Robotics and upper limb rehabilitation after stroke Thierry Lejeune Collaborator: S. Dehem, M. Gilliaux, G. Stoquart, C. Detrembleur Institut de Recherche Expérimentale et Clinique, Neuro Musculo Skeletal Lab NMSK, Brussels, Belgium Cliniques universitaires Saint-Luc, Service de médecine physique et réadaptation, Brussels, Belgium This talk will be divided in two parts: a general introduction about robotic assisted upper limb rehabilitation after stroke; and, the presentation of the result of a RCT. Introduction/Background Upper limb (UL) robotic assisted therapy (RAT) is promising for stroke rehabilitation. It allows to apply motor relearning principles. Systematic reviews show effectiveness at the at the early and chronic stage. In addition to conventional therapy (CT), RAT has been shown to decrease UL impairment. When RAT is provided as partial substitution of CT, it is supposed to be at least as effective as CT. A few studies demonstrated that RAT might be more effective than CT. Its effectiveness has been less studied in the acute phase than in the chronic one. Assessments have usually been restricted to the two first domains of the International Classification of Functioning framework (ICF). RCT Aim: To perform a pragmatic multicentric single-blind randomized controlled trial to evaluate UL RAT effectiveness, used as partial substitution to CT, in acute stroke patients, following the three ICF domains. Methods: Forty-five acute stroke patients were randomized into two groups (CT group, n=22 and RAT group, n=23). Both interventions were dose-matched about treatment duration and lasted nine weeks. The CT group followed a standard rehabilitation. In the RAT group, four sessions of CT (25%) were substituted by RAT each week. RAT consisted of moving the paretic UL along a reference trajectory while the robot provided an assistance as needed. A blinded assessor evaluated patients before, just after intervention and at 6-month post-stroke according to ICF. Results: Twenty-eight patients were assessed after intervention. UL motor control (Fugl Meyer +16%; p=0.05) and gross manual dexterity (Box and Block test +10 blocks; p=0.02) improved significantly more in RAT group than in CT group at 6-month post-stroke. Ability to perform activities and social participation tended also to improve more in RAT group (p=0.06-0.16). Conclusion: RAT during acute stage of stroke rehabilitation is more effective than CT to improve UL motor control and gross manual dexterity, and is at least as effective as CT for activity and participation.