The use of electrohysterogram (EHG) - uterine muscular activity signals - along with artificial intelligence models would help in the prediction of preterm delivery and thereby save the lives of many early-delivered infants through the necessary early medical care. The informative portions in the EHG recording which could aid in the prediction of this risky delivery are the ones related to the uterine muscular contractions. Hence, the more we perform a precise segmentation of these contraction signals, the more valuable and discriminative are the data provided to the AI models, which will in turn reflect a better learning process and prediction outcome. This paper presents a new algorithm called Slope of Tangent (SOT) for an enhanced segmentation of uterine muscular contraction signals. The method is further compared to the up-to-date uterine contraction automatic segmentation methods. The results showed that the method allowed a higher number of full detections ($F D=90$) and partial detections ($P D=263$) in comparison to the wavelet $H 2$ nonlinear correlation method (wavh2) ($F D=42$ and $P D=145$) and the sample entropy method ($F D=13$ and $P D=138$). Further work should be done in order to reduce the number of false detections given by the SOT method. In addition, the method should be further validated on open-source databases.

Premature delivery is a leading cause of fetal death and morbidity, making the prediction and treatment of preterm contractions critical. The electrohysterographic (EHG) signal measures the electrical activity controlling uterine contraction. Analyzing EHG features can provide valuable insights for labor detection. In this paper, we propose a framework using simulated EHG signals to identify features sensitive to uterine connectivity. We focus on EHG signal propagation during delivery, recorded by multiple electrodes. Simulated EHG signals were generated using electrical diffusion (ED) and mechanotransduction (EDM) to identify which connectivity methods and graph parameters best represent uterine synchronization. The signals were simulated in two scenarios: using only ED by modifying tissue resistance, and using both ED and EDM by varying mechanotransduction model parameters. A matrix of 16 surface electrodes was used for the simulations. Our results show that a simplified electromechanical model can monitor uterine synchronization. Feature selection using Fscore on real and simulated EHG signals highlighted that the best features for detecting mechanotransduction shifts were H2 alone or combined with Str, R2(PR), and ICOH(Str). The best features for detecting electrical diffusion shifts were H2, Eff, PR, and BC.

Preterm labor represents the prominent cause of mortality and morbidity, highlighting the important need for improved preterm contraction prediction and management. One promising approach to resolving this challenge is to analyze the electrohysterographic (EHG) signal, which records the electrical activity regulating uterine contractions. Analyzing the features of the EHG signal contributes valuable data to detect labor. In this paper, we propose a new framework using simulated EHG signals to identify features sensitive to uterine connectivity. We focus on EHG signal propagation during labor, recorded by multiple electrodes. We simulated EHG signals in different groups to determine which connectivity methods and graph parameters best represent the two main factors driving uterine synchronization: short-distance propagation (via electrical diffusion, ED) and long-distance synchronization (via mechanotransduction, EDM). Using the uterine model, signals were first simulated using just electrical diffusion by modifying the tissue resistance; second, signals were simulated using ED and mechanotransduction by holding the tissue resistance constant and varying the model parameters that affect mechanotransduction. We used the bipolar technique to construct our simulated EHGs by modeling a matrix of 16 surface electrodes organized in a 4 × 4 matrix placed on the pregnant woman’s abdomen. Our results show that even a simplified electromechanical model can be useful for monitoring uterine synchronization using simulated EHG signals. The differences seen between the selection performed by Fscore on real and simulated EHG signals show that when employing the mean function, the best features are H2(Str), FW_h2 alone, and in combination with PR, BC, and CC. The best characteristics that demonstrate a shift in the mechanotransduction process are H2 alone or in combination with Str, R2(PR), and ICOH(Str). The best characteristics that demonstrate a shift in electrical diffusion are H2 alone and in combination with Eff, PR, and BC.

