Fig 1.
Cable-robot setup (a) and design (b) for applying lateral perturbations to the CoM (adopted from Major et al 2018 [26]), and measured variable definitions (c) according to temporal medial-lateral COM position.
Table 1.
BKPU participant characteristics (Median (interquartile range)) [26].
Fig 2.
Representative plots of temporal medial-lateral CoM position for a single BKPU subject perturbed toward the impaired limb (positive CoM position values) when the perturbation direction was known and the perturbation timing was either unknown (blue) or known (red). The vertical dotted black line denotes the onset time of the perturbation (time 0) and the y-axis zero denotes the treadmill belt center.
Fig 3.
Average CoM displacement during the 400 msec exposure period (ΔCoMexposure) separated by group, perturbation direction, and perturbation timing knowledge.
The main effect of perturbation direction was significant (p = 0.033) with displacement greater with perturbations towards the impaired/non-dominant side. Error bars denote 95% confidence intervals.
Fig 4.
Average maximum CoM displacement (ΔCoMpeak) separated by group, perturbation direction, and perturbation timing knowledge.
The main effect of timing knowledge was significant (p = 0.010), with less displacement when perturbation timing was known. Error bars denote 95% confidence intervals.
Fig 5.
Average time to reversal (Treversal) separated by group, perturbation direction, and perturbation timing knowledge.
The main effect of timing knowledge was significant (p = 0.043), with a greater Treversal when perturbation timing was known. Error bars denote 95% confidence intervals.
Fig 6.
Plots of ΔCoMexposure and average CoM medial-lateral velocity against gait cycle for control (A and B, respectively) and BKPU (C and D, respectively) participants. The connected red diamonds in plots A and C represent average displacement values for each 20% of the gait cycle to demonstrate how displacement varies over the gait cycle and has a similar shape to CoM velocity (B and D). Start of the gait cycle (0%) represents non-dominant or impaired limb initial contact. A representative scatterplot generated from the repeated-measures correlation analysis (E) demonstrates the positive linear relationship between CoM velocity and displacement for each able-bodied participant (denoted by different colors).
Table 2.
Repeated-measures correlation analysis results evaluating relationships between ΔCoMexposure and CoM medial-lateral velocity.