DETECTION OF SEIZURE ACTIVITY IN EEG BY AN ARTIFICIAL NEURAL-NETWORK - A PRELIMINARY-STUDY

Neural networks, inspired by the organizational principles of the huma n brain, have recently been used in various fields of application such as pattern recognition, identification, classification, speech, visio n, signal processing, and control systems. In this study, a two-layere d neural network has been trained for the recognition of temporal patt erns of the electroencephalogram (EEG). This network is called a Learn ing Vector Quantization (LVQ) neural network since it learns the chara cteristics of the signal presented to it as a vector. The first layer is a competitive layer which learns to classify the input vectors. The second, linear,layer transforms the output of the competitive layer t o target classes defined by the user. We have tested and evaluated the LVQ network. The network successfully detects epileptiform discharges (EDs) when trained using EEG records scored by a neurologist. Epochs of EEC containing EDs from one subject have been used for training the network, and EEGs of other subjects have been used for testing the ne twork. The results demonstrate that the LVQ detector can generalize th e learning to previously ''unseen'' records of subjects. This study sh ows that the LVQ network offers a practical solution for ED detection which is easily adjusted to an individual neurologist's style and is a s sensitive and specific as an expert visual analysis. (C) 1996 Academ ic Press, Inc.