HOME  ABOUT   PERSONNEL   PROJECTS   PUBLICATIONS   LINKS   CONTACT

FMG

Force Myography

Introduction and Background

Comparison of filtered and unfiltered FMG (top traces) to EMG for isometric extension (left) and flexion (right) tasks.

Whereas electromyography (EMG) is essential to many applications such as assessment of injury or disease, control of external devices, biofeedback, and human performance measurements, and does so by direct measurement of the electrical activation of muscles, it may not be the ideal tool for all muscle assessment applications. Muscle activation can be assessed in alternate ways by measuring either the movement of or the force produced by the muscle during contraction.

FMG consists of an array of pressure sensors mounted to a cuff and worn on the leg. Theses sensors record the outward pressure exerted on the skin, which changes at different locations along the leg as muscles contract and relax. FMG works on the principle that as limb muscles contract, certain regions on the muscle bellies bulge outward, while others recede inward from the surface, resulting in positive or negative pressure changes below the surface of the skin. These pressures are detected by sensors within the sleeve which register the pressures as activation and cessation of muscle activity, which can be displayed in real-time and recorded for post-hoc analysis.

Comparison to Electromyography

Enhanced Repeatability & Compatibility

Comparison of filtered FMG (left) and filtered & rectified EMG (right) for ambulatory tasks (v=1.3m/s, n=8). Traces normalized temporally and for magnitude.

A single female subject (25 years old), an experienced treadmill walker, ambulated on a treadmill at 1.3 and 2.7 m/s while data were collected simultaneously from the upper and lower legs, as were footswitch data for temporal landmark measurement controls. Footswitch data were captured at 100Hz for a series of 10-second trials. Final EMG and FMG signals used for analysis were composites (sums of sensors), rectified and then summed. Signals were smoothed with a low-pass filter and analyzed for muscle onset and cessation during gait, as well as their waveform similarity.

Results show that FMG records temporal landmarks of muscle activity with as much fidelity as EMG, but with less waveform variability from stride to stride and from trial to trial: mean variance ratios were calculated to be an order of magnitude greater for that of EMG as compared to FMG. Intraclass correlation coefficients of temporal landmarks typically exceeded 0.8 for the inter-rater comparison of FMG to EMG.

Additionally, a single male subject, 53 years old, and with no significant limitations, participated in a single session of isometric protocols. Seated ina Biodex Dynamometer and using the right (dominant) leg, 10-15 seconds of isometric flexion and extension tasks were recorded simultaneously by EMG and FMG. Data was smoothed and processed for evaluation of temporal landmarks and waveform fidelity.

During isometric contractions, FMG and EMG waveforms were highly correlated, with r>0.8 for flexion and r>0.95 for extension. These results, together with those of the ambulatory data, suggest the utility of FMG as a simpler alternative to EMG for some applications.