[Channel selection for EEG-monitoring in anesthesia].

The conventional multichannel electroencephalogram is quite inconvenient for long-term monitoring in the operating theatre or intensive care unit. Recording of the EEG would be easier if a small number of channels was sufficient. Aiming at reduction of channels, leads from different regions of the scalp were analysed visually and with regard to their spectral content. METHODS. Electrode placements corresponded to the International 10/20 System (Fig. 1). EEG recordings were made with a conventional device (ES 12,000), a personal computer, and a spectral analyser. TWO-CHANNEL RECORDINGS. Retrospective analysis was performed on data from 392 patients (age 14-90 years) whose anaesthesia was induced with various anaesthetics/narcotics, for instance thiopental, ketamine, etomidate, halothane, and enflurane. The EEG was recorded using C3-P3 and Cz-A1. For each patient the changes of spectral parameters during the course of the induction were plotted and visually analysed. For statistical analyses a 30-s epoch of each patient was randomly selected from the first few minutes after the beginning of induction. TEN-CHANNEL RECORDINGS. In ten gynaecological patients (age 26-55 years) EEG recordings were performed during induction of anaesthesia with thiopental in combination with fentanyl, N2O and O2. The set of channels consisted of Fz'-Cb1, F3'-Cb1, Cz-Cb1, C3-Cb1, P3-Cb1, Oz-Cb1, Fz'-F3', F3'-C3, C3-P3, and P3-Oz. The electrodes Fz' and F3' were positioned on the forehead near to Fz and F3, respectively. These sites were chosen because they allow easy application of electrodes. The relationship between channels was calculated with Bravais-Person's coefficient of correlation for the power and the absolute power in the frequency bands delta (0.5-3.5 Hz), theta (3.5-7.5 Hz), alpha (7.5-12.5 Hz), and beta (> 12.5 Hz). RESULTS. In visual and statistical analyses of the two- and ten-channel recordings under the influence of anaesthetics/narcotics, similar changes of EEG activity could be observed in all channels. Although differences in the absolute power of the frequency bands were present, there was high conformity in the composition of the spectral content of the different channels. Classification of the EEG into stages of anaesthesia by means of a single channel led to consistent results for all channels. Alpha activity as leading feature of the awake state predominated occipitally. In channels including the region around the ears, contamination with EKG artifacts was observed. CONCLUSIONS. EEG patterns under the influence of different anaesthetics/narcotics are adequately represented by a reduced number of channels. For the choice of an appropriate set of channels the following aspects should be considered. Contamination with artifacts should be as low as possible, electrode sites should easily be accessible, and special features of the awake state should be identifiable. Experience with routinely conducted EEG recordings in the operating theatre and the intensive care unit showed that the channels C3-P3 or C4-P4 provide a sufficient basis for automatic staging of the depth of anaesthesia, which is implemented in the EEG monitor Narkograph.