Decoding 3-D reach and grasp kinematics from high-frequency local field potentials in primate primary motor cortex

Area of Science:

• Neuroscience
• Neural Engineering
• Biomedical Engineering

Background:

• Intracortical microelectrode arrays record neural signals for neural interfaces.
• Previous research focused on single/multiunit activity and low-frequency LFPs.
• High-frequency LFPs (>200 Hz) and their decoding potential for 3-D kinematics remain underexplored.

Purpose of the Study:

• To investigate the information content of various LFP frequency bands for 3-D reach and grasp kinematics.
• To demonstrate the feasibility of decoding 3-D reach and grasp movements using LFPs.
• To identify optimal LFP frequency bands for neural interface applications.

Main Methods:

• Recorded LFPs using 96-microelectrode arrays in the primary motor cortex (M1) of two monkeys.
• Utilized mutual information and decoding analyses to assess information content across seven LFP frequency bands (0.3-400 Hz).
• Employed Kalman filters for decoding reach and grasp kinematics.

Main Results:

• Higher frequency LFP bands (100-200 Hz and 200-400 Hz) contained the most kinematic information.
• Broad-band high-frequency LFPs, potentially reflecting multiunit activity, yielded the best decoding performance.
• Accurate reconstruction of reach kinematics, grasp aperture, and aperture velocity was achieved.

Conclusions:

• High-frequency LFPs are valuable signals for neural interfaces.
• LFPs, particularly in higher frequency bands, can effectively control 3-D reach and grasp kinematics.
• This study highlights the potential of LFPs for sophisticated neural control systems.