The proposed research aims to combine the advantages of human and robotic post-stroke gait rehabilitation in a new approach called cyberphysical therapy (CBT). This paradigm allows therapists to remain actively involved in gait training while removing barriers limiting the effectiveness of traditional human-provided therapy: high effort and poor real-time feedback. In this innovative approach a therapist interacts with patients through coupled robotic arms that act as an extension of the therapist’s arm; this rehabilitative collaboration can take place either face-to-face or remotely by teleoperation. The central hypothesis is that combining the advantages of human and robotic gait rehabilitation approaches in CBT will significantly improve patient learning outcomes compared to traditional therapist-only and automated robotic gait training.