The experimental setup. Upper panel: The forearm mockup with implanted magnets (MMs). The mockup aimed at reproducing the natural position and orientation of forearm muscles, in addition to their deformation due to contraction. The mockup replicated the movements of 17 degrees of freedom of the hand and wrist33, for a total of 17 wires, albeit only four were used. Muscles were modelled using a wire attached on one side to a servo motor (housed in a remote actuation unit) used to actively contract the muscle. Four cylindrical MMs were implanted on the abductor pollicis longus (MM0), flexor pollicis longus (MM1) and compartments 1 and 2 of the flexor digitorum superficialis actuating the index (MM2) and middle fingers (MM3). Lower panel: The localizer, composed of two custom PCBs and a host PC, was used to track the position of the MMs solving the magnetic inverse problem. Emodel indicates the errors due to the magnetic dipole model. Ecross-talk indicates the error due to cross-talk between MMs. Esensor indicates the error introduced by the six magnetic field sensors (i.e. electronic noise). P0 indicates the rest position of the MMs (i.e. relaxed muscle); P9 corresponds to maximum muscular contraction and depends on the muscle where the MM is implanted in. In order to assess Emodel, Ecross-talk, Esensor, affecting the myokinetic interface, each MM, one at a time, was moved along the entire range of motion (ROM) of the emulated muscle, and data was collected at the ten equidistant points in space P0, …, P9.