Surface Electromyography Characteristics of Pelvic Floor Muscles in Healthy Women, Pelvic Floor Dyssynergia, and Urinary Incontinence: A Retrospective Comparative Study

Introduction

Pelvic floor dysfunctions constitute a major clinical and public health concern, leading to urinary incontinence, pelvic organ prolapse, chronic pain, and sexual dysfunction [1]. Despite the widespread use of surface electromyography (sEMG) to assess pelvic floor muscle (PFM) activity [2,3], existing reference ranges are based on limited cohorts and provide insufficient diagnostic clarity, particularly in differentiating between overactivity, underactivity, and impaired relaxation. sEMG signals are the time-varying electrical potentials recorded within a muscle, in the surrounding tissue, or at the skin surface during muscle activity; they represent the summation of the extracellular action potentials generated by muscle fibers in the vicinity of the electrode and constitute the active component of muscle tone [4]. sEMG provides critical insights into neuromuscular control, both at rest and during dynamic tasks such as contraction and relaxation [5]. In contrast, the passive component (viscoelastic component) results from the biomechanical properties of muscles, including connective tissue tension and osmotic pressure within cells. It is important to note that sEMG signal recordings cannot differentiate between active muscle tone (resulting from muscle contraction) and passive muscle tone (stemming from the elastic properties of tissues) [6]. As emphasized in recent reviews, sEMG measures electrical excitation rather than functional activation, and its interpretation must be cautiously contextualized [7]. No universally accepted normative values exist for PFM sEMG signals, and comparisons between individuals or studies require strict standardization and normalization procedures [8]. Some authors, such as Muallem et al [9], have proposed reference thresholds (eg, ≥3 μV) to indicate elevated resting activity; however, such cut-offs should be interpreted with caution given methodological variability [7,8]. Elevated resting sEMG signals can suggest pelvic floor overactivity, a neuromuscular phenomenon that can also manifest during voluntary tasks or as impaired relaxation. While the mechanisms underlying overactivity remain poorly understood, it is recognized as a neuromuscular phenomenon that can most directly be observed through sEMG [10]. In various conditions, such as vestibulodynia [10], altered PFM sEMG signals can manifest either as increased resting excitation (often referred to as overactivity) or reduced recruitment capacity, as observed in cases of stress urinary incontinence [11]. Similarly, studies comparing individuals with pelvic floor dyssynergia (paradoxical contraction or impaired relaxation of the pelvic floor during attempted defecation or muscle activity [12]) and healthy controls have reported elevated resting sEMG signals in the dyssynergia group, suggesting a potential pattern of persistent muscle excitation [13]. These alterations are commonly interpreted as markers of overactivity or underactivity in the PFM (reduced signal amplitude or peak microvolts, or shorter duration of sustained contraction suggests decreased myoelectrical activity [12]); however, their physiological meaning remains complex and context-dependent. Nevertheless, identifying overactivity based on sEMG signals alone remains challenging and must be supported by comprehensive clinical evaluation. As emphasized by Aw et al [14], the presence of overactivity has different therapeutic implications than does underactivity. Elevated EMG values may reflect compensatory patterns of motor unit recruitment in response to muscle weakness or impaired coordination, rather than effective function. In such cases, the application of conventional strengthening exercises can exacerbate the problem, whereas relaxation techniques and “down-training” are more appropriate. These perspectives highlight the need for caution when applying sEMG thresholds for diagnostic purposes and support the present study’s approach of comparative analysis within and between well-defined clinical subgroups, rather than relying on predefined normative cut-offs [8]. Nonetheless, the search for clinically meaningful sEMG reference points remains relevant. Despite growing interest in establishing diagnostic criteria, to date, no consensus sEMG cut-off point exists for differentiating between PFM dysfunctions, and the current evidence base is insufficient to determine which sEMG signal values most accurately distinguish healthy from impaired PFM function [15]. Understanding the relationship between different states of muscle tone as a whole is essential for diagnosing and treating motor dysfunctions [16]. This gap has been recently emphasized in large-scale analyses [17], further underscoring the importance of subgroup-specific comparative studies such as the present one.

