Pitfalls in the analysis of electrogastrographic recordings

Abstract

Electrogastrography (EGG) is a noninvasive method to study gastric myoelectrical activity in humans. Because frequency characteristics are the most reliable parameters and visual analysis of the EGG recordings is notoriously difficult, automated frequency analysis, especially running spectrum analysis, is often used. However, EGG frequency spectra can be misinterpreted easily. Movement artifacts and noise from various sources can result in abnormal frequency spectra with significant power in the low-frequency and high-frequency range, or even make the EGG completely uninterpretable. Signals that differ from a sinusoid waveform have harmonics in the high-frequency range of the spectrum and may be interpreted as abnormal. Visual inspection of raw signals and frequency spectra remains essential in the analysis of EGG signals. The value of computerized analysis should not be overrated. EGG is an important research tool, but a clinical role still needs to be established.