Reinforcement learning of self-regulated sensorimotor β-oscillations improves motor performance

Area of Science:

• Neuroscience
• Rehabilitation Medicine
• Human-Computer Interaction

Background:

• Self-regulation of sensorimotor oscillations is explored for neurorehabilitation, particularly in stroke patients.
• The potential for brain self-regulation to yield motor gains requires further investigation under specific training conditions.

Purpose of the Study:

• To investigate the conditions under which brain self-regulation training can lead to motor improvements.
• To determine the efficacy of contingent neurofeedback and adaptive classifier thresholding in learning brain self-regulation.

Main Methods:

• Thirty-two healthy subjects underwent a three-day intervention using a brain-robot interface (BRI) for kinesthetic motor-imagery.
• Four training conditions compared adaptive/non-adaptive thresholding and contingent/non-contingent feedback.
• Electroencephalography (EEG) monitored cortical activity, and behavioral performance was assessed in a subsequent motor task.

Main Results:

• Contingent neurofeedback and adaptive classifier thresholding were crucial for learning brain self-regulation.
• Successful self-regulation of sensorimotor beta-band desynchronization correlated significantly with motor performance gains.
• The intervention led to measurable behavioral improvements after training.

Conclusions:

• Operant learning of brain self-regulation using a BRI is effective in enhancing motor skills.
• This neurofeedback approach holds therapeutic potential for stroke patients with limited hand function.