Therefore, the aim of this study was to evaluate a range of sEMG signal characteristics, including baseline activity and parameters reflecting contractile function (baseline rest period, quick flicks, tonic contractions, static contractions, and post-contraction relaxation) based on the Glazer Protocol [3,18], in 3 groups of women: healthy controls, those with pelvic floor dyssynergia, and those with urinary incontinence. By comparing these subgroups, the study sought to explore whether distinct sEMG signals may help distinguish between different types of pelvic floor dysfunctions and provide clinically relevant insights to inform assessment and rehabilitation strategies. We hypothesized that women with dyssynergia would demonstrate signals consistent with pelvic floor “overactivity”, operationalized as elevated sEMG amplitudes across tasks, while women with urinary incontinence would present lower amplitudes indicative of “underactivity”. Importantly, these hypotheses are exploratory, as the study is not intended to establish diagnostic thresholds but rather to generate data that may, in the future, support clinical practice.

Material and Methods

STUDY DESIGN AND PARTICIPANTS:

This retrospective observational cross-sectional study was based on datasets from 3 independent projects conducted between 2018 and 2022 in the Laboratory of Functional Diagnostics at Opole University, Poland. Ethical approval was obtained for all participant groups prior to data collection, and additional variables were extracted to address new research questions. sEMG data were collected using the same equipment (Noraxon Ultium system with an intravaginal bipolar probe) and the standardized Glazer protocol, with all assessments performed by a single experienced examiner. The study followed the Consensus for Experimental Design in Electromyography checklist [19] to ensure methodological rigor and reproducibility. Sample size was determined a priori from the available datasets. To enhance robustness despite unequal group sizes, Bayesian multivariate regression with covariate adjustment was applied. Thus, while limited, the sample was considered adequate for exploratory analyses and consistent with prior literature. As a retrospective study, no new data collection was undertaken, and group sizes were dictated by the number of eligible participants. No randomization was applied, and although the examiner was not blinded to group allocation, measurement bias was minimized by having all assessments performed by the same researcher. Participants were allocated to 3 groups: healthy controls (H group, n=28), urinary incontinence (UI group, n=22), and pelvic floor dyssynergia (DS group, n=18). Dyssynergia was confirmed by medical diagnosis using anal manometry. The UI group allocation was verified with the International Consultation on Incontinence Questionnaire-Short Form [20] (ICIQ-SF; the Polish version of the ICIQ-UI SF was used with permission from the ICIQ group), while the H group completed the Pelvic Floor Distress Inventory-20 (PFDI-20) [21] only women scoring “0” across all sections were included. All participants were additionally screened for factors potentially affecting PFM function, including parity, menopausal status, physical activity, age, and body mass index (BMI). These characteristics are shown in Table 1.

ELIGIBILITY CRITERIA:

The inclusion criteria were as follows: female sex, 18–80 years, ability to understand and follow instructions, and no neurological disorders affecting PFM.

Group-specific inclusion criteria were as follows: H group: PFDI-20=0 in all 3 sections; UI group: self-reported urinary incontinence, confirmed with the ICIQ-SF; and DS group: pre-existing diagnosis of dyssynergia (anal manometry by medical doctor).

The exclusion criteria were as follows: age >80 years, gynecological/pelvic surgery in past 10 years, lower extremity, pelvic, or spinal injuries affecting function, pelvic organ prolapse or third-degree incontinence, ongoing pelvic floor physiotherapy, pregnancy, medications affecting neuromuscular function, menstruation, acute infection, nickel allergy during assessment, pain during sEMG, and refusal to participate.

ETHICAL CONSIDERATIONS AND CONSENT:

All participants provided written informed consent prior to participation after receiving both verbal and written information about the study aims, procedures, potential discomforts, and their right to withdraw at any time without consequences. The study was approved by the Local Ethics Committee of the Wrocław Medical University (approval No. 145/2018) for the UI group, the Bioethics Committee of the Polish Chamber of Physiotherapists (approval No. 6/2022) for the DS group, and the Ethics Committee of the Opole Medical School (approval No. KB/205/FI2019) for the H group. All procedures were conducted in accordance with the Declaration of Helsinki. The project has been registered on the OSF.io platform as a secondary analysis: https://doi.org/10.17605/OSF.IO/5JDUT.

