Date of Award

6-2026

Degree Name

MS in Biomedical Engineering

Department/Program

Biomedical Engineering

College

College of Engineering

Advisor

Robert Szlavik

Advisor Department

Biomedical Engineering

Advisor College

College of Engineering

Abstract

Electrically evoked electromyography records the muscle's response to nerve stimulation and has been used for clinical assessment of neuromuscular function, as well as treatments such as functional electrical stimulation or stroke rehabilitation. In every muscular contraction, an electrical signal is generated. These signals come from the summation of simultaneously recruited motor units, enabling decomposition techniques to reveal information about motor unit recruitment.  By improving the ability to decompose and analyze these signals, clinical treatments involving neuromuscular electrical stimulation can be optimized and better understood. Much of the research on EMG analysis focuses on voluntary EMG, leaving a gap in the number of validated techniques for decomposing evoked EMG. In this thesis, Szlavik’s perturbation-based decomposition has been extended to synthetically modeled evoked EMGs and evaluated against the generalized Fourier series. These models were constructed using Associated Hermite basis functions to represent motor unit action potentials, which were summed to form a compound M-Wave and an M-Wave train. The following models were decomposed: a single M-Wave, a 50 Hz M-Wave train, and each of their noisy counterparts, evaluated at varying signal-to-noise ratios (SNR). Szlavik's algorithm successfully decomposed both noiseless models and maintained high accuracy estimates at SNRs of 15 and 20 dB. Additionally, the perturbation-based method outperformed the generalized Fourier series in decomposing all simulations.

Download

Share

COinS