Definition of a depth of anaesthesia index using fuzzy inductive reasoning (FIR)

The objective of this study was the definition of a depth of anaesthesia index using measured data of Auditory Evoked Potentials (AEP). Middle Latency AEP are described in a number of articles as a promising measure of anaesthetic depth. However both extraction and interpretation of AEP have several unsolved problems. The extraction is usually done by Moving Time Averaging (MTA) over a vast number of sweeps, which produces a delay of typically one-minute. This problem was addressed by applying a system identification model, an Auto Regressive Model with Exogenous input (ARX) which enables extraction of the AEP within 15 sweeps. The interpretation is frequently done by visual inspection of the latency of the Nb peak of the AEP. Instead of this a FIR model, which includes the changes in both amplitudes and latencies, was applied. The results showed that it was possible to classify the AEP into five classes corresponding to the states, awake, drowsy, light anaesthesia, anaesthesia and deep anaesthesia.