Motor deficit is a prominent feature among stroke survivors. Robot-assisted therapies have been proposed as a strategy to boost rehabilitation, byallowing therapy to be provided in a more reproducible and intense manner,while quantitatively monitoring patient's improvement. However, thoseapproaches have so far not shown superiority over conventional treatments. Onepotential solution to reach better outcomes would be to personalize the treatment.In this regard, a better understanding of the mechanisms underlying motorrecovery is pivotal to tailor therapy to each patient. Here, we explored the corticalchanges occurring during robotic training. We recorded resting-state fMRI beforeand after the treatment in three sub-acute post-stroke survivors, and we investi-gated the functional connectivity between motor regions. We observed a corticalreorganization following training, consistent with motor improvements