Improving the function of artificial arms remains a challenge, especially for amputations at the elbow or higher, where the disability is greatest. Motorised hooks, hands, wrists, and elbows are available, but existing methods of control are inadequate. Currently, most powered artificial limbs are controlled with the surface electromyogram (myoelectric signals) from a remaining pair of agonist-antagonist muscles in the amputated limb.1 This method allows only a single motion to be controlled at a time; operation of the prosthetic elbow, wrist and hand, or hook must be done sequentially. Furthermore, current methods of myoelectric control do not have a natural feel because proximal muscle functions (eg, shoulder, bicep, or triceps muscles) are not normally used to direct wrist or hand movements. Thus, these methods are frustratingly slow and awkward. Furthermore, current prostheses have no intrinsic sense of touch and provide little sensory feedback to the user. They are instead operated only with visual feedback.