A deep learning framework for noninvasive fetal ECG signal extraction.

Introduction: The availability of proactive techniques for health monitoring is essential to reducing fetal mortality and avoiding complications in fetal wellbeing. In harsh circumstances such as pandemics, earthquakes, and low-resource settings, the incompetence of many healthcare systems worldwide in providing essential services, especially for pregnant women, is critical. Being able to continuously monitor the fetus in hospitals and homes in a direct and fast manner is very important in such conditions. Methods: Monitoring the health of the baby can potentially be accomplished through the computation of vital bio-signal measures using a clear fetal electrocardiogram (ECG) signal. The aim of this study is to develop a framework to detect and identify the R-peaks of the fetal ECG directly from a 12 channel abdominal composite signal. Thus, signals were recorded noninvasively from 70 pregnant (healthy and with health conditions) women with no records of fetal abnormalities. The proposed model employs a recurrent neural network architecture to robustly detect the fetal ECG R-peaks. Results: To test the proposed framework, we performed both subject-dependent (5-fold cross-validation) and independent (leave-one-subject-out) tests. The proposed framework achieved average accuracy values of 94.2% and 88.8%, respectively. More specifically, the leave-one-subject-out test accuracy was 86.7% during the challenging period of vernix caseosa layer formation. Furthermore, we computed the fetal heart rate from the detected R-peaks, and the demonstrated results highlight the robustness of the proposed framework. Discussion: This work has the potential to cater to the critical industry of maternal and fetal healthcare as well as advance related applications.

Investigating the association of maternal heart rate variability with fetal birth weight.

Maternal heart rate (HR) was reported to affect birth weight and birth outcomes. Low birth weight constitutes a major health problem, and it is estimated that around 15% to 20% of births worldwide are low weight. In our previous study, we discussed the presence of similarities between maternal and fetal HRs, therefore, here, we propose to develop a parameter based on maternal and fetal HR variability (HRV) to divide data into two patterns to investigate the association of fetal birth weight with maternal HR and HRV. The parameter was derived from non-invasive records of maternal and fetal electrocardiograms (ECGs) that were collected from 78 subjects (age: 22 - 44 years old, gestational age (GA): 19 - 40 weeks). The HRV parameter was calculated by first evaluating the standard deviation (SD) of the number of R peaks occurring per 2 seconds (snRpp2s). Then, the difference between maternal and fetal snRpp2s (dmf) was calculated. The correlation between our derived parameter [dmf] with GA revealed a significant correlation that suggested the dmf's association with fetal development. The association analysis results between birthweight with maternal HR and HRV per pattern showed that significant negative correlations exist between them in one pattern. Still, the same correlations were not observed in the other pattern. This study's findings emphasise maternal health's role in fetal development assessment. In addition, this study highlights the importance of developing novel factors for properly assessing fetal development and birth outcomes.

Prediction of fetal RR intervals from maternal factors using machine learning models.

Previous literature has highlighted the importance of maternal behavior during the prenatal period for the upbringing of healthy adults. During pregnancy, fetal health assessments are mainly carried out non-invasively by monitoring fetal growth and heart rate (HR) or RR interval (RRI). Despite this, research entailing prediction of fHRs from mHRs is scarce mainly due to the difficulty in non-invasive measurements of fetal electrocardiogram (fECG). Also, so far, it is unknown how mHRs are associated with fHR over the short term. In this study, we used two machine learning models, support vector regression (SVR) and random forest (RF), for predicting average fetal RRI (fRRI). The predicted fRRI values were compared with actual fRRI values calculated from non-invasive fECG. fRRI was predicted from 13 maternal features that consisted of age, weight, and non-invasive ECG-derived parameters that included HR variability (HRV) and R wave amplitude variability. 156 records were used for training the models and the results showed that the SVR model outperformed the RF model with a root mean square error (RMSE) of 29 ms and an average error percentage (< 5%). Correlation analysis between predicted and actual fRRI values showed that the Spearman coefficient for the SVR and RF models were 0.31 (P < 0.001) and 0.19 (P < 0.05), respectively. The SVR model was further used to predict fRRI of 14 subjects who were not included in the training. The latter prediction results showed that individual error percentages were (≤ 5%) except in 3 subjects. The results of this study show that maternal factors can be potentially used for the assessment of fetal well-being based on fetal HR or RRI.

