Assessing the influence of SNR and pre-processing filter bandwidth on the extraction of different muscle co-activation indexes from surface EMG data.

Muscle co-activation is the mechanism that regulates simultaneous activity of agonist and antagonist muscles crossing the same joint. During functional movements, robust measurement techniques are required for an accurate determination of muscle co-activation, since various environmental and processing factors in the surface electromyography (sEMG) measurement process might influence the estimation of linear envelope profiles, and therefore the outcome of co-activation evaluated from the signal envelope. The aim of this study is to verify the performance of the co-activation indexes introduced in six different techniques used to assess muscle co-activation. A sensitivity analysis with respect to both noise and pre-processing choices for envelope estimation has been done by using a data-set of simulated sEMG signals. The results show how the indexes are affected by both the level of noise and pre-processing choices. The Vector Coding Technique and the Time-varying Multi-muscle approach perform better than the others in terms of both sensitivity to varying levels of co-activation and robustness to noise.