We aim to identify key quantitative parameters (robotic, kinematic, and physiological) that have the biggest impact on motor recovery. In other words, the purpose is to discover which aspects related to movement characteristics and heart activity undergo the greatest changes when a patients takes part to a specific rehabilitation therapy that includes also a robotic device. This will be achieved by analyzing data collected from the robotic rehabilitation device itself, Inertial Measurement Unit sensors, and wearable devices. 

Identifying and understanding the role of these key factors in the context of motor recovery prediction may help recognizing early signs of improvement and design therapies that better match each patient's unique needs.

Every stroke survivor’s recovery journey is different, depending on the severity of the motor issue and on the person's overall health condition and response to therapy. Using insights from AI predictions, we plan to tailor rehabilitation programs to each individual. This means adjusting how the robotic device interacts with the patient (such as changing resistance, speed, or range of motion), as well as the quantity, type and frequency of rehabilitation exercises to propose, based on the patient’s specific recovery patterns.

Our goal is to make therapy sessions more adaptive and effective, ensuring that each patient receives the right level of support at the right time.

We are developing decision-support tools that will help doctors and therapists make more informed choices about treatment. These tools, powered by interpretable AI models, will provide clear insights into a patient’s progress and suggest possible next steps in the rehabilitation plan. Rather than replacing human expertise, the AI will act as a trusted assistant, helping healthcare professionals make data-driven decisions that improve patient outcomes.

We will test whether wearable sensors can effectively be correlated to a patient’s progress in terms of upper limb motor recovery outside of the clinic.
Wearable sensors are devices that can be comfortably worn on the wrist and that can track everyday movements, stress, sleep quality, and heart-related features, providing continuous information about how recovery may be progressing in real life, not just during therapy sessions.
By integrating data collected at home with data from clinical environments, we will evaluate how reliable and useful home monitoring is in predicting recovery outcomes in addition to data collected in clinical settings.