Due to the influence of methodological factors, the reliability of the sEMG analysis needs to be determined for adequate clinical use and interpretation. The aim of the study was to investigate the test-retest and interrater reliability of the surface electromyography (sEMG) of the female pelvic floor muscles (PFM) through visual determination of the onset and offset of the maximum voluntary contraction (MVC) of female MAP.

Study of test-retest and interrater reliability of sEMG during maximal voluntary contraction (MVC) of female PFM, using vaginal probe. Three repetitions of the MVC were requested, recorded for 5 seconds (s) each, separated by 60s of resting baseline. Test-retest and interrater reliability were evaluated by two independent raters through visual inspection, considering for statistical analysis the variables: integral and median frequency (MDF) of electromyographic signals. The visual determination of the onset and offset of each MVC was analyzed through the Intraclass Correlation Coefficient (ICC2,1, ICC2,3), analysis of variance with repeated measures (ANOVA) and effect size), the agreement of the qualitative visual analysis was classified as “high quality”, “moderate quality” or “low quality” was analyzed through Cohen's k coefficient and quantitative analysis of the signals was analyzed through the signal noise-ratio - SNR: 10log10(Icontração/Ibaseline), using R Project version 4.0.2, IC 95%.

68 women were selected to perform the sEMG examination of the PFM, of which 10 were excluded (05 due to the impossibility of carrying out the sEMG with a vaginal probe, and 05 excluded during the analysis of the signs due to the impossibility of identifying the MVC onset and offset), resulting in 58 women, with an average age of 54.68 (± 10.72). Of the total, 22 (37.9%) data were considered of high quality by the two raters, while only 5 (8.6%) were considered of low quality; the cross analysis between raters showed a fair agreement (Cohen’s κ coefficient = 0.306 [0.148; 0.463]). It was also observed that the SRN decreases with the signal quality category, from 34.3-32.9 dB to 27.2-22.2 dB for raters 1-2 respectively. Regarding test-retest reliability, MDF showed excellent values of ICC2,1 (0.963) and ICC2,3 (0.987) higher than the integral, which showed reliability from acceptable to excellent (ICC2,1: 0.675; ICC2,3 0.861). Both variables investigated here exhibited higher values of ICC2,3 than ICC2,1, with estimates ranging from excellent to acceptable, table 1. For interrater reliability, no statistical evidence of difference between the 3 repetitions was observed for both the integral (ω2 = -0.006, p = 0.558) and the MDF (ω2 = -0.009, p = 0.992). An excellent reliability was observed for both variables, being higher for MDF (ICC2,1 = 0.997 [0.996; 0.998]) than for the integral (ICC2,1 = 0.793 [0.697; 0.861), and higher values of I ICC2,3 than ICC2,1, table 2.

Our main results show reliability ranging from acceptable to excellent, both for single or aggregate measures (average of 3 repetitions) of integral and MDF during PFM MVC in women through the method of visual determination of the onset and offset of a contraction. Our findings also show that the use of only one contraction of the three is acceptable, so that both the average of the three contractions and the use of a contraction can be used to assess the PFM myoelectric activity. Based on our findings, to assist in the evaluation of signal quality, quantitative tools should be used instead of qualitative evaluation due to fair agreement between raters.

Qualitative visual analysis of the EMG signal is not recommended because it presents a fair agreement interrater, although it reflects the signal-to-noise ratio. Although the method of visual practice of the onset and offset of contraction is subjective, it is a practical method for use in clinical practice. Based on our findings can be recommended the use the visual determination of the PFM MVC onset and offset using the standard interface of the sEMG software for excellent reliability.