SEMG DATA ACQUISITION AND SIGNAL PROCESSING PROTOCOL:

sEMG signals were recorded using a high-resolution electromyography system (Noraxon Ultium USA, Inc, Scottsdale, AZ, USA), with a sampling rate of 1000 Hz and 16-bit resolution. The signal was acquired at 1500 Hz and processed using MyoResearch XP software v1.0. To remove movement artifacts and external noise, a band-pass filter (20–450 Hz) was applied. Data analysis included root mean square (RMS) values (μV), normalized to each participant’s maximum voluntary contraction (MVC) to account for inter-individual variability. Following the Surface Electromyography for the Non-Invasive Assessment of Muscles (SENIAM) [22] recommendations, self-adhesive bipolar Ag/AgCl electrodes were placed bilaterally on synergistic muscles (rectus abdominis, hip adductors, and gluteus maximus) to assess PFM co-activation. Prior to placement, the skin was cleaned with alcohol wipes to reduce impedance, and electrode spacing was standardized at 2 cm for consistency.

MEASUREMENT PROTOCOL:

Participants emptied their bladders before assessment. A wireless intravaginal probe (76 mm length, 28 mm diameter; Lifecare PR-02, Everyway Medical Instruments Co, Ltd, Taiwan) with 2 longitudinal metal plates (positioned at 3–9 o’clock) was used, connected to the Noraxon Ultium. The physiotherapist verified electrode placement. Measurements were performed in a supine position with slight hip and knee flexion. Participants were trained using sEMG biofeedback to ensure correct PFM contractions, avoiding engagement of abdominal, hip, or thigh muscles and preventing breath-holding. They were instructed to squeeze and lift the vaginal probe, simulating the action of holding urine or feces, with clear verbal commands to contract and relax fully. Five sEMG signal recordings were obtained, following the protocol proposed by Oleksy et al [3], which is based on the Glazer protocol [18], and included a series of functional PFM activities, as follows.

(1) A 60-second baseline rest period was taken prior to the start of the examination. (2) Five rapid (phasic/quick) contractions, each lasting 2 seconds, with 10-second rest intervals between them were performed. (3) Five tonic contractions, each held for 10 seconds, also with 10-second rest intervals, were performed; participants were instructed to contract their PFM as strongly as possible, maintain the contraction for the full 10 seconds, and then fully relax. (4) Experimentally, a post-tonic relaxation phase following each tonic contraction was introduced to assess the participants’ ability to return to baseline activity. This modification aligns with the rationale discussed by Bø [15], who emphasized the clinical relevance of distinguishing between contraction and relaxation capacity in PFM training – particularly in the context of dyssynergia and functional bladder outlet obstruction[15]. (5) One 60-second endurance contraction was performed, during which participants were asked to contract their PFM at a submaximal intensity that could be sustained throughout the entire duration without noticeable fluctuations in tension. (6) A 60-second post-endurance relaxation (post-baseline) was performed, during which participants were instructed to fully relax their PFM.

INVESTIGATED PARAMETERS:

The primary outcome measures included RMS values (μV) and the percentage of maximum voluntary contraction (%MVC) of sEMG signals recorded during various activities. The RMS value represents the square root of the mean square value estimated over a defined time window or epoch, providing an estimate of the signal’s amplitude variability. As RMS is the square root of the power of the signal, it serves as a reliable indicator of muscle activation [23]. For normalization, the highest sEMG amplitude recorded during 1 of 3 MVCs was used, following the SENIAM guidelines [22].

DATA ANALYSIS:

Normalized sEMG values (%MVC) could not be calculated for 7 participants from the DS group due to invalid or missing MVC recordings. These cases were excluded from analyses involving %MVC normalization but were retained in analyses based on raw RMS amplitude (μV). No missing data were identified in the RMS μV dataset; therefore, no imputation procedures were applied. To estimate the mean differences in sEMG recordings between the groups, we used Bayesian multivariate robust linear regression, with all 5 PFM activities treated as a multivariate dependent variable and the group index as categorical predictor. In Bayesian estimation, the uncertainty of estimates is a direct function of the number of observations; thus, the small sample size in our study is reflected in the width of the posterior probability of the mean differences. Furthermore, the multivariate model naturally accounts for the interdependencies between the measurements, and the robust component (ie, the t-distribution) makes the posterior estimates of means insensitive to outliers. Bayesian estimation results in posterior distributions of parameters (eg, mean microvolts value in each group); to summarize the distributions, we report the medians of the posterior with 95% Bayesian confidence intervals (BCI). For pairwise comparisons between groups, we used the 95% BCI of the posterior difference between mean sEMG values. When the 95% BCI excluded zero, we inferred that the data provides evidence for a difference between the groups. Full model outputs are available in Tables 2 and 3. Data processing, figures, and analyses were conducted using R 4.4.3 [24], with the brms [25] package used to estimate the models and ggplot2 [26] package used to create the figures.

