Parallel Spatial-Temporal Self-Attention CNN-Based Motor Imagery Classification for BCI.

Area of Science:

• Neuroscience and Artificial Intelligence
• Brain-Computer Interface (BCI) technology
• Signal Processing

Background:

• Motor imagery (MI) electroencephalography (EEG) classification is crucial for brain-computer interfaces (BCIs), enabling communication for individuals with mobility impairments.
• Decoding EEG signals is challenging due to signal complexity, dynamic nature, and low signal-to-noise ratio.
• Existing frameworks struggle to fully extract high-level features from EEG signals.

Purpose of the Study:

• To develop an end-to-end deep learning framework for robust four-class MI EEG signal classification.
• To introduce a novel spatial-temporal representation of EEG signals utilizing self-attention mechanisms.
• To enhance feature extraction for improved BCI performance.

Main Methods:

• A parallel spatial-temporal self-attention-based convolutional neural network (CNN) was designed.
• Spatial self-attention module captures inter-channel dependencies, improving accuracy and reducing artifacts.
• Temporal self-attention module extracts high-level temporal features from EEG signals.

Main Results:

• The proposed method achieved superior performance in both intra-subject and inter-subject MI EEG classification compared to state-of-the-art approaches.
• Qualitative analysis confirmed the effectiveness of the learned spatial-temporal representation.
• The system demonstrated feasibility for real-time drone control via EEG signals.

Conclusions:

• The developed parallel spatial-temporal self-attention network offers a robust and effective solution for MI EEG classification.
• This approach significantly advances BCI capabilities by improving EEG signal decoding.
• The framework shows promise for real-world applications, including drone control systems.