Demonstration of a neural network expert system for recognition of glaucomatous visual field changes.

Expert systems (ESs) are computer programs designed to make decisions in a manner similar to the way human experts make them. There are many forms of ESs, ranging from those that are given specific If-Then rules to those that are capable of deriving their own rules. Some ESs use neural network simulations to analyze patterns of data. In this project, a commercially available neural network shell program was used to demonstrate the development of an ES capable of analyzing the visual field loss patterns produced by glaucoma. Potential applications of ESs for the diagnosis of glaucoma are discussed.