Detection of interictal epileptic events in EEG using ANN
Yusuf Uzzaman Khan · 1997
Describes a system for the detection of interictal spikes in an EEG using artificial neural networks (ANNs). The input layer of the ANN, which is a multilayer perceptron (MLP), utilises a feature vector which quantifies the slope, sharpness and autoregressive parameters extracted from the EEG every second. There are two classes, namely normal and epileptic. The MLP classification error rates evaluated for two subjects (referred to as A and B) are 6.04% and 7.33%, respectively. It is clear that the problem of subject specificity requires further work.