1) Introduction: Preterm labor (PL) has globally become the leading cause of death in children under the age of 5 years. One of the most significant keys to preventing preterm labor is its early detection. 2) Objectives: The primary objectives of this study are to address the problem of PL by providing a new approach by analyzing the electrohysterographic (EHG) signals, which are recorded on the mother's abdomen during labor and pregnancy. 3) Methods: The EHG signal reflects the electrical activity that induces the mechanical contraction of the myometrium. Because EHGs are known to be non-stationary signals, and because we anticipate connectivity to alter during contraction (due to electrical diffusion and the mechanotransduction process), we applied the windowing approach on real signals to identify the best windows and the best nodes with the most significant data to be used for classification. The suggested pipeline includes: i) dividing the 16 EHG signals that are recorded from the abdomen of pregnant women in N windows; ii) apply the connectivity matrices on each window; iii) apply the Graph theory-based measures on the connectivity matrices on each window; iv) apply the consensus Matrix on each window in order to retrieve the best windows and the best nodes. Following that, several neural network and machine learning methods are applied to the best windows and best nodes to categorize pregnancy and labor contractions, based on the different input parameters (connectivity method alone, connectivity method plus graph parameters, best nodes, all nodes, best windows, all windows). 4) Results: Results showed that the best nodes are nodes 8, 9, 10, 11, and 12; while the best windows are 2, 4, and 5. The classification results obtained by using only these best nodes are better than when using the whole nodes. The results are always better when using the full burst, whatever the chosen nodes. 5) Conclusion: The windowing approach proved to be an innovative technique that can improve the differentiation between labor and pregnancy EHG signals.

1) Objectives: Preterm birth caused by preterm labor is one of the major health problems in the world. In this article, we present a new framework for dealing with this problem through the processing of electrohysterographic signals (EHG) that are recorded during labor and pregnancy. The objective in this research is to improve the classification between labor and pregnancy contractions by using a new approach that focuses on the connectivity analysis based on graph parameters, representative of uterine synchronization, and comparing neural network and machine learning methods in order to classify between labor and pregnancy.2) Material and methods: after denoising of the 16 EHG signals recorded from pregnant women abdomen, we applied different connectivity methods to obtain connectivity matrices; then by using the graph theory, we extracted some graph parameters from the connectivity matrices; finally, we tested different neural network and machine learning methods on the features obtained from both graph and connectivity methods in order to classify between labor and pregnancy.3) Results: The best results were obtained by using the logistic regression method. We also evidence the power of graph parameters extracted from the connectivity matrices to improve the classification results.4) Conclusion: The use of graph analysis associated with machine learning methods can be a powerful tool to improve labor and pregnancy classification based on the analysis of EHG signals. (c) 2021 AGBM. Published by Elsevier Masson SAS. All rights reserved.

This paper presents the sensitivity analysis (SA) of a model developed to simulate the uterine electrical activity recorded on the woman's abdomen (the electrohysterogram, EHG). This model contains different sub-models that permit to simulate the electro-chemical behavior of the uterine muscle during a contraction (EHG), the abdominal conducting volume as well as a personalized geometry of the uterus. We based our sensitivity analysis on the Morris elementary effects method, a well-known screening method suited for large dimension and complex models. We adapted the classical Morris sensitivity measures to deal with the non-uniform distribution of the elementary effects. The SA tested the effect of the 32 parameters of the model on 5 classical features computed from the simulated EHG. The results indicate a nonlinear influence of the parameters on the EHG features. They permit to evidence the most important parameters as well of the negligible ones for the further use of the uterine model.

In order to study the anatomical variability of the uterus induced by pregnancy, a parametrization of gravid uterine geometry based on principal component analysis (PCA) is proposed. Corresponding meshes used for PCA are created by a ray description technique applied to a reference mesh. A smoothed voxel-based methodology is applied to determine the reference mesh from a database of 11 real shapes produced by the FEMONUM project. The ray-based correspondence technique is compared to two existing methods (He, Giessen) as well as a proposed mixed method. Principal component analysis results are based on a database of 11 existing shapes. Results of the parametrization show that 90% of the total variance of the database can be represented with four new shape parameters and that a wide spectrum of shapes can be generated. Graphical Abstract Proposed correspondence technique compared to existing methods.

