Decoding the rat forelimb movement direction from epidural and intracortical field potentials.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Neural Engineering

Background:

• Brain-machine interfaces (BMIs) utilize neural signals for device control, with various signal types explored.
• Intermediate-level signals like subdural field potentials show promise for BMI applications.
• Understanding the information content of different signal types is crucial for specific BMI functions.

Purpose of the Study:

• To evaluate the performance of epidural field potentials (EFPs) and local field potentials (LFPs) in decoding reach direction.
• To compare the efficacy of EFP and LFP signals for brain-machine interface applications.

Main Methods:

• Rats were trained to control a joystick for a two-dimensional reach task.
• Forelimb reach direction was decoded from intracortical (LFPs) and epidural (EFPs) field potentials using linear discriminant analysis.
• Signal quality and decoding accuracy were assessed over time.

Main Results:

• Epidural field potentials (EFPs) achieved a mean decoding accuracy of 69 ± 3%.
• Local field potentials (LFPs) achieved a mean decoding accuracy of 57 ± 2%.
• Both signal types demonstrated performance significantly better than chance, with good signal quality maintained for up to 13 months.

Conclusions:

• Epidural field potentials (EFPs) provide high-quality inputs for brain-machine interfaces.
• EFPs offer a potentially less invasive and effective alternative to intracortical recordings for BMI applications.
• This study highlights the potential of epidural signals for robust neural decoding.