Background

Two-thirds of stroke survivors are permanently disabled and require extensive rehabilitation. There is an urgent need for personalized robot-aided rehabilitation systems to improve the therapy provided to stroke survivors and relieve caregivers from physically demanding work.

After a stroke, patients exhibit stereotypical synergies, resulting in a preference for specific postures that they are unable to overcome independently. Other patients suffer from coordination and dexterity issues during movements.

One of my main research interests is to identify quantifiable measures of human arm motions. I have a long-term interest of using these measures to 1) quantify the recovery state of a patient and 2) using robots to training care recipients in restoring spatial actions or to serving as “hands-on” assistants to physiotherapists.

Decoding “healthy” arm motions

While traditional neuromotor control studies use basic mathematical tools to quantify neuromotor behavior, I want to use Differential Geometry where simple (Euclidean) metrics fail. My current experiments assess whether healthy subjects rely on optimized arm motions, such as minimum energy trajectories.

What we learn from these experiments can be used for robot-aided stroke rehabilitation. By statistically assessing the differences between healthy and impaired arm motions, a measure of the recovery state of the patient can be established. This measure will serve as an objective indicator of the patient’s progress throughout the rehabilitation process.

---