Motor unit recruitment behavior of continent and incontinent women’s pelvic floor muscles while running: A wavelet approach

Koenig I1, Eichelberger P2, Moser H1, Leitner M2, Kuhn A3, Taeymans J1, Radlinger L2

Research Type


Abstract Category

Research Methods / Techniques

Abstract 320
Pelvic Floor Muscle Assessment and Treatment
Scientific Podium Short Oral Session 20
Thursday 30th August 2018
12:50 - 12:57
Hall C
Biomechanics Female Outcomes Research Methods Physiotherapy Stress Urinary Incontinence
1. Bern University of Applied Sciences, Department of Health Professions, Division of Physiotherapy, Bern, Switzerland; Vrije Universiteit Brussel, Faculty of Physical Education and Physiotherapy, Brussels, Belgium, 2. Bern University of Applied Sciences, Department of Health Professions, Division of Physiotherapy, Bern, Switzerland, 3. University Hospital and University of Bern, Women’s Hospital, Urogynaecology, Bern, Switzerland

Irene Koenig



Hypothesis / aims of study
High impact activities such as running require involuntary, reflexive pelvic floor muscle (PFM) contractions to avoid urinary leakage. Structural and histochemical characteristics showed that the levator ani muscle is composed of 54.9-70.3% slow type I and 29.7-45.1% fast type II fibers in healthy women. The predominance of type I fibers suggests that the levator ani muscle is suited to maintain pelvic floor organs in an optimal position. However, the high proportion of the fast type II fibers refers also to rapid PFM contractions. Additionally, a significantly smaller diameter of the levator ani muscle fibers was found and the percentage of type II fibers was decreased in women with stress urinay incontinence (SUI) [1]. Therefore, SUI is structurally and histochemically related to the proportion of type I and type II fibers. However, the knowledge about fiber recruitment behavior of PFM during high impact sport activities is still inadequate. It was shown that continent women produce faster, however, voluntary contractions than incontinent women. While running, a rapid increase to maximum PFM electromyography (EMG) activity within the first milliseconds after heel strike was observed in young continent women.
Wavelet analysis of EMG signals allows to extract EMG frequencies among time with fine time resolution. Therefore, task-specific muscular activation in terms of frequency and magnitude during short time periods, like for example the pre-initial contact phase of the gait cycle, can be extracted with wavelets. During such a muscular event, muscle fiber groups can selectively be activated. These groups generate a specific EMG spectrum [2]. EMG spectra are associated with the conduction velocity of the motor unit action potential. Hence, fiber type recruitment can be estimated with wavelets since slow type I fibers generate lower frequencies in the signal and fast type II fibers higher frequencies. In this study it was hypothesized that different PFM fiber types are activated in the pre- and post initial contact phase during running. The aim of the study was to illustrate spectral changes of the pre- and post-initial contact phase during running at three different speeds and to compare continent and incontinent women with a wavelet approach.
Study design, materials and methods
This cross-sectional study included 28 continent (CON) women, age: mean 38.9 (SD:10.3) years, Oxford grading [scale: 0-5]: median 5 (IQR:1) and ICIQ UI-sf [questionnaire to screen level and impact of incontinence; scale: 0-21]: median 0 (IQR:0) and 21 women with SUI, age: mean 46.1(SD:9.9) years, Oxford grading: median 5(IQR:1), ICIQ UI-sf: median 6 (IQR:2) and self-reported SUI. PFM EMG was recorded during 10 s at 7, 11 and 15 km/h treadmill running. The raw EMG signals were 20 Hz high-pass and 500 Hz low-pass filtered. PFM EMG was analyzed with a continuous wavelet transform using Morse wavelets. Wavelet patterns were normalized to the total power (100% = sum of all intensities) of the signal within the frequency band 20-500 Hz. The relative distribution of power (%) over the six selected frequency bands of interest (Table 1) were extracted within six time intervals of 30 ms, from -30 ms before to 150 ms after initial contact [3], (Figure 1). The frequency band between 0 and 20 Hz was not analyzed due to possible movement artifacts. An analysis of variance for repeated measures was performed to identify power spectra differences within and between the two groups. Additionally, a post-hoc t-test was used. The significance level was set at p ≤ 0.05.
Group differences:
The power spectra of each time interval showed no statistically significant group differences. Nevertheless, wavelet patterns of women with SUI showed a systematic trend towards more intensity in the frequency band between 20 and 50 Hz but less intensity between 80 and 200 Hz than continent women.
Pre- and post-initial contact spectral differences:
EMG intensity was significantly lower in the 20-50 Hz frequency band before initial contact than in the post-initial contact phase for all running speeds (group CON) and also for 7 and 11 km/h in group SUI. Group SUI showed no difference in this frequency band between the pre-initial contact phase and the first 60 ms after initial contact during running at 15 km/h. However, running at 15 km/h showed significantly higher intensities before initial contact in the 20-50 Hz frequency band than 60 to 150 ms after initial contact. Group CON: Intensities between 110 and 200 Hz were higher before initial contact for all running speeds. Group SUI: During running at 7 km/h there was more intensity before initial contact than 60-120 ms after initial-contact (110-200 Hz). At 11 km/h more intensity was found before initial contact than 0-150 ms after initial contact (80-200 Hz). During running at 15 km/h higher intensities were found in the pre-initial contact phase than 60-120 ms after initial contact (140-200 Hz), and furthermore, significantly higher intensities were found between 0 and 30 ms than between 30 and 150 ms after initial contact (110-140 Hz).
Interpretation of results
Group differences:
Although there was no significant group difference, a trend towards lower frequencies in women with SUI was found. Women with SUI have been demonstrated to present relatively less intensity in high frequency bands and more in the lower frequencies. These findings could be related to the structurally and histochemically change of the proportion of type I and type II fibers in women with SUI. Therefore, a possible explanation could be a selective atrophy of type II muscle fibers, as previously investigated in a histological study [1]. However, the histological study included only a small sample. In this study, group SUI had a low ICIQ UIsf score and did not differ from group CON in terms of the Oxford grade. This might be an explanation that no significant group differences were found.
Pre- and post-initial contact spectral differences:
The evaluated specific differences of EMG spectra in the pre-initial and the post-initial contact phase seem to indicate two different muscle activation events. Power spectra shifts towards higher frequency bands in the pre-initial contact phase could indicate a feed-forward anticipation and a muscle tuning for the expected impact of initial contact event in order to maintain continence. However, higher intensities in high frequencies during preparation and adaptation were more consistently in continent women than in women with SUI. Especially when running at 15 km/h, highest intensity production of the frequency band between 110 and 140 Hz was in the time window from initial contact to 30 ms after initial contact (group SUI). These findings could refer to a delayed PFM activation and could indicate an inconsistency in muscle preparation.
Concluding message
Since running is an activity of daily life, the often-related urinary leakage can cause a reduction in the quality of life of many women. Therefore, a deeper insight into continence mechanisms while running is needed. In this study, differences of pre- and post-initial contact activation behavior of PFM during different running speeds as well as spectral changes towards high or low frequencies could be extracted with wavelets. The potential of wavelet analyses in the field of PFM research should be used in further studies.
Figure 1
Figure 2
  1. Zhu (2005), Int Urogynecol J
  2. von Tscharner (2006), J Electromyogr Kinesiol
  3. Fleischmann et al. (2011), Appl Physiol
Funding None Clinical Trial No Subjects Human Ethics Committee The study was approved by the ethics committee of the Canton of Bern, Switzerland (No 391/14) Helsinki Yes Informed Consent Yes