Predication of premature neonates prognosis based on their electroencephalogram using artificial neural network

The electroencephalogram (EEG) is a signal that measures the electrical activity of the brain. In this paper, we proposed an artificial neural network (ANN) having as output the category of the newborn (healthy, sick or risky) and as input 14 parameters taken from inter-burst intervals of EEG signal. These parameters are detected using a Java application called EEGDiag dedicated to the analysis of EEG. We used a dataset of 397 EEG records detected at birth of premature newborns and their classification two years after birth: healthy, sick or risky. The aim of our work is to provide an automated predication of their prognosis based on their EEG using an ANN. We obtained satisfying results concerning sick class (performance 85.5%) and risky class (performance 90.3%), and we demonstrated the need of extracting new characteristics concerning healthy ones.