A Consecutive Multi-Day High-Density Surface Electromyography Dataset Comprising 7 Grasps and 11 Gestures.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Rehabilitation Technology

Background:

• Surface electromyography (sEMG) measures muscle electrical activity, crucial for understanding neuromuscular function.
• High-density sEMG (HD-sEMG) offers advanced non-invasive insights into motor unit action potential trains (MUAPTs) and neural drive.
• Limited availability of multi-day HD-sEMG datasets hinders long-term studies of motor neuron activity.

Purpose of the Study:

• To present the CEMHSEY dataset, a comprehensive 320-channel HD-sEMG recording spanning 11 consecutive days.
• To provide a valuable resource for research in human-machine interfaces and neuromuscular modulation.
• To facilitate long-term investigations into motor neuron activities and muscle physiological behaviors.

Main Methods:

• Acquisition of 320-channel HD-sEMG data from forearm muscles over 11 consecutive days.
• Creation of two sub-datasets: GRASP (isometric contractions, 13 subjects, 7 grasps, 3 force levels) and GESTURE (6 subjects, 11 hand gestures).
• Validation of the dataset's utility through force regression, hand gesture recognition, and motor unit decoding.

Main Results:

• The CEMHSEY dataset offers extensive multi-day HD-sEMG recordings for diverse forearm tasks.
• Demonstrated usability for force regression, hand gesture recognition, and motor unit decoding.
• The dataset supports the development of robust human-machine interfaces and analysis of neuromuscular modulation.

Conclusions:

• The CEMHSEY dataset is a significant contribution to the field, addressing the scarcity of multi-day HD-sEMG data.
• It provides a robust platform for advancing research in motor control, neuroprosthetics, and rehabilitation.
• Enables deeper understanding and analysis of long-term neuromuscular adaptation and neural drive variations.