The extraction of neural strategies from the surface EMG: an update

Abstract

A surface EMG signal represents the linear transformation of motor neuron discharge times by the compound action potentials of the innervated muscle fibers and is often used as a source of information about neural activation of muscle. However, retrieving the embedded neural code from a surface EMG signal is extremely challenging. Most studies use indirect approaches in which selected features of the signal are interpreted as indicating certain characteristics of the neural code. These indirect associations are constrained by limitations that have been detailed previously (Farina D, Merletti R, Enoka RM. J Appl Physiol 96: 1486-1495, 2004) and are generally difficult to overcome. In an update on these issues, the current review extends the discussion to EMG-based coherence methods for assessing neural connectivity. We focus first on EMG amplitude cancellation, which intrinsically limits the association between EMG amplitude and the intensity of the neural activation and then discuss the limitations of coherence methods (EEG-EMG, EMG-EMG) as a way to assess the strength of the transmission of synaptic inputs into trains of motor unit action potentials. The debated influence of rectification on EMG spectral analysis and coherence measures is also discussed. Alternatively, there have been a number of attempts to identify the neural information directly by decomposing surface EMG signals into the discharge times of motor unit action potentials. The application of this approach is extremely powerful, but validation remains a central issue.

Differential Motor Unit Changes after Endurance or High-Intensity Interval Training

“…However, this earlier work also showed an increase of EMG amplitude at 10 and 30% MVC after 3 weeks of END training. Since a decrease in motor unit discharge rate was simultaneously observed after END training, this result was interpreted as an increase of motor unit recruitment at these force levels, although EMG amplitude depends on multiple influencing factors (11,12). In this study, we attempted to limit the variability in EMG amplitude estimates by averaging across all electrodes of the grid (14, 24).…”

Experimental muscle pain increases variability of neural drive to muscle and decreases motor unit coherence in tremor frequency band

“…The magnitude of the coherence function increases quadratically with the number of motor unit spike trains considered in the CST, resulting in better estimates (Farina et al 2014a;Negro and Farina 2011a). In the present study, the number of decomposed motor units was only four in some cases.…”

New Perspectives for Computer-Aided Discrimination of Parkinson’s Disease and Essential Tremor

“…By summing motor unit spike trains together into CST, the intrinsic noise components of many motoneurons average out, revealing their time-varying common input [21,36,44,45]. Autocorrelation function of neural drive further emphasizes the rhythmic (synchronous) motor neuron activities and suppresses the asynchronous ones.…”

“…CST captures the temporal variability of neural drive to a muscle, while averaging out the differences among identified motor units [36]. Thus, the CST ignores how the tremor input is distributed across the motoneurons innervating the muscle.…”