Reinforcement Learning is Impaired in the Sub-acute Post-stroke Period.

Area of Science:

• Neuroscience
• Motor Control
• Stroke Rehabilitation

Background:

• Spontaneous motor recovery after stroke primarily occurs within 3 months.
• Heightened neural plasticity in early recovery may overlap with motor learning plasticity.
• Animal models suggest increased training responsiveness in early post-stroke periods.

Purpose of the Study:

• To investigate if motor learning sensitivity, specifically reinforcement and error-based learning, is enhanced early after stroke.
• To compare learning processes in individuals tested within 3 months (early group) versus over 6 months (late group) post-stroke.
• To understand how different motor learning mechanisms are affected during the critical early recovery phase.

Main Methods:

• A cross-sectional design comparing two groups of stroke survivors (early vs. late).
• Assessment of reinforcement and error-based motor learning using a visuomotor rotation task.
• Ensured matched motor execution, cognitive function, and demographic factors between groups.

Main Results:

• Reinforcement learning was significantly impaired in the early stroke group (within 3 months).
• Error-based learning remained unimpaired in both early and late stroke groups.
• Findings were independent of lesion characteristics, cognitive status, or demographics.

Conclusions:

• A deficit in reinforcement motor learning exists in the early subacute phase after stroke.
• Rehabilitation strategies may need modification in the early period, potentially by adjusting reinforcement feedback or training type.
• Consideration should be given to delaying therapies reliant on reinforcement learning, like constraint-induced movement therapy, in the early subacute phase.