Multi-Task Assessment of Context-Specific Gait Changes to Virtual Reality-Based Visual Perturbations

Conventional gait assessments often fail to detect early, subclinical balance abnormalities that may become apparent under sensory conflict, such as visual perturbations. We investigated whether instrumented gait analysis combined with virtual reality (VR)-delivered visual perturbations can reveal task-dependent gait adaptations in healthy adults exposed to discordant sensory information.

Ten healthy adults completed a purpose-designed battery of 10 walking tasks spanning steady-state and balance-challenging conditions (e.g., tandem gait, obstacle negotiation, Timed Up and Go) on an instrumented walkway under four visual conditions: no virtual reality, unperturbed VR, optic flow-perturbed VR, and flickering light-perturbed VR. Spatiotemporal gait parameters were quantified to assess task- and condition-specific changes in dynamic gait control.

Visual perturbations elicited task-dependent gait adaptations rather than uniform effects across the battery. Compared to unperturbed walking, optic flow was associated with selective increases in stride width during tasks involving rapid reorientation and rotational head movements, including pivot turning (0.10 vs. 0.06, p = 0.045) and horizontal head turns (0.12 vs. 0.08, p = 0.027). Temporal and spatial gait variability increased during tandem gait under optic flow (step time variability: 32.23 vs. 12.75, p = 0.036; stride length variability: 27.96 vs. 8.08, p = 0.041), reflecting impaired cycle-to-cycle consistency under heightened precision demands.

VR-delivered visual perturbations revealed task-specific alterations in mediolateral stability, temporal consistency, and spatial gait scaling that were absent during unperturbed walking. A multi-task, perturbation-based assessment framework may therefore enhance sensitivity to subtle gait instability by exposing context-dependent limits of balance control.