A Novel Hybrid Model for Drawing Trace Reconstruction from Multichannel Surface Electromyographic Activity

Area of Science:

• Biomedical Engineering
• Neuroscience
• Machine Learning

Background:

• Surface electromyography (sEMG) shows promise for reconstructing movements.
• Clinical application of sEMG techniques is limited by practical challenges.

Purpose of the Study:

• To propose a novel three-step hybrid model for reconstructing drawing traces from multichannel sEMG.
• To enhance the accuracy and clinical feasibility of sEMG-based movement reconstruction.

Main Methods:

• A hybrid model combining coordinate state transition, sEMG feature extraction (Root Mean Square), and Gene Expression Programming (GEP) prediction.
• Recording 7-channel sEMG signals and corresponding drawing trace coordinate data.
• Developing prediction models using GEP to approximate original drawing traces.

Main Results:

• The hybrid model achieved a mean accuracy of 74% in within-group design and 86% in between-group design for reconstructing drawing traces.
• Successfully converted arm muscle sEMG activity into smooth reconstructions of drawing traces.
• Demonstrated feasibility of the model for improving drawing trace reconstruction from sEMG.

Conclusions:

• The proposed three-step hybrid model is effective for reconstructing drawing traces from sEMG.
• The model shows potential for advancing clinical applications of sEMG-based movement analysis.
• Further development could enhance the precision and scope of sEMG-driven motion reconstruction.