Results

GROUP COMPARISONS WITH RESPECT TO MICROVOLTS:

RMS values, expressed in microvolts (μV) and thus non-normalized, are presented in Figure 2.

60-SECOND BASELINE REST:

No significant differences in sEMG signals were found among the groups after adjusting for BMI and age. All pairwise differences included zero: DS group vs H group (Δ=0.26, 95% BCI: [−0.81, 1.32]), DS group vs UI group(Δ=0.17, 95% BCI: [−0.79, 1.21]), and H group vs UI group (Δ=−0.07, 95% BCI: [−0.91, 0.82]).

RAPID CONTRACTIONS:

PFM sEMG signals during rapid contractions showed no statistically significant differences between groups. All pairwise differences included zero: DS group vs H group (Δ=3.07, 95% BCI: [−0.06, 6.09]), DS group vs UI group (Δ=2.15, 95% BCI: [−0.54, 4.81]), and H group vs UI group (Δ=−0.90, 95% BCI: [−3.53, 1.67]).

TONIC CONTRACTIONS:

Significant group differences were found in sEMG amplitudes during tonic contractions. The highest activity was found in the DS group (=11.23, 95% BCI: [8.74, 13.59]), followed by the UI group (=7.79, 95% BCI: [6.06, 9.59]) and the H group (=6.99, 95% BCI: [5.19, 8.73]). Pairwise comparisons indicated that the DS group had significantly higher sEMG signals than the H group (4.20, 95% BCI: [0.99, 7.29]) and UI group (3.44, 95% BCI: [0.48, 6.20]). In contrast, the difference between the H and UI groups was not statistically meaningful (−0.78, 95% BCI: [−3.53, 1.87]), as the 95% credible interval included zero.

POST-TONIC RELAXATION:

Following tonic effort, significant differences were found in resting sEMG signals. The DS group exhibited higher post-contraction values (=3.59, 95% BCI: [2.76, 4.45]) than the UI group (=3.0, 95% BCI: [2.38, 3.66]) and H group (=2.46, 95% BCI: [1.98, 2,97]). The difference between the DS group and H group reached statistical significance (Δ=1.13, 95% BCI: [0.10, 2.15]), while differences between the DS group and UI group (Δ=0.58, 95% BCI: [−0.42, 1.57]) and UI group and H group (Δ=−0.54, 95% BCI: [−1.40, 0.35]) were not significant. Age was negatively associated with post-contraction activity (=−0.74, 95% BCI: [−1.13, −0.33]). BMI showed no meaningful association.

ENDURANCE CONTRACTION (60-SECOND HOLD):

While the DS group again showed the highest sEMG values (=9.48, 95% BCI: [7.39, 11.44]), followed by the UI group and H group, none of the between-group differences were statistically significant (DS vs H: Δ=2.73, 95% BCI: [−0.03, 5.26]; DS vs UI: Δ=1.73, 95% BCI: [−0.73, 3.97]; H vs UI: Δ=−1.00, 95% BCI: [−3.05, 1.14]). Results suggest a trend but with considerable uncertainty.

POST-ENDURANCE RELAXATION (REST AFTER 60 SECONDS):

Group differences in sEMG signals following in the 60-second post-endurance contraction relaxation did not reach statistical significance. All pairwise comparisons included zero (DS vs H: Δ=0.51, 95% BCI: [−0.51, 1.58]; DS vs UI: Δ=−0.14, 95% BCI: [−1.16, 0.87]; H vs UI: Δ=−0.65, 95% BCI: [−1.55, 0.31]). Age again showed a modest but significant negative association with resting activity (=−0.48, 95% BCI: [−0.91, −0.06]). BMI was not meaningfully associated (=−0.15, 95% BCI: [−0.57, 0.28]).

GROUP COMPARISONS WITH RESPECT TO %MVC:

sEMG amplitude, presented in %MVC (normalized), are illustrated in Figure 3.

60-SECOND BASELINE REST:

At the 60-second baseline rest period, none of the pairwise comparisons reached statistical significance, with all credible intervals including zero: DS group vs H group (Δ=−7.1, 95% BCI: [−16.57, 2.35]), DS group vs UI group (Δ=−4.50, 95% BCI: [−14.60, 6.27]), and H group vs UI group (Δ=2.67, 95% BCI: [−6.25, 11.80]).

