Analysis of phasic and tonic electromyographic signal characteristics: Electromyographic synthesis and comparison of novel morphological and linear-envelope approaches

https://doi.org/10.1016/j.jelekin.2007.02.018Get rights and content

Abstract

The pattern of tonic and phasic components in an EMG signal reflects the underlying behaviour of the central nervous system (CNS) in controlling the musculature. One avenue for gaining a better understanding of this behaviour is to seek a quantitative characterisation of these phasic and tonic components. We propose that these signal characteristics can range between unvarying, tonic and intermittent, phasic activation through a continuum of EMG amplitude modulation. In this paper, we present two new algorithms for quantifying amplitude modulation: a linear-envelope approach, and a mathematical morphology approach. In addition we present an algorithm for synthesising EMG signals with known amplitude modulation. The efficacy of the synthesis algorithm is demonstrated using real EMG data. We present an evaluation and comparison of the two algorithms for quantifying amplitude modulation based on synthetic data generated by the proposed synthesis algorithm. The results demonstrate that the EMG synthesis parameters represent 91.9% and 96.2% of the variance of linear-envelopes extracted from lumbo-pelvic muscle EMG signals collected from subjects performing a repetitive-movement task. This depended, however, on the muscle and movement-speed considered (F = 4.02, p < 0.001). Coefficients of determination between input and output amplitude modulation variables were used to quantify the accuracy of the linear-envelope and morphological signal processing algorithms. The linear-envelope algorithm exhibited higher coefficients of determination than the most accurate morphological approach (and hence greater accuracy, T = 8.16, p < 0.001). Similarly, the standard deviation of the coefficients of determination was 1.691 times smaller (p < 0.001). This signal processing algorithm represents a novel tool for the quantification of amplitude modulation in continuous EMG signals and can be used in the study of CNS motor control of the musculature in repetitive-movement tasks.

Introduction

In addition to the more familiar measures of signal amplitude, median frequency or onset timing, surface electromyographic (EMG) signals can be described by their tonic and phasic characteristics. Typically, a “tonic” EMG signals exhibits sustained electrical activity with little modulation of EMG amplitude. “Phasic” signals, on the other hand, exhibit distinct on and off periods of electrical activity. Between these two extremes, signals exhibit varying degrees of amplitude modulation (Fig. 1). These differences reflect the functional differentiation in CNS control of the musculature. For example, tonic activation in the deep lumbo-pelvic musculature by the CNS during repetitive-movements (such as cyclic arm movement or walking) has been linked to “stabilisation” of the lumbar spine (Hodges et al., 1999, Hodges and Gandevia, 2000b, Saunders et al., 2004b). Another example of the relevance of observing phasic, on–off, activation patterns in the superficial trunk musculature is the linkage to CNS control of overall trunk orientation and posture (Hodges et al., 1999, Hodges and Gandevia, 2000b, Hodges et al., 2003, Moseley et al., 2002, Saunders et al., 2004b).

Existing signal processing methods only allow quantification of a limited scope of tonic and phasic EMG signal characteristics, and do not allow measurement of absolute EMG signal amplitude modulation. Measurement of onset and offset timing of muscle activity (such as during rapid arm movement or walking) (e.g. Hodges et al., 1999, Saunders et al., 2004b) characterise the extremes of phasic and tonic muscle activation: whether the CNS activates a muscle in either a phasic (on–off) or tonic (sustained) capacity, but not amplitude modulation of the EMG signal. Relative change in EMG signal amplitude modulation (such as between speeds of repetitive-movement) can be measured using frequency domain methods. Such approaches quantify the proportional change in the amplitude of the power spectrum at the movement frequency at each movement speed (Hodges et al., 1999, Hodges and Gandevia, 2000a, Saunders et al., 2004b). This approach cannot be used, however, to quantify the absolute amount of EMG signal amplitude modulation.

The ability to quantify absolute amplitude modulation is important in the study of motor control for several reasons. Firstly, it is then possible to measure whether the CNS drives a particular muscle in a more tonic or more phasic capacity than other muscles; something not possible with existing algorithms. This in turn may imply functional differences between the muscles. Secondly, quantification can facilitate the study of motor control dysfunction. In “dysfunction” (such as pain (Hodges et al., 2003, Saunders et al., 2004a), spinal cord transection (Alaimo et al., 1984) and animal hindlimb suspension (Blewett and Elder, 1993, Riley et al., 1990)) a shift from tonic to phasic muscle activation is commonly observed. However, existing research has been limited to either time-domain approaches (onset/offset timing and/or duration of activity) or qualitative observations of the EMG signals. Quantification of absolute EMG signal amplitude modulation further opens up understanding of both function and dysfunction of CNS motor control of the musculature.

In this paper we proffer two algorithms (linear-envelope and morphological filtering) for quantifying absolute EMG signal amplitude modulation. In addition we present an evaluation and comparison of the two methods based on synthetically generated EMG signals. The reason for using synthetic data is that it can be generated with known amplitude modulation; i.e. we know the “ground truth”. This would not be the case for real data collected from test subjects. The efficacy of the synthesis algorithm is demonstrated using real EMG data. In this paper we define an index of amplitude modulation as the ratio of maximum (burst) signal amplitude to the underlying sustained (tonic) amplitude (burst-to-tonic ratio [BTR]).

