Using a noninvasive decoding method to classify rhythmic movement imaginations of the arm in two planes.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-computer interfaces (BCIs) aim to restore movement for individuals with spinal-cord injury.
• Existing BCIs successfully use invasive methods for control and noninvasive methods for movement reconstruction.
• The noninvasive decoding of detailed imagined movements using electroencephalography (EEG) remains an open challenge.

Purpose of the Study:

• To investigate the feasibility of noninvasively decoding detailed imagined arm movements using EEG.
• To classify horizontal and vertical imagined rhythmic movements of the right arm.
• To advance BCI technology for potential application in neuroprosthetics.

Main Methods:

• Utilized electroencephalography (EEG) to record brain activity during imagined movements.
• Developed an experimental design to isolate imagined movements from muscle and eye artifacts.
• Employed partial least squares for decoder training and trajectory correlation analysis.
• Classified imagined movements based on the correlation between decoded and assumed movement trajectories.

Main Results:

• Successfully classified horizontal and vertical imagined rhythmic arm movements in 7 out of 9 healthy subjects using EEG.
• Demonstrated indirect decoding of imagined movements through a trajectory-based classification approach.
• Identified significant involvement of the supplementary motor area in imagined movement classification.

Conclusions:

• Noninvasive EEG-based BCI can classify detailed imagined movements in specific planes.
• This research represents a significant step towards developing advanced neuroprosthetics for movement restoration.
• The findings suggest the potential for sophisticated control of robotic limbs through imagined movements.