{"paper_id":"1590d51f-a938-4602-8b1a-8e8aae3d6ff6","body_text":"Human uterine activity changes dynamically across the menstrual cycle. Menses\nbegins when serum concentrations of the hormones progesterone and estrogen drop, signaling\nthe uterus to shed blood and epithelial cells through the cervix. In the proliferative\nphase, the uterine epithelium grows in thickness to prepare for potential embryo\nimplantation as a follicle develops on one or both ovaries to release an oocyte. During the\nperi-ovulatory phase, an oocyte is released and travels down the fallopian tube. If\nunprotected sexual intercourse occurs during this time, fertilization may occur. During the\nsecretory phase, the uterine epithelium continues to thicken in preparation for potential\nembryo implantation.\nMost research on the menstrual cycle has focused on hormones and their effects on\nthe epithelium. However, some evidence indicates that the smooth muscle layer, the\nmyometrium, also contributes to uterine functions by generating slow, low-magnitude,\nspontaneous contractions, termed uterine peristalsis 1 – 10 . Unlike labor\ncontractions, in which the entire myometrium produces faster and stronger contractions,\nuterine peristalsis only involves the inner layer of the myometrium, the stratum\nsubvasculare. Uterine peristalsis, first observed on ultrasound 5 , has been shown to vary in direction and frequency\nthroughout the phases of the menstrual cycle 1 . During menses, peristalsis waves travel from the fundus to the cervix and\nhelp expel blood and tissue. Conversely, peristalsis waves travel from the cervix toward the\nfundus during the peri-ovulatory phase and help transport sperm toward the fallopian\ntubes.\nSeveral studies have suggested that uterine peristalsis plays an essential role in\nuterine pathology. Disruptions in uterine peristalsis may occur in women who experience\ninfertility 9 , dysmenorrhea 4 , and endometriosis 11 , 12 , a painful\ncondition in which cells from the uterine epithelium implant and grow outside of the uterus,\ncommonly in the peritoneal space. In addition to causing chronic pelvic pain, endometriosis\nmay also cause dysmenorrhea, irregular bleeding, and subfertility 13 . Evidence that disrupted uterine peristalsis contributes\nto endometriosis comes from studies using ultrasound and intrauterine pressure catheters.\nThese studies demonstrated that patients with endometriosis had dysperistalsis and higher\nuterine tone, and more frequent Cervix-Fundus contractions than normal women 8 , 14 , 15 .\nAlthough previous studies provided measurements of uterine peristalsis, the\navailable data have been limited by the capabilities of the four main technologies used to\nassess uterine peristalsis 1 , 16 . First, intrauterine pressure catheters are invasive,\nand a catheter placed inside the uterus could alter peristalsis patterns. Second,\ntransvaginal ultrasound (TVUS) 17 – 19  is invasive and is not sensitive enough to\nidentify the site of peristalsis initiation. Additionally, the quality of TVUS measurement\ndepends on the orientation of the ultrasound transducer, making this method highly\nsubjective and operator- and time-dependent 20 – 25 . Third,\nhysterosalpingography (HSSG) is a procedure in which X-rays are used to detect a\nradiographic contrast dye injected into the uterus and fallopian tubes. Although HSSG\nmeasures are objective, HSSG cannot be used to measure peristalsis amplitude or frequency,\nand radiation exposure limits the imaging time. Fourth, cine magnetic resonance imaging\n(MRI) 26 – 29  can be used to detect uterine peristalsis by acquiring\nsequential images for an extended period of time and playing the MRI frames 12 times faster\nthan the actual speed 26 . However, extended\ncine MRI is expensive, time-consuming, and operator-dependent, and it cannot reveal the\ninitiation and termination sites of uterine peristalsis. Moreover, all of the above\nmodalities can be uncomfortable for the participant and cannot be used for long-term\nobservation.