Livrables
An automatic sleep analyser based on polygraph signals (EEG, EOG, EMG, ECG) has been developed that describes sleep as a continuum of several process with a resolution of 1 second. It has been proven that the outputs of the analyser contain more information about normal and disturbed sleep than the traditional sleep staging standard by Rechtschaffen & Kales. Furthermore, it is completely data-driven and automatic and thus is reproducible and robust.
The analyser is based on the assumption that sleep can be described by three processes (wakefulness, slow-wave sleep, and non-slow-wave sleep including REM) plus spindle activity. Validation has been performed on a large database of healthy subjects of all adult ages and sleep-disturbed patients (result 01), including subjective and objective measures of sleep quality.
The analyser comprises state-of-the-art methodologies for signal processing and statistical data modelling and includes estimates of uncertainty providing information about possible artefacts or novelties in the input.
The sleep analyser is in its first prototype state. Although proven to contain novel information about sleep, a variety of improvements (e.g. concerning its sensitivity to signal artefacts and its dependency on recording laboratory settings) are necessary to turn it into an exploitable product. Furthermore, the software code will need to be validated and embedded into a commercial tool before being offered on a commercial basis.
Recherche de données OpenAIRE...
Une erreur s’est produite lors de la recherche de données OpenAIRE
Aucun résultat disponible