In 1963, Stan Lee introduced a new villain to the Spiderman Universe: Dr Octopus – a human equipped with multiple robotic arms that can be controlled seamlessly in coordination with his natural limbs. Throughout the last decades, turning such fiction into real-life applications gave rise to the research field of human motor augmentation, ultimately aiming to enable humans to perform motor tasks that are sheer impossible with our natural limbs alone. While a significant process was made in designing artificial supernumerary limbs, a central problem remains: identifying adequate bodily signals that allow moving supernumerary degrees-of-freedom together with our natural ones. So far, neural activity in the brain seems to hold the greatest potential for providing all the flexibility needed to ensure such coordination between natural and supernumerary degrees-of-freedom. However, accessing neural populations in the cortical regions is accompanied by an unacceptable risk for most users. A different group of neural cells can be found in the outmost layer of the motor pathway, driving the contraction of muscles and generation of force – spinal motor neurons. The development of novel neural interfaces has made it possible to study single motor neuron activity with minimal harm to the user. This allows a direct and non-invasive window into the neural activity orchestrating human movement. In this dissertation, I investigate whether these neurons innervating our muscles could provide supernumerary control signals. The results indicate, in essence, that features extracted non-invasively from motor neuron activity have the potential to overcome current limitations in supernumerary control and thus could significantly advance human motor augmentation.