\nWe recently developed an electrophysiological imaging system called\nElectromyometrial Imaging (EMMI) 30 – 33  to quantitatively measure the electrical\nactivity underlying uterine contractions during labor. Here, we adapted this system to\nlongitudinally image the 4-dimensional (4D) electrical waves of uterine peristalsis over\neach phase of the menstrual cycle in healthy, nonpregnant participants with normal menstrual\ncycles and in participants with endometriosis. With this uterine peristalsis imaging (UPI)\nsystem, we can image human uterine peristalsis in a safe, comfortable, and accurate way. UPI\ncan provide precise quantitative electrophysiological evidence that uterine peristalsis\nchanges in frequency, direction, duration, magnitude, and power throughout the menstrual\ncycle and is disrupted in endometriosis patients.\n\nOur uterine peristalsis imaging (UPI) system is further developed based on the\nEMMI system and is illustrated in  Fig. 1 . First, a\nwoman underwent a one-time, fast, anatomical MRI scan ( Fig.\n1A ) to acquire the patient-specific uterus-body surface geometry ( Fig. 1B ,  C ), while wearing\nMRI-compatible fiducial markers around the abdomen and lower back. Second, customized\npin-type electrode patches were applied to the same locations on the body surface as the\nMRI fiducial markers ( Fig. 1D ). Body surface\nelectrical signals ( Fig. 1E ) were recorded for 20\nminutes, and electrical signals (peristalsis wave signals  Fig. 1F ) were generated using a band-pass filter (0.01–0.1 Hz) 25 , 34 , 35 . Third, UPI software was used to generate\nelectrical signals at each point on the entire 3D uterine surface ( Fig. 1G ,  H ). These electrical\nsignals were used to derive activation sequences, uterine potential maps, and uterine\nisochrone maps ( Fig. 1I – K ). Finally, the uterine surface data were automatically analyzed\nto define the peristalsis direction (Cervix-Fundus, Fundus-Cervix, or other), initiation\nand termination sites (cervix area, fundus area, and other areas), and their distributions\n( Fig. 1L ). Other UPI electrophysiological indices\nof uterine peristalsis include duration, magnitude, and power of peristalsis waves. See\ndetailed descriptions in the  Method  section.\nWe used the UPI system to image uterine peristalsis during each menstrual cycle\nphase in 17 nonpregnant women with regular menstrual cycles. In total, we imaged 4968\nuterine peristalses over 34 hours. In  Fig. 2 , we\npresent representative uterine peristalsis waves of a 26-year-old participant. During the\nmenses phase, 65% of waves traversed from near the fundus toward the cervix, and 35%\ntraversed from near the cervix toward the fundus ( Fig.\n2A ). During the proliferative phase, 52.8% of waves were Fundus–Cervix and\n44.4% were Cervix-Fundus ( Fig. 2B ). During the\novulatory phase, 75.8% of waves were Cervix-Fundus, and 24.2% were Fundus-Cervix ( Fig. 2C ). In the secretory phase, 60% of waves were\nCervix-Fundus, and 34% were Fundus-Cervix ( Fig. 2D ).\nIn all cases in which we were able to determine the direction of peristalsis in TVUS\nimages (n = 111), the direction of peristalsis imaged by UPI matched the direction\nobserved by TVUS. Overall, uterine peristalsis waves during menses were significantly\nlonger in duration and had greater magnitude and power than those during the ovulatory\nphase ( Fig. 2F – I ).\nWe used our UPI system to image uterine peristalsis during each phase of the\nmenstrual cycle in five nonpregnant women with surgically confirmed endometriosis. In\ntotal, we imaged 679 peristalses over 12.5 hours throughout the menstrual cycle. In  Fig. 3 , we present representative uterine peristalsis\nwaves of a 30-year-old participant with endometriosis. During the menses phase ( Fig. 3A ), 44.2% of waves were Fundus-Cervix, and 48.8%\nwere Cervix-Fundus. During the proliferative phase ( Fig.\n3B ), 36.3% of waves were Fundus-Cervix, and 42.2% were Cervix-Fundus. During the\novulatory phase ( Fig. 3C ), 59.9% of waves were\nCervix-Fundus, and 25.4% were Fundus-Cervix. During the secretory phase ( Fig. 3D ), 47.8% of waves were Cervix-Fundus, and 50% were\nFundus-Cervix. In all cases in which we were able to determine the direction of\nperistalsis in TVUS images (n = 126), the direction of peristalsis imaged by UPI matched\nthe direction observed by TVUS. Overall, uterine peristalsis waves during menses were\nsignificantly shorter in duration than those during the ovulatory phase and had greater\nmagnitude and power than those during the secretory phases ( Fig. 3F – I ).