Dimension reduction of RCE signal by PCA and LPP for estimation of the sleeping

Abstract

Irregular hour and suffering from stress cause driver doze and falling asleep during important situations. Therefore, it is necessary to know the mechanism of the sleeping. In this study, we distinct the sleep conditions by the rhythmic component extraction (RCE). By using this method, a particular EEG component is extracted as the weighted sum of multi-channel signals. This component concentrates the energy in a certain frequency range. Furthermore, when the weight of a specific channel is high, this channel is thought to be significant for extracting a focused frequency range. Therefore, the sleep conditions are analyzed by the power and the weight of RCE. As for weight analysis, the principal component analysis (PCA) and the locality preserving projection (LPP) are used to reduce the dimension. In the experiment, we measure the EEG in two conditions (before and during the sleeping). Comparing these EEGs by the RCE, the power of the alpha wave component decreased during the sleeping and the theta power increased. The weight distributions under two conditions did not significantly differ. It is to be solved in the further study. © 2011 Springer-Verlag.