Towards hand grip force assessment by using EMG estimators

The purpose of this study was to propose a method to assess individual regression (calibration) curves to establish a relationship between an isometric grip force and surface electromyography (EMG) estimator. In this study, 18 healthy volunteers (12 male (23.0 ± 2.0 years) and 6 female (23.2 ± 0.7 years)) had been examined. Ten EMG estimators (mean absolute value, root mean square, entropy, energy, turns per second, mean of zero crossings, median power frequency, mean of mean power frequency, and Hilbert transforms), two EMG locations (brachioradialis and flexor digitorum superficialis) and five types of regression curves (linear, exponential, power, logarithm, and quadratic polynomial) have been tested. Using a quadratic polynomial regression curve and energy EMG estimator of the flexor digitorum superficialis muscle, it was obtained the lowest dispersity of EMG estimator and the most accurate prediction of grip isometric force among all tested EMG estimators.