Design, Fabrication and Evaluation of a Stretchable High-Density Electromyography Array

Rejin John Varghese, Matteo Pizzi, Aritra Kundu, Agnese Grison, Etienne Burdet, Dario Farina

TL;DR

This paper introduces a novel, stretchable, high-density electromyography (HD-EMG) array using dry electrodes on a flexible PCB. This design overcomes practical limitations of existing systems, offering a comfortable, easy-to-apply solution for real-world human-machine interfaces with performance comparable to traditional wet electrodes.

What Problem Does This Paper Address?

Traditional high-density EMG (HD-EMG) systems, crucial for understanding muscle activity and enabling advanced human-machine interfaces (HMIs), face significant hurdles for widespread adoption. These include cumbersome electrode application requiring skin preparation (shaving, abrasive gels), the need for conductive pastes, and susceptibility to noise and motion artifacts. These issues make them impractical for everyday use, limiting their transition from laboratory research to consumer applications. The paper aims to bridge this gap by developing a wearable, stretchable HD-EMG array that is easy to apply, comfortable, and maintains high signal quality without the drawbacks of wet electrodes.

Methods

The core innovation lies in the design and fabrication of a stretchable HD-EMG array. The process begins with a flexible printed circuit board (PCB) populated with dry electrodes. This PCB is then integrated into a multi-layered, stretchable silicone substrate, reinforced with fabric for structural integrity. The fabrication involves pouring silicone layers into custom 3D-printed molds, folding the flexible PCB, and encasing it within the silicone and fabric layers. This method allows for the creation of a seamless, wearable sleeve.

For evaluation, the researchers employed several methods:

  • Baseline Noise Characterization: Comparing the root mean square (RMS) of signals from their dry electrodes against traditional wet electrodes to quantify noise levels.
  • Electrochemical Characterization: Using Electrochemical Impedance Spectroscopy (EIS) and Cyclic Voltammetry (CV) to assess the electrical properties of the gold-coated electrodes.
  • EMG Model and Data Acquisition: Recording EMG signals during dynamic gestures and isometric contractions using their stretchable array.
  • Gesture Classification: Training a Convolutional Neural Network (CNN) to classify 7 different dynamic wrist gestures using the recorded EMG data.
  • Motor Unit (MU) Decomposition: Analyzing the ability of the dry electrodes to capture signals suitable for decomposing into individual motor unit activities, comparing this to wet electrodes.

Datasets

The paper does not explicitly mention or introduce new public datasets. Instead, it relies on data collected from human participants during controlled experimental sessions. These sessions involved recording EMG signals from participants performing specific tasks:

  • Dynamic Wrist Gestures: 7 distinct gestures (flexion, extension, radial deviation, ulnar deviation, pronation, supination, and hand closing) were performed by two participants.
  • Isometric Contraction: Tibialis Anterior (TA) muscle activity was recorded during sustained isometric contractions at 20% MVC from one participant.

The data collected from these experiments served as the basis for evaluating the performance of the developed HD-EMG array in gesture classification and motor unit decomposition tasks.

Key Insights and Ideas

  • Dry Electrodes Can Compete: The study demonstrates that dry electrodes, when integrated into a well-designed stretchable array, can achieve performance comparable to traditional wet electrodes, significantly reducing application complexity.
  • Stretchability is Key for Wearability: The fabric and silicone-based stretchable sleeve design is crucial for comfort, adaptability to different body shapes, and maintaining consistent electrode-skin contact during movement.
  • Fabric Reinforcement Enhances Durability: Incorporating fabric into the silicone substrate not only improves stretchability but also enhances the structural integrity and durability of the array, preventing fatigue failure.
  • DIY Fabrication is Achievable: The described fabrication process uses readily available materials and standard lab equipment (like 3D printers), making it accessible for other researchers and makers.
  • Signal Quality is Sufficient for Complex Tasks: The dry HD-EMG array proved capable of capturing high-quality signals for both gesture recognition (achieving >95% accuracy) and motor unit decomposition, indicating its potential for advanced HMI applications.

Key Results

  • Noise Performance: The dry-electrode grid showed a slight increase in baseline RMS noise (average 7.2%) compared to wet electrodes, as expected, but this difference did not significantly hinder performance in key tasks.
  • Gesture Classification: The stretchable dry-EMG array achieved high accuracy in classifying 7 dynamic wrist gestures, with test-set accuracies above 95.9% for all gestures and subjects. Wet electrodes offered a marginal improvement of 1.61%, but the practical advantages of dry electrodes outweigh this.
  • Motor Unit Decomposition: The dry-electrode configuration performed comparably to wet electrodes in decomposing motor unit signals from the Tibialis Anterior muscle, detecting 12 MUs compared to 13 with wet electrodes. The dry grid even showed a slightly higher average silhouette (SIL) value (0.922 vs. 0.918).
  • Electrochemical Properties: Gold-coated electrodes exhibited good conductivity and a well-defined capacitive behavior, suitable for EMG signal acquisition.

Limitations and Future Work

  • Slightly Higher Noise: While performance was comparable, the dry electrodes did exhibit slightly higher noise than wet electrodes. Future work could explore further noise reduction techniques or electrode material optimizations.
  • Limited Subject Pool: The validation experiments were conducted on a small number of participants. Expanding testing to a larger and more diverse population would be beneficial.
  • Inter-electrode Distance: The paper notes that the larger inter-electrode distance in their configuration, while beneficial for coverage, might limit spatial resolution compared to some ultra-high-density arrays.
  • Fabric Reinforcement Material: While conductive stretch fabric was investigated for EMI shielding, it did not yield significantly improved results. Further investigation into its effectiveness and other EMI mitigation strategies is suggested.
  • Generalizability: While the technique is presented as generalizable to other modalities, further research would be needed to adapt and validate it for ultrasound, IMUs, or FES electrodes.

The authors emphasize the potential for this stretchable HD-EMG array to significantly advance the real-world translation of high-density EMG for human-machine interfaces, particularly in areas like rehabilitation, prosthetics, and wearable health monitoring.

Summary updated

Abstract

The adoption of high-density electrode systems for human-machine interfaces in real-life applications has been impeded by practical and technical challenges, including noise interference, motion artefacts and the lack of compact electrode interfaces. To overcome some of these challenges, we introduce a wearable and stretchable electromyography (EMG) array, and present its design, fabrication methodology, characterisation, and comprehensive evaluation. Our proposed solution comprises dry-electrodes on flexible printed circuit board (PCB) substrates, eliminating the need for time-consuming skin preparation. The proposed fabrication method allows the manufacturing of stretchable sleeves, with consistent and standardised coverage across subjects. We thoroughly tested our developed prototype, evaluating its potential for application in both research and real-world environments. The results of our study showed that the developed stretchable array matches or outperforms traditional EMG grids and holds promise in furthering the real-world translation of high-density EMG for human-machine interfaces.
BibTeX
@inproceedings{varghese2024design,
  author = {Rejin John Varghese and Matteo Pizzi and Aritra Kundu and Agnese Grison and Etienne Burdet and Dario Farina},
  title = {{Design, Fabrication and Evaluation of a Stretchable High-Density Electromyography Array}},
  booktitle = {Sensors},
  year = {2024},
  doi = {10.3390/s24061810},
}
Figure