Performance of Different Average Methods for the Automatic Detection of Evoked Potentials

Abstract

The evoked potentials can be auditory, visual or somatosensory. Noise reduction is the first step in most biomedical signal processing systems. The quality and accuracy of the rest of the operations carried out on the signal depend to a large extent on the quality of the noise reduction algorithms that have been used in the preprocessing of the signal. The method commonly used to enhance the signal of interest is the coherent average, however, this technique has some limitations that justify the search for alternatives to detect or extract the characteristics of these signals. The weighted average is a possible alternative; however, this is still not appropriate enough when the signal may have outliers within the epoch. Trimmed average techniques have a better solution when in the presence of impulsive noise. The modified trimmed average is adapted for use in auditory evoked potentials. In this work, we compare different techniques of trimmed average, with the weighted average and the coherent average; using quality measures in the frequency domain for detect Auditory Brainstem Evoked Potentials. The results showed that the proposed method and his variants are superior to the rest of the used ones. The Q-Sample Modified measure offers the best result.