Knowledge-based flash evoked potential recognition system

摘要

A knowledge-based evoked potential monitoring and recognition system has been developed to automatically monitor patients' flash evoked potentials and recognize the signals. It also automatically labels the peaks and troughs of the evoked potentials, and calculates the latencies and amplitudes of the evoked potentials. This knowledge-based system was implemented using object-oriented programming. Due to the variability of flash evoked potentials, the system uses temporal and spatial contexts to achieve reliability. The overall accuracy (3 correct of 3 waves, plus 2 correct of three in the presence of noise or low amplitude) is around 95%.