Utilizing Reinforcement Learning to Overcome the Challenge of Muscle-Specific EMG Placements for Musculoskeletal Modeling

Key Points

• This research aims to develop a reinforcement learning approach to map electromyography recordings to neural excitation effectively.
• Developed a novel reinforcement learning-based mapping solution.
• Focused on muscle-specific electromyography (MSp EMG) recordings and their challenges.
• Evaluated the effectiveness in modeling musculoskeletal systems.
• Demonstrated that the RL-based method effectively maps NMSp EMG to MSp neural excitation.
• Highlighting potential increases in application for musculoskeletal models in challenging scenarios.