\nWe next compiled all our data from the healthy and endometriosis participants.\nThe length of each participant’s menstrual cycle was normalized to 28 days. We\nplotted each participant’s overall frequency and dominant direction ratio (the\npercentage of Cervix–Fundus peristalsis waves over the percentage of\nFundus–Cervix peristalsis waves) ( Fig. 4\nA – B ). We also graphed the average\nmagnitude, duration, and power of peristalsis waves from each participant, with data from\nthe Fundus–Cervix waves plotted separately from the data from Cervix–Fundus\nwaves ( Fig. 4 C – H ). We observed significant differences in multiple uterine peristalsis indices\nbetween healthy participants and those with endometriosis ( Fig. 4 I – X ). During the menses\nphase, peristalsis waves were significantly more likely to be Fundus–Cervix in\nhealthy participants than in those with endometriosis ( Fig.\n4J ). The Fundus–Cervix waves were longer ( Fig. 4R ) and had a higher magnitude ( Fig.\n4N ) in healthy participants than in those with endometriosis. Conversely, the\nCervix–Fundus waves were longer duration ( Fig.\n4Q ) and higher magnitude ( Fig. 4M ) and power\n( Fig. 3U ) in the participants with endometriosis\nthan in the healthy patients. In the peri-ovulatory phase, peristalsis waves were more\nlikely to be Cervix–Fundus in the healthy participants than in the participants\nwith endometriosis ( Fig. 4K ), and the\nCervix–Fundus waves were longer ( Fig. 4S ) and\nhigher magnitude ( Fig. 4O ) and power ( Fig. 4W ) in the healthy participants than in those with\nendometriosis. Conversely, the Fundus–Cervix waves in the peri-ovulatory phase were\nlonger duration ( Fig. 4T ) and higher magnitude ( Fig. 4P ) in the participants with endometriosis than in\nthe healthy participants.\nFinally, we found that Cervix–Fundus peristalsis waves during the\nperi-ovulatory phase tend to move preferentially toward one fallopian tube. In nine of the\nhealthy participants and two of the participants with endometriosis, we were able to\ndetermine which ovary had a dominant follicle by clinical TVUS and then observe whether\nthe peristalsis propagated in the direction of the dominant follicle.  Fig. 5A  shows an example of UPI from a healthy participant with a\ndominant follicle in the right ovary. In this patient, 5 of 8 Cervix–Fundus\nperistalsis episodes moved toward the right ovary. The other 3 waves showed no\npreferential direction.  Fig. 5B – D  show additional examples of healthy participants in which\nperistalsis patterns propagated toward the ovary with the dominant follicle.  Fig. 5E  shows an example of a participant with\nendometriosis and a dominant follicle in the left ovary. In this participant, 4 out of 5\nperistalsis cycles progressed toward the right fallopian tube and 1 progressed toward the\nleft fallopian tube.  Fig. 5F  shows a second\nparticipant with endometriosis and a dominant follicle in the left ovary. In this\nparticipant, 6 out of 13 Cervix–Fundus peristalsis waves moved in the direction of\nthe right fallopian tube, while none moved toward the left fallopian tube.\nIn the eight healthy participants for whom we had TVUS imaging demonstrating the\ndominant follicle, peristalsis waves during the ovulatory phase more often moved toward\nthe side with the dominant follicle than toward the side with no dominant follicle. In two\nparticipants with endometriosis for whom we had data regarding the dominant follicle, the\nperistalsis waves during the ovulatory phase more often moved toward the side without the\ndominant follicle than toward the side with the dominant follicle (Table 1).\n\nThe UPI imaging data presented herein suggest that UPI can provide objective and\nquantitative measures of uterine peristalsis throughout the human menstrual cycle.