Brain-Computer Interfaces (BCIs) constitute a promising tool for communication and control. However, mastering non-invasive BCI systems remains a learned skill difficult to develop for a non-negligible proportion of users. Even though similarities have been shown between MI-based BCI learning and motor sequence learning our understanding of the dynamical processes, and their reflection on brain signals during BCI performance is still incomplete. In particular, whole-brain functional imaging is dominated by a 'bursty' dynamics, "neuronal avalanches", with fast, fat-tailed distributed, aperiodic perturbations spreading across the whole brain. Neuronal avalanches evolve over a manifold during resting-state, generating a rich functional connectivity dynamics. In this work, we evaluated to which extent neuronal avalanches can be used as a tool to differentiate mental states in the context of BCI experiments.