From linear to nonlinear gait measures: validity of a multi-view markerless motion capture system

Markerless motion capture technology allows for the analysis of human gait without the need for physical markers attached to the body, presenting a practical alternative to traditional marker-based optoelectronic systems. However, most validation studies have focused on linear gait measures, leaving the validation of nonlinear measures, which are sensitive to the fine temporal structure of locomotor signals, largely unexplored.

This study aimed to evaluate the concurrent validity of a three-camera markerless system against a 15-camera optoelectronic system across four treadmill speeds in 22 healthy adults. The researchers examined spatiotemporal parameters, joint angle and trunk acceleration time series, and sagittal-plane joint kinematic waveforms to assess the performance of the markerless system.

Inter-system agreement was found to be excellent for spatiotemporal parameters, indicating that the markerless system could provide reliable measurements comparable to the optoelectronic system. Spatial measures also showed good to excellent agreement with minor positive biases at different treadmill speeds. However, the agreement for sagittal-plane joint kinematic waveforms was lower, with root-mean-square errors ranging from 1.2 to 5.5 degrees for hip, knee, and ankle joints.

The Attractor Complexity Index (ACI) demonstrated moderate to good agreement across all joints and trunk acceleration directions, further supporting the validity of the markerless system for nonlinear gait analysis.

The study also investigated the performance of Maximum Lyapunov exponents and sample entropy, which are commonly used to assess the complexity and predictability of dynamic systems. While these measures showed moderate agreement at the hip and knee joints, preserving inter-individual ranking, they collapsed at the ankle level and for trunk sample entropy, indicating a need for further refinement.

Overall, this study establishes a tiered scope for markerless gait analysis, confirming that both spatiotemporal and nonlinear measures of gait dynamics can be reliably assessed using markerless systems, with the exception of ankle joint kinematics and trunk sample entropy.