Similarities between maternal and fetal RR interval tachograms and their association with fetal development.

An association between maternal and fetal heart rate (HR) has been reported but, so far, little is known about its physiological implication and importance relative to fetal development. Associations between both HRs were investigated previously by performing beat-by-beat coupling analysis and correlation analysis between average maternal and fetal HRs. However, studies reporting on the presence of similarities between maternal and fetal HRs or RR intervals (RRIs) over the short term (e.g., 5-min) at different gestational ages (GAs) are scarce. Here, we demonstrate the presence of similarities in the variations exhibited by maternal and fetal RRl tachograms (RRITs). To quantify the same similarities, a cross-correlation (CC) analysis between resampled maternal and fetal RRITs was conducted; RRITs were obtained from non-invasive electrocardiogram (ECG). The degree of similarity between maternal and fetal RRITs (bmfRRITs) was quantified by calculating four CC coefficients. CC analysis was performed for a total of 330 segments (two 5-min segments from 158 subjects and one 5-min from 14 subjects). To investigate the association of the similarity bmfRRITs with fetal development, the linear correlation between the calculated CC coefficients and GA was calculated. The results from the latter analysis showed that similarities bmfRRITs are common occurrences, they can be negative or positive, and they increase with GA suggesting the presence of a regulation that is associated with proper fetal development. To get an insight into the physiological mechanisms involved in the similarity bmfRRITs, the association of the same similarity with maternal and fetal HR variability (HRV) was investigated by comparing the means of two groups in which one of them had higher CC values compared to the other. The two groups were created by using the data from the 158 subjects where fetal RRI (fRRI) calculation from two 5-min ECG segments was feasible. The results of the comparison showed that the maternal very low frequency (VLF) HRV parameter is potentially associated with the similarity bmfRRITs implying that maternal hormones could be linked to the regulations involved in the similarity bmfRRITs. Our findings in this study reinforce the role of the maternal intrauterine environment on fetal development.

Correlation between maternal and fetal heart rate increases with fetal mouse age in typical development and is disturbed in autism mouse model treated with valproic acid.

Introduction: Autism spectrum disorder (ASD) is considered a significant behavioral problem that is characterized by impairment in social interaction and communication. It is believed that some cases of ASD originate in the intrauterine maternal environment. Therefore, we hypothesized that there might be qualitative changes in the interaction between the mother and fetus in ASD during the prenatal period, hence, we investigated the similarity patterns between maternal and fetal heart rate (HR). Methods: In this study, we first demonstrate the presence and formation of similarities between maternal and fetal RR interval (RRI) collected from typical developmental mice at different embryonic days (EDs), ED13.5, ED15.5, ED17.5, and ED18.5. The similarities were quantified by means of cross-correlation (CC) and magnitude-squared coherence (MSC) analyses. Correlation analysis between the CC coefficients and EDs and between MSC coefficients and EDs showed that the same coefficients increase with EDs, suggesting that similarities between maternal and fetal RRI are associated with typical fetal development. Next, because maternal and fetal similarities were indicative of development, a comparison analysis between the autism mouse model (injected with valproic acid (VPA)), and the control group (injected with saline) was performed for ED15.5 and ED18.5. Results: The results of the comparison showed that the CC and MSC coefficients of VPA fetuses were significantly lower than that of the control group. The lower coefficients in VPA-treated mice suggest that they could be one of the features of ASD symptoms. The findings of this study can assist in identifying potential ASD causes during the prenatal period.

Deep learning identifies cardiac coupling between mother and fetus during gestation.

In the last two decades, stillbirth has caused around 2 million fetal deaths worldwide. Although current ultrasound tools are reliably used for the assessment of fetal growth during pregnancy, it still raises safety issues on the fetus, requires skilled providers, and has economic concerns in less developed countries. Here, we propose deep coherence, a novel artificial intelligence (AI) approach that relies on 1 min non-invasive electrocardiography (ECG) to explain the association between maternal and fetal heartbeats during pregnancy. We validated the performance of this approach using a trained deep learning tool on a total of 941 one minute maternal-fetal R-peaks segments collected from 172 pregnant women (20-40 weeks). The high accuracy achieved by the tool (90%) in identifying coupling scenarios demonstrated the potential of using AI as a monitoring tool for frequent evaluation of fetal development. The interpretability of deep learning was significant in explaining synchronization mechanisms between the maternal and fetal heartbeats. This study could potentially pave the way toward the integration of automated deep learning tools in clinical practice to provide timely and continuous fetal monitoring while reducing triage, side-effects, and costs associated with current clinical devices.

