Task-Dependent Reorganization of Ankle–Knee Mechanical Coordination Revealed by Moment–Moment Phase Space Analysis

Background: Human locomotion requires coordinated torque production across multiple joints, yet conventional gait analysis typically evaluates joint behavior independently, limiting insight into inter-joint coordination. This study aimed to quantify task-dependent reorganization of ankle–knee mechanical coordination using a moment–moment phase space framework. Methods: A normative dataset of healthy adults (N = 50) performing natural-speed walking, toe walking, heel walking, stair ascent, and stair descent was analyzed. Sagittal-plane external ankle and knee moments were extracted from time-normalized stride cycles and z-score normalized within each stride to emphasize coordination topology. Ankle–knee trajectories were represented in moment–moment space and characterized using three geometric metrics: loop magnitude (|Area|), principal axis orientation, and anisotropy. Metrics were aggregated within subject and analyzed using linear mixed-effects models with planned contrasts against walking. Results: Loop magnitude differed significantly across tasks (p < 0.001), with the largest increases observed during toe walking (+3.45 relative to walking) and stair descent (+2.41). Principal axis orientation also showed a significant task effect (p = 0.026), with stair descent producing the largest rotation of the coordination axis (−29.8°). Anisotropy varied significantly across tasks (p < 0.001), indicating systematic changes in the dimensionality and strength of inter-joint torque coupling. Conclusions: Locomotor tasks induce structured, task-dependent reorganization of ankle–knee coordination topology. Moment–moment phase space analysis provides a compact and interpretable framework for quantifying inter-joint torque coupling, with potential applications in biomechanics research and the development of activity-aware assistive technologies.