Comparing objective MLAEP detection techniques in time- and frequency-domain for anaesthesia monitoring

The mid-latency auditory evoked potential (MLAEP) has been used to monitor anaesthesia by means of objective methods, such as magnitude squared coherence (MSC), an objective response detection (ORD). This work compares MSC with a time-domain ORD technique, the evoked potential detector (DPE), using simulation and real EEG under auditory stimulation during propofol-induced anaesthesia. Using simulation, MSC distinguishes the changes in the amplitude and stretching better than DPE. For EEG under auditory stimulation, both MSC and DPE reflect infusion protocol. For the first, the choice of maximal response harmonic before infusion is a good criterion (even for simulated signal). Besides, the well-established distribution for the alternative hypothesis of existence of response allows the estimation of confidence limits to MSC and so the comparison with a pre-established threshold in order to suggest the transition from consciousness to unconsciousness and vice-versa. On the other hand, DPE is a much simpler technique, involving considerably fewer computational calculations. Therefore, both MSC and DPE can be pointed out as potential techniques to be used in anaesthesia monitoring based on MLAEP.