Development of functional electrical stimulation system using mechanomyography as muscle state feedback for paraplegics / Jannatul Naeem

• Main Author: Jannatul , Naeem
• Format: Thesis
• Published: 2021
• Subjects: RA Public aspects of medicine; TA Engineering (General). Civil engineering (General)
• Online Access: http://studentsrepo.um.edu.my/14909/1/Jannatul_Naeem.pdf; http://studentsrepo.um.edu.my/14909/2/Jannatul_Naeem.pdf; http://studentsrepo.um.edu.my/14909/

Functional electrical stimulation (FES) supported exercises aid purposeful muscle contractions to restore the lost motor functions with associated health benefits in individuals with spinal cord injury (SCI). The use of this device globally is majorly still restricted to the laboratory, especially in developing countries such as countries in Asia. This challenge is principally due to the high cost of the device for home care use and even for clinical deployment and the complexity of the commercially available options. One limitation of the available options is the inefficient muscle stimulation outcome due to the early onset of muscle fatigue. The reversal of motor unit recruitment in FES-evoked muscle contractions is one of the reasons for quicker muscle fatigue than the natural muscle contractions that occur from the central nervous system in a healthy person. The FES systems with sensor feedback are used to lessen this limitation with improved functional outcomes. The use of muscle contraction signals such as electromyography marred with stimulation artefacts, and the removal of these artefacts often presents another challenge. The studies reported in this thesis proposed and developed an FES system to monitor muscle condition using mechanomyography (MMG) without the limitation of stimulation artefacts. Before the implementation, a unique Mel-frequency cepstral coefficient (MFCC) feature of the MMG signal was introduced to classify the fresh muscle and muscle fatigue contractions in cycling exercise. Using this MFCC feature of MMG, 90.7% average classification accuracy was achieved, while the root mean square (RMS) feature of MMG had an accuracy of 74.5%. However, the computational cost needed for this research is supported using the RMS amplitude feature of MMG. This is also based on the literature that the MMG-RMS has a positive and primarily linear relationship with torque generated from FES-evoked muscle actions. The developed FES system created sufficient stimulation power to support standing exercise and monitor muscle condition using MMG signal related to knee-buckling, which indicates muscle fatigue. Furthermore, the developed FES system was applied to detect muscle fatigue in realtime using the MMG-RMS feature during isometric knee extension. The torque generated when tested on the isometric dynamometer was related to 70% drop in MMG-RMS threshold as actual muscle fatigue. Finally, an investigation of different modes of user control strategies of FES standing was simulated as the increase of amplitude by button pressing to sustain standing and detect muscle fatigue using the MMG-RMS feature. The outcome of this investigation showed that the single press 10 s mode provides optimal standing duration compared to other modes. Overall, this research shows the development of real-time muscle monitoring and prevention FES system with MMG sensor.