RAPID CONTRACTIONS:

The highest estimated value was observed in the H group (%MVC=58.13, 95% BCI: [51.96 to 64.83]), followed by the UI group (%MVC=52.62, 95% BCI: [46.92 to 58.80]), and the DS group (%MVC=47.71, 95% BCI: [41.87 to 53.52]). Neither age (%MVC=3.02, 95% CI: −1.15 to 7.33) nor BMI (%MVC=0.075, 95% CI: −3.89 to 4.00) had a statistically significant effect on the parameter, as the 95% credible intervals for both variables included zero. The DS group vs H group comparison was statistically significant (Δ=−10.49, 95% BCI: [−19.46, −1.86]), indicating lower rapid contraction activity in the DS group was comparable to that of the H group. Comparisons between the DS group and UI group, and H group and UI group were non-significant.

TONIC CONTRACTIONS:

The H group had the highest estimated value for the tonic contraction (%MVC=62.18, 95% BCI: [57.22, 67.12]), indicating superior endurance capacity during sustained pelvic floor contractions. Lower values were observed in the DS group (%MVC=54.81, 95% BCI: 46.39, 62.95) and UI group (%MVC=51.83, 95% CI: 46.19 to 57.31). Pairwise comparisons revealed a statistically significant difference between the H and UI groups (Δ=10.36, 95% BCI: [2.67, 18.48]), indicating better tonic contraction performance in healthy participants. The difference between the DS and H groups (Δ=−7.41, 95% BCI: [−17.11 to 2.04]) was not statistically significant. Similarly, the comparison between the DS and UI groups (Δ=2.87, 95% BCI: [−7.01 to 13.26]) showed no significant difference.

POST-TONIC RELAXATION:

The relaxation phase following each tonic contraction showed no statistically significant differences, with pairwise comparisons all including zero within credible intervals.

ENDURANCE CONTRACTION (60-SECOND HOLD):

During the 60-second endurance contraction, the H group once more exhibited the highest activity, significantly exceeding that of the DS group, as indicated by a meaningful pairwise difference (Δ=−11.04, 95% BCI: [−22.69, −0.15]).

POST-ENDURANCE RELAXATION (REST AFTER 60 SECONDS):

During the 60-second post-endurance contraction rest period, sEMG activity did not significantly differ between groups. Age and BMI adjustments showed no consistent significant associations across conditions, as all credible intervals overlapped with zero.

Discussion

INTERPRETATION OF RESULTS EXPRESSED IN MICROVOLTS (μV):

