Abstracts

Rationale: Wearable EEG headsets that require minimum skin preparation, are easy to use, are wireless and also equipped with active dry-electrodes, are potentially changing the paradigm for the diagnosis and management of epilepsy. Such equipment could provide mobility for the patients during routine EEG and long term monitoring, reduce the set up time, increase comfort, and limit the need for the assistance of trained EEG technicians. We tested Zeto’s 6-channel easy-to-use headset (zEEG) equipped with 8 plug-in dry electrodes, including a reference and a DRL (driven right leg) electrode to reduce Common-mode interference (Figure A). In order to evaluate whether zEEG meets the technical requirements of clinical grade EEG or not, we performed simultaneous recordings of clinical EEG (Xltek, a clinical grade EEG system by Natus®) and zEEG on patients during clinical epilepsy monitoring sessions and compared signal quality and diagnostic value. Methods: Simultaneous recordings using Xltek and zEEG were performed on 11 children (age between 7 and 16 years old, mean=12.28 years, 8 female) over a 30-minute period per subject. In contrast to the clinical EEG, zEEG used dry electrodes, active amplification and wireless signal transmission by Bluetooth to a mobile device using an Android OS. The 6 commensurable electrode positions were Fp1, Fp2, C3, C4, O1 and O2.  The left mastoid was used as reference for zEEG and T3, T1 or A1 for the clinical EEG. The sampling rate of zEEG was 500 Hz while the clinical EEG was sampled at 256 or 512 Hz. We interpolated, normalized and time aligned the two types of signals. The quantification of the difference was done with respect to signal correlation, spectral correlation and signal-to-noise (SNR) ratio. In addition, we compared the usability of the systems by an EEG technologist. Results: The zEEG system demonstrated highly significant signal amplitude correlations with the Xltek (r>0.53 p < 0.00001; Figure B,C) when tested between corresponding channels. The spectral correlations between the two systems, including a broad range of frequencies (between 0.5 and 100 Hz), exceeded r>0.95 (Figure D). The signal to noise ratio (SNR) of zEEG was found to be on average 100 dB better than that of clinical EEG, based on comparing the SNR from spectral density between corresponding channels. Additionally, we observed significant reduction of set up time (average=20 minutes and 1 minute with Xltek and zEEG, respectively). The clinical evaluationof the two types of recordings is in progress. Conclusions: The comparisons of signals not only validated the quality of zEEG system, but the dry electrode system consistently outperformed the clinical system with regards to SNR in every comparison for all subjects in our cohort. Although dry electrode EEG systems have already demonstrated signal quality comparable to clinical systems (Fridman et al., 2016; Halford et al., 2016; Wyckoff, Sherlin, Ford, & Dalke, 2015), zEEG further improved the technical parameters. In addition, zEEG, for the first time, was able to integrate all necessary features and provided one of the fastest installation times reported combined with fully wireless operation. Moreover, zEEG’s dry electrodes with the active drive achieved a superior signal resolution and noise attenuation leading to a better SNR than that of the Xltek EEG system. These results validate the design principles guiding the development of a 19-electrode headset as per the 10-20 system. This type of validation is critical and timely for improving the design of dry-electrode headsets, as well as to prepare the clinical market for the acceptance of wireless EEG systems that are aiming to improve diagnosis and management of epilepsy. Funding: Zeto Inc.