Johnnie W. Huang

Depth of anesthesia estimation and control [using auditory evoked potentials]

A fully automated system was developed for the depth of anesthesia estimation and control with the intravenous anesthetic, Propofol. The system determines the anesthesia depth by assessing the characteristics of the mid-latency auditory evoked potentials (MLAEP). The discrete time wavelet transformation was used for compacting the MLAEP which localizes the time and the frequency of the waveform. Feature reduction utilizing step discriminant analysis selected those wavelet coefficients which best distinguish the waveforms of those responders from the nonresponders. A total of four features chosen by such analysis coupled with the Propofol effect-site concentration were used to train a four-layer artificial neural network for classifying between the responders and the nonresponders. The Propofol is delivered by a mechanical syringe infusion pump controlled by Stanpump which also estimates the Propofol effect-site and plasma concentrations using a three-compartment pharmacokinetic model with the Tackley parameter set. In the animal experiments on dogs, the system achieved a 89.2% accuracy rate for classifying anesthesia depth. This result was further improved when running in real-time with a confidence level estimator which evaluates the reliability of each neural network output. The anesthesia level is adjusted by scheduled incrementation and a fuzzy-logic based controller which assesses the mean arterial pressure and/or the heart rate for decrementation as necessary. Various safety mechanisms are implemented to safeguard the patient from erratic controller actions caused by external disturbances. This system completed with a friendly interface has shown satisfactory performance in estimating and controlling the depth of anesthesia.

Multiple-drug hemodynamic control using fuzzy decision theory

A fuzzy-logic-based, automated drug-delivery system has been developed and validated on a nonlinear canine circulatory model for managing hemodynamic states. This controller features: (1) a fuzzy decision analysis module for patient status determination by assessing cardiac index, systemic vascular resistance index, and pulmonary vascular resistance index and (2) a fuzzy hemodynamic management module utilizing dopamine, phenylephrine, nitroprusside, and nitroglycerin for regulating mean arterial pressure, mean pulmonary arterial pressure, and cardiac output. A rule-based drug delivery scheduling program has been devised and incorporated to execute the therapeutic strategy as recommended by the decision-analysis module. Compared to the existing controllers, this system is able to achieve a faster response time with a more secured and effective regulation. The simulation results have demonstrated the feasibility of the decision analysis process for automated management of the arterial and venous circulation with an expanded arsenal of pharmacological agents.

Closed-loop system for total intravenous anesthesia by simultaneously administering two anesthetic drugs

Total Intravenous Anesthesia (TIVA) for general anesthesia using Propofol (a hypnotic) and Fentanyl (a narcotic) is a popular technique. When designing computer control of TIVA, the main difficulty is how to determine the infusion rate of two interacting drugs. In this paper, the authors mainly address the simulation results on the interaction model of two drugs, how to use the model to simultaneously control infusion rates and how to incorporate it into an Automatic Anesthesia Management System (AAMS) for TIVA. The paper provides a novel idea regarding how to design closed-loop system for simultaneously administrating two drugs in TIVA.

Control of a nonsquare drug infusion system: A simulation study.

A model predictive control strategy was developed and tested on a nonlinear canine circulatory model for the regulation of hemodynamic variables under critical care conditions. Different patient conditions such as congestive heart failure, post‐operative hypertension, and sepsis shock were studied in closed‐loop simulations. The model predictive controller, which uses a different linear model depending on the patient condition, allowed constraints to be explicitly enforced. The controller was initially tuned on the basis of a linear plant model, then tested on the nonlinear physiological model; the simulations demonstrated the ability to handle constraints, such as drug dosage specifications, commonly desired by critical care physicians.

Depth of Anesthesia Control with Fuzzy Logic

The anesthetic management of a surgical patient is a process that relies on the experience of an anesthesiologist, since currently there is no direct means of assessing a patient’s level of consciousness during surgery. The decision for the initial anesthetic level is generally made by using the recommended drug dosages based on various patient characteristics, such as age and weight. The anesthesiologist determines any subsequent alteration in the anesthetic level by observing signs from the patient. These signs, the indirect indicators of the depth of anesthesia (DOA), may include changes in blood pressures or heart rate, lacrimation, facial grimacing, muscular movement, spontaneous breathing, diaphoresis, and other signs that may predicate awareness. However, they are not reliable indicators of changes in a patient’s level of consciousness. Although an anesthesiologist can adjust recommended anesthetic dosages based on individual patient characteristics, these adjustments cannot always account for variability in patient responses to anesthesia or changes in anesthetic requirements during the course of surgery.

Estimation of depth of anesthesia using the midlatency auditory evoked potentials by means of neural network based multiple classifier system

This paper proposes a neural network based multiple classifier system (MCS) for monitoring the depth of anesthesia by assessing the characteristics of the middle latency auditory potentials (MLAEP) and the Propofol effect-site concentration. The system is composed of three individual neural network based classifiers with different sets of features. Discrete wavelet transformation (DTWT) and power spectrum estimation (PSD) were utilized to extract the MLAEP features. A Bayesian combination rule was then applied to evaluate the final decision by combining the results of the three individual classifiers. From total of 113 data samples only one was incorrectly classified and the misclassified sample belonged to a positive response. The system achieved a 99% accuracy rate for classifying anesthesia depth.

Modeling and Control of a Nonsquare Drug Infusion System

Abstract A model predictive control strategy is developed and tested on an nonlinear canine circulatory model for the regulation of hemodynamic variables under critical care conditions. Several cases are studied, including congestive heart failure, post-operative hypertension and a patient that moves from hypertensive to hypotensive conditions. The “nonsquare” (more process inputs than outputs) control system allows the independent management of the hemodynamic and venous circulation. The model predictive controller, which uses a different linear model depending on the patient condition, allows constraints to be explicitly enforced. The controller is initially tuned based on linear plant responses, then tested on the nonlinear plant model; the simulations verify the robustness of the control strategy

Multiple Input-Multiple Output Hemodynamic Control using Fuzzy Decision Theory

Abstract A fuzzy-logic based, automated drug-delivery system for managing hemodynamic states has been developed and validated on a non-linear canine circulatory model. This controller features: (1) a fuzzy decision analysis module for patient status detennination by assessing cardiac index, systemic vascular resistance index, and (2) a fuzzy hemodynamic management module utilizing dopamine, phenylephrine, nitroprusside and nitroglycerine for regulating mean arterial pressure, mean pulmonary arterial pressure, and cardiac output. A rule-based drug delivery scheduling program has been devised and incorporated to execute the therapeutic strategy as recommended by the decision analysis module.

Drug delivery control system challenge

The automated control of haemodynamic variables such as cardiac output and mean arterial pressure has been a goal of many research projects. While the design of such drug infusion controllers has often been based upon simplified linear models, the actual validation procedure requires a more comprehensive nonlinear model that is representative of the circulatory system. This phase is in fact often followed by animal experiments. Because of challenges involved in the direct design of controllers suitable for such a representative model, C language code for one such model is now publicly available through World Wide Web and FTP (File Transfer Protocol) access. Relevant performance specifications and constraints are presented as guidelines for developing suitable drug infusion controllers.