Open Access Dissertation
Date of Award
Electrical and Computer Engineering
Le Yi Wang
Modern anesthesia management is a comprehensive and the most critical issue in medical care. During the past dacades, a large amount of research works have been focused on the problems of monitoring anesthesia depth, modeling the dynamics of anesthesia patient for the purpose of control, prediction, and diagnosis.
Monitoring the anesthesia depth is not only for keeping the patient in adquate anesthesia level but also for preventing the patient from overdosing. Several EEG based indexes have been developed such as the BIS, and Entropy etc. for measuring depth. However, reports mentioned that those indexes in some cases fail in detecting the awareness of the the patient. In this research work, a new EEG based parameter, beta_2/theta-ratio, was introduced as a potential enhancement in measuring anesthesia depth. It was compared to the relative beta-ratio which had been commercially used in the BIS monitor and proved that the beta_2/theta-ratio has improved reliability and sensitivity in detecting the awareness than the beta-ratio does.
Traditional modeling, diagnosis and control in anesthesia focus on a one-drug one-outcome scenario. In fact, Anesthesia drugs have impact on multiple outcomes of an anesthesia patient. Due to limited real-time data, real-time modeling in multi-outcome modeling requires low complexity model structures. A method of decision-oriented modeling which employs simplified and combined model functions in a Wiener structure to reduce model complexity was introduced. This model structure was implemented in device level and tested in operation room for real-time anesthesia monitoring, diagnosis, and prediction.
Furthermore, the impact of wireless channels on patient control in anesthesia applications was also investigated. Such a system involves communication channels which introduce noises due to quantization, channel noises, and have limited communication bandwidth resources. Usually signal averaging can be used effectively in reducing the noise effects. However, when feedback was intended, we showed that signal averaging will lose its utility substantially. To explain this phenomenon, we analyzed stability margins under signal averaging and derived some optimal strategies for selecting window sizes.
Finally, a mathematical model for the auditory system was introduced to characterize the impact of anesthesia on auditory systems, and analyze and diagnose hearing damage. The auditory system was represented by a black-box input-output system with external sound stimuli as the input and the neuron firing rates as the output. Two parallel subsystem models were developed for modeling the auditory system dynamics: an ARX (Auto-Regression with External Input) model for the auditory system under external stimuli and an ARMA (Auto-Regression and Moving Average) model for the spontaneous activities of the neurons on primary auditory cortex. These models provide a quantitative characterization of anesthesia's impacts and diagnosis of hearing loss on auditory transmission channels.
Tan, Zhibin, "Monitoring, diagnosis, and control for advanced anesthesia management" (2011). Wayne State University Dissertations. 297.