reply: In response to our recent paper (Neto and Christou 2010), Dr. Boonstra raised primarily two arguments in his letter. The first argument indicates that periodic input to the muscle may be evident not only at lower frequencies of muscle activity (sub-60 Hz) but also as the modulation of higher frequencies (e.g., modulation of 100–200 Hz). The secondary argument suggests that rectification of the EMG may be an appropriate method to capture such high frequency modulation. Although we agree with the first argument that the modulation of motor units may be derived from higher frequencies, the following questions remain with respect to his secondary argument:
1. Is rectification of the EMG signal an appropriate preprocessing step to identify the modulation of motor units?
Modulation of motor units exists at various frequency bands. For example, motor units typically discharge from 6 to 15 Hz (Christou et al. 2007) and can be modulated from 0 to 1 Hz (Christou et al. 2007; DeLuca 1982) and from 13 to 30 Hz (Christou et al. 2007; Farmer et al. 1993). There is no evidence that we can extract the 0–1 Hz from either the interference or rectified EMG signal and, although peaks occur from 13 to 60 Hz in both the interference and rectified EMG, the origins of such peaks remain unclear (Christou and Neto 2010).
The strongest evidence for identification of motor unit modulation occurs at the frequency band that is associated with the mean discharge rate. There is evidence that both the interference and rectified EMG signals can identify the mean discharge rate using simple simulations (Christou and Neto 2010; Myers et al. 2003). However, during more realistic simulations only the interference EMG accurately identifies the mean discharge rate of the motor units (see Fig. 4 in Farina et al. 2004). The rectified EMG impairs the identification of the mean motor unit discharge rate when the mean discharge rate increases or when physiological variability in motor unit discharge is introduced. Therefore simulations that use biophysical models, as suggested by Dr. Boonstra in his letter, support our recent article (Neto and Christou 2010) that rectification of the EMG impairs the identification of oscillatory input to the muscle.
2. Is the sub-60 Hz power spectra in the rectified EMG sensitive enough to identify physiologically relevant changes to the modulation of higher frequencies?
As we stated earlier, rectification of the EMG may even impair the identification of the mean discharge rate that typically occurs around 6–15 Hz (Farina et al. 2004). Furthermore, if the sub-60 Hz oscillations in the rectified EMG signal represent the modulation of higher frequencies (e.g., the motor unit shape ∼100–200 Hz), then manipulation of higher frequencies in the original signal should result in changes to the low-frequency oscillations in the rectified EMG. Our recent findings (Neto and Christou 2010), which were based on real EMG signals, suggest that sub-60 Hz oscillations in the rectified EMG signal are insensitive to dramatic changes to the original signal from 100 to 200 Hz. For example, doubling or removing the power from 100 to 150 Hz in the original signal results in similar sub-60 Hz power spectra in the rectified EMG (see Fig. 5B in Neto and Christou 2010). Finally, we have recently demonstrated that when subjects voluntarily increased force from 15 to 50% maximal voluntary contraction, the oscillations from 12 to 60 Hz increased in the interference EMG but not in the rectified EMG (OP Neto, HS Baweja, and EA Christou, unpublished data). Thus the results from the above-cited studies, which include simulations and experimental data, suggest that the rectified EMG may not be a very sensitive or consistent method to identify physiologically relevant changes to the modulation of higher frequencies.
In summary, we agree with Dr. Boonstra that further computational research is necessary to identify the various frequency bands of motor unit modulation using the interference EMG signal. Nonetheless, based on the evidence we provide in the preceding text, rectification of the EMG signal may not be a precise method to determine modulation of motor units from surface EMG signals.
GRANTS
This work was supported by National Institute on Aging Grant R01 AG-031769 to E. A. Christou.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
ACKNOWLEDGMENTS
Present address: E. A. Christou and O. P. Neto, Neuromuscular Physiology Laboratory, Department of Applied Physiology and Kinesiology, University of Florida 32511; and O. P. Neto, Department of Biomedical Engineering, Universidade Camilo Castelo Branco, Brazil.
REFERENCES
- Christou and Neto 2010.Christou EA, Neto OP. Identification of oscillations in muscle activity from surface EMG: reply to Halliday and Farmer. J Neurophysiol 103: 3548–3549, 2010 [Google Scholar]
- De Luca and Erim 2002.De Luca CJ, Erim Z. Common drive in motor units of a synergistic muscle pair. J Neurophysiol 87: 2200–2204, 2002 [DOI] [PubMed] [Google Scholar]
- Farina et al. 2004.Farina D, Merletti R, Enoka RM. The extraction of neural strategies from the surface EMG. J Appl Physiol 96: 1486–1495, 2004 [DOI] [PubMed] [Google Scholar]
- Farmer et al. 1993.Farmer SF, Bremner FD, Halliday DM, Rosenberg JR, Stephens JA. The frequency content of common synaptic inputs to motoneurones studied during voluntary isometric contraction in man. J Physiol 470: 127–155, 1993 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Halliday and Farmer 2010.Halliday DM, Farmer SF. On the need of rectification for surface EMG (Letter to the Editor). J Neurophysiol 103: 3547, 2010 [DOI] [PubMed] [Google Scholar]
- Myers et al. 2003.Myers LJ, Lowery M, O'Malley M, Vaughan CL, Heneghan C, St Clair Gibson A, Harley YX, Sreenivasan R. Rectification and non-linear pre-processing of EMG signals for cortico-muscular analysis. J Neurosci Methods 124: 157–165, 2003 [DOI] [PubMed] [Google Scholar]
- Neto and Christou 2010.Neto OP, Christou EA. Rectification of the EMG signal impairs the identification of oscillatory input to the muscle. J Neurophysiol 103: 1093–1103, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]