\nAdditionally, we developed novel indices to quantitatively characterize uterine peristalsis\npatterns automatically. Finally, we used UPI to provide evidence that uterine peristalsis\npatterns differ in women with normal anatomy and menstrual cycles and in women with\nendometriosis.\nIn the normal participants, the predominant peristalsis pattern in menses was\nFundus-Cervix. This pattern has been seen by others and postulated to facilitate the\nexpulsion of blood and endometrial tissue while protecting against ascending\npathogens 36 . In the peri-ovulatory\nphase, the predominant peristalsis pattern was Cervix-Fundus. Kunz et al. used serial HSSG\nto follow labeled macrospheres the size of sperm and observed that they were transported\nfrom the cervix into the uterus and fallopian tubes 37 , suggesting that the Cervix-Fundus peristalsis pattern facilitates the\ntransport of sperm toward the oocyte. We observed no predominant pattern in the\nproliferative and secretory phases. The duration and magnitude of contractions differed in\neach phase. The rise in oxytocin and estrogen in the follicular phase may explain why the\nmagnitude of the peristalsis pattern is increased during menses 1 , 38 , 39 . After ovulation, during the secretory phase,\nprogesterone (a known muscle relaxant) contributes to the decrease in the magnitude of\nperistalsis by antagonizing the estrogen and oxytocin receptors 40 .\nEndometriosis has long been hypothesized to be caused by retrograde\nmenstruation 13 , 41 – 46 .\nHowever, as all reproductive-age women have some amount of retrograde menstruation, it is\nunclear why only 10–15% of females would develop endometriosis 42 , 45 , 47 – 49 . We\nfound that all healthy participants had at least some Cervix–Fundus peristalses,\nwhich could cause retrograde menstruation. Our data suggested that Cervix-Fundus peristalsis\nwaves were less frequent and weaker than the Fundus–Cervix waves in subjects without\nendometriosis. Therefore, the strong and frequent Fundus-Cervix waves may have effectively\nexpelled blood vaginally and left a small amount of blood in the uterine cavity. Although\npart of the blood could still be transported retrogradely to the peritoneal space by the\nweak Cervix–Fundus waves, the level may not be sufficient to cause endometriosis in\nhealthy people. On the contrary, in participants with endometriosis, a higher percentage of\nwaves were Cervix–Fundus, and these were stronger and had longer durations than the\nCervix–Fundus waves in normal patients. More importantly, in healthy subjects, the\nFundus-Cervix peristalsis waves were less frequent and weaker in endometriosis patients than\nthe Fundus-Cervix peristalsis waves, which impair normal expulsion and leave more blood in\nthe uterine cavity. Therefore, retrograde menstruation is more likely to push much more\nblood and tissue into the peritoneal space in women with endometriosis than in women without\nendometriosis 8 , 12 , 50 , 51 . Our work suggests that a comprehensive evaluation of 4D\nuterine peristalsis direction distribution, frequency, magnitude, duration, and power during\nthe menses phase could be used to stratify the risk of developing endometriosis and assess\nthe severity of endometriosis.\nOur data may also provide clues to infertility in women with endometriosis. In\nhealthy participants during the peri-ovulatory phase, uterine peristalsis waves most\nfrequently traveled Cervix–Fundus, with most peristalsis waves traveling toward the\ndominant follicle. These patterns could assist sperm in transit to ensure interaction with\nan oocyte. Conversely, in participants with endometriosis during the peri-ovulatory phase,\nuterine peristalsis waves most frequently traveled Fundus–Cervix, and those that\ntraveled Cervix–Fundus traveled toward the ovary without a dominant follicle more\noften than toward the ovary with a dominant follicle. These patterns could limit the number\nof spermatozoa that reach the oocyte 20 , 21 , 52 , 53 .