A postural unloading task to assess fast corrective responses in the upper limb following stroke

Area of Science:

• Neuroscience
• Robotics
• Biomechanics

Background:

• Robotic technologies offer novel methods for assessing brain function through human behavior analysis.
• Previous robot-based postural Load Tasks revealed significant arm corrective response impairments in stroke survivors but suffered from high variability in healthy controls.
• This variability in healthy performance complicated the identification of stroke-related deficits.

Purpose of the Study:

• To develop and validate an intuitive "Unload Task" using robotic technology.
• To hypothesize that the Unload Task would decrease performance variability in healthy individuals.
• To improve the detection of motor impairments following stroke.

Main Methods:

• Compared performance on Load and Unload Tasks using the KINARM exoskeleton robot in 107 healthy controls and 31 stroke participants.
• Participants maintained a cursor in a target while the robot applied or removed unexpected loads.
• Kinematic parameters quantified performance; impairment was defined as performance outside the 95% confidence interval of controls, adjusted for covariates.

Main Results:

• The Unload Task demonstrated reduced performance variability and less skewed distributions in healthy controls compared to the Load Task.
• Fewer outliers were removed for normative model development with the Unload Task (3.7%) versus the Load Task (7.4%).
• Stroke participants showed significantly higher impairment rates on the Unload Task (68%) compared to the Load Task (23%) for the affected arm, with more impairments detected at the parameter level (median 52% vs. 29%).

Conclusions:

• The Unload Task serves as a superior method for evaluating rapid arm corrective responses.
• Stroke survivors exhibit impaired corrective responses, often affecting both arms.
• Deficits in rapid motor correction generation may increase fall risk in individuals with stroke.