Despite differences in the participants age and BMI by group (Figure 1), sEMG analysis of PFM revealed no statistically significant differences in EMG PFM signals at baseline rest or following endurance contraction among the 3 groups, although higher variability was observed within the DS group. However, muscle activity during post-tonic relaxation demonstrated significant differences among groups, with higher values and greater variability observed in the DS group. These results may suggest disturbances in PFM relaxation mechanisms within the DS group, which could have significant clinical implications. In the context of the review by Bø et al [15], the absence of clear differences in resting muscle tone across various pelvic floor dysfunctions aligns with the existing research issue of inadequate standardization and unclear definitions of so-called PFM overactivity assessed via sEMG. Moreover, the review of Bø et al [15] highlights a scarcity of studies explicitly evaluating the ability of the PFM to relax after contraction – an aspect addressed in our study through the analysis of post-tonic relaxation (which showed a significant difference between the DS and H groups) and 60-second post-endurance relaxation (which showed no differences between groups). Although evidence suggests that PFM contractions can enhance muscle relaxation, the underlying mechanisms remain unclear. Our findings, highlighting relaxation differences between the DS and H groups, contribute significantly to future research in this area. The hypothesis concerning the UI group was only partially confirmed. While participants with urinary incontinence demonstrated reduced performance in certain parameters, their sEMG signals did not consistently differ from those of the H group. In some aspects – such as baseline rest activity or relaxation – their patterns more closely resembled those observed in the DS group. A limitation in our study is that the ICIQ-SF does not differentiate between subtypes of urinary incontinence. Espuña-Pons et al [27] demonstrated that while the ICIQ-SF is a reliable screening tool, additional procedures, such as a stress test or voiding diary, are required to distinguish between stress, urgency, and mixed urinary incontinence. Consequently, in our study, the UI group should be interpreted as women reporting urinary incontinence symptoms in general, rather than exclusively stress urinary incontinence. This distinction may have influenced the results and needs to be considered in their interpretation, as the observed variability may reflect the heterogeneity of PFM function in individuals with urinary incontinence, who can present a spectrum of impairments ranging from underactivity to compensatory overactivation. This issue has also been highlighted in previous research. For instance, in a study by Ptaszkowski et al [11], lower PFM EMG activity was associated with stress urinary incontinence, particularly in postmenopausal women. The authors concluded that resting sEMG values might serve as diagnostic indicators, with lower resting PFM sEMG signals suggesting an increased risk of stress urinary incontinence [11]. This variability in sEMG patterns is not limited to incontinence but has also been reported in other pelvic floor disorders, including pain syndromes and dyssynergia, in which distinct alterations in resting activity and contractility have been described. These findings resemble observations made in women with pain syndromes, such as interstitial cystitis/bladder pain syndrome, in which altered neural activity and reduced motor unit action potential amplitudes have been reported, suggesting neuromuscular control disturbances and structural changes within the pelvic floor musculature [28]. Additionally, women with provoked vestibulodynia have been shown to exhibit elevated resting PFM sEMG activity and altered contractility, including increased tissue stiffness and baseline muscle tone, in comparison to asymptomatic women. Such findings may hold clinical relevance in the development of personalized physiotherapeutic strategies based on sEMG profiling [29]. Conversely, analysis of resting sEMG activity in patients with pelvic floor dyssynergia has demonstrated higher sEMG values compared with those in healthy individuals [13], a pattern that was not as clearly observed in our data. Our results on resting PFM sEMG (approximately 5 μV) are consistent with the results of Zhang et al [30], who reported similar values in stress urinary incontinence and non–stress urinary incontinence groups, with no significant differences between populations. This suggests that resting activity may not be a reliable discriminator, a conclusion supported by a systematic review by Audrain et al [31], which found that although sEMG is widely used to assess strength, endurance, tone, and coordination, its utility in distinguishing continent from incontinent women remains inconclusive due to methodological heterogeneity. The review underscores the need for standardized protocols and a more comprehensive approach that integrates biopsychosocial factors, as sEMG alone does not directly quantify functional contribution of the PFM to continence mechanisms. Resting sEMG values in our study were comparable to those reported by Muallem et al [9], who proposed a 3-μV threshold for “pelvic floor hypertension”. However, elevated values also occurred in women without pain symptoms, suggesting that such cut-offs may be arbitrary and not specific to pain syndromes. This suggests that elevated resting PFM sEMG activity may not be specific to pain syndromes, and that the 3-μV threshold may be too arbitrary for diagnostic use. Further validation is warranted in larger cohorts, taking into account the type of dysfunction, clinical symptoms, and the specific muscles contributing to resting tone [4]. Similarly, the 2-μV upper limit suggested by Voorham-van der Zalm et al [32] highlights the lack of consensus and need for further validation in larger cohorts. However, the lack of standardized reference values for sEMG hampers the interpretation of findings and the classification of “overactivity” or “hypertonicity”, as results are influenced by factors such as body position, electrode type, and signal processing [29]. In our study, resting sEMG values aligned with those reported by Muallem et al [9], but did not clearly indicate overactivity in the DS group. Instead, dysfunctions appeared more related to impaired neuromuscular control and coordination. Greater variability in the rest period after the 60-second contraction suggested difficulties in regulating tone and achieving relaxation, resembling patterns described in pain syndromes. Women in the DS group also showed elevated post-tonic sEMG activity compared with the H group, indicating impaired deactivation of muscle excitation. Although it was less evident after endurance contraction, variability again suggested unstable relaxation, consistent with the view that relaxation depends on active central inhibitory mechanisms. As demonstrated by Kato et al [33] and De Havas et al [34], cortical structures such as the primary motor cortex and supplementary motor area, together with short-interval intracortical inhibition and “active braking”, play a key role in terminating contraction. Dysfunction of these pathways may contribute to prolonged post-contraction sEMG activity in dyssynergia. Clinically, this inability to efficiently relax the PFM represents a characteristic feature of the dyssynergia phenotype and supports recommendations by Bø et al [15] that therapeutic strategies should emphasize training of relaxation and coordination rather than strengthening alone.

