Objectives. - Our study focuses on the analysis of the uterine electromyogram recorded on women during pregnancy. We were interested in evaluating the synchronization of this electrical signal at various terms in order to follow evolution of synchronization as labor approaches. This study attempts to deepen our understanding of the myometrial maturation close to labor and to provide reliable parameters for improving preterm labor diagnosis.Patients and methods. - We performed a prospective study by recording the electrical signals of physiological uterine contractions (causing no delivery) on 16 pregnant women. We then calculated the non-linear correlation coefficient h(2) to estimate synchronization between EMG signals collected for each contraction. We expressed the results by grouping synchronization values by class of term in order to study the evolution of this coefficient along gestation. This study has been approved by the ethical committee of our hospital.Results. - Our results show a non-significant increase of the h2 value along term. There is however a trend towards an increase of the synchronization of EMG signals as labor approaches but not enough to conclude definitively.Discussion and conclusion. - With a confirmation of the increase of h(2) along term, the study of the synchronization of uterine electrical activity could be an important clue to support the notion of myometrial maturation close to labor. Synchronization analysis could also be a promising parameter for reliable diagnosis of preterm labor. (C) 2012 Elsevier Masson SAS. All rights reserved.

Most of the studies on the synchronization between EHG signals, recorded during the same contractions at different locations, are limited to the use of only two channels. Multichannel techniques have however been widely applied to EEG signals but rarely to EHG. In this paper, we investigate the use of multichannel uterine EMG signals for classifying contractions. We compare the performance of phase synchronization in distinguishing between labor and normal pregnancy contractions by using either only two channels or a 4x4 matrix positioned on the woman's abdomen. We used two indexes to measure the phase synchronization: mean phase coherence and phase entropy. ROC curves indicate that the use of multichannel signals can significantly improve the classification rate of pregnancy and labor contractions.

Labor prediction using the electrohysterogram has immediate clinical applications and has been the aim of several studies in recent years. Studies using various linear methods such as classic spectral analysis do not give clinically useful results. In this paper we present the use of two methods that investigate nonlinearity to predict normal labor. We show the comparison between a linear method that is known from the literature (mean power frequency) and two nonlinear methods (approximate entropy and time reversibility) using ROC analysis. The comparison indicates that the best method for pretreatment to classify pregnancy and labor signals is time reversibility. The results indicate that time reversibility is a very promising tool for distinguishing between labor and physiological contractions during pregnancy. This could be the first step in developing a clinical application method to predict preterm labor.

The analysis of the electrical activity of the uterus recorded externally, the electrohysterogram (EHG), may find an application in the prediction of labor. In the literature parameters that are supposed to be related to the excitability of the uterine cells have almost exclusively been used for this purpose. In the present paper we evaluate the possible use of synchronization parameters for EHG measured in several places on the uterus for this prediction. The EHG is mainly composed of two frequency components called FWL and FWH. These components may be related to the synchronization and the excitability phenomenon respectively. In order to study independently these frequency components, we extracted the corresponding ridges of the wavelet transform of the EHG using the GVF-snake method. For each frequency component we computed parameters that are related to the frequency content of the signals and parameters that consider the synchronization relationship between signals. The synchronization parameters used were the mean phase coherence and the phase entropy. The values of the different parameters were compared during pregnancy and labor by statistical analysis. The detection quality of each parameter was evaluated using ROC curves. Our results suggest that synchronization parameters could be used for the detection of labor in addition to the classical previously published parameters. Another important result of our study is that both FWH and FWL seem related to excitability but only FWL seems to be related to the synchronization of the uterus at term.

Analysis of synchronization between biological signals is often used for the characterization of biological functions giving rise to these signals. In this paper we present the application of a method based on the wavelet transform to detect the coherence between two uterine activity bursts recorded at different places on the pregnant abdomen, during the same uterine contraction. The method used in this work is referred to as the wavelet coherence. The results of this study show that the wavelet analysis can successfully detect temporal and spectral interactions between uterine contractions. The results indicate that the coherence is higher in the lower frequencies of the EHG signal. We find the method to give promising results and to put in evidence the coherence present during a uterine contraction.

