Short-term changes in protective stepping for lateral balance recovery in older adults

Abstract

Background: Fall prevention for older adults is dependent on the ability to maintain protective balance. This study measured the short-term changes of protective stepping following waist-pull perturbations in the medio-lateral direction, to identify what, if any, properties of protective stepping are improved with repeated perturbation exposures.

Methods: Sixty waist-pulls (2 directions × 5 intensities × 6 repetitions) from a single session were analyzed separately as early, middle, and late testing periods, for a comparison over time of typical responses. Outcome measures included the number of evoked steps, type of step, incidence of interlimb collisions, and kinematic and kinetic properties of the first step in frequently used crossover-type responses.

Findings: Improvements were evident as significantly reduced number of steps and collisions. However, these improvements could not be completely accounted for by significant changes in first step kinematic or kinetic properties.

Interpretation: We infer that older individuals experiencing repeated lateral waist-pull perturbations optimize the predictive or feed-forward motor control for balance recovery through stepping.

Figure 1

Illustration of Experimental Setup and Step Strategy Types. Pull direction indicated by arrows at waist-level. A: Position-controlled motor pulls waist belt in one of two lateral directions at one of five magnitudes. Subjects are instructed to hold a lightweight object at approximately waist height and to react naturally to prevent a fall. B: Lateral Step Strategy (LSS). C: Crossover Step to the Front (CSF). D: Crossover Step to the Back (CSF). E: Medial Step (MS).

Figure 2

Representative trial data from one subject at perturbation intensity one level above BTL, resulting in a single CSF step to the right. From top to bottom: Vertical ground reaction force under stepping (left) foot; Shear ground reaction force under stance (right) foot (impulse during stepping in shaded area); M-L location of the stepping (left) foot; waist-pull perturbation displacement. Vertical lines indicate, from left to right: Pull onset, Step Onset, and Step Cessation.

Figure 3

Mean Number of Steps during Early, Middle, and Late testing periods, mean of subject means ± S.E. Significant decreases in the number of steps evoked by a waist-pull were observed between early and late periods at all intensity values except at the Level 3 intensity (p<0.01).

Figure 4

Mean Number of Steps at Normalized Pull Intensity, mean of subject means ± S.E. Pull intensities aligned according to minimum value at which subjects used multiple steps during early testing. Significant decreases across all testing periods (p<0.01) were observed at and above the BTL, but only at late testing at the intensity below the BTL (p<0.05).

Figure 5

First Step Kinematics in Crossover type responses (mean ± S.E.). A: Step Distance in M-L direction, as a percentage of leg length. B: Mean Step Velocity in M-L direction, normalized to leg length. C: Step Duration (sec). Significant differences were only found in step durations between early and late testing.

Figure 6

Shear Impulse (Force × time) in M-L direction under Stance Leg during Crossover stepping, (mean ± S.E.). Results shown as a fraction of shear impulse during Early testing. Forces were oriented such that positive impulses acted as a braking force, towards the direction of the pull. Middle testing had significantly greater shear impulse than the early period, with minimal change between middle and late testing.