BACKGROUND:Recent years have seen an increased interest in electrohysterogram (EHG) signals as a means to evaluate the synchronization of uterine contractions. Several studies have pointed out that the quality of signal processing - and hence the interpretation of measurement results - is affected significantly by the choice of measurement technique and the presence of non-stationary frequency content in EHG signals. To our knowledge, the effect of time variance on the quality of EHG signal processing has never been fully investigated. How best to process EHG signals with the goal of distinguishing labor-induced contractions from their harmless, pre-labor cousins, remains an open question.METHOD:Our methodology is based on three pillars. The first consists of a new method for EHG preprocessing in which we apply a second-order Butterworth filter to retain only the EHG fast-wave, low-frequency band (FWL), then use a bivariate piecewise stationary pre-segmentation (bPSP) algorithm to segment the EHG signal into stationary parts. The second pillar addresses the estimation of connectivity and directionality using three methods: nonlinear correlation coefficient (h2), general synchronization (H), and Granger causality (GC). The third pillar is related to signal classification and discrimination between pregnancy and labor using receiver operating curves (ROC) and connectivity and direction maps. For this purpose, we analyze the impact of four factors on data processing efficiency: i) method of connectivity detection, ii) effect of piecewise stationary segmentation preprocessing, iii) retained frequency content and iv) electrode configuration used for EHG recording (bipolar vs. unipolar).RESULTS:Our results show that piecewise signal segmentation and filtering considerably improves classification performance and statistical significance for some connectivity methods, in particular the h2. To this end we propose a new approach (detailed below) for h2 called Filtered-Windowed (FW) h2 that better highlights the differences between pregnancy and labor in the connectivity matrix and directionality maps.CONCLUSIONS:This is the first comparative study of the effects of multiple processing factors on connectivity measurement efficiency. Our results indicate that appropriate preprocessing can improve the differentiation of pregnancy and labor-induced contraction signals and may lead to innovative applications in the prevention of preterm labor.

In this paper, we propose an automatic prediction system allowing to predict high probability of birth within 1–2 weeks from the EHG measurement [1–3]. Despite continuous clinical routine improvements, the preterm rate remains steady. With the aim of avoiding long hospitalization for pregnant women we propose an embedded system which acquires and processes EHG signals. We have already proposed a detection and recognition system for intrauterine contractions using directly the EHG signal obtained from a matrix of 16 electrodes. Since the measurements must be made at home, the decrease of computation power is an important constraint. In this work, we compare the results of the preterm birth prediction algorithm using a filtering step and with only the raw signals. The filtering step is applied directly on the raw signals or only on the automatically detected contractions to reduce computation time. We have applied in this, different filtering methods as denoising step to analyse their influence on the global classification performances. Two types of filtering are evaluated separately or combined: Canonical Correlation Analysis (CCA) and Empirical Mode Decomposition (EMD). The EMD decomposes a signal into a collection of oscillatory modes, called IMFs, which represent fast to slow oscillations in the signal. The CCA is a Blind Source Separation (BSS) method which assumes that the observed multichannel signals reflect a linear combination of several sources which are associated to underlying physiological processes, artefacts, and noise. The global classification results are compared between filtered and not filtered signals.

Over-segmentation in non-stationary signals remains a challenge in several automatic segmentation methods. In this context, most of automatic segmentation methods of contractions in multichannel uterine Electromyogram signals (Electrohysterogram signals, EHGs) encounter the problem of detecting other events which are not corresponding to contractions labeled by expert. In addition, the great advantage of using wavelet manifests by its ability to separate the fine details in a signal. Many studies applied wavelets for sake of detecting and analyzing abrupt changes in non-stationary signals. This study is another step of our project focused on the automatic contraction segmentation of EHGs. Indeed, we will focus on the application of dynamic cumulative sum method (DCS) with over-segmentation elimination techniques on bipolar EHGs and details after wavelet decomposition of those signals. Detected events are then compared to contractions identified by expert using Margin validation test with and without application of over-segmentation techniques. Regarding the obtained sensitivity and other events rate of methods we find that DCS with over-segmentation elimination techniques application on bipolar EHGs gives the highest sensitivity 100% and lowest other event rate 39.77%.

