Gesture Decoding Using ECoG Signals from Human Sensorimotor Cortex: A Pilot Study.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electrocorticography (ECoG) shows promise for brain-machine interfaces (BMIs).
• Large craniotomies for ECoG grid implantation pose clinical translation challenges.
• Developing minimally invasive ECoG-based BMIs is crucial for wider adoption.

Purpose of the Study:

• To investigate the feasibility of using a reduced and clustered set of ECoG channels for real-time hand gesture decoding.
• To assess the decoding accuracy of a brain-machine interface system using ECoG signals.
• To determine optimal channel selection strategies for ECoG-based BMI.

Main Methods:

• Collected clinical ECoG signals from the sensorimotor cortex of three epilepsy patients performing hand gestures.
• Extracted ECoG power spectrum in hybrid frequency bands to build a real-time BMI system.
• Utilized a greedy algorithm for channel selection and analyzed channel distribution.

Main Results:

• Achieved high offline decoding accuracy for three hand gestures (85.7%, 84.5%, 69.7%).
• Demonstrated significant online decoding accuracy in two participants (80%, 82%).
• Found that decoding performance was maintained with a subset of channels, primarily clustered along the central sulcus.

Conclusions:

• Reduced and clustered ECoG channel distribution supports the clinical implementation of ECoG-based BMIs for hand gesture control.
• Minimally invasive ECoG BMI systems are feasible for controlling hand movements.
• This research paves the way for more practical ECoG BMI applications.