RESUMO
We used a laboratory software package for data capture and data analysis (Spike 2, CED) to develop a computerised sleep staging. The modular software is programmed to extract the following data: power spectral density in the alpha and delta frequency bands with fast Fourier Transformation, K-complexes using pattern recognition on the basis of signal amplitude and zero level crossings, sleep spindles using auto-correlation of the EEG signal, rapid eye movements with pattern recognition of the bipolar EOG signal using amplitude and time differences, and muscle tone from the rectified and integrated submental EMG signal. These results are displayed together with the raw signals of EEG, EOG, and chin EMG on a user-defined time scale. Thus visual scoring of sleep stages is easter. Furthermore a decision table, containing the sequence and weighting factors for the extracted parameters, serves to perform an automated sleep scoring.