This paper introduces the use of a method based on wavelet transform to detect the correlation between two uterine electrical activity bursts, recorded at different places on the pregnant abdomen during the same uterine contraction. The method used in this work is called wavelet coherence. The results of this study show that the wavelet analysis can successfully detect and quantify the temporal and spectral interactions between uterine bursts of electrical activity. They also indicate that the coherence is higher in the lower frequencies of the uterine electromyogram signal (EHG), and that it is possible to apply the method to non-segmented uterine signals. We find the method to give promising results permitting to evidence the coherence present in EHGs during a uterine contraction. (C) 2009 Elsevier Masson SAS. All rights reserved.

This study investigates phase synchronization in the time-frequency domain between uterine signals recorded at different sites during the same contraction from women during pregnancy and women in labor. We used the complex Morlet wavelet transform to estimate the phase synchronization between the uterine signals. The method was applied on a set of uterine bursts during pregnancy and in labor. The results indicated that the uterine bursts are more synchronized in phase during pregnancy than during labor. This phase desynchronization during labor may be a tool to differentiate between contractions during pregnancy and labor and could therefore be used in the prediction of preterm labor.

Surrogates are commonly used to test a particular hypothesis on time series. The parameter commonly used in the literature to test these hypotheses is the z score. The z score assumes that the distribution of the statistics obtained on the surrogates is Gaussian. In this paper, we propose the use of a more general parameter than the z score that will also work in the case of non-Gaussian distribution of the statistics. We also derive a statistical test, based on the fitting of the distribution of the surrogate measure profile, in order to test the initial hypothesis. We validate the proposed approach on both synthetic signals and real uterine EMG signals by using the nonlinear correlation coefficient as initial statistic. We further show that this corrected nonlinear correlation coefficient can discriminate between pregnancy contractions and labor in a monkey, but the uncorrected nonlinear correlation coefficient cannot. This makes the corrected nonlinear correlation coefficient a promising candidate in a future application for preterm labor prediction in humans.

Evaluation of synchronization between signals can give new insights into the functioning of the related systems. Methods that can detect synchronization or coupling between signals can be divided two types: linear and non linear methods. In this paper we use the non linear correlation coefficient (h2) to show the difference in synchronization between efficient uterine contractions during labor and normal physiological contractions during pregnancy, in uterine activity bursts recorded at different places on the pregnant abdomen. Our interest in the non linear correlation coefficient is based on the fact that the propagation mechanism of uterine EMG signal may be strongly non linear. The results obtained from estimating the synchronization between 16 uterine EMG channels indicate that synchronization between contractions as measured by h2 is stronger in labor than in pregnancy. Limited data indicates that the h2 value increases markedly with term when expressed in duration before spontaneous labor.

The empirical mode decomposition is an iterative method able to decompose a signal into several modes or intrinsic mode functions (IMF). An algorithm for the selection of modal components of interest has recently been proposed. This algorithm is based on the statistical analysis of the noise contained in each IMF. A mathematical model of the noise repartition between each IMF is personalized for the signal under analysis by estimating its noise content from the energy of the first IMF, which is supposed to contain a certain part of the noise only. A mode mixing can however be present and give rise to an over estimation of the noise in the original signal. Thus several IMF will be considered as containing only noise and would be erroneously discarded. We propose a general method of mode mixing detection based on a stationary test applied to the first IMF. In case of mode mixing, we propose to correct the noise estimation on the first IMF by extracting from this IMF the parts corresponding to signal and the parts corresponding to noise. The results obtained on synthetic signals, as well as mechanical or biomedical ones, show the good performances of the proposed approach.

The analysis of the synchronization between biological signals can be helpful for the characterization of biological functions. Biological signals are however often strongly non stationary. This is in contradiction to the assumption of commonly used synchronization measures which assume that the signal is stationary. We propose to use a piecewise stationary pre-segmentation (PSP) of the signals of interest, before the computation of synchronization measures. We show on synthetic as well as real biological signals (EEG and uterine EMG) that the proposed piecewise stationary pre-segmentation approach increases the accuracy of the measures by making a good tradeoff between the stationary assumption and length of the analyzed segments, when compared to the classical windowing method.

The efficiency of uterine contraction is usually measured by the intra uterine pressure (IUP) clinically measured through the deformation of the woman's abdomen. This indirect measurement can be inaccurate sometimes. The objective of this research is to identify in the electrical bursts, generated by the uterus during contraction electrohysterogram (EHG), some parameters that can give direct information on the IUP. In this paper, we explore the correlation that can exist between temporal, energetic as well as frequency EHG parameter and the IUP curve. Preliminary results show the existence of strong correlations between several evaluated parameters and the IUP. It suggests that non invasive recording of the electrical activity of the uterus is a promising tool for pregnancy monitoring giving thus, non invasively, direct information on the mechanical activity of the uterus.

