The Identification and Tracking of Uterine Contractions Using Template Based Cross-Correlation

Unjumbling Procedure in the Algorithmic Analysis of Biomechanical Torques Induced by Electrical Stimulation: Case Study of the Lower Limb

Functional Electrical Stimulation (FES) uses electrical pulses to cause muscles to contract synthetically. When muscles contract due to stimulation, torques are produced around joints attached to the muscle. It is important to understand torques being produced, for reasons such as safety and indirect fatigue measurement. Given the translatability of FES devices into the home for therapy, having ways to non-invasively measure muscle responses to stimulation is essential for understanding and diagnosing the biomechanical response of the human body. Here, we present data from a stimulation experiment examining knee joint torques (KJTs) arising when quadriceps are subjected to electrical stimulation. A novel algorithm for computing and summarizing KJT data into a series of simplified parameters was developed using MATLAB software. From this, we draw some conclusions about the effect of changing the stimulation duty cycle on the resultant KJT. We believe this method will provide researchers with a tool to measure torque in a semi-automated, convenient fashion.

Uterus Modeling from Cell to Organ Level: towards Better Understanding of Physiological Basis of Uterine Activity

The relatively limited understanding of the physiology of uterine activation prevents us from achieving optimal clinical outcomes for managing serious pregnancy disorders such as preterm birth or uterine dystocia. There is increasing awareness that multi-scale computational modeling of the uterus is a promising approach for providing a qualitative and quantitative description of uterine physiology. The overarching objective of such approach is to coalesce previously fragmentary information into a predictive and testable model of uterine activity that, in turn, informs the development of new diagnostic and therapeutic approaches to these pressing clinical problems. This article assesses current progress towards this goal. We summarize the electrophysiological basis of uterine activation as presently understood and review recent research approaches to uterine modeling at different scales from single cell to tissue, whole organ and organism with particular focus on transformative data in the last decade. We describe the positives and limitations of these approaches, thereby identifying key gaps in our knowledge on which to focus, in parallel, future computational and biological research efforts.

Estimation of 8-Electrode Configuration for Recognition of Uterine Contraction with Electrohysterogram

As the representative of electrical activity from uterine muscle, electrohysterogram (EHG) is recorded non-invasively by multiple electrodes positioned on the abdominal surface. The purpose of our paper is to estimate different electrode configurations for recognizing uterine contractions (UCs) with EHG signals. 8-electrode configuration was taken as an example to show our novel method with convolutional neural network (CNN) classification and score. The open accessed Icelandic 16-electrode EHG database was adopted in our study. With 8-electrode configuration, EHG signals corresponding to UCs and non-UCs were segmented and saved as image patches. The CNN was established and trained by thousands of EHG segments. The performance of CNN was evaluated by the area under curve (AUC) and accuracy of recognizing UCs and non-UCs. Seven different 8-electrode configurations were scored and ranked. It was found the 8-electrode configuration with 4 on the uterine fundus, 2 on the body and 2 on the cervix achieved the AUC of 0.766 and the highest score of 2.197. Among the configurations we have tried, it is concluded that the 8 electrodes in 4-2-2 configuration placed along the uterus as an upside-down pear could provide the most important information for recognition of UC based on our experiments.

Preliminary Study on the Efficient Electrohysterogram Segments for Recognizing Uterine Contractions with Convolutional Neural Networks

Background

Uterine contraction (UC) is the tightening and shortening of the uterine muscles which can indicate the progress of pregnancy towards delivery. Electrohysterogram (EHG), which reflects uterine electrical activities, has recently been studied for UC monitoring. In this paper, we aimed to evaluate different EHG segments for recognizing UCs using the convolutional neural network (CNN).

Materials and Methods

In the open-access Icelandic 16-electrode EHG database (122 recordings from 45 pregnant women), 7136 UC and 7136 non-UC EHG segments with the duration of 60 s were manually extracted from 107 recordings of 40 pregnant women to develop a CNN model. A fivefold cross-validation was applied to evaluate the CNN based on sensitivity (SE), specificity (SP), and accuracy (ACC). Then, 1056 UC and 1056 non-UC EHG segments were extracted from the other 15 recordings of 5 pregnant women. Furthermore, the developed CNN model was applied to identify UCs using different EHG segments with the durations of 10 s, 20 s, and 30 s.

Results

The CNN achieved the average SE, SP, and ACC of 0.82, 0.93, and 0.88 for a 60 s EHG segment. The EHG segments of 10 s, 20 s, and 30 s around the TOCO peak achieved higher SE and ACC than the other segments with the same duration. The values of SE from 20 s EHG segments around the TOCO peak were higher than those from 10 s to 30 s EHG segments on the same side of the TOCO peak.

Conclusion

The proposed method could be used to determine the efficient EHG segments for recognizing UC with the CNN.

Development of Electrohysterogram Recording System for Monitoring Uterine Contraction

Uterine contraction (UC) is an important clinical indictor for monitoring uterine activity. The purpose of this study is to develop a portable electrohysterogram (EHG) recording system (called PregCare) for monitoring UCs with EHG signals. The PregCare consisted of sensors, a signal acquisition device, and a computer with application software. Eight-channel EHG signals, the tocodynamometry (TOCO) signal, and maternal perception were recorded simultaneously by the signal acquisition device controlled by the computer via Bluetooth. PregCare was firstly evaluated by a signal simulator. Its relative error (RE) and coefficient of variation (CV) were calculated, and its agreement with the commercial instrument PowerLab was assessed by Bland-Altman plots. After that, PregCare was applied to 20 pregnant women in a hospital to record their EHG signals. These EHG signals were preprocessed and segmented into UCs and non-UCs. Then, the EHG features corresponding to UCs and non-UCs were extracted, respectively, including power spectral density (PSD), root mean square (RMS), peak frequency (PF), median frequency (MDF), and sample entropy (SamEn). One-way ANOVA was employed to assess the difference between UCs and non-UCs. The results show that RE and CV were less than 8% and 0.03%, respectively, which indicated the high accuracy and repeatability of PregCare. The small differences of mean and standard deviation indicated the high agreement between PregCare and PowerLab. Besides, the PSD of UCs was much larger than non-UCs between 0 and 0.7 Hz. RMS of UCs was significantly larger than non-UCs (p < 0.05). PF and SamEn of UCs were significantly smaller than non-UCs (p < 0.05). In conclusion, the developed EHG recording system was able to record EHG signals reliably. It has the advantages of portability, low power consumption, and wireless transmission, which can be used for long-term monitoring of UCs and prediction of the preterm delivery.