Roughly 5 million people worldwide are left with long-term impairments each year following a stroke. Motor rehabilitation, essential for regaining mobility and independence, typically requires sustained, repetitive sessions with a human therapist — a resource increasingly scarce as physical therapist shortages widen relative to patient demand.
Addressing that gap is the focus of work by Johannes Lachner, a former MIT postdoctoral researcher now an assistant professor, who is developing a system that pairs robotics with artificial intelligence to support, rather than replace, human-led therapy. The underlying logic is to extend a therapist's reach: a robotic device could handle repetitive rehabilitation exercises, freeing clinicians to focus their attention on cases requiring nuanced clinical judgment.
The project fits within a broader trend in medical robotics research, where AI is used to adjust the intensity and type of movement exercises in real time, based on a patient's progress and physical constraints. The technical challenge lies as much in the mechanical precision of the robotic actuators as in the system's ability to interpret often ambiguous physiological signals, while avoiding the kind of imprecision or bias that can affect automated health assessments.
While full technical specifics of the project are not yet detailed, the initiative points to a plausible direction for digital health: using AI not to fully automate a medical procedure, but to ease the logistical and human burden on an overstretched care system. Such approaches will still need to prove their clinical effectiveness at scale before they can meaningfully complement, let alone relieve, current rehabilitation practices.