Hugo Magalhães

Drug Delivery for Neuromuscular Blockade With Supervised Multimodel Adaptive Control

A major issue in drug delivery systems is the high level of uncertainty due to inter- and intrapatient variations in the dynamics of drug absorption and metabolism. This paper proposes an approach to tackle this problem based on supervised multimodel adaptive control (SMMAC). Although the specific case of neuromuscular-blockade-level control of patients subject to general anesthesia is considered, the overall procedure can be applied to the control of other physiological variables. Design guidelines to implement SMMAC are presented, together with clinical cases of patients undergoing general anesthesia, where atracurium is used as the blocking agent. The important role played by the selection of the observer polynomial in the supervisor is demonstrated.

Control of neuromuscular blockade in the presence of sensor faults

The problem of embedding sensor fault tolerance in feedback control of neuromuscular blockade is considered. For tackling interruptions of feedback measurements, a structure based upon Bayesian inference as well as a predictive filter is proposed. This algorithm is general and can be applied to different situations. Here, it is incorporated in an adaptive automatic system for feedback control of neuromuscular blockade using continuous infusion of muscle relaxants. A significant contribution consists in the experimental clinical testing of the algorithm in patients undergoing surgery.

IDENTIFICATION AND CONTROL OF POSITIVE AND COMPARTMENTAL SYSTEMS APPLIED TO NEUROMUSCULAR BLOCKADE

Abstract The paper presents a positive control law for feedback stabilization of a particular compartmental system. The features of this system allow to show that the application of the control law leads the system not only to a mass balance but also to a state of equilibrium, thus enabling the tracking of a constant reference. This new approach is used in the framework of neuromuscular blockade control and is illustrated by a simulation study. The problem of embedding identification methods in this compartmental model control approach is also considered and analysed

Observer dynamics in switching control: application to neuromuscular blockade

The effect of the choice of observer dynamics in switching control is illustrated through an application to neuromuscular blockade of patients undergoing surgery. The neuromuscular blockade model is subject to a high level of uncertainty due both to inter-patient variability as well as time variations. In order to achieve good results adaptation using switching control is used. The paper studies the effect of the selection of the observer dynamics in order to make the controller robust. A comparison with controller localization is made. It is shown that both methods allow to control a wide range of patient models, the inclusion of an observer has the advantage, with respect to localization, of not restricting the bank of models/controllers, thereby keeping the adaptation potential.

Statistical analysis of neuromuscular blockade response: contributions to an automatic controller calibration

Muscle relaxant drugs are currently given during surgical operations. The design of controllers for the automatic control of neuromuscular blockade benefits from an individual tuning of the controller to the characteristics of the patient. A novel approach to the characterization of the neuromuscular blockade response induced by an initial bolus at the beginning of anaesthesia is proposed. This approach is based on the statistical analysis of the data using principal components and Walsh-Fourier spectral analysis. These methods provide information about the patients dynamics, allowing the on-line autocalibration of the controller, using multiple linear regression techniques. Observed and simulated data are used to compare different approaches to the characterization of the bolus response.

Predictive adaptive control of the bispectral index of the EEG (BIS): using the intravenous anaesthetic drug propofol

The problem of controlling the level of unconsciousness measured by the Bispectral Index of the EEG (BIS) of patients under anaesthesia, is considered. It is assumed that the manipulated variable is the infusion rate of the hypnotic drug propofol, while the drug remifentanil is also administered for analgesia. Since these two drugs interact, the administration rate of remifentanil is considered as an accessible disturbance. In order to tackle the high uncertain present on the system, the predictive adaptive controller MUSMAR is used. The performance of the controller is illustrated by means of simulation with 45 patient individual adjusted models, which incorporate the effect of the drugs interaction on BIS. This controller structure proved to be robust to the remifentanil disturbance, different reference values and noise. A reduction of propofol consumption was also observed when comparing to the real clinical dose used for a similar BIS trend.

Observer design in switching control of neuromuscular blockade: clinical cases

This paper concerns the application of multiple model switched methods to the control of neuromuscular blockade of patients undergoing anaesthesia. Since the model representing the neuromuscular blockade process is subject to a high level of uncertainty due both to inter-patient variability and time variations, switched methods provide the adaptation capability needed to achieve the desired performance. The paper contributions are twofold: first, it is shown that, for the type of process control problem considered, the design of the associated observer must be carefully performed. Guidelines are provided for adequate selection of the characteristic polynomial defining the observer error dynamics. Second, clinical results using atracurium as blocking agent are reported in order to illustrate the use of the proposed control structure in actual clinical practice

A hybrid method for parameter estimation

In this paper a method is presented for plant model parameter estimation. The method combines the artificial neural networks ability for function approximation with a nonlinear least-squares regression technique using the Levenberg-Marquardt optimization method. This combination intends to overcome problems that arise when artificial neural networks or nonlinear least-squares regression are separately applied to parameter estimation, which is accomplished by means of potentiating each of the methods advantages. The estimation of atracurium effect concentration model parameters is used as a case study to show the efficiency of the proposed method.

Predictive adaptive control of unconsciousness - exploiting remifentanil as an accessible disturbance

The problem of controlling the level of unconsciousness measured by the BIS index of patients under anesthesia, is considered. It is assumed that the manipulated variable is the administration rate of propofol, while remifentanil is also administered for analgesia. Since these two drugs interact, the administration rate of remifentanil is considered as an accessible disturbance. A predictive adaptive controller structure that explores this fact is proposed and illustrated by means of simulation.

MULTIPLE STRATEGIES FOR PARAMETER ESTIMATION VIA A HYBRID METHOD: A COMPARATIVE STUDY

Abstract In this paper, two different strategies are considered for application of a previously proposed hybrid method designed to parameter estimation. This method combines the feedforward neural networks ability to produce initial parameter estimates close to the true values with the fast convergence of the Levenberg-Marquardt method using such estimates. The first strategy is of general applicability, while the second one is intended for models having a structure defined by various blocks in series. The neuromuscular blockade model parameter estimation problem is taken as the case study for comparing the performances of the two strategies.