EVOKED POTENTIAL PROCESSING AND PATTERN RECOGNITION.

Two approaches to the classification of brain stem auditory evoked potentials are described. In the first approach, the conventional averaged evoked potential is zero-phase bandpass filtered and classified using syntactic methods. Training and test runs of 70 and 60 subjects, respectively, resulted in a classification accuracy of 83% with few false negatives. The second approach, using single stimulus brain stem auditory evoked potentials extracted by adaptive noise cancellation, was tested by simulation.