated by the stance limb during the perturbed step. They also did not increase leading limb angular impulse or decrease the forward trailing limb angular impulse using their paretic limb during the first recovery step. Lastly, post-stroke individuals who scored poorer on clinical assessments of balance and had greater motor impairment made less use of the paretic limb to reduce forward momentum. Overall, these results suggest that paretic deficits limit the ability to recover from forward loss of balance. Future perturbation-based balance training targeting reactive stepping response in stroke populations may benefit from improving the ability to modulate paretic ground reaction forces to better control whole-body dynamics.

Does ankle push-off correct for errors in anterior-posterior foot placement relative to center-of-mass states?

“…Wang & Srinivasan, (2014) showed that CoM states at mid-stance and heel strike predict about 33% of the variance in AP foot placement. Recently, Liu, (2021) replicated this finding in steady-state walking and additionally in perturbed walking. This study by Liu revealed that a different foot placement model is needed to describe responses to slip-like and trip-like treadmill perturbations, suggesting that the control of foot placement may be different between steady-state and perturbed walking.…”

Gait speed and individual characteristics can be used to predict specific gait metric magnitudes in neurotypical adults

“…The independent dataset used to validate the final models consisted of 23 neurotypical adults (age: 61 ± 15 years (24-77 years); leg length: 0.78 ± 0.06 meters (0.67-0.88 meters); mass: 72 ± 18 kg (37-103 kg)) walking at their self-selected speed (0.48 to 1.29 m/s) for 2-5 minutes. These data were collected as part of another study, and the kinetic and kinematic data collection procedures have been previously reported [31]. Kinematic data were recorded at 100 Hz and kinetic at 1000 Hz.…”

“…We defined leg length as the vertical distance from the greater trochanter marker to the lateral malleolus marker while the participant stood upright. Please see the previous publications [30,31] for additional information about the data collection procedures.…”

“…For the training dataset [30], kinematic and kinetic data were lowpass filtered with a cutoff frequency of 6 Hz. For the validation dataset [31], kinematic data were lowpass filtered with a cutoff frequency of 6 Hz, while kinetic, at 20 Hz [32]. Vertical ground reaction forces were then used to identify kinetic gait events.…”

Gait speed and individual characteristics are related to specific gait metrics in neurotypical adults

“…The independent dataset used to validate the final models consisted of 22 neurotypical adults (age: 61 ± 15 years (24-77 years); leg length: 0.78 ± 0.06 m (0.67-0.88 m); mass: 73 ± 18 kg (37-103 kg)) walking at their self-selected speed (0.48 to 1.29 m/s) for 2-5 min. These data were collected as part of another study, and the kinetic and kinematic data collection procedures have been previously reported 47 . Kinematic data were recorded at 100 Hz and kinetic at 1000 Hz.…”