Modular neuromuscular control of human locomotion by central pattern generator.

Area of Science:

• Neuroscience
• Biomechanical Engineering
• Robotics

Background:

• Spinal cord central pattern generators (CPGs) produce rhythmic motor patterns for locomotion.
• Locomotion involves complex, redundant muscle systems with nonlinear muscle dynamics.
• Existing CPG models struggle with the complexity of biological muscle actuation.

Purpose of the Study:

• To propose a reduced neural control strategy for CPGs based on muscle synergies.
• To develop and validate a CPG model that integrates muscle synergies for motor control.
• To assess the model's ability to reproduce complex leg muscle activation patterns during locomotion.

Main Methods:

• Extracted four muscle synergies from EMG data using non-negative matrix factorization.
• Utilized a Matsuoka's four-neuron CPG model with mutual inhibition.
• Integrated CPG output with muscle synergies, using hip flexion and foot contact as sensory inputs.
• Tuned model parameters with experimental gait data and validated against EMG data.

Main Results:

• Achieved good fitting accuracy (RMSEs 0.0491-0.1399) between simulated and experimental synergy activations.
• Successfully reproduced characteristic features of individual leg muscle activation patterns during locomotion.
• Demonstrated model's effectiveness across different gait trials and subjects.

Conclusions:

• The proposed CPG and muscle synergy-based model offers a simplified yet powerful approach to neuromuscular control.
• The model shows potential for resolving muscular redundancies and enabling fast, distributed control.
• This strategy allows for modulation of locomotion via simple control signals.