\nThe UPI system potentially has a wide range of possible clinical research and\ntherapeutic applications. Based on the initial work presented in this work, UPI can be used\nto further establish reference baseline parameters of uterine peristalsis in normal\nmenstrual cycles. These baseline values could be used to create a composite score to\nidentify patients with abnormal gynecological conditions such as endometriosis, ovulatory\ndysfunction, abnormal uterine bleeding, or amenorrhea. Additionally, UPI could be used to\ncorrelate the dominant follicle with uterine peristalsis direction in the peri-ovulatory\nphase and to develop a predictive biomarker for successful natural conception. With the\ndetailed 4D electrical activation patterns imaged by UPI, we can longitudinally evaluate the\ntreatment effects of various clinical interventions and optimize the treatment plan for an\nindividual patient. In addition, UPI may facilitate the development of nonpharmaceutical\ntreatments to electrically correct abnormal uterine peristalsis underlying various\ngynecological conditions, such as endometriosis, etc., using electronic devices similar to\ncardiac pacemakers.\nUPI has several advantages over other modalities used to image uterine\nperistalsis. First, UPI is noninvasive, which is optimal for long-duration uterine\nmonitoring. Additionally, modalities using invasive monitoring may iatrogenically cause\nnon-physiologic perturbations of peristalsis. Second, UPI provides high spatial-temporal\nresolution information, including the initiation sites, direction, frequency, and duration\nof uterine peristalsis waves. Third, UPI provides 4D data that considers the\nindividual’s unique uterine anatomy in both space and time domains. Fourth, UPI\nsoftware allows automatic, objective, and real-time electrophysiological quantification of\nuterine peristalsis. Future work will focus on developing a portable, low-cost, wearable UPI\nsystem to enable larger UPI studies. To make UPI more accessible to patients, we will\nreplace the current short anatomical MRI scan with a low-cost ultrasound measurement to\ngenerate the patient-specific body-uterus geometry. Wearable, low-cost, printed\nelectrodes 54 , 55  will also be integrated into the UPI system to minimize the costs.\n\nThis study was performed in the Division of Reproductive Endocrinology &\nInfertility at Washington University School of Medicine. This study was approved by the\nWashington University Institutional Review Board, and all participants signed informed\nconsent documents. Participants were included if they were female at birth, between the\nages of 18 and 37 years. Normal participants were included if they had regular,\npredictable menstrual cycles every 24–35 days. Participants with endometriosis were\nincluded if they had surgically confirmed endometriosis. Potential participants were\nexcluded if they were post-menopausal, pregnant, or breastfeeding; had a uterine anomaly;\nhad exposure to medications known to affect uterine contractility (e.g., magnesium,\nopioids, beta antagonists, nifedipine); were non-English speaking; had abdominal\ncircumference > 55 cm; or had MRI contraindications (pacemaker, metal implants,\netc.). Potential participants for the normal group were excluded if they had documented or\nself-reported histories of infertility, ovulatory dysfunction, or endometriosis. Potential\nparticipants for the endometriosis group were excluded if they were currently using female\nbirth control. Seventeen out of them finished the longitudinal data acquisition and MRI\nstudy. Participants with regular menstrual cycles and five patients with endometriosis\nwere enrolled in this study. Demographics and obstetric and gynecologic history of\nenrolled participants are shown in Supplemental Table 1. Each participant was imaged with\nthe UPI system four times during one menstrual cycle, once during menses, early\nproliferative, late proliferative (peri-ovulatory), and secretory phases. Blood was\ncollected at each visit to measure concentrations of the hormones estradiol, progesterone,\nand testosterone to confirm the menstrual phase.