Model-based estimation of QT intervals of mouse fetal electrocardiogram.

BACKGROUND: Abnormal prolongation in the QT interval or long QT syndrome (LQTS) is associated with several cardiac complications such as sudden infant death syndrome (SIDS). LQTS is believed to be linked to genetic mutations which can be understood by using animal models, such as mice models. Nevertheless, the research related to fetal QT interval in mice is still limited because of challenges associated with T wave measurements in fetal electrocardiogram (fECG). Reliable measurement of T waves is essential for estimating their end timings for QT interval assessment. RESULTS: A mathematical model was used to estimate QT intervals. Estimated QT intervals were validated with Q-aortic closure (Q-Ac) intervals of Doppler ultrasound (DUS) and comparison between both showed good agreement with a correlation coefficient higher than 0.88 (r > 0.88, P < 0.05). CONCLUSION: Model-based estimation of QT intervals can help in better understanding of QT intervals in fetal mice.

Effect of Valproic Acid on Maternal - Fetal Heart Rates and Coupling in Mice on Embryonic day 15.5 (E15.5).

Prenatal uptake of valproic acid (VPA) was associated with increased risk of fetal cardiac anomalies and autism spectrum disorder (ASD), but uptake of VPA is considered the only effective treatment for epilepsy and other neurological disorders. Up until now, little is known about the effect of VPA on maternal - fetal heart rate (HR) coupling patterns; therefore, this study aims at studying such patterns in mice on embryonic day 15.5 (E15.5). At E12.5, 8 mothers were injected with VPA (VPA group) and another 8 mothers were injected with saline (control group). At E15.5, electrocardiogram (ECG) records of 15 minutes were collected from the 16 mothers and 25 fetuses. A maximum of 5-minutes and a minimum of 1-minute were selected from the ECG data for analysis. Mean RR intervals and coupling ratios and their occurrence percentages were calculated per 1minute. 1-minute analysis was done for periods with no arrhythmia and clear R peaks. The total number of 1-minute segments that were analyzed was 56 for the saline group and 54 for the VPA group. The correlation analysis between the 1:3 and 2:6 coupling ratios and RR intervals revealed that the ratios were significantly correlated in the saline group, whereas no significant correlations were observed in the VPA group. The results further revealed that fetal RR intervals are strongly correlated with maternal RR intervals in the saline group, but the same correlation is different in the VPA group. The presented results imply that maintaining certain coupling patterns are important for proper fetal cardiac development and maternal uptake of VPA may affect maternal-fetal HRs interactions.

Low Back Pain Exacerbation Is Predictable Through Motif Identification in Center of Pressure Time Series Recorded During Dynamic Sitting.

Background: Low back pain (LBP) is a common health problem - sitting on a chair for a prolonged time is considered a significant risk factor. Furthermore, the level of LBP may vary at different times of the day. However, the role of the time-sequence property of sitting behavior in relation to LBP has not been considered. During the dynamic sitting, small changes, such as slight or big sways, have been identified. Therefore, it is possible to identify the motif consisting of such changes, which may be associated with the incidence, exacerbation, or improvement of LBP. Method: Office chairs installed with pressure sensors were provided to a total of 22 office workers (age = 43.4 ± 8.3 years) in Japan. Pressure sensors data were collected during working days and hours (from morning to evening). The participants were asked to answer subjective levels of pain including LBP. Center of pressure (COP) was calculated from the load level, the changes in COP were analyzed by applying the Toeplitz inverse covariance-based clustering (TICC) analysis, COP changes were categorized into several states. Based on the states, common motifs were identified as a recurring sitting behavior pattern combination of different states by motif-aware state assignment (MASA). Finally, the identified motif was tested as a feature to infer the changing levels of LBP within a day. Changes in the levels of LBP from morning to evening were categorized as exacerbated, did not change, or improved based on the survey questions. Here, we present a novel approach based on social spider algorithm (SSA) and probabilistic neural network (PNN) for the prediction of LBP. The specificity and sensitivity of the LBP inference were compared among ten different models, including SSA-PNN. Result: There exists a common motif, consisting of stable sitting and slight sway. When LBP level improved toward the evening, the frequency of motif appearance was higher than when LBP was exacerbated (p < 0.05) or the level did not change. The performance of the SSA-PNN optimization was better than that of the other algorithms. Accuracy, precision, recall, and F1-score were 59.20, 72.46, 40.94, and 63.24%, respectively. Conclusion: A lower frequency of a common motif of the COP dynamic changes characterized by stable sitting and slight sway was found to be associated with the exacerbation of LBP in the evening. LBP exacerbation is predictable by AI-based analysis of COP changes during the sitting behavior of the office workers.

