SSTAF: Spatial-Spectral-Temporal Attention Fusion Transformer for Motor Imagery Classification

2025-04-17

AI Summary

To address poor generalizability in cross-subject motor imagery EEG classification—caused by signal non-stationarity and inter-subject variability—this paper proposes an end-to-end attention-driven Transformer architecture. We innovatively design a spatial-spectral-temporal triple-attention fusion mechanism and, for the first time, integrate the short-time Fourier transform (STFT) directly into the Transformer encoder to jointly model and discriminatively enhance EEG time-frequency representations. Evaluated on the EEGMMIDB and BCI Competition IV-2a datasets, our method achieves cross-subject classification accuracies of 76.83% and 68.30%, respectively—outperforming state-of-the-art CNNs and existing Transformer-based approaches. This work establishes a novel paradigm for robust and transferable brain–computer interface (BCI) decoding, advancing the practical deployment of subject-independent EEG-based neural interfaces.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Brain-computer interfaces (BCI) in electroencephalography (EEG)-based motor imagery classification offer promising solutions in neurorehabilitation and assistive technologies by enabling communication between the brain and external devices. However, the non-stationary nature of EEG signals and significant inter-subject variability cause substantial challenges for developing robust cross-subject classification models. This paper introduces a novel Spatial-Spectral-Temporal Attention Fusion (SSTAF) Transformer specifically designed for upper-limb motor imagery classification. Our architecture consists of a spectral transformer and a spatial transformer, followed by a transformer block and a classifier network. Each module is integrated with attention mechanisms that dynamically attend to the most discriminative patterns across multiple domains, such as spectral frequencies, spatial electrode locations, and temporal dynamics. The short-time Fourier transform is incorporated to extract features in the time-frequency domain to make it easier for the model to obtain a better feature distinction. We evaluated our SSTAF Transformer model on two publicly available datasets, the EEGMMIDB dataset, and BCI Competition IV-2a. SSTAF Transformer achieves an accuracy of 76.83% and 68.30% in the data sets, respectively, outperforms traditional CNN-based architectures and a few existing transformer-based approaches.

Problem

• Addresses EEG signal non-stationarity in motor imagery classification
• Overcomes inter-subject variability for robust cross-subject models
• Enhances feature distinction via multi-domain attention mechanisms

Innovation

• SSTAF Transformer integrates spatial-spectral-temporal attention mechanisms
• Uses short-time Fourier transform for time-frequency feature extraction
• Outperforms CNN and existing transformer models in accuracy