Background: As an essential indicator of labour and delivery, uterine contraction (UC) can be detected by manual palpation, external tocodynamometry and internal uterine pressure catheter. However, these methods are not applicable for long-term monitoring.
Objective: This paper aims to recognize UCs with electrohysterogram (EHG) and find the best electrode combination with fewer electrodes.
Methods: 112 EHG recordings were collected by our bespoke device in our study. Thirteen features were extracted from EHG segments of UC and non-UC. Four classifiers of the decision tree, support vector machine (SVM), artificial neural network, and convolutional neural network were established to identify UCs. The optimal classifier among them was determined by comparing their classification results. The optimal classifier was applied to evaluate all the possible electrode combinations with one to eight electrodes.
Results: The results showed that SVM achieved the best classification capability. With SVM, the combination of electrodes on the right part of the uterine fundus and around the uterine body's median axis achieved the overall best performance.
Conclusions: The optimal electrode combination with fewer electrodes was confirmed to improve the clinical application for long-term monitoring of UCs.
Keywords: Electrode combination; classifier; electrohysterogram (EHG); uterine contraction.