The effect of physiological and measurement noise on the estimate of individual muscle force from indirect measurements of muscle activity

Several forward dynamics estimators have been proposed to quantify individual muscle force using non-invasive measurements of muscle activity. None of them, however, addresses the inaccuracy that arises when measurements are available only from a subset of the muscles involved in the movement under analysis. We present a novel estimator that integrates a forward dynamics estimation approach with knowledge of the optimal contraction strategy to obtain accurate estimates of individual muscle force when measurements of muscle activity are not available for all muscles. A following in-silico characterization showed that when trying to estimate forces form the forearm muscles acting around the wrist joint, our novel estimator is able to decrease the mean estimation bias by about 25% of the true value of muscle force. With a sensitivity analysis, we show that the model-based estimator is robust against physiological variability in muscle co-contraction strategy.