\nPatients were determined to be in one of four menstrual phases (menses, early\nproliferative, late proliferative, and secretory) by using a combination of\npatient-reported bleeding, cycle length, ultrasound findings, ovulation predictor kit\n(Clearblue, Geneva, Switzerland) results, and hormonal measurements. Serum blood\n(5–10 ml) was collected and sent to the Core Laboratory for Clinical Studies at\nWashington University in St. Louis to measure concentrations of the hormones (estradiol,\nprogesterone, and testosterone). The menses phase was assigned when a patient-reported\nbleeding. The early proliferative phase was assigned after the patient had stopped\nbleeding, ultrasound demonstrated early follicular activity (largest follicle size\n<16 mm), serum estradiol <200 pg/ml, and serum progesterone <3 ng/ml.\nThe late proliferative (peri-ovulatory) phase was defined by a positive result on an\novulation predictor kit, serum estradiol >200 pg/ml, serum progesterone <3\nng/ml, and/or a dominant follicle on ultrasound (⩾16mm). The secretory phase was\nassigned when serum progesterone was >3 ng/ml.\nFirst, a woman underwent a one-time, fast, anatomical (T2W sequence) 3T Siemens\nPrisma MRI scan (~10 mins) to acquire the patient-specific uterus-body surface\ngeometry while wearing up to 8 patches containing up to 128 MRI-compatible fiducial\nmarkers around the abdomen and lower back ( Fig. 1A ).\nUterus and body geometry were generated ( Fig. 1\nB & C ). Second, after the MRI scan,\ncustomized BioSemi pin-type electrode patches were applied to the same locations on the\nbody surface as the MRI fiducial markers. An ADC box was used to record the body surface\nelectrical signals ( Fig. 1D & E ) for 20 minutes. The body surface electrical signals were\nprocessed with a band-pass filter (0.01–0.1 Hz) 25 , 34 , 35  to generate wave electrical signals (peristalsis waves) over the\nentire abdomen surface ( Fig. 1F ). Third, the\nparticipant underwent another 10-minute electrical recording while simultaneously\nundergoing transvaginal ultrasound (TVUS). TVUS scans of the uterus were performed by the\nsonographer holding the transducer probe while the patient was lying in a lithotomy\nposition, and TVUS clips were recorded on a GE Voluson S8 ultrasound machine. The duration\nof each clip was 20 seconds on average, and 30–35 clips were acquired in total. A\nregistered sonographer independently (without knowledge of the UPI results) examined the\nTVUS recordings to determine the uterine peristalsis direction.\nWith the electro-quasi-static assumption of the bioelectric field, the inverse\ncomputation combines the patient-specific uterus-abdomen surface and electrical potentials\nmeasured on the abdominal surface to reconstruct the potential distribution over the\nentire 3D uterine surface. We assume that the medium is homogeneous between the uterine\nsurface and abdominal surface without any primary electrical source. Then, the inverse\nproblem could be mathematically described by the Cauchy problem for Laplace’s  equation (1)  with boundary conditions ( 2 , 3 ) on the abdominal surface.\nDirichlet ( 2 ) and Neumann ( 3 ) conditions for the abdominal surface\npotentials are:  \n (2) \n ϕ ( x ) = ϕ A ( x ) , x ∈ Γ A \n \n \n (3) \n ∂ ϕ ( x ) ∂ n = 0 , x ∈ Γ A\nHere,  n  is the normal vector on the abdominal\nsurface at location  x  and Γ A \nrepresents abdominal surface.\n ϕ A ( x )  is the\npotential measured on the abdominal surface and\n ϕ ( x ) is the potential on the uterine\nsurface.\nAs a mesh-free method robust to noise, a method of fundamental solutions\n(MFS) 56  was deployed to discretize\nthe Laplace’s equation and boundary conditions, which is accurate for solving the\nbioelectric field inverse problem in both electrocardiographic imaging (ECGI) 56  and electromyometrial\nimaging(EMMI) 30 , 32 , 33  systems.\nThis problem cannot be solved directly as it is an ill-posed inverse problem. Therefore,\nTikhonov-based inverse computation with a fixed regularization value of 0.01 was used to\nobtain the solution.\nHere, Φ A  is a M * T matrix of measuring\nsurface potentials, Φ U  is a N * T matrix of uterine\nsurface potentials, where M is the number of measuring electrodes applied on the abdominal\nsurface and N is the number of discrete points on the uterine surface, and T is the number\nof recording time points.  