Until recently, many segmentation research trials on uterine EMG have been recorded for the sake of not only automatic detection of contractions but also curtailment of other events presents in the electrohysterogram (EHG). In this study, we use an online segmentation method, which has proven its efficiency and known by Dynamic Cumulative Sum (DCS). The method is first applied on real bipolar EHGs signals then on their details obtained by wavelet transform. The detected rupture instants are driven through an enhanced technique of faulty rupture instants elimination, dynamic selection of multichannel bipolar EHG signals and its details after wavelet transform and event tracking algorithm. Therefore, enhanced method sensitivity and “other events” rate of bipolar EHGs with and without wavelet are computed using Margin validation test in order to classify among events as contractions or not referring to contractions identified by expert. Indeed, enhanced technique of event tracking, proposed in this study, seems to be more efficient comparing to previous techniques. Further studies should be achieved for the sake of fully identifying the uterine contractions from other events and then decreasing the “other events” rate.

Until recently, many studies have been achieved for the sake of automatically segmentation of the Electrohysterogram (EHG) in order to identify the efficient uterine contractions but the most of them encountered the presence of other events such as motion artifacts and other kind of contractions despite of the use of efficient filtering methods. In this study, we apply an online method which is developed previously and known by Dynamic Cumulative Sum (DCS) on monopolar EHG signals acquired through a 4x4 electrodes matrix with and without Canonical Correlation Analysis and Empirical Mode Decomposition (CCA-EMD) denoising method, then on monopolar EHG after wavelet decomposition. The detected segments are driven through an automatic concatenation technique of detected event time from all channels in order to reduce the unwanted segments, the obtained segments then undergo to implemented Margin validation test in order to classify among them. Sensitivity of detected contractions and other detected events rate referring to identified contractions by expert have been calculated in order to track the efficiency of the fully automated multichannel segmentation method. Additional EHG filtering techniques like CCA-EMD method seems to be better but effective time cost. Further studies should be achieved in order to decreasing the other events rate for the sake of fully identifying the uterine contractions.

Preterm birth (PTB) is one of the most important complications in pregnancy. Reliable diagnosis means are lacking and the underlying physiological mechanisms are unclear. Determining the location of various correlated and simultaneously active uterus sources from abdominal ElectroHysterogram (EHG) recordings and extracting the corresponding uterus signals is a challenging problem. The use of the EHG for imaging the sources of the uterine electrical activity could be a new and powerful diagnosis technique. In this paper we compare the ability of six distributed source localization methods to recover extended sources of uterus activity from abdominal EHG. As no gold standard to evaluate source localization methods, exists, we perform our evaluation by using a well-controlled realistic simulations of EHG signals, involving several locations. Simulated data were corrupted by physiological EHG noise. The performance of several state-of-the-art techniques for extended source localization is evaluated using a detection accuracy index using the Dipole Localization Error, the Area Under the Receiver Operating Characteristic (ROC) Curve AUC, and the Correlation Coefficient.

In this paper, we propose a new framework to analyze the electrical activity of the uterus recorded by electrohysterography (EHG), from abdominal electrodes (a grid of 4x4 electrodes) during pregnancy and labor. We evaluate the potential use of the synchronization between EHG signals in characterizing electrical activity of the uterus during pregnancy and labor. The complete processing pipeline consists of i) estimating the correlation between the different EHG signals, ii) quantifying the connectivity matrices using graph theory-based analysis and iii) testing the clinical impact of network measures in pregnancy monitoring and labor detection. We first compared several connectivity methods to compute the adjacency matrix represented as a graph of a set of nodes (electrodes) connected by edges (connectivity values). We then evaluated the performance of different graph measures in the classification of pregnancy and labor contractions (number of women=35). A comparison with the already existing parameters used in the state of the art of labor detection and preterm labor prediction was also performed. Results show higher performance of connectivity methods when combined with network measures. Denser graphs were observed during labor than during pregnancy. The network-based metrics showed the highest classification rate when compared to already existing features. This network-based approach can be used not only to characterize the propagation of the uterine contractions, but also may have high clinical impact in labor detection and likely in the prediction of premature labor.

