On-line EEG classification and sleep spindles detection using an adaptive recursive bandpass filter

This paper presents a novel adaptive filtering approach for the classification and tracking of the electroencephalogram (EEG) waves. In this approach, an adaptive recursive bandpass filter is employed for estimating and tracking the center frequency associated with each EEG wave. The main advantage inherent in the approach is that the employed adaptive filter only requires one coefficient to be updated. This coefficient represents an efficient distinct feature for each EEG specific wave and its time function reflects the nonstationarity of the EEG signal. Extensive simulations for synthetic and real world EEG data for the detection of sleep spindles show the effectiveness and usefulness of the presented approach.