Mechanomyography for neuromuscular electrical stimulation feedback applications in persons with spinal cord injury / Ibitoye Morufu Olusola

• Main Author: Ibitoye, Morufu Olusola
• Format: Thesis
• Published: 2017
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://studentsrepo.um.edu.my/7337/1/All.pdf; http://studentsrepo.um.edu.my/7337/9/ibitoye.pdf; http://studentsrepo.um.edu.my/7337/
• Institution: Universiti Malaya

Neuromuscular Electrical Stimulation (NMES)-evoked muscle contractions confers therapeutic and functional gains on persons with Spinal Cord Injury (SCI). However, the optimal efficacy of commercial NMES systems’ application is inhibited by the imprecision in muscle force/torque production and rapid muscle fatigue. Evidence suggests that the application of a muscle mechanical response (force/torque) as a feedback to modulate the administration of NMES could optimize the efficacy of the technology by enabling muscle force regulation, and delaying the onset of muscle fatigue. Currently, a direct muscle force measurement is impractical and there is also lack of a reliable, electrical stimulus artifact-free and non-invasive proxy of muscle force to drive the NMES systems for enhanced controllability and clinical use. Attempts on the application of evoked-electromyography for this purpose remain debatable and clinically limited. As a viable alternative, this thesis proposes a non-invasive muscle force/torque measurement technique based on the mechanical activity of contracting muscles (Mechanomyography or MMG). This investigation was motivated by the knowledge that mechanomyography is immune from certain limitations of evoked-electromyography and provides direct information on muscle’s mechanical responses to the electrical stimulation. Systematic literature survey revealed a lack of clear understanding of the relationship between mechanomyography and NMES-evoked torque production in a paralyzed muscle. Therefore, the present research introduces mechanomyography as a proxy of NMES-evoked torque in persons with SCI. At the outset, a hybrid procedure was developed to establish mechanomyography as a proxy of muscle force/torque in healthy volunteers and persons with SCI. This was used to investigate the pattern of incremental torque production and subsequently facilitated the estimation of the torque from mechanomyography using a computational intelligent technique based on Support Vector Regression (SVR) modelling. This thesis also demonstrated, in a clinical setting, the validity of the mechanomyography as a relevant parameter for studying muscle fatigue during critical knee buckling stress i.e. standing-to-failure challenge in persons with SCI. Due to the peculiarity of the study participants/target population and the intended clinical application of NMES-supported standing, the quadriceps muscle group, widely reported for its relevance in studying the knee torque dynamics, was selected as the study site. Findings from these studies revealed that the mechanomyographic amplitude is highly correlated (r> 0.95; P< 0.05) to the muscle force in persons with SCI as it reliably tracked the muscle’s motor unit recruitment pattern during NMES contractions. The SVR modelling results demonstrated a good predictive accuracy (R2≥ 89%) with generalization capacity and suggested that the quadriceps’ mechanomyography is a good indicator of NMES-evoked torque during knee extension tasks. Thus, the signal might be deployed as a direct proxy of muscle torque during leg exercise and functional movements in SCI populations. Additionally, the reliability (intraclass correlation coefficient range: 0.65-0.79; P> 0.05) of the mechanomyography during force production might be useful to evaluate the recovery or deterioration of motor unit activities following NMES supported exercise and as an alternative technique for monitoring the NMES-evoked muscle activity for practical control applications. Together, this thesis lays a foundation for the future implementation of MMG-driven NMES technologies.