A  is a M * N linear transform\nmatrix encoding the relationship between abdominal surface potential\nΦ A  and uterine surface potential\nΦ U .\nThe inverse computation described above was employed to compute the uterine\nsurface electrical signals ( Fig. 1 G & H ) on the three-dimensional uterine surface. The times\nwhen the uterine surface electrical signals at various uterine surface areas reached the\nsteepest negative slope 57 – 61  were extracted and defined as electrical\nactivation times at those uterine areas during peristalsis waves (red dots in  Fig. 1 G & H ).\nDuring each peristalsis wave, sequential time frames were generated as the activation\nsequences ( Fig. 1I ) to reflect the detailed 4D\nspatial-temporal activation patterns of the uterine peristalsis. Within each time frame,\nthe red region indicated the electrically activated myometrium areas currently\nexperiencing peristalsis, and the blue region indicated the inactive areas of the uterus.\nThe isochrone map was generated as a color-coded 3D map to summarize the electrical\nactivation sequence ( Fig. 1J ). In the isochrone map,\nwarm and cool colors denote regions of the uterus that activated early and late,\nrespectively, during the peristalsis wave. The UPI isochrone maps contained rich\nspatial-temporal information of uterine activation, including the activation and\ntermination sites, propagation direction, and duration. In addition, uterine potential\nmaps were generated to reflect the 4D electrical potential distribution during peristalsis\nwaves: 1D electrical signals ( Fig. 1 G & H ) over the entire 3D uterine surface ( Fig. 1K ). The distributions of uterine peristalsis propagation\ndirection, initiation, and termination sites ( Fig.\n1L ) were automatically calculated as the number of peristalsis waves with a\nspecific propagation direction (Fundus-Cervix, Cervix-Fundus or other), initiation, and\ntermination site (cervical, fundal or other regions) divided by the total number of\nperistalsis waves in the 20-minute electrical mapping session, respectively.\nFive UPI electrophysiological indices were defined to qualitatively and\nquantitatively describe uterine peristalsis patterns. First, the propagation direction was\ndetermined from the uterine peristalsis activation maps. Uterine peristalsis directions\nwere classified into three categories: Fundus-Cervix, Cervix-Fundus, and others including\nAnterior-Posterior, Posterior-Anterior, Left-Right, and Right-Left. Second, the initiation\nand termination sites were defined as the region experiencing the earliest and latest\nactivation during uterine peristalsis. The initiation and termination sites were\nidentified on the isochrone maps and were classified into three categories: Cervical\nregion, Fundal region, and Other regions. Third, the duration (Sec.) was defined as the\nduration of a complete peristalsis wave measured in the isochrone map of the uterine\nperistalsis wave. A small fraction of uterine peristalsis waves only involve the partial\nactivation of the uterus and has a relatively shorter duration. Fourth, uterine\nperistalsis magnitude (mV) was defined as the average peak amplitude of electrical\npotential over the uterine region experiencing activation during the entire peristalsis\nwave. Finally, uterine peristalsis power (mV*sec) was defined as the product of magnitude\nand duration for each uterine peristalsis.\nThe distance between the latest fundus-activated uterine site and the left\nfallopian tube insertion site was defined as\n d left ,, the distance between the latest\nfundus-activated uterine site and the right fallopian tube insertion site was defined as\n d right ,,. The ratio between these two\ndistances was defined as  R = d left d right .\nIf  R  < 0.8, the cervix-fundus uterine peristalsis was left\ndominant; if  R  > 1.25, the cervix-fundus uterine peristalsis was\nright dominant; if 0.8 <  R  < 1.25, the cervix-fundus\nuterine peristalsis was middle dominant with no side preference.","source_license":"CC0","license_restricted":false}