Time-frequency identification of the slow wave in external electroenterogram

Abstract

The aim of present study is to identify slow wave (SW) from small bowel myoelectrical activity (electroenterogram) recorded on abdominal surface. Breath interference can confuse the SW identification because both signals are closed in frequency. Recordings from external abdomen would be a noninvasive method to obtain the myoelectrical signal. Internal and external recordings were acquired simultaneously from three Beagle dogs (in eight recording sessions). Six time-frequency distributions were calculated for abdominal myoelectrical signal in one minute periods: Wigner-Ville distribution (WVD), Choi-Williams distribution (CWD), Zhao-Atlas-Marks (ZAM) and spectrograms with three window lengths (SPsw, SPmw, SPIw). Results show that CWD is useful for identifying small intestine myoelectrical signal at low frequencies; i.e., SW could be defined from time-frequency analysis. Breath interference has less energy in time-frequency domain than slow wave components. Furthermore, the SW rapid phase can be detected as an energy distribution in high frequencies and precisely located in time. Spectrogram is conditioned to select between short window (SPsw), which can not discriminate SW and respiration artifacts; or large window (SPIw), which can not detect SW rapid transition. In conclusion, time-frequency studies show that SW can be identified in abdominal recording of small bowel myoelectrical activity.