Classification of multichannel uterine EMG signals using a reduced number of channels

Biocybernetics and Biomedical Engineering

Qiu

Zhou

et al. 2019

The aims of this study were to apply decision tree to classify uterine activities (contractions and non-contractions) using the waveform characteristics derived from different channels of electrohysterogram (EHG) signals and then rank the importance of these characteristics. Both the tocodynamometer (TOCO) and 8-channel EHG signals were simultaneously recorded from 34 healthy pregnant women within 24 h before delivery. After preprocessing of EHG signals, EHG segments corresponding to the uterine contractions and non-contractions were manually extracted from both original and normalized EHG signals according to the TOCO signals and the human marks. 24 waveform characteristics of the EHG segments were derived separately from each channel to train the decision tree and classify the uterine activities. The results showed the Power and sample entropy (SamEn) extracted from the un-normalized EHG segments played the most important roles in recognizing uterine activities. In addition, the EHG signal characteristics from channel 1 produced better classification results (AUC = 0.75, Sensitivity = 0.84, Specificity = 0.78, Accuracy = 0.81) than the others. In conclusion, decision tree could be used to classify the uterine activities, and the Power and SamEn of un-normalized EHG segments were the most important characteristics in uterine contraction classification.

Section: Discussionsupporting

confidence: 90%

Section: Discussionsupporting

confidence: 90%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Application of decision tree in determining the importance of surface electrohysterography signal characteristics for recognizing uterine contractions

Biocybernetics and Biomedical Engineering

Qiu

Zhou

et al. 2019

“…The competitive neural network, recurrent neural network (RNN), and convolutional neural network (CNN) [18,19] have been tested in image recognition and segmentation without additional feature extraction and selection. In particular, CNNs have been used in the recognition of physiological signals, including electromyograms [21,22], electrocardiograms (ECGs), and electroencephalography [23,24].…”

Section: Introductionmentioning

confidence: 99%

Evaluation of convolutional neural network for recognizing uterine contractions with electrohysterogram

Computers in Biology and Medicine

Jin

Wang

et al. 2019

Uterine contraction (UC) activity is commonly used to monitor the approach of labour and delivery. Electrohysterograms (EHGs) have recently been used to monitor UC and distinguish between efficient and inefficient contractions. In this study, we aimed to identify UC in EHG signals using a convolutional neural network (CNN). An open-access database (Icelandic 16-electrode EHG database from 45 pregnant women with 122 recordings, DB1) was used to develop a CNN model, and 14000 segments with a length of 45 s (7000 from UCs and 7000 from non-UCs, which were determined with reference to the simultaneously recorded tocography signals) were manually extracted from the 122 EHG recordings. Five-fold cross-validation was applied to evaluate the ability of the CNN to identify UC based on its sensitivity (SE), specificity (SP), accuracy (ACC), and area under the receiver operating characteristic curve (AUC). The CNN model developed using DB1 was then applied to an independent clinical database (DB2) to further test its generalisation for recognizing UCs. The EHG signals in DB2 were recorded from 20 pregnant women using our multi-channel system, and 308 segments (154 from UCs and 154 from non-UCs) were extracted. The CNN model from five-fold cross-validation achieved average SE, SP, ACC, and AUC of 0.87, 0.98, 0.93, and 0.92 for DB1, and 0.88, 0.97, 0.93, and 0.87 for DB2, respectively. In summary, we demonstrated that CNN could effectively identify UCs using EHG signals and could be used as a tool for monitoring maternal and foetal health.

Effect of electrode configuration on recognizing uterine contraction with electrohysterogram: Analysis using a convolutional neural network

Int J Imaging Syst Tech

Song

Qiu

et al. 2020

This paper aimed to evaluate the effect of various electrode configurations on applying a convolutional neural network (CNN) to recognize uterine contraction (UC) with Electrohysterogram (EHG) signals. Seven 8‐electrode configurations and thirteen 4‐electrode configurations were selected from the 4 × 4 electrode grid in the Icelandic 16‐electrode EHG database. EHG signals were divided into UC and non‐UC sections of 45 seconds and saved as images. Each 8‐electrode configuration with 7152 images and 4‐electrode configuration with 3576 images were applied to train CNN to recognize UCs. A scoring method was proposed based on the area under the curve (AUC) and the accuracy to evaluate the effect of electrode configurations on recognizing UCs. The EHG signals from the 4 electrodes on the upper left of the uterus showed the best classification performance (AUC = 0.79, Accuracy = 0.72, Score = 2.30).