Repetitive and restricted behaviours represent a common problem for various psychiatric syndromes, especially in autistic spectrum disorders, and they include a wide range of heterogeneous behavioural manifestations. An accurate and standardized description of these behaviours is needed to advance the understanding of this complex and heterogeneous clinical dimension of autism. The present article reports the reliability and validity studies of a new assessment scale: the repetitive and restricted behaviour scale. 145 subjects with autism spectrum disorders were assessed using the RRB scale. The RRB scale has good interrater reliability, internal consistency and content validity. Factorial analysis produced four clinically meaningful factors, i.e. "sensorimotor stereotypies", "reaction to change", "restricted behaviours" and "modulation insufficiency". The RRB scale has good psychometric qualities and constitutes a real breakthrough towards a neurofunctional approach to autistic disorders. It should be valuable for research and treatment, and in clinical practice.