Neural correlates of skill acquisition with a cortical brain-machine interface

Area of Science:

• Neuroscience
• Biomedical Engineering
• Robotics

Background:

• Brain-machine interfaces (BMIs) allow real-time control of prosthetic devices through neural signal modulation.
• Cortical BMI studies indicate performance improvements require learning and are linked to altered neuronal tuning.
• Long-term practice leads to sustained improvements in BMI task performance.

Purpose of the Study:

• To investigate long-term skill acquisition in brain-controlled (BC) prosthetic devices.
• To characterize the neural correlates associated with improved task performance over time.
• To assess the consolidation of cortical representations for neuroprosthetic control.

Main Methods:

• Monkeys performed a continuous-control, multistep center-out task using a BC computer cursor.
• Daily performance trends and neural correlates were monitored under varying experimental conditions.
• Researchers compared performance with daily recalibration versus a fixed neural transform across days.

Main Results:

• Variable daily performance was observed under conditions with daily recalibration of the neural transform.
• Consistent long-term skill acquisition was evident when a fixed transform was applied to stable neural recordings.
• Skill acquisition correlated with the "crystallization" of a cortical map for prosthetic control.

Conclusions:

• The primate motor cortex demonstrates the capacity for skilled control of neuroprosthetic devices.
• Long-term skill acquisition is facilitated by the consolidation of cortical representations.
• Stable neural recordings and fixed transforms support the development of robust neuroprosthetic control.