Electromyographic signals during rapid (quick flicks) and tonic contractions showed significant differences between groups, with the DS group exhibiting higher mean amplitudes and greater variability, compared with the H group. These findings suggest compensatory mechanisms or neuromuscular discoordination, particularly during phasic and endurance contractions. Elevated sEMG signals in the DS group, compared with both the H and UI groups, may reflect abnormal motor unit recruitment or impaired ability to relax PFM after contraction, while the lack of significant differences between the H and UI groups indicates that impaired phasic activation is more specific to dyssynergia. The lack of significant differences between the H and UI groups suggests that impaired phasic activation patterns might be more specific to dyssynergia than to urinary incontinence. In contrast, a study [35] comparing women with polycystic ovary syndrome and healthy controls found no significant differences in sEMG amplitudes during quick flicks (P=0.62), 8-second MVC (P=0.74), or 60-second MVC (P=0.68). This discrepancy may stem from differences in participant populations, in which polycystic ovary syndrome–related hormonal alterations did not appear to substantially affect PFM contractions, whereas in the present study, neuromuscular dysfunction in the DS group likely contributed to altered activation patterns. It is important to note that higher sEMG activity does not necessarily reflect superior muscle function. In the context of pelvic floor dysfunction, elevated sEMG signals may indicate a less efficient or compensatory contraction. If the muscles are weakened or unable to sustain contraction over time, a greater number of motor units may be recruited, increasing the recorded sEMG activity without translating into functional muscular endurance. Therefore, the higher sEMG activity observed in the DS group during contractions of varying duration in our study may reflect a disturbed muscle activation strategy, which is clinically relevant for pelvic floor function assessment and physiotherapeutic strategy. The relationship between sEMG signal amplitude and muscle fatigue has been confirmed in previous studies [36] that demonstrated sEMG amplitude gradually increases with progressive muscle fatigue. This phenomenon results from the need to recruit additional motor units and increase their firing rate to maintain the required contraction force. Muscle fatigue leads to a reduced ability of muscle fibers to generate force, even under continued motor neuron excitation [37]. For instance, a study showed that women with sexual dysfunction exhibited higher peak sEMG amplitudes during rapid contractions (quick flicks) compared with healthy controls, suggesting a greater susceptibility of their PFMs to fatigue [37]. As noted by the Consensus for Experimental Design in Electromyography publication [23], muscle fatigue alters the relationship between sEMG and force, with reduced conduction velocity and motor unit action potential changes leading to higher amplitudes even when strength does not increase. In our study, elevated sEMG signals in the DS group may therefore reflect muscle fatigue and altered recruitment strategies, with important implications for functional PFM assessment.

INTERPRETATION OF RESULTS EXPRESSED IN %MVC:

There are several important considerations when interpreting sEMG signals. Baseline and contractile sEMG amplitudes are influenced by electrode characteristics, electrode configuration, recording systems, and individual patient-specific factors. Raw amplitude values cannot be directly compared between individuals, due to variability in factors such as cutaneous or mucosal tissue thickness, vaginal lubrication, the positioning and orientation of the electrodes relative to the muscle and its fibers, and the technical specifications of the detection system. Consequently, normalization of sEMG amplitude is considered essential for inter-individual comparisons [12]. Nevertheless, as indicated in the literature, a critical requirement for the effective use of sEMG in clinical practice is the establishment of a reliable reference value – most commonly the MVC [38]. In the present study, values expressed in microvolts, particularly during tasks such as the tonic contractions and the endurance contraction, were higher in the DS group than in the H group. However, once normalized to %MVC, the DS group demonstrated relatively lower values, often falling below those of the H group. A similar pattern was also observed in the UI group. This discrepancy may stem from an underestimation of the true MVC, suggesting that standard protocols (eg, standard verbal encouragement) may not fully reflect the actual maximum contraction capacity of the PFMs [38]. In light of our findings, alternative strategies for eliciting MVC, such as the Reinforcement-based Maximum Voluntary Isometric Contraction protocol [39], may provide more accurate assessment. This approach, shown to yield higher sEMG amplitudes than does standard verbal encouragement, could be particularly relevant for populations with impaired coordination, such as the DS group, in which high absolute activity but low %MVC suggested that traditional MVC assessment may have underestimated true muscle capacity. Overall, these findings provide a more nuanced understanding of PFM neuromuscular behavior across clinical populations. Considering both %MVC and microvolt values emphasizes not only the magnitude of muscle activity but also its functional efficiency. Since raw sEMG amplitudes are affected by technical and anatomical factors, normalization – typically to MVC – is essential for meaningful inter-individual comparisons [12].

