Decoding Synergy-Based Hand Movements using Electroencephalography.

Area of Science:

• Neuroscience
• Robotics
• Biomechanics

Background:

• Understanding the neural basis of human movement is crucial for developing advanced prosthetics and brain-computer interfaces.
• Electroencephalography (EEG) offers a non-invasive method to capture brain activity related to motor control.
• Kinematic synergies represent coordinated muscle activations underlying complex movements like grasping.

Purpose of the Study:

• To investigate if EEG signals can decode kinematic synergies during hand grasping.
• To establish a predictive model linking EEG features to movement reconstruction weights.
• To assess the feasibility of using EEG for real-time decoding of hand movements.

Main Methods:

• Recorded scalp EEG signals and CyberGlove kinematic data from 10 subjects performing various hand grasps.
• Determined kinematic synergies and their reconstruction weights from a training data subset.
• Employed multivariate linear regression to train EEG features (power spectral densities) on synergy weights, followed by 3-fold cross-validation for decoding.

Main Results:

• EEG features were successfully used to decode kinematic synergy weights from unseen grasp data.
• The decoding accuracy significantly surpassed chance levels, indicating a genuine relationship between EEG and movement generation.
• Multivariate linear regression effectively modeled the connection between spectral EEG features and kinematic parameters.

Conclusions:

• Scalp EEG signals contain sufficient information to decode synergy-based hand movement generation.
• The developed model demonstrates the potential for EEG-based control of robotic or prosthetic limbs.
• Future research should explore more complex movements and advanced decoding algorithms for enhanced brain-computer interface applications.