Comparison of adaptive filtering in time-, transform- and frequency-domain: an electrogastrographic study.

Abstract

Adaptive cancellation of motion artifacts in the electrogastrogram (EGG) is presented in this paper. The EGG is a surface measurement of gastric electrical activity. Like other noninvasive electrophysiological measurements, the EGG contains motion artifacts. A number of papers have been published on the adaptive cancellation of motion artifacts or interferences in biomedical signals. Adaptive filtering was performed in time domain in almost all of the previous publications. In this paper, however, three different sorts of adaptive filters were investigated and their efficiencies in cancellation of motion artifacts were compared with each other. These include time-domain, transform-domain, and frequency-domain adaptive filters. A series of simulations were conducted to investigate the performance of these adaptive filters in cancellation of respiratory and motion artifacts. The results show that the frequency-domain adaptive filter has superior performance over the time- and transform-domain adaptive filters in the cancellation of stationary respiratory artifacts in the EGG. Although results focus on the EGG, this paper provides useful information for adaptive filtering of other biomedical signals.