A Basic Study on an Application of Empirical Mode Decomposition to Electrogastrogram Extraction

Abstract

Normal Stomach has a pacemaker which generates three cycle-per-minute of rhythm. Monitoring the rhythm can detect the signs of stomach disorders which disturb periodical activity of the pacemaker. Our final goal is intracutaneous non-invasive measurement of gastric electrical activity. In order to approach the goal rejections of the noise and artifact are necessary. In this study, empirical mode decomposition(EMD) was applied to extract a signal in a frequency band of electrogastrogram(EGG). First of all, we measured the EGG before and after meal load. The measured signal was purified by EMD. EGG signal extraction process by using EMD is; 1) Signal decomposition by EMD. 2) Selection of the decomposed signals in the EGG frequency band. 3) Signal reconstruction using the selected decomposition signal. A comparison between the results by EMD and conventional frequency filtering, shows that extracted signal quality by EMD is better than the conventional technique.