A new frequency based automatic sleep analysis - description of the healthy sleep

A new system for complete automated sleep analysis has been developed (Somnoscript2000). Eighteen healthy young volunteers (ages 21-28 years, 9 females and 9 males) underwent a complete polysomnographic (PSG) registration. EEG from 17 electrode positions was recorded simultaneously with the polygraphic parameters (CATEEM(R)). The characterization of the recorded bioelectrical signals is based on the spectral frequency analysis by Fast Fourier Transformation (FFT). A new sleep-staging algorithm was developed, which allows for an objective presentation of the sleep phenomena. This system is characterized by a high reproducibility of the individual polysomnographic data throughout the recording with no significant difference between different all-night recordings of the subject. The sleep architecture shows considerable intersubject variability leading to differentiation of two main subpopulations: "good" and "poor sleepers".