Decoding grasp types with high frequency of local field potentials from primate primary dorsal premotor cortex

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Local field potentials (LFPs) offer stable signals for brain-machine interfaces (BMIs).
• Previous studies utilized LFPs for arm and hand movement decoding.
• The efficacy of different LFP frequency bands for decoding hand grasp types remains underexplored.

Purpose of the Study:

• To investigate the decoding performance of LFPs in various frequency bands for hand grasp types.
• To identify optimal LFP characteristics for enhancing BMI functionality in grasp decoding.

Main Methods:

• Collected LFPs from monkey dorsal premotor cortices (PMd) using microelectrode arrays during grip-specific tasks.
• Employed a K-nearest neighbor classifier on LFP power spectrum to decode grasping movements.
• Analyzed decoding performance across different LFP frequency bands, channels, and training trials.

Main Results:

• The broad high-frequency band (200-400Hz) LFPs demonstrated superior performance in decoding grasp types.
• Achieved classification accuracy exceeding 0.9 using high-frequency LFPs.
• Identified specific LFP frequency bands, channels, and trial configurations impacting decoding accuracy.

Conclusions:

• High-frequency LFPs in the dorsal premotor cortex are a viable and effective control signal source for functional BMIs.
• These findings suggest potential for improved hand grasp control in future BMI applications.
• Further research into LFP frequency band optimization can advance BMI technology.