Neural interfacing by decoding spinal motoneuron activities in SCI patients
File(s)
Author(s)
Yimyam, Manasavee
Type
Thesis
Abstract
This thesis presents the development of a high-density dry surface electromyography (HD-sEMG) wearable system, specifically designed for individuals with SCI. The primary objective of this work was to overcome the limitations of traditional systems, such as reliance on wet electrodes, signal degradation, and the need for stationary setups, by developing a system capable of detecting residual muscle activity in paralysed limbs. This advancement offers enhanced control for assistive devices and rehabilitation approaches in dynamic environments where continuous monitoring is crucial.
The NeuroSens system merges innovative hardware and software features. The hardware includes a custom-designed flexible sensory strip with dry sEMG units, embedded in a lightweight, form-fitting sleeve. This design eliminates the need for gels or adhesives, ensuring reliable electrode-skin contact. The modular hardware supports scalable signal acquisition, and a built-in real-time processing unit enables the use of decomposition and classification algorithms for motor unit firing patterns. These advancements significantly surpass traditional systems, which often struggle with immobility and low signal resolution. On the software side, a MATLAB-GUI interface was developed for data recording, transmission, feedback, and real-time classification. This real-time classification system, validated through various experiments, showed high accuracy in decoding hand gestures, contributing to understanding how residual muscle activity can control virtual interfaces or assistive devices.
Results indicated that offline classification accuracy reached up to 100% for specific gestures like making and holding a fist. However, during real-time control, individuals with SCI successfully performed only three gestures: two-finger pinch, three-finger pinch, and fist. Performance discrepancies arose from cognitive load, muscle fatigue, and gesture complexity. Despite challenges in real-time EMG-based control, the system holds promise for restoring multi-degree-of-freedom movements in paralysed limbs. The innovations presented here offer a practical solution for enhancing independence and functionality for individuals with motor impairments.
The NeuroSens system merges innovative hardware and software features. The hardware includes a custom-designed flexible sensory strip with dry sEMG units, embedded in a lightweight, form-fitting sleeve. This design eliminates the need for gels or adhesives, ensuring reliable electrode-skin contact. The modular hardware supports scalable signal acquisition, and a built-in real-time processing unit enables the use of decomposition and classification algorithms for motor unit firing patterns. These advancements significantly surpass traditional systems, which often struggle with immobility and low signal resolution. On the software side, a MATLAB-GUI interface was developed for data recording, transmission, feedback, and real-time classification. This real-time classification system, validated through various experiments, showed high accuracy in decoding hand gestures, contributing to understanding how residual muscle activity can control virtual interfaces or assistive devices.
Results indicated that offline classification accuracy reached up to 100% for specific gestures like making and holding a fist. However, during real-time control, individuals with SCI successfully performed only three gestures: two-finger pinch, three-finger pinch, and fist. Performance discrepancies arose from cognitive load, muscle fatigue, and gesture complexity. Despite challenges in real-time EMG-based control, the system holds promise for restoring multi-degree-of-freedom movements in paralysed limbs. The innovations presented here offer a practical solution for enhancing independence and functionality for individuals with motor impairments.
Version
Open Access
Date Issued
2024-04
Date Awarded
2024-11
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Farina, Dario
Del Vecchio, Alessandro
Publisher Department
Bioengineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)