A pilot study of two degrees of freedom dynamic EMG-force at the wrist using a minimum number of electrodes

Grasp force estimation from the transient EMG using high-density surface recordings

“…So far researchers have investigated this relationship using either intra-muscular [9][10][11][12] or surface EMG recordings [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]. Several estimation techniques, exploiting a wide range of methodologies, were proposed (Table 1).…”

“…a classifier or a regressor) that provides an estimate of the force for each segment. By using several different algorithms, such as regressors [9,11,16,17,22,24,25] or artificial neural networks [9,[11][12][13][14]21,26,27], these studies demonstrated notable estimation accuracies up to 0.95 R 2 (coefficient of determination) or 4.21% absolute error (AE) from wrist, finger and trunk movements, and from a wide range of forces (e.g. from 0% to 100% of muscle activation [10,11,13,[15][16][17]19,22,23,25,26], or up to 300N of output force [9,12,14,18,24]).…”

“…By using several different algorithms, such as regressors [9,11,16,17,22,24,25] or artificial neural networks [9,[11][12][13][14]21,26,27], these studies demonstrated notable estimation accuracies up to 0.95 R 2 (coefficient of determination) or 4.21% absolute error (AE) from wrist, finger and trunk movements, and from a wide range of forces (e.g. from 0% to 100% of muscle activation [10,11,13,[15][16][17]19,22,23,25,26], or up to 300N of output force [9,12,14,18,24]). Notably, some of the methods developed for the estimation of the grip force from the EMG also allow for the simultaneous control of up to 6 degrees of freedom of a prosthesis [12,15,17,20,25,27].…”

“…from 0% to 100% of muscle activation [10,11,13,[15][16][17]19,22,23,25,26], or up to 300N of output force [9,12,14,18,24]). Notably, some of the methods developed for the estimation of the grip force from the EMG also allow for the simultaneous control of up to 6 degrees of freedom of a prosthesis [12,15,17,20,25,27]. This is particularly interesting in clinical settings, where multi-articulated hand prostheses are now available, as these solutions also provide a smooth way to switch between motor functions of the prosthetic device (Table 1).…”

“…This indeed was demonstrated to be descriptive of the intended movement [31]. However, thus far, such temporal arrangement was poorly exploited for the estimation of the output force during a movement (Table 1, [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]). Here we hypothesized that the informative content of the transient EMG in humans is likely due to the preplanned nature of grasping, in general, and of the final GF, in particular [32].…”