Day-to-Day Reliability of Nonlinear Methods to Assess Walking Dynamics

Area of Science:

• Biomechanics
• Human Locomotion Analysis
• Nonlinear Dynamics

Background:

• Assessing human locomotion dynamics often employs nonlinear methods.
• Understanding the day-to-day reliability of these nonlinear methods is crucial for accurate gait analysis, especially in clinical populations like those with knee osteoarthritis.

Purpose of the Study:

• To investigate the day-to-day reliability of nonlinear methods for assessing human locomotion dynamics.
• To compare different state space reconstruction methods and nonlinear techniques for their reliability in gait analysis.

Main Methods:

• Twenty-four participants (12 with knee osteoarthritis) underwent treadmill walking on two separate days.
• Lower limb kinematics were recorded, and hip, knee, ankle joint angles, 3D sacrum displacement, and stride time intervals were extracted.
• Nonlinear measures including largest Lyapunov exponent, correlation dimension, sample entropy, and detrended fluctuation analysis (DFA) were calculated using various reconstruction methods.

Main Results:

• The group average state space reconstruction method demonstrated superior relative reliability and lower measurement error for joint angles and sacrum displacement compared to individual or test-retest average methods.
• Detrended fluctuation analysis (DFA) exhibited good day-to-day reliability for stride time intervals.
• Sample entropy showed poor reliability for stride time intervals.

Conclusions:

• The group average state space reconstruction method is recommended for nonlinear gait dynamics analysis of joint angles and sacrum displacement.
• Detrended fluctuation analysis (DFA) is a reliable method for analyzing stride time variability.
• These findings provide a reference for future nonlinear gait dynamics research.