Date of Award
2006
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Electrical and Computer Engineering
Supervisor
Dr. Lyndon Brown
Abstract
This thesis presents a new method to detect and characterize neuronal populations in raw human muscle sympathetic nerve activity (MSNA). The method includes raw MSNA signal recording, wavelet denoising and action potential detection by a new algorithm that can locate and measure the height of spikes. Based on this method, inter-spike interval and neuronal spike height distributions during a held-inspiration reflex were compared with baseline using data collected from a healthy subject. Factors affected spike height such as SNR, filtering and denoising, and measuring distance were assessed using simulated data. The Hypothesis that the complex waveform is the convergence of two or more spikes firing in very close sequence was tested by decomposing it into two single spikes by squared optimization method. The distortions of band pass filter on individual neuron action potential are analyzed and an inverse filter is designed to reduce these distortions. Simulation results show that depolarization peak is recovered by applying designed inverse filter only if the SNR of signal is high enough. In addition, a new digital band-pass minimum phase filter is designed. Simulation results show that with this filter, spike detecting ability of the proposed algorithm is improved.
Recommended Citation
Liu, Yingchun, "Detecting Multi-Neuronal Discharge in Raw Human Muscle Sympathetic Nerve Activity (MSNA)" (2006). Digitized Theses. 5014.
https://ir.lib.uwo.ca/digitizedtheses/5014