2019 年 57 巻 6 号 p. 198-205
In recent years, the application of electrogastrogram (EGG) to functional gastric disorder evaluation has attracted attention. However, in the case of EGG measurement, gastric electrical activity signals included in the EGG are easily contaminated by artifacts and noise from other organs with the same frequency band. Therefore, it is necessary to apply signal processing techniques that can extract only gastric electroactive components. On the other hand, since gastric electrical activity is non-linear and non-stationary, it is impossible to have an absolute answer when we evaluate the performance of signal processing. In order to solve this problem, we propose a model generating EGG mathematically, which can be applied to evaluate signal processing. First of all, we prepared a mathematical model with a white Gaussian noise source and a mathematical model of gastric electrical activity as proposed in previous research. In the next step, various numerical solutions were generated by changing two control parameters. These numerical solutions were compared with measured EGGs. EGGs used for model evaluation were measured from 4 subjects. Evaluations were performed using amplitude distribution and frequency distribution. The results showed that some numerical solutions closely resembled the measured EGGs. These optimal numerical solutions may be used for the evaluation of analytical methods.