"Biological signals recorded from the human body can be translated into actions of external devices to create man-machine interaction. This concept has clinical implications in rehabilitation technologies for replacing or recovering impaired motor functions. Among the possible biosignals for man-machine interaction (brain, nerve, and muscle signals), muscle signals, i.e. electromyography (EMG), are the only that allow applications in routine clinical use within a commercially reasonable time horizon. Although the current efforts in myoelectric interfaces are mainly focusing on decoding EMG signals, myoelectric interaction has the unique and little exploited feature of provoking changes in the neural circuits that are active during the interaction, i.e. of artificially inducing brain plasticity. However, current commercially viable myoelecric interfaces do not implement sensory-motor integration (decoding intentions and at the same time providing a sensory feedback to the patient), which conversely is the basis of plasticity of the central nervous system. This limit reflects the gap between academic research and the clinical and commercial needs. Myoelectric interfacing with sensory-motor integration is indeed feasible now if the knowledge from basic neurophysiology research and signal analysis in the academia is transferred to industrial sectors and if the requirements of and testing for clinical and commercial viability are transferred from the industry to academia. With a consortium of internationally regarded European academic teams and industries, we thus propose the implementation of sensory-motor integration into commercially viable myoelectric devices in two key clinical applications: 1) training for the active control of prostheses; and 2) rehabilitation of stroke patients with robotics. These two areas require a similar technological ground for sensory-motor integration and for artificial induction of neural plasticity, necessary to (re)learn motor tasks."