Cell replacement therapies aim at reestablishment of neuronal circuits after brain injury, stroke or neurodegeneration. Recently, direct reprogramming of resident glial cells into the affected neuronal subtypes has become a feasible and promising option for central nervous system regeneration. Direct reprogramming relies on the implementation of a new transcriptional program defining the desired neuronal identity in fully differentiated glial cells implying the more or less complete down-regulation of the program for the former identity of the glial cell. Despite the enormous progress achieved in this regard with highly efficient in vivo reprogramming after injury, a number of hurdles still need to be resolved. One way to further improve direct neuronal reprogramming is to understand the molecular hurdles which we discuss with the focus on chromatin states of the starting versus the reprogrammed cells.