We present a protocol for conducting electromyometrial imaging (EMMI), including the following procedures: multiple electromyography electrode sensor recordings from the body surface, magnetic resonance imaging, and uterine electrical signal reconstruction.
During normal pregnancy, the uterine smooth muscle, the myometrium, begins to have weak, uncoordinated contractions at late gestation to help the cervix remodel. In labor, the myometrium has strong, coordinated contractions to deliver the fetus. Various methods have been developed to monitor uterine contraction patterns to predict labor onset. However, the current techniques have limited spatial coverage and specificity. We developed electromyometrial imaging (EMMI) to noninvasively map uterine electrical activity onto the three-dimensional uterine surface during contractions. The first step in EMMI is to use T1-weighted magnetic resonance imaging to acquire the subject-specific body-uterus geometry. Next, up to 192 pin-type electrodes placed on the body surface are used to collect electrical recordings from the myometrium. Finally, the EMMI data processing pipeline is performed to combine the body-uterus geometry with body surface electrical data to reconstruct and image uterine electrical activities on the uterine surface. EMMI can safely and noninvasively image, identify, and measure early activation regions and propagation patterns across the entire uterus in three dimensions.
Clinically, uterine contractions are measured either by using an intrauterine pressure catheter or by performing tocodynamometry1. In the research setting, uterine contractions can be measured by electromyography (EMG), in which electrodes are placed on the abdominal surface to measure the bioelectrical signals generated by the myometrium2,3,4,5,6,7. One can use the magnitude, frequency, and propagation features of electrical bursts8,9,10,11,12 derived from EMG to predict the onset of labor in the preterm. However, in conventional EMG, the electrical activity of uterine contractions is measured from only a tiny region of the abdominal surface with a limited number of electrodes (two13 and four7,14,15,16 at the center of the abdominal surface, and 6417 at the lower abdominal surface). Furthermore, conventional EMG is limited in its ability to study the mechanisms of labor, as it reflects only the averaged electrical activities from the entire uterus and cannot detect the specific electrical initiation and activation patterns on the uterine surface during contractions.
A recent development called electromyometrial imaging (EMMI) has been introduced to overcome the shortcomings of conventional EMG. EMMI enables noninvasive imaging of the entire myometrium's electrical activation sequence during uterine contractions18,19,20,21. To acquire the body-uterus geometry, EMMI uses T1-weighted magnetic resonance imaging (MRI)22,23,24, which has been widely used for pregnant women during their second and third trimesters. Next, up to 192 pin-type electrodes placed on the body surface are used to collect electrical recordings from the myometrium. Finally, the EMMI data processing pipeline is performed to combine the body-uterus geometry with the electrical data to reconstruct and image electrical activities on the uterine surface21. EMMI can accurately locate the initiation of uterine contractions and image propagation patterns during uterine contractions in three dimensions. This article aims to present the EMMI procedures and demonstrate the representative results obtained from pregnant women.
All methods described here have been approved by the Washington University Institutional Review Board.
1. MRI-safe marker patches, electrode patches, and rulers (Figure 1)
2. MRI scan
NOTE: The MRI scan is scheduled at a gestational age (GA) of 36-40 weeks, prior to the mother's expected delivery date, determined based on the subject's schedule and her nurse's recommendation. The estimated time duration for this step is 2 h.
3. Bioelectricity mapping and 3D optical scan
NOTE: Conduct bioelectricity mapping after the subject has been admitted to the labor and delivery unit, and her cervix has dilated to around 4 cm. The estimated time duration for this step is 2 h.
4. Generation of the body-uterus geometry
5. Electrical signal preprocessing
6. Uterine electrical signal reconstruction and characterization
Representative MRI-safe patches and electrode patches are shown in Figure 1B,C, created from the template shown in Figure 1A. The bioelectricity mapping hardware is shown in Figure 1C, with the connections of each component marked in detail. Figure 2 shows the entire EMMI procedure, including an MRI scan of the subject wearing MRI patches (Figure 2A), 3D optical scanning (Figure 2B), bioelectricity mapping (Figure 2C), the generation of body-uterus geometry (Figure 2D), and a schematic of the EMMI data (Figure 2E).
