Spectrum analysis of gastric emptying

The application of spectral analysis methods to radiologically gathered gastric data is discussed. The data is obtained by sampling counts of gamma rays present in the antrum section of the stomach of gastrointestinal patients after each patient was fed two scrambled eggs injected with a radioactive material. The primary goal of this investigation is to provide a noninvasive diagnostic tool to determine the stomach contraction pattern for normals and abnormals. The frequency components of contraction cycle are computed using three methods: (1) Fourier transform, (2) maximum entropy method (MEM), and (3) multiple signal classification (MUSIC). Statistical tests are performed on the frequencies of the normal and abnormal groups, to assess the discrimination powers of these methods. It is shown that Fourier transform performance is not adequate, while MEM and MUSIC produce very good results. The frequency components estimated agree with physiological findings for the normal and abnormal patients