MODELING TRAJECTORIES USING FUNCTIONAL LINEAR DIFFERENTIAL EQUATIONS.

Area of Science:

• Biomechanics
• Neuroscience
• Functional Data Analysis

Background:

• Locomotion involves complex interactions between muscle activity and limb movement.
• Existing models often struggle to capture the dynamic, time-evolving nature of this relationship.

Purpose of the Study:

• To develop and validate an innovative regression method for modeling the dynamic relationship between muscle activation and paw position during locomotion.
• To analyze gait cycles in mice, integrating muscle activation data with paw position tracking.

Main Methods:

• Proposed a novel general regression method combining ordinary differential equations (ODEs) and functional data analysis.
• Simultaneously estimated ODE parameters across all gait cycle curves, borrowing strength across observations.
• Validated the approach through simulations and cross-validated predictive accuracy for paw position.

Main Results:

• The proposed model successfully captured the dynamic system relating muscle activation (biceps, triceps) to paw position and speed.
• Muscle activation was found to dynamically influence paw speed and position during mouse locomotion.
• The influence of muscle activation on paw movement was observed to persist beyond the period of activation itself.

Conclusions:

• The novel ODE-based functional data analysis method provides a powerful tool for understanding dynamic biological systems.
• Muscle activation plays a crucial, time-delayed role in controlling limb movement during locomotion.
• This approach offers improved modeling capabilities for complex biological processes involving continuous functional data.