Determining User Intent of Partly Dynamic Shoulder Tasks in Individuals With Chronic Stroke Using Pattern Recognition

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Biomedical Engineering

Background:

• Stroke is a primary cause of long-term disability, often leading to impaired arm function and abnormal muscle synergies.
• Current therapies offer limited improvement for paretic arm use, highlighting the need for advanced assistive technologies.
• Wearable exoskeletons offer potential for powered limb support but require sophisticated control systems.

Purpose of the Study:

• To investigate the efficacy of pattern recognition algorithms in identifying user intent for shoulder movements in individuals post-stroke.
• To assess the feasibility of using sensor data for controlling multi-degree-of-freedom shoulder tasks.
• To evaluate factors influencing classifier performance, including arm type, load, and sensor data modality.

Main Methods:

• Participants (stroke survivors and healthy controls) performed shoulder movement tasks (abduction, adduction, rotation) using a robotic system.
• Sensor data, including electromyography (EMG) and load cell data, were collected.
• Pattern recognition classifiers were trained and tested to identify user intent across various shoulder movement combinations.

Main Results:

• Classifier accuracy was reduced when using data from the paretic arm, lower load levels, or single sensor types (EMG or load cell alone).
• Despite these challenges, the pattern recognition approach demonstrated potential for identifying user intent.
• Further research is needed to optimize sensor selection and refine classifier performance during active device control.

Conclusions:

• Pattern recognition of sensor data holds promise for developing intuitive control schemes for arm exoskeletons in stroke rehabilitation.
• Optimizing sensor configurations and exploring user-classifier interactions are crucial next steps.
• This technology could significantly enhance functional recovery and independence for individuals affected by stroke.