Mapping of uterine contractions by recording the electrohysterogram (EHG) in many places on the abdominal wall is a new way of investigating the electrical activity of the uterus. Good spatial resolution is an important issue for this technique to provide new information. The use of monopolar recordings is an obvious way of increasing resolution, in spite of their having a lower signal to noise ratio (SNR) than bipolar measurements. We explored the use of the LMS and RLS filter as well as Laplacian filtering for the increase in monopolar EHG SNR. The best results for monopolar signals were obtained using the RLS algorithm but the SNR is still lower than that obtained on bipolar signals. The resulting EHG signals are nevertheless sufficiently improved to clearly identify EHG bursts during contractions. A precise selection of the different methods parameters, as well as an increase in the number of studied contractions, has to be done in order to confirm these preliminary results.

Numerous studies have observed and analyzed the external electrical activity of the uterus, the so-called electrohysterogram (EHG), associated with contractions during pregnancy and labor. The EHG is mainly composed of two distinct frequency components, FWL (Fast Wave low, low frequency component) and FWH (Fast Wave High, high frequency component). It has been suggested that FWH is mainly associated with uterine cell excitability and FWL with the propagation of this activity. This hypothesis is still unproven. We compared two procedures for ridge extraction/reconstruction of the EHG scalogram, with the objective of analyzing the propagation of the EHG on FWH and FWL separately. The performance of the methods under investigation was tested on both synthetic and real signals. The results indicate that the EHG can be characterized by two distinct continuous ridges, supposed to be FWH and FWL, with a low reconstruction error. We have also shown that the extracted ridges have different energy, temporal characteristics and bandwidths.

The uterine electromyogram is a signal that can be used to follow, during pregnancy, the uterine contractility. Moreover, we can use this signal to predict a potential risk of preterm delivery in women. On the other hand, we know that the placenta is a source of hormonal secretion which can modulate the uterine contractility. Thus, our hypothesis is that the placenta could exert a local influence on the characteristics of uterine electromyogram and may bias the prediction of preterm delivery. In this study, we tried to explore the placental influences by recording internally contractions, in the monkey. Contractions were analysed by the way of their absolute energy spectrum calculated from their time-frequency representation. Our results showed significant differences between contractions, in regards of the placenta, whatever the parturition period. Moreover, we found that these differences evolved during parturition. However, more studies are needed to really understand the physiological phenomena which could explain these differences and explore their influence on the prediction of the preterm delivery risk.

Several methods have been proposed to investigate the relationship between signals recorded externally from several sites on the pregnant uterus. A promising recent method is the multivariate autoregressive (MVAR) model. In this paper we proposed a windowing (time varying) version of the multivariate autoregressive model, called W-MVAR, to investigate the connectivity between signals while still respecting their nonstationary characteristics. The proposed method was tested on synthetic signals as well as applied to real signals. The comparison between the two methods on synthetic signals indicated the superiority of W-MVAR to detect connectivity even if nonstationarity is present. The application of W-MVAR on multichannel real uterine signals show that the proposed method is a good tool to distinguish non-labor and labor signals. These results are very promising and can very possibly have important clinical applications in labor detection and preterm labor prediction. Résumé : Plusieurs méthodes ont été proposées pour étudier la relation entre les signaux enregistrés en plusieurs sites sur l'abdomen de femmes gravides. Une méthode prometteuse récente est le modèle autorégressif multivarié (MVAR). Dans cet article nous proposons une version fenêtrée (variable dans le temps) du modèle autorégressif multivarié, appelée W-MVAR, pour détecter la connectivité entre les signaux, tout en respectant leurs caractéristiques non stationnaires. La méthode proposée a été testée sur des signaux synthétiques ainsi que sur des signaux réels. La comparaison entre les deux méthodes sur des signaux synthétiques a indiqué la supériorité de la W-MVAR pour détecter la connectivité, même si des non-stationnarités sont présentes. L'application de W-MVAR sur des signaux utérins multicanaux montre que la méthode proposée est un bon outil pour distinguer les signaux de grossesse des signaux de travail. Ces résultats sont très prometteurs et peuvent avoir des applications cliniques importantes dans la détection du travail et la prédiction du travail prématuré