Automated interpretation of brainstem auditory evoked potentials: a prototype system

A description is given of a preliminary version of an expert decision-support system for interpretation of brainstem auditory evoked potentials. The design combines an analytical signal-processing module with a rule-based module that incorporates the kinds of heuristic criteria used by human interpreters. The prototype system determines whether a sensory response is present in the waveforms; if so, it identifies peak V, the most prominent peak. A level of confidence is associated with the identification. Results of applying the system to a test set of 20 waveforms, including some waveforms from comatose patients that showed no sensory response, are described. The system performed well when a response was present, but it was not as effective in interpreting waveforms with no response present.