Modeling of neuromuscular blockade system using neural networks

Delivering drugs for muscle relaxation is known to be a delicate process, which is highly nonlinear in nature. On of the most commonly-used drugs to create neuromuscular blockade is atricurium. Here, the authors develop a dynamic neural network to model neuromuscular blockade system when atricurium is used for paralysis. The minimum-complexity neural modeling algorithm used here is based on the PAC learning theory and is shown to perform similarly on the testing and training data. The resulting model is proved to be stochastically stable and can be used as a reliable and accurate model of neuromuscular blockade system.