Distinct neural representations during a brain-machine interface and manual reaching task in motor cortex, prefrontal...

Area of Science:

• Neuroscience
• Neural Engineering
• Primate Motor Control

Background:

• Brain-machine interfaces (BMIs) traditionally focus on local neural populations.
• Distributed cortical and subcortical networks are crucial for BMI learning and control.
• The prefrontal cortex's role in motor BMI control remains understudied despite its known functions in action planning and learning.

Approach:

• Simultaneous local field potential recordings from the primary motor cortex (M1), dorsolateral prefrontal cortex (DLPFC), and caudate nucleus (Cd) in nonhuman primates.
• Comparison of neural activity during a 2D center-out task performed under both BMI and manual control.
• Analysis of distinct neural representations and effective connectivity between brain regions.

Key Points:

• Distinct neural representations for BMI and manual control were identified in M1, DLPFC, and Cd.
• DLPFC activity best distinguished control types at the go cue, while M1 activity was key at target acquisition.
• Effective connectivity was observed from DLPFC to M1 across both control types, and from Cd to M1 during BMI control.

Conclusions:

• The findings suggest a distributed network involving M1, DLPFC, and Cd is active during BMI control.
• This network's activity is similar yet distinct compared to manual control.
• Highlights the importance of considering prefrontal cortex and striatal contributions to motor BMI control.