What Is Force Myography?

Force myography is a non-invasive muscle sensing technique that infers movement intent from the mechanical deformation of soft tissue, measured as pressure at the skin surface. When a muscle contracts it shortens and bulges, redistributing its volume and increasing its stiffness, and that change presses outward against anything wrapped around the limb. An array of force sensors held in a band, a cuff, or a prosthetic socket records the resulting radial pressure pattern, and a classifier maps that pattern onto a posture or a control command.

The technique belongs to force measurement instrumentation rather than to bioelectric recording. It registers a mechanical consequence of muscle activation instead of the electrical activity that produces it, which places it alongside methods such as mechanomyography and residual limb pressure mapping. Interest in the approach grew through the 2010s as low-cost force sensors and embedded classifiers made wearable pressure arrays practical, and a 2024 survey in Medical and Biological Engineering and Computing traces the consolidation of the terminology and the sensing conventions now in common use.

Sensor Arrays and Signal Acquisition

Most systems are built from force sensitive resistors, thin polymer elements whose resistance drops as normal force rises. Piezoelectric, capacitive, and optical alternatives exist, but resistive elements dominate because they are inexpensive, thin enough to sit inside a socket liner, and simple to read with a voltage divider and an analog-to-digital converter. Channel counts range from a single sensor placed over a specific muscle belly to bands of sixteen or more distributed around the forearm circumference. A low-density armband evaluated in Sensors showed that a small number of well-placed channels can approach the performance of denser arrays, which matters for cost and for battery life in a worn device.

Signal quality depends heavily on mechanical conditions that have no analogue in electrical recording. Strap tension sets the baseline pressure, so the offset shifts each time the band is removed and replaced. Limb posture changes the contact geometry, producing the limb-position effect, and slow tissue creep introduces drift over a session. Practical designs address these with recalibration routines, differential features that discard the common baseline, and inertial measurement units that supply arm orientation as an extra input to the classifier.

Pattern Recognition and Control

Raw sensor vectors are usually reduced to simple time-domain features and passed to a supervised classifier. Linear discriminant analysis, support vector machines, k-nearest neighbors, random forests, and small neural networks all appear in the literature, and offline accuracies above 90 percent for six to ten discrete hand gestures are routine. Regression models offer an alternative output, estimating continuous grip force or joint angle so that a prosthetic hand can close proportionally rather than snapping between preset grips. Research on force-myographic control of upper limb prostheses has emphasized that real-time performance, rather than offline classification accuracy, is the measure that determines whether a scheme is usable.

Relationship to Surface Electromyography

Surface electromyography remains the reference technique for prosthetic control, and force myography is best understood in contrast with it. Pressure sensing is unaffected by electrode-skin impedance, sweat, and power line interference, and it needs no instrumentation amplifier or high-gain analog front end. The trade is temporal: electrical activity precedes mechanical deformation by the electromechanical delay, so a pressure signal arrives a few tens of milliseconds later and carries less information about fine, low-force activation. Studies of assistive hand orthosis control report that the two modalities are complementary, and hybrid systems that fuse both often outperform either alone.

Applications

Force myography has applications in a range of fields, including:

• Control of transradial and transhumeral hand prostheses
• Robotic hand orthoses and rehabilitation exoskeletons
• Gesture recognition for wearable human-computer interfaces
• Clinical assessment of grip strength and motor recovery
• Ergonomic and sports monitoring of muscle loading