The influence of contraction type, prior performance of a maximal voluntary contraction and measurement duration on fine-wire EMG amplitude

https://doi.org/10.1016/j.jelekin.2021.102566Get rights and content

Abstract

We aimed to investigate the impact of time on fine-wire (fw) electromyography (EMG) signal amplitude, and to determine whether any attenuation is confounded by task type. Twenty healthy participants were instrumented with fw and surface (s) EMG electrodes at the biceps brachii bilaterally. Participants held a weight statically with one arm and with the other arm either repeated the same task following a maximum voluntary contraction (MVC) or repeated dynamic elbow flexion/extension contractions. Each task was repeated for 30 s every five minutes over two hours. EMG amplitude was smoothed and normalized to time = 0. Stable median power frequency of the s-EMG ruled out the confounding influence of fatigue. Repeated-measures ANCOVAs determined the effect of electrode type and time (covariate) on EMG amplitude and the confounding impact of task type. During the isometric protocol, fw-EMG amplitude reduced over time (p = 0.002), while s-EMG amplitude (p = 0.895) and MPF (p > 0.05) did not change. Fw-EMG amplitude attenuated faster during the dynamic than the isometric protocol (p = 0.008) and there was evidence that the MVC preceding the isometric protocol impacted the rate of decline (p = 0.001). We conclude that systematic signal attenuation of fw-EMG occurs over time and is more pronounced during dynamic tasks.

Introduction

In kinesiological studies, electromyography (EMG) can be recorded using sensors located on the skin overlying the muscle of interest (i.e. surface EMG; s-EMG) or with electrodes inserted directly into the muscle. The resulting interference EMG signal represents the summation of the voltage potentials generated by the active motor units within the pick-up area of the electrode, and can be used to infer global activation of a muscle using features such as root mean square (RMS) amplitude (Besomi et al., 2019, Del Vecchio et al., 2017). Both s-EMG and intramuscular EMG recording techniques have limitations and the choice of electrode most appropriate to use will depend on the application. Recording s-EMG is considered valid for large and superficial muscles, while it is necessary to insert an electrode directly into the muscle to record with specificity from deep or small muscles that are located in close proximity to other muscles. When studying human movement, bipolar fine-wire (fw) electrodes are typically used for intramuscular recordings, as they are flexible and remain fixed within in the muscle while allowing for pain-free movement after any initial insertional pain has subsided. Unlike s-EMG, where electrodes are adhered to the skin surface, fw-EMG is invasive, and it requires additional training that is not widely available. Relative to s-EMG, fw-EMG has less potential for crosstalk, but may be less representative of overall muscle activity due to recording from a smaller number of active motor units from a more localized region within the muscle (Besomi et al., 2019).

The quantification of EMG amplitude does not directly reflect the extent of neural drive to a particular muscle, since neural drive impacts both motor unit recruitment and rate modulation (Besomi et al., 2019, Del Vecchio et al., 2017). Fatigue and motor-unit synchronisation can also lead to increases in RMS amplitude (Dideriksen et al., 2010, Lippold et al., 1960, Merletti et al., 1990, Yao et al., 2000). The recorded EMG signal will further depend on factors unrelated to muscle activation level, such as changes in the electrode-tissue interface and electrode movement (Besomi et al., 2019).

The validity of any measurement made during an experimental protocol is underpinned by the stability of that measurement over the duration of the data collection session. Studies which involve dynamic activation may be particularly problematic. While fw-EMG was successfully used to evaluate changes in regional and synergistic activation in the triceps surae complex during sustained, low-level contractions lasting one hour without evidence of signal degradation (McLean and Goudy, 2004), in one study using fw-EMG recorded from the tibialis posterior muscles during a running task, the authors reported that signal degradation over time resulted in complete data sets for only 4/10 participants (O'Connor et al., 2006). Similarly the fw-EMG signal from the tibialis posterior during walking was disrupted or noisy in 2/5 participants during a protocol lasting 60–90 min (Stacoff et al., 2007), and more recently Reeves et al. (2020) reported evidence of a systematic reduction in fw-EMG signal amplitude recorded from the tibialis anterior muscle after approximately 20–30 min of walking. Despite this preliminary evidence of signal degradation over time, particularly when fw-EMG is recorded during dynamic tasks, recently published studies have involved fw-EMG recordings during dynamic tasks, with protocols lasting over one hour (Kingston and Acker, 2018) and even up to four hours (Raven et al., 2018). The time course and causes of potential systematic decreases in fw-EMG signal amplitude have not been explored.

Several processes may contribute to fw-EMG signal attenuation. Potential causes include physiological responses such as oedema, which could reduce signal amplitude by distancing the electrode from the active tissue (Geddes and Roeder, 2003) and electrode fouling in which the adhesion of biomolecules to an implanted sensor can result in changes in the characteristics of the recorded signals (Hanssen et al., 2016, Harreither et al., 2016). Signal attenuation could also be a consequence of the intrinsic properties of the electrode. Modern silver/silver-chloride electrodes, typically used in s-EMG (e.g. Bortec Biomedical Ltd., Canada), take very little time to achieve a stable impedance (Searle and Kirkup, 2000). However, silver/silver-chloride is toxic (Geddes and Roeder, 2003), so cannot be used for intramuscular recordings. Fine-wire electrodes (e.g. Chalgren Enterprise) are often manufactured from stainless steel insulated with Teflon, due to the low toxicity and corrosion resistance of stainless steel (Geddes and Roeder, 2003, Merrill et al., 2005). Conversely, stainless steel takes longer to reach a constant impedance than silver-chloride (Searle and Kirkup, 2000) and has been shown to be unstable over time (Lykken, 1959, Merrill et al., 2005).