Model based estimation of QT intervals in non-invasive fetal ECG signals.

The end timing of T waves in fetal electrocardiogram (fECG) is important for the evaluation of ST and QT intervals which are vital markers to assess cardiac repolarization patterns. Monitoring malignant fetal arrhythmias in utero is fundamental to care in congenital heart anomalies preventing perinatal death. Currently, reliable detection of end of T waves is possible only by using fetal scalp ECG (fsECG) and fetal magnetocardiography (fMCG). fMCG is expensive and less accessible and fsECG is an invasive technique available only during intrapartum period. Another safer and affordable alternative is the non-invasive fECG (nfECG) which can provide similar assessment provided by fsECG and fMECG but with less accuracy (not beat by beat). Detection of T waves using nfECG is challenging because of their low amplitudes and high noise. In this study, a novel model-based method that estimates the end of T waves in nfECG signals is proposed. The repolarization phase has been modeled as the discharging phase of a capacitor. To test the model, fECG signals were collected from 58 pregnant women (age: (34 ± 6) years old) bearing normal and abnormal fetuses with gestational age (GA) 20-41 weeks. QT and QTc intervals have been calculated to test the level of agreement between the model-based and reference values (fsECG and Doppler Ultrasound (DUS) signals) in normal subjects. The results of the test showed high agreement between model-based and reference values (difference < 5%), which implies that the proposed model could be an alternative method to detect the end of T waves in nfECG signals.

Detection of End of T-wave in Fetal ECG Using Recurrence Plots.

Automatic detection of fetal ECG features can assist in diagnosis of fetal cardiac complications and may reduce the time required for diagnosis. Detection of the end of the repolarization period wave in ECG has been proven challenging due to its low amplitude and low frequency range. The prolongation of end of T-wave is associated with sudden cardiac death, thus, methods that can accurately pinpoint it is highly desirable for early diagnosis of cardiac diseases. In this paper, a technique based on recurrence plots is developed for the detection of end of T-wave. The developed technique was tested on maternal ECG (mECG), fetal scalp ECG (fsECG) and non-invasive fetal ECG (nfECG) records. The technique was able to detect end of T-waves in all of the mECG beats, 75% of the non-invasive fECG beats (verified by simultaneously captured doppler ultrasound signals) and 78% of the fsECG beats. Detection of fECG signals were more challenging than mECG signals due to the noise and their low amplitude T-waves.

Quantification of sterol-specific response in human macrophages using automated imaged-based analysis.

BACKGROUND: The transformation of normal macrophage cells into lipid-laden foam cells is an important step in the progression of atherosclerosis. One major contributor to foam cell formation in vivo is the intracellular accumulation of cholesterol. METHODS: Here, we report the effects of various combinations of low-density lipoprotein, sterols, lipids and other factors on human macrophages, using an automated image analysis program to quantitatively compare single cell properties, such as cell size and lipid content, in different conditions. RESULTS: We observed that the addition of cholesterol caused an increase in average cell lipid content across a range of conditions. All of the sterol-lipid mixtures examined were capable of inducing increases in average cell lipid content, with variations in the distribution of the response, in cytotoxicity and in how the sterol-lipid combination interacted with other activating factors. For example, cholesterol and lipopolysaccharide acted synergistically to increase cell lipid content while also increasing cell survival compared with the addition of lipopolysaccharide alone. Additionally, ergosterol and cholesteryl hemisuccinate caused similar increases in lipid content but also exhibited considerably greater cytotoxicity than cholesterol. CONCLUSIONS: The use of automated image analysis enables us to assess not only changes in average cell size and content, but also to rapidly and automatically compare population distributions based on simple fluorescence images. Our observations add to increasing understanding of the complex and multifactorial nature of foam-cell formation and provide a novel approach to assessing the heterogeneity of macrophage response to a variety of factors.