Detection of gastric dysrhythmia using WT and ANN in diabetic gastroparesis patients

Gastric myoelectrical activity can be measured by a noninvasive technique called electrogastrography where surface electrodes are placed on the epigastric area of the abdomen. The electrogastrogram (EGG) signal is by nature a nonstationary signal in terms of its frequency, amplitude and wave shape. Unlike the other methods discrete wavelet analysis (DWT) was designed for nonstationary signals. For automatic assessment of EGG, we used artificial neural networks (ANNs) that have been widely employed in pattern recognition due to their great potential of high performance, flexibility, robust fault tolerance, cost-effective functionality and capability for real-time applications. So we developed a new method for classification of EGG based on DWT and ANN.