2.2 Experimental Protocol

During the experiment, participants stood on a translating platform used to present stimuli (Figure 1A). Participants were blindfolded and wore headphones that played white noise to eliminate auditory and visual information, thus limiting the subject to proprioceptive, vestibular and somatosensory cues. Stance width was standardized for participants by having the middle of their heels positioned at a distance that matched their inter-anterior superior iliac spine (ASIS) distance. The participant’s stance width was marked in order to maintain consistency throughout the experiment. The first platform translations, or perturbation, was always straight back (Θ = 270°) while the second perturbation ranged from 245–295° (ΔΘ = ± 25° from the first perturbation) where 0° represents rightward perturbation and 180° represents leftward perturbation. The actual platform movement was analyzed relative to the desired platform movement (Figure 1B) to quantify the accuracy of the platform motion. The error between the actual ΔΘ and desired ΔΘ increased with the amplitude of the desired Δ (ΔΘ = 15°, mean error = 0.71°), but the precision of the platform (interquartile range = [0.04, 0.25] °) was sufficient to present functionally distinct stimulus perturbations at a resolution of 0.5°.

A) Participants stood on a translating platform without auditory or visual feedback. Each trial consisted of two perturbations that varied only in the direction of the perturbation (Displacement = 7.5 cm, Velocity = 15 cm/s, Peak Acceleration = 0.1 m/s²). The first perturbation was straight back (270°), and the direction of the second perturbation varied relative to the first perturbation (255 - 295°, Δθ = ±25°). After the second perturbation, the participants reported whether the perturbations were in the ‘same’ or ‘different’ direction in each trial. The trial was completed when the platform then returned to a starting position. The session ended when a threshold was reached through either the PEST or psychometric method. B) The platform movement error scaled with the amplitude of the desired stimulus (Δθ), but the variability of the platform movement was less than 1°.

Each trial in the experiment consisted of two perturbations in which the participant was asked to determine whether or not the two perturbations were in the same or different direction. Each perturbation had a displacement of 7.5 cm, a velocity of 15 cm/s, and a peak acceleration of 0.1 m/s². The direction, or lateral deviation, of the second perturbation (ΔΘ) for each trial was set by the psychophysical method used to determine the sensory threshold: the psychometric method or the parameter estimation by sequential testing (PEST) method [10]. The subject was not informed that the first perturbation was always in the backward direction (Θ = 270°), nor when the first perturbation would occur (the time between the two perturbations was fixed at 0.5 seconds). After the second perturbation, the subject responded ‘same’ or ‘different’, by pressing a button on a response box that was held throughout each trial, to indicate whether the two perturbations were in the same or different direction. A trial was excluded if the subject took a step or did not respond in the given amount of time (3 seconds). Excluded trials were repeated at a later time. Due to limitation in the total travel of the perturbation platform, after each subject response, the platform was translated in the forward direction to prepare for the next trial (time between trials was randomized). This return movement was always in the 90° direction (straight forward) without any lateral deviation to prevent the subject from receiving feedback about the lateral movement of the two previous perturbations. Thus the initial absolute position of the platform varied in the lateral direction by a small amount. However subjects were blindfolded and thus did not receive any external cues about absolute position in the room.

Each experiment consisted of either the psychometric method alone (n = 14 subjects) or 2+ PEST runs and one full psychometric run in the following order: first PEST run, half of the psychometric stimulus set, second PEST run, and last half of the psychometric stimulus set (n = 11 subjects). If time permitted, a third PEST run was completed (n = 8 of 11 subjects). Subjects were required to rest after each section of the experiment or by request. The psychophysical methods were interleaved to prevent long-term adaptation effects, such as fatigue, which could bias the data. The estimated threshold from the first PEST run provided a relevant stimulus range for each subject such that the ΔΘ stimulus set chosen for the psychometric curve would include the threshold.