Procedures

Participants wore a standardized shoe (Excelsior training shoe; Adidas AG, Herzogenaurach, Germany) for all testing conditions to ensure that they had the same shoe–surface interface during testing. To minimize movement within the shoe, we instructed participants to tie their shoes tightly. An electrogoniometer (model SG110/A; Biometrics Ltd, Newport, United Kingdom) was placed on the right lateral ankle just proximal to the distal fibula, and the distal axis was placed on the shoe at the subtalar joint. We secured it with hook-and-loop tape and medical tape as recommended by the manufacturer. The right ankle was assessed in all participants. Before testing each participant, we used the software to calibrate the electrogoniometer with a standard goniometer at 0° and 30°. Participants walked along a custom-built walkway in 3 conditions: brace, tape, and control. The order of prophylactic condition was counterbalanced. Participants rested for 30 seconds between trials and 2 to 5 minutes between conditions. They completed at least 5 walking trials to provide 3 acceptable test trials for each condition. After each condition, we instructed participants to rate their perceived ankle stability during the condition using a visual analog scale. They marked a dash across a vertical 10-cm line, with 0 cm (top) indicating that the ankle felt completely stable and 10 cm (bottom) indicating that the ankle felt very unstable during the condition.

All data collection was completed on the custom-built 7.2-m-long walkway, which was modeled after the device used by Hopkins et al.^31,32,38 It included four 1.2-m active sections with a set of doors on the right and left that opened to a 30° angle. An industrial-strength electromagnet held each door closed. When triggered by a control panel, the voltage supplied to 1 electromagnet decreased to a set point at which it supported only the weight of the door. The instant a force greater than the weight of the door was applied, the door fell open (Figure 1).

When triggered, the active door opened to approximately 30° of inversion when participants stepped on it. Participants were instructed to keep walking after the sudden perturbation occurred.

We instructed participants to walk along the marked nonslip path at the pace of a metronome (model MA-1; KORG Inc, Tokyo, Japan) set to 110 beats per minute while focusing on a target mounted at eye level on the wall at the end of the walkway. During each walking trial, 1 random door was triggered to open. Participants were instructed to keep walking and to take the next step or steps when a door opened. We collected data from the time the participants were instructed to begin walking until they reached the end of the walkway, which was approximately 10 seconds. Whereas data were recorded only for the right side, the randomization of doors included the right and left sides, so participants were unaware of which door would open. All trials were video recorded (LifeCam Studio webcam; Microsoft Corporation, Redmond, WA); if we questioned whether the participant was not completely within the footpath or a door did not trigger, the trial was flagged for video review before the data were included in the analysis.