Adaptable preprocessing units and neural classification for the segmentation of EEG signals

In this contribution, a methodology for the simultaneous adaptation of prep rocessing units (PPUs) for feature extraction and of neural classifiers tha t can be used for time series classification is presented. The approach is based upon an extension of the backpropagation algorithm for the correction of the preprocessing parameters, In comparison with purely neural systems, the reduced input dimensionality improves the generalization capability an d reduces the numerical effort, In comparison with PPUs with fixed paramete rs, the success of the adaptation is less sensitive to the choice of the pa rameters. The efficiency of the developed method is demonstrated via the us e of quadratic filters with adaptable transmission bands as preprocessing u nits for the segmentation of two different types of discontinuous EEG: disc ontinuous neonatal EEG (burst-interburst segmentation) and EEG in deep stag es of sedation (burst-suppression segmentation).