In this paper, an experimental framework to perform closed-loop experiments is presented, in which information processing (i.e., coding and decoding) and learning of neuronal assemblies are studied during the continuous interaction with a robotic body.