Detecting low-to-moderate isometric muscle activity through a generalized CWT-based technique

A recently introduced generalization of a Continuous Wavelet Transform-based (gCWT) algorithm for the detection of muscular activation from EMG recordings, previously tested in silico on synthetic and semi-synthetic data, is here applied for the detection of experimental data. Healthy young adults were requested to isometrically contract the upper trapezius while seated, and a visual bio-feedback was implemented to maintain the level of the upper trapezius activity between 5% and the 10% of the maximum voluntary contraction. Difference of duration between the estimated muscular activity and the force exertion captured by a pressure switch was calculated for both the gCWT technique and an amplitude-based technique, using the Teager-Kaiser Energy Operator (TKE). Difference of duration resulted in line with uncertainties associated with the variations in electromechanical delay (EMD) and relaxation-EMD reported in the literature, and they were found lower than those obtained with TKE.