Autoregressive Analysis Applied to Surface and Serosal Measurements of the Human Stomach

A comparative study has been made on the surface and serosal data recorded from the human stomach. Autoregressive technique has been used for frequency analyses. It is observed that slow waves in both the surface and the serosal data can be detected with reasonable success using a moderate model order. An examination of the frequency spectra reveals that while the serosal signals have a well-defined harmonic pattern, the surface signals are noiselike. Some 9-10 c/min spike activity seems to be present in both kinds of recordings. Attempts have been made to single out the "surface effects" making use of the linear properties of the autoregressive filters and our results indicate that the main surface interferences seem to be related to respiration artifacts and the 9-10 c/min spike activity.