A Lower Limb Phantom for Simulation and Assessment of Electromyography Technology.

Abstract

Electromyography signal processing approaches have traditionally been validated through computer simulations. Electromyography electrodes and systems are often not validated or have been validated on human subjects where there is no clear ground truth signal for comparison. We sought to develop a physical limb phantom for validation of electromyography hardware and signal processing approaches. We embedded pairs of wires within a conductive gelatin surrounding an artificial bone such that the antennae could broadcast identified ground truth signals. The ground truth signals can be simple sinusoids or more complex representations of muscle activity. With the phantom and surface electromyography electrodes, we were able to show varying levels of crosstalk between nearby recording electrodes as we altered the amplitude of the antennae signals. We were also able to induce motion artifacts in our recordings by lightly dropping the phantom on a surface while antennae broadcast signals. High-density electromyography recordings of the trials showed that traditional filtering techniques fail to fully eliminate relatively small motion artifacts. The results suggest that the electrical limb phantom could be a valuable tool for testing potential effects of muscle crosstalk and motion artifacts on different electromyography systems and signal processing approaches.