Preterm labor (PL) is a major health issue worldwide. In this paper, we present a new framework to deal with this problem through processing of the electrohysterographic (EHG) signals which are recorded during labor and pregnancy. This new method is based on the neural network analysis to enhance the classification between labor and pregnancy. The proposed pipeline includes; i) using 16 EHG signals recorded from pregnant women's abdomen; ii) Using network measures in order to characterize the connectivity matrices through graph-theory based analysis; iii) and classifying between labor and pregnancy contractions based on neural network methods. Results showed that the logistic regression method gives the best classification between pregnancy and labor. Neural network based analysis can be a beneficial tool to enhance the classification between labor and pregnancy EHG signals.

Objective: Preterm birth is the first cause of perinatal morbidity and mortality. Despite continuous clinical routine improvements, the preterm rate remains steady. Moreover, the specificity of the early diagnosis stays poor as many hospitalized women for preterm delivery threat finally deliver at term. In this context, the use of electrohysterograms may increase the sensitivity and the specificity of early diagnosis of preterm labor. Methods: This paper proposes a clinical application of electrohysterogram processing for the classification of patients as prone to deliver within a week or later. The approach relies on non-linear correlation analysis for the contraction bursts extraction and uses computation of various features combined with the use of Gaussian mixture models for their classification. The method is tested on a new dataset of 68 records collected on women hospitalized for preterm delivery threat. Results: This paper presents promising results for the automatic segmentation of the contraction and a classification sensitivity, specificity, and accuracy of, respectively, 80.7%, 76.3%, and 76.2%. Conclusion: These results are in accordance with the gold standards but have the advantage to be non-invasive and could be performed at home. Significance: Diagnosis of imminent labor is possible by electrohysterography recording and may help in avoiding over-medication and in providing better cares to at-risk pregnant women.

Automatic segmentation of contractions in multichannel uterine EMG signals is the main issue of interest in many recent studies. Most of them are faced with the problem of the detecting other events than contractions. Wavelets have the great advantage of being able to separate the fine details in a signal. Many studies applied wavelets for sake of detecting and analyzing abrupt changes in non-stationary signals. This study is another step of our project focused on the automatic contractions segmentation of uterine EMG signals. Indeed, we will focus on the application of dynamic cumulative sum method (DCS) with over-segmentation elimination techniques on details after wavelet decomposition of bipolar uterine EMG records. DCS will be applied in mono- and multi-dimensional. Detected events are then compared to contractions identified by expert using Margin validation test. Regarding the obtained sensitivity and other events rate of methods we find that DCS with multidimensional application give the highest sensitivity 100% and lowest other events rate 49.6%.

Preterm labor is the primary driver of infant mortality. These days, ponders manage the electrical sign of the uterus which can be utilized by investigation strategies so as to concentrate highlights to describe compression and their job in labor. In this investigation, a few non-straight strategies are utilized to concentrate highlights from the sign so as to portray pregnancy and work flag and afterward to arrange them utilizing diverse sort of order apparatuses.All information was recorded in France and Iceland during pregnancy and labor by utilizing of 4x4 matrix of electrodes. Highlights of this sign was extricated utilizing nonlinear techniques, LZC (Lempel-Ziv Complexity), FD (Fractal Dimension), Hjorth parameter (mobility, complexity). At that point these highlights were nourished into classifiers, for example, support vector machine, artificial neural network (single layer perceptron and multilayer perceptron (MLP)); MLP gives the best grouping (rate) utilizing the extracted features and in this manner encourages us to separate among pregnancy and labor contraction signals.