Dynamic MRI Reveals How the Young Pelvic Floor Moves From Rest to Strain to Squeeze

Deep inside the pelvis, a hammock of muscle and connective tissue performs a quiet, constant balancing act. It holds up the bladder, the uterus, and the rectum against the pull of gravity, and it adjusts its shape every time we breathe, cough, strain, or squeeze. For most of medical history, this hidden architecture has been studied indirectly, through physical examination or ultrasound. Now, a team of researchers in Kunming, China, has used high-resolution dynamic magnetic resonance imaging to watch the pelvic floor of young, healthy women as it moves through three distinct states: rest, maximal straining, and maximal voluntary contraction. The results, published in BMC Medical Imaging, offer one of the most detailed portraits yet of what a normal, never-pregnant pelvic floor actually looks like in motion.

The study, led by Ling Li and Daoming Tian of the Center for Pelvic Floor Disorders at the First Affiliated Hospital of Kunming Medical University, recruited sixteen healthy Asian women between the ages of eighteen and thirty who had never been pregnant. This nulligravid population is a crucial reference point. Pelvic floor dysfunction, including prolapse and incontinence, affects a large share of women worldwide, but clinicians still lack a firm baseline for what the uninjured, youthful pelvic floor does when asked to perform. Without that baseline, it is difficult to say whether a given measurement in a patient is abnormal or simply part of natural variation.

To capture the pelvic floor in action, the team used a 3.0 Tesla MRI scanner, a field strength high enough to resolve fine soft-tissue detail in the midsagittal plane, the side-on view that slices the pelvis down the middle. Each participant was imaged three times in quick succession: once at rest, once while performing a maximal Valsalva maneuver, meaning a forced exhalation against a closed airway that raises intra-abdominal pressure, and once while performing a maximal voluntary pelvic floor contraction, the deliberate squeeze familiar to anyone who has practiced Kegel exercises. The researchers then measured a battery of anatomical landmarks in each state, from the position of the bladder neck and anorectal junction to the angle of the levator plate, the length of the urethra and vaginal walls, and the area of the levator hiatus, the opening through which the urethra, vagina, and rectum pass.

One of the study’s technical strengths lies in how those measurements were made. Rather than relying on raw scanner coordinates, which shift depending on how a patient is positioned on the table, the team applied a Pelvic Inclination Correction System, or PICS, coordinate framework. This approach corrects for pelvic tilt and establishes a reproducible reference line running from the pubis to the sacrococcygeal joint, allowing landmarks measured in different women and different states to be compared on equal footing. All measurements were performed independently by two radiologists using specialized segmentation software, and the average of their readings was used in the final analysis.

The reliability of those readings was generally strong. Intraclass correlation coefficients, which quantify how closely two observers agree, ranged from 0.858 to 0.999 across the three states, values that in most imaging studies would be considered excellent. The authors were careful, however, to note that some of the confidence intervals around those estimates were wide and dipped below 0.75, particularly during the Valsalva and contraction maneuvers. In a cohort of only sixteen participants, a single outlier can stretch those intervals considerably, a limitation the researchers acknowledge openly.

When the three states were compared statistically using repeated-measures analysis of variance, with a false discovery rate correction applied across thirty-three endpoint tests to guard against spurious findings, thirty-two of the thirty-three endpoints showed evidence of state-related differences. In other words, nearly every measurable feature of pelvic floor morphology changed significantly depending on whether the woman was resting, straining, or squeezing. The largest differences emerged between the two active maneuvers, Valsalva and voluntary contraction, which turned out to be far from mirror images of each other.

The numbers tell a striking story. During voluntary contraction compared with Valsalva, the anorectal junction moved upward by an average of 18.50 millimeters, the upper levator plate angle decreased by about 30.44 degrees, the M-line displacement, a measure of pelvic floor descent, was smaller by 12.69 millimeters, and the levator hiatus area shrank by 2.41 square centimeters. In plain terms, when these women actively squeezed their pelvic muscles, the entire pelvic floor lifted, the muscular plate behind the anus tilted into a more horizontal, supportive orientation, and the opening in the levator muscle narrowed. During straining, the opposite occurred: the hiatus widened, the organs descended, and the pelvic floor flattened under pressure.

Perhaps the most intriguing observation concerns the uterovaginal axis, the line running from the uterine fundus down through the cervix and vagina. The researchers sampled points along this axis and tracked their displacement across states, finding hints that different segments of the axis move by different amounts during contraction. This suggests that the pelvic floor does not behave as a single rigid sheet but as a segmented, coordinated system in which the upper and lower portions may respond differently to muscular effort. The authors describe this as a hypothesis-generating finding, one that will require three-dimensional imaging and concurrent functional measurements to confirm.

The clinical implications could be substantial. Dynamic MRI of the pelvic floor is already used to diagnose prolapse and obstructed defecation, but interpretation depends on reference values, and most existing norms come from older or parous women. By documenting how a healthy young pelvic floor responds to maximal effort, this study provides a physiological anchor for distinguishing normal function from early dysfunction. It also highlights a subtlety that matters in the clinic: straining and squeezing are not simply opposite ends of one spectrum but distinct neuromuscular configurations, and a patient’s ability to perform each correctly can reveal different problems. A woman who bears down when she intends to squeeze, a common error in pelvic floor rehabilitation, would show a pattern on dynamic MRI that is unmistakably different from a correct contraction.

The authors are appropriately cautious about how far these findings can be pushed. Sixteen women, all from a single ethnic group and a single hospital, cannot define the full range of human pelvic floor anatomy, and the nonparametric sensitivity analyses were not fully concordant with the primary results, a reminder that some findings may not survive in larger samples. Still, as an exploratory map of the living pelvic floor in motion, the study is a valuable first step. It shows that with careful coordinate correction, rigorous observer agreement, and modern 3.0 Tesla imaging, the invisible mechanics of pelvic support can be measured with precision, opening the door to better baselines, better diagnostics, and perhaps one day, imaging-guided pelvic floor training tailored to each woman’s own anatomy.

Subject of Research: Dynamic MRI assessment of pelvic floor morphology in young healthy nulligravid women during rest, Valsalva, and voluntary contraction

Article Title: Dynamic MRI of pelvic floor morphology in young, healthy Asian nulligravid women: an exploratory study across rest, Valsalva, and voluntary contraction

Article References: Li, L., Tian, D., Zeng, Y., Wang, X., Luo, Q., & Shen, J. (2026). Dynamic MRI of pelvic floor morphology in young, healthy Asian nulligravid women: an exploratory study across rest, Valsalva, and voluntary contraction. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02853-8

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02853-8

Keywords: pelvic floor, dynamic MRI, nulligravid women, Valsalva maneuver, pelvic floor contraction, levator hiatus, pelvic organ prolapse, PICS coordinates, 3.0 Tesla MRI, pelvic floor dysfunction, levator ani, urogynecology