[Proper orthogonal decomposition and its application in EEG signal analysis].

In this paper is proposed the use of proper orthogonal decomposition (POD) for the decomposition and reconstruction of electroencephalograph (EEG) signal. The EEG signal can be denoted as the composition of eigenfunction modes and main coordinate components. The eigenfunction mode is a deterministic function of spatial variables, which represents the spatial information of electrodes. And the main coordinate component is a random function of time variables, which represents the real time signal of EEG. So a new way is pointed out for eigenvector space compression and prediction of electro potential values for EEG signal in any random scalp position. Orientation of further studies is also discussed.