Targeted muscle reinnervation (TMR) is a surgical procedure which allows to restore myoelectric control sources in people with proximal upper-limb amputations. However, the large physical displacement generally provoked by the reinnervated muscles following contraction can represent a drawback for the use of surface electrodes, which are affected by movement artifacts. In this regard, the viability of directly exploiting the physical displacement of muscles as control source would be beneficial. We recently introduced the so called <i>myokinetic interface</i>, aimed at transducing muscle movements into decipherable signals for artificial hands by tracking magnetic markers implanted inside the muscles. This work features the combination of the TMR procedure with such interface, in a non-invasive way. Two participants who underwent TMR surgery following above-the-elbow amputation were enrolled in this study. During two experimental sessions, we assessed the feasibility of: (i) disentangling muscle displacements associated with different degrees of freedom (DoFs) of the missing limb through video analysis, and (ii) discriminating different DoFs by tracking the displacement of five magnetic markers placed on the skin, over the reinnervated sites. A simple logistic regressor proved able to discriminate three different DoFs (six movements), with an average F1-score among classes and testing conditions of 0.84 (0.65) and 0.69 (0.60) for the video and the myokinetic data, for the first (second) participant, respectively. The presented outcomes encourage further investigations, and pave the way towards novel control strategies for artificial hands in TMR patients.