An optimal feature set for seizure detection systems for newborn EEG signals

A novel automated method is applied to Electroencephalogram (EEG) data to detect seizure events in newborns. The detection scheme is based on observing the changing behavior of the wavelet coefficients (WCs) of the EEG signal at different scales. An optimizing technique based on mutual information feature selection (MIFS) is employed. This technique evaluates a set of candidate features extracted from the WCs to select an informative subset. This subset is used as an input to an artificial neural network (ANN) classifier. The classifier organizes the EEG signal into seizure or non-seizure activities. The training and test sets are obtained from EEG data acquired from 1 and 5 other neonates, respectively, with ages ranging from 2 days to 2 weeks. The optimized results show an average seizure detection rate of 94%.