N100 auditory potential and electroencephalogram discriminate propofol-induced sedation levels

In the present study, we evaluated the electroencephalogram (EEG) and auditory N100 potential (N100) before and during propofol-induced sedation. The aim was to test whether using EEG and N100 the level of sedation may be evaluated.

Twenty-nine cardiac surgery patients were studied. The EEG signal and the N100 potential were recorded at awake one day before the cardiac operation and two times after the operation, when the clinical level of postoperative propofol sedation was considered deep (Ramsay Score 6) and moderate (Ramsay Score 4). Discriminant analysis was used to select those spectral EEG and/or N100 variables which would predict the correct level of sedation best. The final classification was based on canonical discriminant functions and Mahalanobis' distance.

The spectral EEG variables (slow/fast-ratio, delta, and beta2 powers) predicted the correct level of sedation with 81% (canonical discriminant functions) and 80% (Mahalanobis' distance) accuracy. Similarly, the N100 (amplitude, latency, and the first principal component) predicted the correct level of sedation with 91% and 92% accuracy, and the combination of the EEG and N100 with 96% and 93% accuracy.

Our findings suggest that the combined use of EEG and N100 may help to differentiate the propofol-induced sedation levels, and thus be a useful compliment to clinical sedation scales.