The accuracy of rapid treadmill-belt movements as a means to deliver standing postural perturbations

Treadmill-induced postural perturbations are a promising tool in assessing and reducing the risk of falls. We evaluated the accuracy with which two treadmills (Simbex ActiveStep^® and an AMTI instrumented treadmill) achieved commanded displacements, peak velocities, and average initial accelerations. To do so, we included a range of perturbation magnitudes (20, 30, and 40 cm displacements) applied in unweighted and weighted (body mass = 46–84 kg) conditions. Across treadmills and perturbation magnitudes, absolute errors in displacement (< 0.5 cm) and peak velocity (< 4 cm/s) were small (relative error < 5%). Between-treadmill differences in displacement and peak velocity were marginal (< 3%), regardless of the perturbation magnitude and participant body mass. Observed accelerations were more than 5% smaller than commanded values. The front, but not back, AMTI belt demonstrated less acceleration accuracy than the ActiveStep^® (≈ 5% difference). In summary, both treadmills demonstrated a reasonable, consistent level of accuracy in delivering postural perturbations.