Abstracts

Rationale: Providing routine EEG is still an unmet challenge today in electrically hostile and difficult environments such as the emergency department (ED) or intensive care units. The reasons for this are cost, space and time constraints, the difficulty of finding the expertise needed to apply electrodes and use the EEG machines, and the electrical environments that are prone to extracerebral as well as nonphysiological artifacts. We have tested a novel recording device, microEEG, which is designed to overcome these limitations and make routine EEG possible in such environments. We describe the microEEG and characterize its performance in relation to that of standard clinical equipment.Methods: We executed 30 minute volunteer recordings in three distinct ways. All recordings were institutionally approved and written informed consent was obtained from subjects. We compared the performance of (1) microEEG using Electro-Cap to that of a standard system, Nicolet Monitor ICU Monitor System by Carefusion using cup electrodes, in the EEG Laboratory at Downstate Medical Center; (2) microEEG using Electro-Cap in the ED to that of randomly selected standard recordings from a database of EEGs in ED and the ICU; and (3) microEEG using cup electrodes to that of the standard system recorded in parallel using a signal splitter. For analysis we used a diverse set of quantities in the time and frequency domains generally considered to be useful in automated algorithms for detection, prediction, and/or classification of EEG abnormalities (e.g. Delorme et al. (2007), Mormann et al. (2007), Minasyan et al. (2010), Wendling et al. (2009)). In the time domain, we computed the short time correlations between the two signals, as well as higher order statistics, and the Hjorth parameters of each signal. In the frequency domain, we computed the power spectral density, the band power for distinct frequency bands, the mean frequency, spectral edge frequencies (SEF), and spectral entropy.Results: In the microEEG recordings in the EEG Lab and the ED, there was good agreement in all frequency bands between microEEG and the standard system, except in the lowest frequencies, 0-4 Hz (due to differences between the two systems' hardware filters), and near 60 Hz (due to microEEG's greater immunity to ambient electrical noise). In the parallel recording, the short time correlations between microEEG and standard signals showed a relatively greater suppression of high frequency noise by microEEG. Quantities derived from the time and frequency domain illustrated a wide range of sensitivities to high frequency noise.Conclusions: We have found that microEEG's characteristics in the ED are noninferior to standard equipment in terms of a wide range of criteria of signal quality. The performance of microEEG suggests that it is suitable for EEG recordings in electrically hostile environments.