Mathias Guillard1, Pascal Van Beers1, Marie Coroenne, Laurent Lely, Fabien Sauvet, Mounir Chennaoui2
10:15 - 11:00 | Thu 11 May | Einstein Auditorium Foyer | ThPoS.12
For the ecological context, it is relevant to collect electroencephalographic (EEG) signals and to evaluate their technical quality before long acquisition. A four EEG channels acquisition unit (AU) (Actiwave®) was implemented with a specific interface (SI). The available tools to manage signal quality are "impedance check" module or the complete signal processing software tools for viewing and checking the signal quality. No "impedance check" module was available for the AU, that is why a simple software called "EET" (EEg Testing) was developed in order to visualize recorded data, to compute signal spectral density on manually selected epoch and to check EEG technical quality. For signal verification, technical artifacts ("base-line drift" and "electrode problem") are calculated using judgment criteria on frequency bands of 0-0.5 Hz and 0.5-4 Hz. Then the physiological status is evaluated with a spectral density analysis in the alpha (8-12 Hz) frequency band, which is higher when a subject closes his eyes with respect to the open eyes, for 30 seconds each. Ten healthy voluntary pilots from the Maritime Patrol Squadrons were monitored with the device and the SI and the technician tested the EET "off line" software for two EEG derivations (C3-M2 and O1-M2) at 128 Hz frequency rate. In the case of detection of technical artifacts, for O1-M2, the detection accuracy (quality of the technical signal compared to the diagnostic of electroencephalographers) was 81.9 % (resp. 86.2%) and the area under the curve of the characteristics of the receiver operating was 0.87 (resp. 0.94) for "electrode problem" (resp. "baseline drift problem"). Concerning the physiological status, detection accuracy was 86.2 % and the area under the curve of the characteristics of the receiver operating was 0.94 for O1-M2. These results are interesting and promising for the development of ambulatory systems to collect the most accurate EEG signals.