In our study, normalization influenced the interpretation of group differences. No significant differences were found in sEMG signals at rest or during most contractions once the data were normalized by MVC. However, a significant difference in rapid contractions persisted between the H and DS groups, with the H group demonstrating higher normalized values (95% BCI excluded 0). This suggests that the ability to generate rapid, phasic contractions may remain a distinguishing functional feature between these populations, even after accounting for individual variability in maximal capacity. While normalized sEMG data (eg, %MVC) offer practical advantages – especially during active tasks – caution is warranted in their interpretation. Without accurate and meaningful determination of MVC, %MVC values may lead to misleading conclusions, particularly in patient groups with pelvic floor dysfunction. In our study, the divergence between raw microvolt amplitudes and %MVC values – especially in the DS group – underscores how overactivity at rest or during suboptimal contraction does not necessarily reflect functional strength or coordination. This supports the interpretation that normalized values are more informative in assessing true neuromuscular efficiency, especially when distinguishing between compensatory overactivation and effective muscle recruitment. Previous research has shown that normalized and non-normalized sEMG amplitudes tend to be uncorrelated at rest but demonstrate moderate correlation during maximal effort tasks, suggesting that normalization yields more reliable and comparable information under load [10]. Therefore, future studies should consider implementing more comprehensive MVC protocols (eg, repeated maximal voluntary isometric contractions), which account for motivational and task-specific factors. Such approaches could improve both clinical and research validity when assessing PFM function [39].

The findings of this study highlight the importance of assessing not only strength but also relaxation and coordination of the PFMs in clinical practice. Elevated and variable post-contraction sEMG activity in women with dyssynergia indicates impaired relaxation mechanisms, suggesting that therapeutic strategies should emphasize relaxation training, proprioceptive feedback, and motor control rather than strengthening alone. Resting sEMG values proved insufficient to distinguish between dysfunctions, underscoring the need to interpret baseline activity in combination with functional tasks. The heterogeneity observed in the UI group further supports the necessity of additional diagnostic procedures to differentiate urinary incontinence subtypes and tailor individualized rehabilitation programs. Finally, differences between raw amplitudes and normalized %MVC values emphasize the need for accurate MVC assessment protocols to avoid misinterpretation of muscle function and to inform evidence-based physiotherapeutic interventions.

STUDY LIMITATIONS:

This study has several important limitations. First, the retrospective design introduces the possibility of biases that may be difficult or impossible to correct post hoc. Additionally, selection bias may limit the generalizability of the findings to broader populations. Control over variables that could have influenced PFM sEMG signals – such as menopausal status, estrogen levels, quality of life, or physical activity – was limited or incomplete, which may have affected the results. The H group was significantly younger and had a lower BMI than the DS and UI groups. Although age and BMI were statistically controlled as covariates, their group-level differences remained significant and could have influenced sEMG outcomes. The extent to which this covariate adjustment mitigated confounding depends on the strength of correlations between these factors and outcome measures; thus, residual confounding cannot be ruled out. This limits the ability to draw definitive conclusions regarding the direct effect of clinical group affiliation on PFM sEMG activity. Furthermore, this study focused solely on objective electrophysiological parameters, without incorporating subjective clinical indicators. Future research should consider combining sEMG with patient-reported outcome measures. The integration of subjective and objective data may offer a more comprehensive understanding of PFM dysfunction and enhance the clinical relevance of sEMG-based assessments.

STRENGTHS:

Despite its limitations, this study offers several methodological and conceptual strengths. It is among the few studies to comprehensively evaluate PFM activity with sEMG across multiple standardized tasks using the Glazer protocol. The combined use of RMS values and normalized (%MVC) data allowed for a more precise assessment of neuromuscular efficiency. Applying Bayesian multivariate regression with adjustment for age and BMI strengthened the robustness of the analysis. Finally, the inclusion of 3 distinct groups – healthy women, those with urinary incontinence, and those with pelvic floor dyssynergia – enabled subgroup comparisons and provided a broader perspective on PFM dysfunction.

Conclusions

Women in the DS group exhibited higher absolute RMS amplitudes in tonic contractions and impaired post-tonic relaxation, yet reduced recruitment efficiency when normalized to %MVC, indicating paradoxical rather than truly overactive muscle activity. The UI group showed heterogeneous patterns, underscoring the multifactorial nature of this condition and highlighting the need for future studies on more homogeneous incontinence subtypes to refine clinical interpretation based on sEMG signals. These findings emphasize that reliance on raw RMS alone may lead to misclassification and support the use of a multiparametric assessment integrating %MVC, relaxation metrics, standardized protocols, and patient-reported outcome measures to better guide clinical decision-making. Validation in larger prospective cohorts is warranted.