Motion Cognitive Decoding of Cross-Subject Motor Imagery Guided on Different Visual Stimulus Materials

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Motor imagery (MI) is crucial for brain-computer interfaces (BCIs) by evoking electroencephalogram (EEG) event-related desynchronization/synchronization (ERD/S) rhythms.
• Subjectivity in MI tasks leads to individual EEG variations, complicating motion cognitive decoding.

Purpose of the Study:

• To investigate the impact of different visual stimuli on MI-EEG responses.
• To develop and evaluate a novel cross-subject MI-EEG classification method.

Main Methods:

• Designed three visual stimuli (arrow, human, robot) for three MI tasks (left arm, right arm, feet).
• Employed covariance matrix centroid alignment for EEG preprocessing.
• Utilized model-agnostic meta-learning for cross-subject MI-EEG classification.

Main Results:

• Robot stimuli outperformed arrow and human stimuli in MI-EEG decoding.
• Achieved an optimal cross-subject motion cognitive decoding accuracy of 79.04% with the proposed method.
• Demonstrated robust classification and superior performance compared to conventional methods.

Conclusions:

• Robot visual stimuli are more effective for MI-based BCIs.
• The proposed cross-subject classification method offers a robust and effective approach for decoding MI-EEG signals.
• This research advances personalized BCI development by addressing individual response variations.