Nonlinear net ankle quasi-stiffness reduces error and changes with speed but not load carried

Area of Science:

• Biomechanics
• Biomedical Engineering
• Orthotics and Prosthetics

Background:

• Natural ankle quasi-stiffness (NAS) is vital for personalizing ankle-foot orthoses and prostheses.
• Traditional NAS definition as linear slope may not capture its complex, nonlinear nature.
• Characterizing nonlinear NAS across diverse tasks is essential for advanced device design.

Purpose of the Study:

• To investigate if nonlinear NAS differs across various intensities of walking, running, and load carriage.
• To compare the accuracy of linear, quadratic, and cubic regressions in modeling NAS.

Main Methods:

• Observed 22 healthy individuals performing walking, running, and load carriage at varying intensities.
• Applied linear, quadratic, and cubic regressions to ankle moment-angle data during stance phases.
• Utilized Root Mean Square Error (RMSE) and multilevel linear models (MLMs) to assess regression accuracy and coefficient differences.

Main Results:

• Quadratic and cubic regressions demonstrated significantly lower RMSE than linear regressions for NAS.
• Quadratic regression coefficients showed trends with walking/running speed but not with load.
• Nonlinear NAS characteristics were identified across different gait intensities.

Conclusions:

• Nonlinear NAS modeling, particularly quadratic, offers superior accuracy over linear models.
• Gait intensity, specifically speed, influences nonlinear NAS, while load carriage does not.
• Findings provide critical insights for developing sophisticated, personalized ankle-foot orthotic and prosthetic devices.