Normalizing shoulder EMG: An optimal set of maximum isometric voluntary contraction tests considering reproducibility.

Normalization of the electromyography (EMG) signal is often performed relatively to maximal voluntary activations (MVA) obtained during maximum isometric voluntary contraction (MVIC). The first aim was to provide an inter-session reproducible protocol to normalize the signal of eight shoulder muscles. The protocol should also lead to a level of activation >90% of MVA for >90% of the volunteers. The second aim was to evaluate the influence of the method used to extract the MVA from the EMG envelope on the normalized EMG signal. Thirteen volunteers performed 12 MVICs twice (one week interval). Several time constants (100ms to 2s) were compared when extracting the MVA from the EMG envelope. The EMG activity was also acquired during an arm elevation. Our results show that a combination of nine MVIC tests was required to meet our requirements including reproducibility. Both the number of MVIC tests and the size of the time constant influence the normalized EMG signal during the dynamic activity (variations up to 15%). A time constant of 1s was a good compromise to extract the MVA. These findings are valuable to improve the reproducibility of EMG signal normalization.