Assessing Stylistic Differences in the Underlying Biomechanical Objectives of Walking Using Simulation-Based...

Area of Science:

• Biomechanics
• Human Movement Science
• Computational Modeling

Background:

• Clinical gait analysis commonly relies on observational methods and categorical ratings to identify pathologies.
• Assessing subtle biomechanical differences in gait through visual observation remains a challenge for clinicians.

Purpose of the Study:

• To evaluate the ability of novice observers to identify specific biomechanical objectives in simulated walking using stylistic labels.
• To determine if visual perspective influences the characterization of gait behavior in physics-based simulations.

Main Methods:

• Generated 75 muscle-driven gait simulations for five subject models, varying five biomechanical objectives and three visual perspectives.
• Utilized computed muscle control and direct collocation to optimize simulations based on objectives like metabolic cost and muscle activation.
• Employed 100 crowd workers to rate simulations using stylistic labels, followed by multinomial logistic regression analysis.

Main Results:

• Observer ratings for loading and activation significantly predicted muscle activation-optimized movements.
• Activation ratings were predictive of movement perspective, while balance ratings were only significant from the frontal view.
• The 'wisdom of crowds' could discern some biomechanical objectives, but motor control redundancy limited observational accuracy.

Conclusions:

• Visual observation of gait simulations can identify some underlying biomechanical objectives.
• The effectiveness of observational gait analysis is constrained by the inherent redundancy in human motor control strategies.
• Further research may explore improved methods for objective gait assessment in clinical and research settings.