Mechanical effects due to length changes and forces generated in a muscle may also lead to changes in fw-EMG signal amplitude. Although hooked (bent) electrode tips help the electrode to stay in place in a contracting muscle, the ends may not maintain a fixed relationship to the active fibres as the muscle changes its length and pennation angle during contraction (Besomi et al., 2019). The electrode tips may be damaged by contractile or shear forces induced on them by muscle shortening (Helton et al., 2011), which may explain why degradation in fw-EMG signals was not observed in a protocol involving static contractions (McLean and Goudy, 2004), but has been observed in gait studies (O'Connor et al., 2006, Reeves et al., 2020, Stacoff et al., 2007). Indeed a recent expert consensus process led to caution being advised when fw-EMG is used to study dynamic tasks (Besomi et al., 2019). Caution was also advised if using a maximum voluntary contraction (MVC) for the purposes of normalisation, because forceful contractions might move the position and/or orientation of recording tips within the muscle and may damage the wire (Besomi et al., 2019). Despite these recommendations, there is little empirical evidence on which to base recommendations around the recording of fw-EMG.

The purpose of this study was to investigate the impact of recording time on the amplitude of fw-EMG signals recorded from the biceps brachii, while also evaluating the potential confounding factors of task type (isometric, dynamic) and maximum voluntary contraction, as it is typically performed for normalization purposes prior to data collection.

Section snippets

Sample size estimate

Previous work (Reeves et al., 2020) found a large effect size (d = 1.4) for a reduction in peak tibialis anterior fw-EMG amplitude after 30 min of walking. A power analysis was conducted (G*Power) (Faul et al., 2007) for a dependent t-test; using 80% power and α = 0.05, the required sample size was estimated at n = 7 to detect a reduction in fw-EMG amplitude over time alone. However, no data were found comparing the difference in the relative reduction in fw-EMG amplitude over time between

Results

A total of 21 participants were recruited and twenty participants completed the study (15 females, 5 males, age = 26 ± 6 years, height = 1.73 ± 0.01 m, mass = 69.71 ± 16.22 kg, mean ± SD), n = 12 were randomized to G1 and n = 8 were randomized to G2. One participant reported feeling faint after the fw electrodes were inserted, and took a break of 20 min before beginning the protocol. At the time that this participant chose to continue, their data were extremely noisy to the extent that it was

Discussion

This study explored some of the potential causes of signal attenuation when recording fw-EMG. There was a significant drop in fw-EMG amplitude recorded from the biceps brachii during isometric contractions over two hours, with no concurrent change in s-EMG. The consistency of fw-EMG amplitude over time appeared to be influenced by task type, where dynamic contractions resulted in more fw-EMG signal attenuation over time than isometric contractions. The number of full data sets was limited, and

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to thank Olena Klahsen (University of Ottawa) for producing the figure of the experimental set up and Dr. Derek Zwambag for his Matlab code to calculate median power frequency.

Linda McLean is Full Professor and Chair in Women’s Health Research in the Faculty of Health Sciences at the University of Ottawa, Canada. She runs an interdisciplinary research program focused mainly on pelvic health rehabilitation, through which she merges her background in both physiotherapy (McGill University, BSc 1990) and engineering (University of New Brunswick, MSc 1995 and PhD 1998). Her lab can be found at www.mfmlab.ca and on Twitter at @mfmlab.

References (36)

  • J. Reeves et al.

    EMG gait data from indwelling electrodes is attenuated over time and changes independent of any experimental effect

    J. Electromyogr. Kinesiol.

    (2020)
  • N. Wisniewski et al.

    Methods for reducing biosensor membrane biofouling

    Colloids Surf., B

    (2000)
  • I.N. Blazquez et al.

    An adequate interset rest period for strength recovery during a common isokinetic test

    J. Strength Condition. Res..

    (2013)
  • G.A. Borg

    Psychophysical bases of perceived exertion

    Med. Sci. Sports Exerc.

    (1982)
  • S. Campuzano et al.

    Antifouling (Bio)materials for Electrochemical (Bio)sensing

    Int. J. Mol. Sci.

    (2019)
  • A. Del Vecchio et al.

    Associations between motor unit action potential parameters and surface EMG features

    J. Appl. Physiol.

    (2017)
  • J.L. Dideriksen et al.

    Influence of fatigue on the simulated relation between the amplitude of the surface electromyogram and muscle force

    Philosoph. Trans. Roy. Soc. A: Mathe., Phys. Eng. Sci.

    (2010)
  • F. Faul et al.

    G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences

    Behav. Res. Methods

    (2007)
  • Cited by (5)

    • Self-Powered Real-Time Wireless Communication System Using Wearable Fabric Based Triboelectric Nanogenerator and Inductor

      2024, Proceedings of the IEEE International Conference on Micro Electro Mechanical Systems (MEMS)

    Linda McLean is Full Professor and Chair in Women’s Health Research in the Faculty of Health Sciences at the University of Ottawa, Canada. She runs an interdisciplinary research program focused mainly on pelvic health rehabilitation, through which she merges her background in both physiotherapy (McGill University, BSc 1990) and engineering (University of New Brunswick, MSc 1995 and PhD 1998). Her lab can be found at www.mfmlab.ca and on Twitter at @mfmlab.

    View full text