ADAPTIVE SPECTRAL ANALYSIS OF GASTRIC MYOELECTRICAL ACTIVITY IN PATIENTS WITH RENAL FAILURE | ||||
| The International Conference on Electrical Engineering | ||||
| Article 19, Volume 2, 2nd International Conference on Electrical Engineering ICEENG 1999, November 1999, Page 158-165 PDF (1.68 MB) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/iceeng.1999.62305 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
| E. EI-Samahy1; Abdalla S. A. Mohamed2; S. Alian1 | ||||
| 1Egyptian Armed Force. | ||||
| 2Department of systems & Biomedical Engineering, Faculty of Engineering, Cairo University, Egypt. | ||||
| Abstract | ||||
| Abnormalities in the gastric electrical activity (GEA) and cutaneous Electrogastrogram (EGG) have been studied in variety of clinical conditions related to digestive system and its motility. Unfortunately, the impact of renal failure on GEA remains in question especially with patients performing regular Haemodialysis. The aim of this study is to determine whether EGG would be changed in patients with renal failure even after Haemodialysis to become as healthy control. An experiment was conducted on two groups: The first group contained (11) healthy volunteers with average age (23) years, free from any diseases related to digestive system and renal failure. The second group was selected from patients suffering from renal failure and performing regular Haemodialysis twice a week (10 patients) with average age of (27.5) years. EGG was measured non-invasively from both groups for 20 minutes and digitized with sampling frequency 2 Hz before pre-processing stage. In the preprocessing stage, the EGG signal becomes noisily suppressed using an adaptive enhancement technique. Based on parametric modeling, and the nonstationarity of EGG, the instantaneous EGG signal is modeled by adaptive Autoregressive (AR) model. The power spectrum of EGG signal can be calculated from those time-varying parameters, which are updated with on-line EGG samples. Therefore, the instantaneous frequency of EGG signal would be tracked. This approach has the ability to provide how far the fluctuations with renal failure from normal will be. The accuracy of classification with adaptive AR coefficients using RLS algorithm shows higher percentage for discrimination than other techniques as LMS. The percentage of classifying patients with renal failure before Haemodialysis from normal persons reaches 98.75%. It is expected that the GEA of the digestive system would become normal after Haemodialysis, but evident differences in EGG power spectrum were found, and patients after Haemodialysis were discriminated easily from normal with percentage 100% using scatter diagram. Finally, these results showed that GEA did not reach the control level even after Haemodialysis (5hours/session). | ||||
| Keywords | ||||
| Electrogastrogram (EGG); Adaptive AR; Renal Failure. Haemodialysis | ||||
|
Statistics Article View: 103 PDF Download: 177 |
||||
