Abstract
It is important to know the human's sleep process. EEG (Electroencephalogram) sleep stage is defined according to the depth of human sleep. We have been also studying and developing the sleep stagers based on APSS 's rules3). However, these rules are insufficient for such systems. Then, extraction of significant information from EEG signal is tried in order to supplement these rules. Therefore, in this paper, more detailed analysis was tried in the frequency domain by using the continuous wavelet analysis. The setting method of the damping coefficient of Gabor function was proposed. Wavelet transform results show the characteristic transition phenomenon of a principal frequency component. And it is confirmed that the automatic discrimination of human EEG sleep stages based on wavelet transform can be executed in the reasonable accuracy.