Figure 3A shows a representative raw body surface electrogram with a sampling rate of 2,048 Hz. The raw signal is significantly affected by baseline drift, maternal electrocardiographic signal, maternal breathing, and other factors. In electrical signal preprocessing (section 5 in the protocol), a Butterworth bandpass filter with cutoff frequencies of 0.34-1 Hz and a downsample of a factor of 20 were applied to generate the filtered signal shown in Figure 3B. Three clear EMG bursts are marked with green lines in Figure 3B.
Figure 4A–F shows six successive uterine surface potential maps 0.2 s apart in anterior, left, posterior, and right views. The warm colors represent positive potentials and the cool colors represent negative potentials. The respective time of each uterine potential is labeled in the electrogram in Figure 4G, which is from the sites indicated with asterisks in Figure 4A–F. A region of high positive potential starts at the site marked with an asterisk (Figure 4A), enlarges (Figure 4B–E), and finally diminishes (Figure 4F). These EMMI-generated potential maps allow investigators to visualize the dynamic progression of uterine contractions in three dimensions.
Figure 5A shows an EMMI-generated isochrone map from four views. In the images, warm colors represent early activation, cool colors represent late activation, and dark blue represents no activation in the specific observation window. This isochrone map displays a uterine contraction sequence in which the uterine activation is initiated at the right fundus and propagated to the anterior and right. No activation occurred in the left posterior. Three representative uterine electrograms from sites a, b, and c are shown in Figure 5B. The red and blue lines mark the start and end times, respectively, of the isochrone map in Figure 5A. The EMG burst at site a occurred before those at sites b and c. These EMMI-generated isochrone maps allow investigators to visualize the uterine contraction sequence.
Figure 1: Design of the electrode patch. (A) Template for making MRI-safe marker patches and electrode patches, with measurements shown in millimeters. (B) MRI-safe marker patch. (C) Electrode holder, pin-type electrode, and electrode patch. (D) Bioelectricity mapping hardware with each component labeled. (E) Patch layout on the abdominal surface. (F) Patch layout on the back surface. Please click here to view a larger version of this figure.
Figure 2: Flowchart of the EMMI system. (A) MRI scan of the lower body. (B) A 3D optical scan of the body surface with electrodes in place. (C) Bioelectricity mapping. (D) Body-uterus geometry and electrical signal preprocessing. (E) Uterine electrical signal reconstruction and characterization. Please click here to view a larger version of this figure.
Figure 3: Representative body surface electrogram. (A) A 375 s raw signal recorded from a pin-type electrode on the body surface. (B) Signal from A after a Butterworth bandpass and downsampling. The green lines mark the times of EMG bursts. Please click here to view a larger version of this figure.
Figure 4: Representative uterine surface potential maps. (A–F) Potential maps shown in four views at times marked in the electrogram in G with red dots. The warm colors represent positive potentials and the cool colors represent negative potentials. (G) Electrogram at the site labeled with an asterisk in A-F. Please click here to view a larger version of this figure.
Figure 5: Representative uterine isochrone map and electrograms. (A) An isochrone map shown in four views, with warm colors representing early activation, cool colors representing late activation, and dark blue representing non-activation. (B) Uterine electrograms from sites a, b, and c. The red and blue vertical lines mark the start and end, respectively, of the observation window for this isochrone map. Please click here to view a larger version of this figure.
Electromyography has indicated that the frequency and amplitude of uterine electrical signals alter during the gestational period2,16,25. Several studies have explored the uterine propagation patterns of uterine contractions in patients in active labor10,17,26,27,28. Still, no conclusive propagation direction has been reported, due to the limited number and coverage, as well as the non-standard configuration of the body surface electrodes. The absence of the predominant propagation direction may also be because of the non-fixed pacemaker in the myometrium16,29, but no convincing direct evidence has been reported. EMMI implements a full coverage of the electrodes on the body surface and applies an inverse computation to reconstruct the electrical activities on the uterine surface. EMMI makes it possible to characterize the electrical propagation of the uterine contraction on the whole uterine surface, displaying where the contractions initiate and how they propagate. In addition, with its high temporal resolution, EMMI can analyze the evolution of uterine contractions as the labor progresses with isochrone maps. A thorough analysis of uterine contractions would hold promise to provide new insights into human myometrium electrical maturation and improve the clinical management of human labor.
