Clinical Validations
The Aseega Technology has been clinically validated on both privately owned and publicly available data.
Privately owned data
This validation involved 30 subjects at the Sleep Laboratory - Department of Physiology and Functional Explorations of the H. Mondor Hospital (Créteil, France). The cohort included :
The 30 subjects underwent a standard polysomnography. The recordings were analysed by Aseega using only one EEG channel, and by 2 independant experts using the conventional channels.
For each recording, the 3 scorings were compared epoch by epoch. The following table shows the agreement between Aseega and the visual scorers when comparing 2-state and 3-state scorings(*). The sensitivity and predictive positive value for the detection of Wake and REM are also shown.
| 15 healthy volunteers(*) | 15 patients (**) | |
| Wake/sleep scoring agreement | > 95% | > 94% |
| Wake/REM/NREM scoring agreement | > 92% | > 88% |
| Wake detection: sensitivity, PVP | > 82%, 89% | > 83%, 93% |
| REM detection: sensitivity, PVP | > 82%, 89% | > 71%, 80% |
Publicly available data
The public-data validation involved the 8 healthy subjects available on the Physionet public database.
The recording characteristics of this database did not comply with the algorithm recommendations, especially in terms of recording sites (Pz-Oz instead of Cz-Pz or C4-O2) and analogical gain (fifty times lower than recommended).
Nevertheless, overall results are similar as reported in the following table:
| 8 healthy subjects(*) | |
| Wake/sleep scoring agreement | > 95% |
| Wake/REM/NREM scoring agreement | > 88% |
| Wake detection: sensitivity, PVP | > 85%, 87% |
| REM detection: sensitivity, PVP | > 63%, 91% |