Effects Of Digital Filtering On The Waveform & Peak Parameters Of The Auditory Brainstem Response

Publisher Website
Google Scholar
Abstract
Since the auditory brainstem response (ABR) has a narrower bandwidth than the spontaneous EEG, filtering can be used to improve the signal-to-noise ratio (SNR) prior to averaging. The purpose of this study is to compare the effects of different digital filters. Filters evaluated included lowpass (40–1600 Hz), bandpass (200–1600 Hz), and time invariant a-posterioriWiener filters. All filters were digitally implemented and had zero phase-shift. The filters were applied to averaged auditory brainstem responses based on different numbers of individual responses (128 to 1,024). The effects of the filters on both root-mean-square (rms) error and on Jewett wave parameters (peak latency and amplitude) were considered. The lowpass filter did not affect the rms error or wave amplitude; it did reduce the wave latency variability, but it did not affect the mean latency value. It greatly facilitated wave identification by removing high frequency noise. The bandpass filter behaved similarly, except that a slow wave activity in the response was removed also, resulting in an increased rms error. The performance of the Wiener filters was essentially the same as the lowpass filters except for a tendency to attenuate the waveform. In evaluating these filtering techniques, both waveform mean square error and wave latency and amplitude variability should be considered.

This publication has 0 references indexed in Scilit: