CLASSIFICATION OF BRAINSTEM AUDITORY EVOKED POTENTIALS USING TIME AND FREQUENCY DOMAIN FEATURES.

Brainstem auditory evoked potentials (BAEPs) are widely used during neurological diagnosis. The development of statistical classifiers to classify the BAEPs is reported. From the time domain the latencies of the initial five peaks and the amplitude of the fifth peak were selected as features, and from the frequency domain the total powers in each of the three principal bands of the power spectra of the BAEPs were used. Classification accuracies of 85% for the classifiers designed with time-domain features and 77% for the classifiers designed with frequency-domain features were obtained.