Section snippets

Overview

This section is divided into two parts. The first part describes the EMG synthesis algorithm (Section 2.2.1), the real-world data used in its implementation (Section 2.2.1.1) and evaluation (Section 2.2.2), and the evaluation methodology (Sections 2.2.3 Algorithm evaluation I: fitting of algorithm parameters to tonic-phasic EMG signals by a skilled operator, 2.2.4 Algorithm evaluation II: data and statistical analysis). The second part describes the algorithms devised to estimate amplitude

EMG synthesis algorithm evaluation

ANOVA of the Rfit2 data showed significant variation between muscles (F = 24.04, p < 0.001) as well as an interaction with movement speed (F = 4.02, p < 0.001, Table 1). The minimum and maximum Rfit2 value were 0.919 and 0.962, respectively, indicating that the algorithm’s parameters explained between 91.9% and 96.2% of the variance in the linear-envelope signals.

For the LES muscle, Rfit2 was slightly (though significantly) less at the 75 and 100 cyc/min speeds compared to the 25 speed. In the IGM

Discussion

This study aimed to evaluate linear-envelope and morphological signal processing methods for the quantification of amplitude modulation in continuous EMG signals. Such quantification has not been performed to date. The linear-envelope algorithm was found to be the most reliable estimator of actual EMG amplitude modulation. In conducting these analyses, it was necessary, however to develop and evaluate an algorithm that synthesised the necessary amplitude-modulated EMG signals.

Conclusion

We evaluated morphological and linear-envelope signal processing algorithms for the quantification of EMG signal amplitude modulation, and found the linear-envelope approach to generate the most accurate estimates of amplitude modulation in synthesised EMG signal. The parameters forming the EMG synthesis algorithm performed quite well in explaining above 91% of the variance in linear-envelopes extracted from signal obtained from subjects during a repetitive-movement task. This work represents

Acknowledgements

The authors wish to thank Dr. Ross Darnell of the School of Health and Rehabilitation at the University of Queensland for expert statistical advice. We also wish to thank Prof. Jörn Rittweger of the Manchester University at Cheshire, UK and Prof. Dieter Felsenberg of the Centre for Muscle and Bone Research at the Charité Benjamin Franklin, Berlin, Germany for leading the “Berlin Bed-Rest Study” which provided part of the source data for this study.

Glossary of Abbreviations

BTR
burst-to-tonic ratio; ratio of phasic burst amplitude to underlying tonic activity amplitude (equivalent to an amplitude modulation index +1).
BTRinput
burst-to-tonic ratio of the modulation signal used in EMG synthesis
Section (2.2.1.2).
BTRfit
burst-to-tonic ratio calculated during fitting of EMG synthesis algorithm parameters to linear-envelopes of signals collected from subjects during repetitive leg movement
Section (2.2.3).
BTRoutput
generalised term for the quantification algorithms’

Daniel Belavý graduated in 1999 from physiotherapy at The University of Queensland and joined the Joint Stability Research Team as a research worker in 2001. Later, he enrolled in post-graduate studies with A/Prof Carolyn Richardson and Dr. Stephen Wilson. In 2003 he relocated to Berlin, Germany to collect data for the research team and his PhD in the Berlin Bed-Rest Study. His thesis focused on the changes of the motor control of lumbo-pelvic stabilisation in bed-rest, as reflected in

References (24)

  • H.J.A.M. Heijmans

    Morphological Image Operators

    (1994)
  • P.W. Hodges et al.

    Activation of the human diaphragm during a repetitive postural task

    Journal of Physiology

    (2000)
  • Cited by (7)

    View all citing articles on Scopus

    Daniel Belavý graduated in 1999 from physiotherapy at The University of Queensland and joined the Joint Stability Research Team as a research worker in 2001. Later, he enrolled in post-graduate studies with A/Prof Carolyn Richardson and Dr. Stephen Wilson. In 2003 he relocated to Berlin, Germany to collect data for the research team and his PhD in the Berlin Bed-Rest Study. His thesis focused on the changes of the motor control of lumbo-pelvic stabilisation in bed-rest, as reflected in electromyography and magnetic resonance imaging. Having been awarded his PhD, he has returned to Berlin to take up a post as coordinator of the 2nd Berlin Bed-Rest Study.

    Andrew Mehnert received his BAppSc degree (Hons 1) in Mathematics in 1990 and his MAppSc degree in Mathematics and Planning in 1994 from Edith Cowan University, Perth, Western Australia. In 2004 he received his PhD in Electrical Engineering from the University of Queensland, Brisbane, Australia. He is currently a Research Fellow in the School of Information Technology and Electrical Engineering at the University of Queensland. His research interests include image and signal processing and analysis, and statistical pattern recognition.

    Stephen Wilson, a medically qualified engineer, is currently Senior Lecturer in the School of Information Technology and Electrical Engineering at the University of Queensland. Instrumentation and imaging for musculo-skeletal measures is one theme of his research. He also participates in sleep and respiratory medicine based research projects. He pursues interests in nonlinear biological signal analysis, biomedical instrumentation generally and engineering teaching at undergraduate level.

    Carolyn Richardson’s research fields are investigations into the most effective and efficient exercise treatment (and countermeasures) for musculoskeletal injuries, especially low back pain, which are linked to inadequate stabilisation and support of the joints. This knowledge has formed the basis of many textbooks and scientific papers. Current research has developed a new focus relating to the function of the human antigravity muscle system, which is severely affected when gravity is minimized (including in microgravity). These changes provide new information on the possible aetiology of low back pain, as well as other conditions such as osteoarthritis of the weightbearing joints and osteoparosis. This line of research has led to extensive research collaboration and consultancies with the European Space Agency (ESA), including scientific consultant to the Toulouse Bedrest study, member of the ESA Topical Team for Low Back Pain, Senior Collaborator and on the Board of the ESA project “Vibration Exercise in Space”, a Bedrest study, undertaken at the Free University of Berlin February 2003 to May 2005.

    View full text