Research reportA note on the probability distribution function of the surface electromyogram signal

The probability density function (PDF) of the surface electromyogram (EMG) signals has been modelled with Gaussian and Laplacian distribution functions. However, a general consensus upon the PDF of the EMG signals is yet to be reached, because not only are there several biological factors that can influence this distribution function, but also different analysis techniques can lead to contradicting results. Here, we recorded the EMG signal at different isometric muscle contraction levels and characterised the probability distribution of the surface EMG signal with two statistical measures: bicoherence and kurtosis. Bicoherence analysis did not help to infer the PDF of measured EMG signals. In contrast, with kurtosis analysis we demonstrated that the EMG PDF at isometric, non-fatiguing, low contraction levels is super-Gaussian. Moreover, kurtosis analysis showed that as the contraction force increases the surface EMG PDF tends to a Gaussian distribution.

► We recorded surface EMG signals with a biofeedback setup at 7 different contraction levels. ► We estimated the PDF, kurtosis and bicoherence index of the measured signals. ► We show that the EMG PDF at low contraction levels is super-Gaussian. ► At higher contraction forces, the EMG PDF tends to a Gaussian distribution.