Aseega is a sleep analysis technology based on EEG signal processing. The main difficulty in EEG analysis is its huge variability : inter-subjects, inter-nights for the same subject, and even during the night for one subject. Aseega copes with this problem with an auto-adaptive approach which is the result of 35 years of ongoing research.
Aseega Technology is a three step procedure: EEG preprocessing, sleep analysis, and classification into sleep stages.
In the preprocessing phase, the EEG signal goes through filtering, artifact rejection and subband extraction via a recording-adapted non-uniform filter bank.
Various signal processing techniques are used in the analysis phase in order to detect sleep microstructures such as spindles, α bursts and K-complexes, and to roughly localize the main wake and REM episodes.
The classification phase assigns sleep stages to epochs by feeding the results of the previous phase to a fuzzy logic iterative algorithm. The final hypnogram is obtained after contextual rule smoothing similar to the visual one.
It is a challenge to score sleep with one EEG channel only, using neither EMG nor EOG. Aseega Technology has met this challenge, as proved by the results of the clinical validations.
Aseega is now being used in sleep linked research domains.