Preterm labor is a condition potentially caused by multiple pathologic processes, such as cervical diseases, infection, a decline in progesterone action, placental pathologies, abnormal uterine contraction, etc.30,31. By providing high temporal and spatial resolution electrical images of uterine contractions, EMMI holds great promise to improve the prediction accuracy of preterm labor/birth caused by abnormal uterine contractions.
There are several critical steps in performing EMMI in pregnant women. Firstly, the electrode patches must be placed in the same locations as the MRI-safe patches. Following the placement instructions (see the protocol) is critical to reducing electrode location errors. Secondly, it is crucial to use the appropriate amount of gel and establish adequate contact between electrodes and skin to ensure optimal electrical signal activity. Thirdly, multiple optical scans may be required to ensure the acquisition of high-quality body surface geometry.
We have two limitations in the current version of EMMI. One limitation is that MRI is expensive and not portable. Because it is challenging for women to undergo MRI after labor starts, MRI is conducted a few days before they are anticipated to go into labor. As for the preterm patients, whose anticipated labor date is more uncertain than that of term patients, we scheduled multiple MRI scans at 24, 28, 32, and 37 weeks (if the patient goes to term) to record the body-uterus geometry as close to the labor as possible. However, for clinical feasibility, a potential enhancement for EMMI is to utilize clinical ultrasound to obtain patient-specific body-uterus geometry at the bedside. This would decrease the overall expense of EMMI and allow real-time geometry measurement right before or during the electrical recording. The other limitation is the large number of electrodes, which increases the cost of the study and may makes it hard for daily clinical usage. Thus, on the one hand, we plan to make a validation test over the accuracy of EMMI with fewer electrodes. On the other hand, we plan to incorporate cheaper, wearable, disposable, printed electrodes that can be mounted on an elastic material32,33,34. Though several enhancements will be made in the future, the core protocol reported in this manuscript would not change. This work would make it possible for other research groups to reproduce our EMMI work.
The authors have nothing to disclose.
We thank Deborah Frank for editing this manuscript and Jessica Chubiz for organizing the project. Funding: This work was supported by the March of Dimes Center Grant (22-FY14-486), by grants from NIH/National Institute of Child Health and Human Development (R01HD094381 to PIs Wang/Cahill; R01HD104822 to PIs Wang/Schwartz/Cahill), by grants from Burroughs Wellcome Fund Preterm Birth Initiative (NGP10119 to PI Wang), and by grants from the Bill and Melinda Gates Foundation (INV-005417, INV-035476, and INV-037302 to PI Wang).
16 G Vinyl 54" Clear | Jo-Ann Stores | 1532449 | |
3 T Siemens Prisma | Siemens | N/A | MRI scanner |
3M double coated medical tape – transparent | MBK tape solutions | 1522 | Width – 0.5" |
Active electrode holders with X -ring | Biosemi | N/A | 17 mm |
Amira | Thermo Fisher Scientific | N/A | Data analysis software |
Bella storage solution 28 Quart clear underbed storage tote | Mernards | 6455002 | |
Extreme-temperature silicone rubber translucent | McMaster-Carr | 86465K71 | Thickness 1.32” |
Gorilla super glue gel | Amazon | N/A | |
LifeTime carbide punch and die set, 9 Pc. | Harbor Freight | 95547 | |
Optical 3D scan | Artec 3D | Artec Eva Lite | |
PDI super sani cloth germicidal wipes | McKesson medical supply company | Q55172 | Santi-cloth |
Pin-type active electrodes | Biosemi | Pin-type | |
REDUX electrolyte gel | Amazon | 67-05 | |
Soft cloth measuring tape | Amazon | N/A | any brand can be used |
Sterilite layer handle box | Walmart | 14228604 | Closed box |
TD-22 Electrode collar 8 mm | Discount disposables | N/A | |
Vida scanner | Siemens | N/A | MRI scanner |
Vitamin E dl-Alpha 400 IU – 100 liquid softgels | Nature made | SU59FC52EE73DC3 |