Related Experiment Videos
Nonlinear model-based predictive control of non-depolarizing muscle relaxants using neural networks
M Lendl1, U H Schwarz, H J Romeiser
1Department of Theoretical Foundations of Electrical Engineering, University of Erlangen-Nuremberg, Erlangen, Germany. lendl@late.e-technik.uni-erlangen.de
Journal of Clinical Monitoring and Computing
|February 6, 2003
Summary
This study introduces a self-adapting control system for neuromuscular blockade using artificial neural networks. The system ensures stable anesthesia with mivacurium, showing improved performance over existing closed-loop controllers.
Area of Science:
- Anesthesiology
- Biomedical Engineering
- Artificial Intelligence in Medicine
Background:
- Neuromuscular blockade monitoring is crucial for patient safety during anesthesia.
- Current closed-loop systems require complex pharmacological models.
- Short-acting neuromuscular blocking agents like mivacurium offer clinical advantages.
Purpose of the Study:
- To develop and evaluate a self-adapting, model-based predictive control system for neuromuscular blockade.
- To achieve stable neuromuscular blockade without prior knowledge of pharmacodynamic processes.
- To assess the system's performance using mivacurium in a clinical setting.
Main Methods:
- A closed-loop control system utilizing artificial neural networks was developed.
- The system measured neuromuscular activity via EMG and controlled mivacurium infusion.
- A clinical study involved 35 patients (ASA I-III) receiving mivacurium.
Main Results:
- The novel system demonstrated favorable performance with a mean T1 error of -0.32 +/- 1.7.
- Performance compared favorably to previously demonstrated closed-loop controllers.
- The system adapted effectively without explicit pharmacodynamic modeling.
Conclusions:
- The self-adapting predictive control system offers a stable and adaptable method for managing neuromuscular blockade.
- Artificial neural networks facilitate parameter adjustment without detailed pharmacodynamic knowledge.
- This technology shows promise for application with other anesthetic drugs, including hypnotics.