Reliability of electromyography detection systems for the pelvic floor muscles

Abstract

The purpose of this study was to investigate the reliability of three different electromyography (EMG) detection systems commonly used to acquire EMG data from the pelvic floor muscles (PFM) at rest, during maximum voluntary contractions (MVCs) and during a coughing task. Twelve nulliparous women between the ages of 24 and 40 participated in the study. EMG data were recorded from each side of the pelvic floor using surface (Femiscan™ and Periform™ vaginal probes), and fine-wire electrodes while subjects performed three repetitions of each task in supine and in standing. RMS amplitudes of baseline and peak PFM activity were computed from the data acquired during the MVC and the coughing tasks. The peak RMS amplitudes from the cough data were also normalized to each subject’s MVC and report as a percent of their maximum voluntary electrical activation (% MVE). Signal to noise ratio (SNR) was calculated for each task. Comparisons were made between the RMS amplitudes and SNR recorded from each side of the PFM. Between-trial and between-day reliability was determined using a variety of measures including intraclass correlation coefficients (ICC), coefficients of variation (CV) and mean absolute difference (MAD). The reliability of the different devices was compared using the MAD normalized to average signal amplitude (nMAD). The results indicated that the EMG data recorded from the right and left sides of the PFM were different, therefore the EMG data acquired from each side of the pelvic floor were analyzed separately. Between-trial reliability assessed by ICC was good for all the devices (left and right average ICC(3,1) = 0.80 – 0.96); the CVs supported these findings (average CV = 12.4 – 17.1%) Between-day reliability was poor and inconsistent across all devices as tasks. When the cough data were normalized however, the between-day RMS amplitudes were very consistent (79.5 – 90.2% MVE). Each device recorded consistent activation amplitudes within a given day. Between-day reliability results indicate that EMG data recorded on separate days with these instruments should not be compared unless data can be normalized. Clinicians and researchers are cautioned to normalize their data if day-to-day comparisons in EMG amplitude are to be made using data recorded with these devices.