Loss of an arm makes it harder to work, to take part in leisure activities, and to do many of the tasks needed to lead an independent life. Using an artificial arm (prosthesis) can help with these activities, and a lot of research has gone into designing advanced prostheses that can replace the many functions of an intact hand and arm. However, a prosthesis is typically attached to the body using a socket, which fits over the residual limb. Sockets can cause problems with skin irritation and can be uncomfortable or hot to wear for long periods. One way around this is to attach the prosthesis directly to the bone of the residual limb. This solves the problems associated with wearing a socket, allows a wider range of movement, and the makes the prosthesis feel more like a part of the user’s body. The prosthesis is attached to a metal rod that is inserted into the arm bone during a two-step surgery. The bone then grows into the metal rod, forming a strong permanent attachment. This process is called osseointegration (OI). Another problem is how to control the prosthesis, i.e., how to make it do what the user wants it to do. Some prostheses are controlled using small electrical signals (called EMG signals) that are produced by muscles when they contract. EMG signals are usually recorded by electrodes placed on the skin over the muscle; the signals are decoded by a computer algorithm and turned into control signals for the prosthesis. EMG signals recorded from the skin surface can be unreliable, so another option is to implant the electrodes onto the muscle. A system called Osseointegration of Prostheses for the Rehabilitation of Amputees (OPRA) uses OI to attach the prosthesis. A new version of this system, e-OPRA, uses OI with implanted electrodes. e-OPRA can also be used to stimulate peripheral nerves so that the user feels some sensory feedback from the prosthesis, which also may help with control of the device. Our objective is to perform two clinical trials to first test whether e-OPRA provides better function and comfort than OPRA, and second to find out whether e-OPRA is better with or without sensory feedback. Our rationale is that directly attaching the prosthesis to the skeleton has many advantages over use of a socket, that implanted electrodes will provide cleaner, more stable and reliable EMG signals than electrodes at the skin surface, and that sensory feedback will help the user control their device. We therefore expect that the e-OPRA system with sensory feedback will provide improved prosthesis control and function compared to conventional socket-based systems or OI without sensory feedback.