Event-Based Control of Depth of Hypnosis in Anesthesia

Abstract Background and Objective: In this paper we propose the use of an event-based control strategy for the closed-loop control of the depth of hypnosis in anesthesia by using propofol administration and the bispectral index as a controlled variable. Methods: A new event generator with high noise-filtering properties is employed in addition to a \{PIDPlus\} controller. The tuning of the parameters is performed off-line by using genetic algorithms by considering a given data set of patients. Results: The effectiveness and robustness of the method is verified in simulation by implementing a Monte Carlo method to address the intra-patient and inter-patient variability. A comparison with a standard \{PID\} control structure shows that the event-based control system achieves a reduction of the total variation of the manipulated variable of 93% in the induction phase and of 95% in the maintenance phase. Conclusions: The use of event based automatic control in anesthesia yields a fast induction phase with bounded overshoot and an acceptable disturbance rejection. A comparison with a standard \{PID\} control structure shows that the technique effectively mimics the behavior of the anesthesiologist by providing a significant decrement of the total variation of the manipulated variable.