Continuous decoding of human grasp kinematics using epidural and subdural signals

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Restoring function in paralyzed individuals is a key goal for brain-machine interfaces (BMIs).
• Regaining hand function is a major objective for patients with paralysis.
• Accurate movement control and minimized invasiveness are crucial for widespread adoption of hand neuroprostheses.

Purpose of the Study:

• To explore the trade-offs between decoding accuracy and invasiveness in hand movement control.
• To compare the efficacy of subdural (electrocorticography; ECoG) and epidural field potentials (EFPs) for decoding continuous hand kinematics.
• To assess the impact of electrode array resolution on decoding performance.

Main Methods:

• Measured continuous hand and finger kinematics during naturalistic grasping motions in five human subjects.
• Recorded both subdural ECoG and epidural field potentials (EFPs) using standard- and high-resolution electrode arrays.
• Analyzed decoding accuracy of kinematics from recorded neural signals.

Main Results:

• Decoding of continuous kinematics significantly exceeded chance for both ECoG and EFPs.
• ECoG decoding accuracy was comparable to previous studies (0.54 ± 0.05 variance accounted for).
• EFP decoding performance was generally comparable to ECoG, with higher-resolution arrays yielding superior results. The 70-115 Hz spectral band was most informative for subtle movements.

Conclusions:

• High accuracy in decoding motor intent, necessary for fine motor control, may be achievable with both ECoG and EFPs.
• Epidural electrode placement is less invasive and carries lower risks than subdural placement.
• Accurate decoding of motor commands using EFPs represents a significant step towards clinically viable BMIs for restoring hand function.