A prosthetic hand replaces a missing hand by translating some other signal, whether shoulder movement, muscle contractions, or even brain activity, into grip and release actions. The specific mechanism depends on the type of prosthesis, but all designs solve the same core problem: reading the user’s intention and converting it into mechanical motion at the fingers.
Body-Powered Prosthetics: Cable and Harness
The simplest and most time-tested design is the body-powered prosthesis. A steel cable runs from the terminal device (the hook or hand at the end) up through the arm and connects to a harness wrapped around the opposite shoulder. When you move your shoulder forward, shrug, or extend your arm, the cable pulls tight and opens or closes the hand. One of those actions is under your voluntary control, while the other happens automatically through a spring that returns the device to its resting position.
This direct mechanical link means every force the hand exerts travels back through the cable to your shoulder. That’s actually an advantage: you can feel resistance when gripping an object, giving you a rough sense of how hard you’re squeezing without any electronic sensors. The tradeoff is that operating the hand requires deliberate body movements, and learning to coordinate those motions smoothly takes practice. Users have to build a new relationship between shoulder positioning and hand function, essentially training their body to map familiar movements onto an unfamiliar tool.
Myoelectric Prosthetics: Reading Muscle Signals
Myoelectric hands use electrical signals from your remaining muscles instead of cables. When you flex a muscle, motor neurons fire in patterns that produce tiny electrical signals at the skin’s surface. Sensors embedded in the prosthetic socket detect these signals and feed them to a processor, which interprets your intention and drives small motors in the hand.
In a basic two-sensor setup, one muscle site controls opening and another controls closing. You might flex the muscles on top of your forearm to open the hand and the muscles underneath to close it. The strength of the contraction can control speed or grip force: a gentle flex closes the hand slowly, while a strong flex snaps it shut.
More advanced systems use pattern recognition. Instead of relying on just two sensor sites, these systems place multiple electrodes across the residual limb and record the full landscape of muscle activity. During a training session, you perform a series of specific movements, and the software learns the unique electrical fingerprint of each one. It calculates features like the average signal strength and waveform shape across all channels, then builds a classification model. After training, the system reads your muscle signals in real time and matches them to the most likely intended movement. This approach lets a single prosthesis respond to many distinct commands rather than just open and close, giving access to different grip types without manually switching modes.
Grip Patterns in Multi-Articulating Hands
Modern powered hands have individually motorized fingers, which means they can form a variety of grip shapes rather than just clamping open and shut. The standard set of pre-programmed grips typically includes:
- Cylindrical grip: all fingers wrap around an object like a water bottle or handrail
- Tripod pinch: thumb, index, and middle finger meet at a point for holding a pen or picking up a marble
- Lateral pinch: the thumb presses against the side of the index finger, the way you’d hold a key or a credit card
- Hook grip: fingers curl without the thumb, for carrying a bag handle
- Pulp pinch: thumb and index fingertip press together for small, precise objects
Users switch between grips through muscle signals, button presses, or app controls depending on the device. With pattern recognition systems, different muscle contractions can trigger different grips directly, making transitions faster and more intuitive.
How the Hand Attaches to the Body
Most prosthetic hands connect through a socket, a custom-molded shell that fits snugly over the residual limb. The socket is the foundation: it holds the prosthesis in place and transfers forces between your body and the device. But sockets come with well-documented problems. Up to three-quarters of lower-extremity amputees experience skin ulcers, excessive sweating, or fit issues due to the residual limb changing size over time. While that statistic comes from leg prostheses, upper-limb socket users face similar skin irritation and comfort challenges.
An alternative called osseointegration skips the socket entirely. In this approach, a metal implant is anchored directly into the residual bone, and a small connector passes through the skin to attach to the prosthesis externally. Users who switch from sockets to osseointegrated attachments consistently report dramatic improvements: in one study, quality-of-life scores nearly doubled, and patients described the prosthesis as feeling like “part of me” rather than something strapped on. Donning and removing the prosthesis becomes quicker, and the direct skeletal connection provides better proprioceptive feedback, a more natural sense of where the limb is in space. The procedure does carry surgical risks, including infection at the skin opening, so it’s typically considered when socket problems become unmanageable.
Sensory Feedback: Feeling Through a Prosthesis
One of the biggest limitations of any prosthetic hand is the loss of touch. Without sensation, you have to watch your hand constantly to avoid crushing a paper cup or letting a slippery object fall. Several feedback technologies aim to close that gap.
Non-invasive methods stimulate the skin on a part of the body you can still feel. Vibrotactile feedback uses small vibrating motors, similar to a phone’s haptic motor, placed against the skin to signal grip pressure. These are compact and energy-efficient but offer limited spatial detail, and users can become desensitized to the vibration over time. Mechanotactile systems apply direct pressure or skin stretch, more closely mimicking natural touch but adding bulk and mechanical complexity. Electrotactile feedback sends controlled electrical pulses through skin electrodes to stimulate sensory nerves, offering precise and adjustable signals but requiring careful calibration to avoid discomfort.
Invasive approaches go further. Electrodes implanted directly into peripheral nerves can deliver sensory information straight to the nervous system. Research has shown that this kind of biomimetic nerve stimulation improves object manipulation and grasp control, because the brain receives signals that more closely resemble natural touch. These systems are still primarily used in research settings, but they represent the closest current technology has come to restoring genuine hand sensation.
Surgical Techniques That Improve Control
For people with above-elbow amputations, there may not be enough remaining forearm muscle to generate clear signals for a myoelectric hand. A surgical procedure called targeted muscle reinnervation (TMR) addresses this by rerouting nerves that originally controlled the hand. The surgeon transfers those nerves to muscles in the residual limb or chest. Once the nerves grow into their new muscle targets, the patient can generate distinct electrical signals simply by thinking about hand movements. Each reinnervated muscle site becomes an independent control channel, giving the prosthesis more commands to work with and enabling more natural, intuitive operation.
Brain-Computer Interfaces
The most experimental control method bypasses muscles altogether. Brain-computer interfaces use microelectrode arrays implanted in the brain’s motor cortex to record neural activity directly. A processor interprets those signals in real time and translates them into commands for the prosthetic hand. This gives users the ability to control individual finger movements with a level of precision that muscle-based systems struggle to match.
The technology has shown real results in clinical trials, with paralyzed patients using implanted arrays to operate robotic arms with multiple degrees of freedom. But significant barriers remain. The surgery is invasive, implanted electrodes can degrade over time as the body’s immune response builds scar tissue around them, and the systems currently require external computing hardware. For now, brain-computer interfaces are a research tool rather than a clinical option for everyday prosthetic use, but they demonstrate that direct neural control of artificial limbs is technically achievable.
Materials and Weight
Weight matters enormously in a device you wear all day. Heavier hands cause faster fatigue and increase the likelihood of abandoning the prosthesis. Modern hands use lightweight materials like carbon fiber for structural components and silicone for cosmetic gloves that mimic the look of natural skin. 3D-printed hands made from PLA plastic have pushed weights remarkably low: a pediatric model developed at UC Davis weighs just 177 grams (about 6 ounces) while still producing enough grip force to handle everyday objects. Another 3D-printed design, the HANDi Hand, comes in at 256 grams. Commercial multi-articulating hands for adults are heavier, typically in the 400 to 600 gram range, but still far lighter than prosthetic hands from a generation ago.