A model of gastric electrical activity in health and disease

The idea of diagnosing gastric dysfunction from noninvasive measurements of gastric electrical activity (GEA) is intuitively appealing, but the predictive accuracy of the cutaneous signal, especially that of its amplitude, is still in question. Mathematical modeling provides a means of investigating, analyzing, and predicting GEA measured percutaneously. In this study, a model of GEA applicable both in health and disease was developed and simulated for a cylindrical body surface. Body-surface maps of the simulated electrogastrogram (EGG) were generated at a 20 by 20 array of sites on the model's surface, and the accuracy of the percutaneous method in detecting simulated gastric electrical abnormalities was determined. The relationship between the amplitude of the simulated surface EGG and the velocity of propagation of the myogenic activity was also investigated. This was compared to a similar investigation of the fluctuations in the amplitude of the surface EGG with the velocity of propagation of the serosal activity measured in humans. The diagnostic accuracy of the measured cutaneous EGG in humans was also determined. The results obtained from the mathematical model show that the amplitude of the electrogastrogram increases with the propagation velocity of GEA. Similar results were obtained from the experimental measurements. The amplitude of the simulated and measured cutaneous signal correlated well (p < 0.05) with the phase shift of the simulated and measured activities, (-0.85, -0.54), respectively. Serosal normal activity, tachygastria, and uncoupling were detected 67%, 90%, and 0% of the time, respectively, at the cutaneous electrode in humans. In simulations, normal activity and tachygastria were accurately detected at all 400 sites on the surface. Uncoupling simulated with 50% of the myogenic sources "diseased" was detected at only 20 of the 400 sites. The results confirm that the amplitude of the cutaneous signal is a function of the velocity of propagation of the myogenic signal. It also confirms that while GEA in health may be accurately predicted from percutaneous recordings, frequency and phase/coupling abnormalities are poorly detected from single-channel electrogastrograms. This suggests the use of multiple-channel surface recordings in clinical electrogastrography.