Related Experiment Video
Updated: Jun 4, 2025
11:15
Applications of EEG Neuroimaging Data: Event-related Potentials, Spectral Power, and Multiscale Entropy
Published on: June 27, 2013
33.6K
A temporal-spectral graph convolutional neural network model for EEG emotion recognition within and across subjects
Rui Li1, Xuanwen Yang1, Jun Lou1
1Brain Cognition and Computing Lab, National Engineering Research Center for E-Learning, Faculty of Artificial Intelligence in Education, Central China Normal University, Wuhan, Hubei, China.
Brain Informatics
|December 18, 2024
Summary
We developed a new AI model, the Temporal-Spectral Graph Convolutional Network (TSGCN), for more accurate electroencephalogram (EEG)-based emotion recognition. This model effectively analyzes brain activity across time, frequency, and spatial domains, improving emotion prediction.
Area of Science:
- Neuroscience
- Artificial Intelligence
- Signal Processing
Background:
- EEG-based emotion recognition faces challenges due to sparse neural information across domains and inter-subject variability.
- Existing methods struggle to effectively integrate temporal, spectral, and spatial information from EEG signals.
Purpose of the Study:
- To propose a novel neural network, the Temporal-Spectral Graph Convolutional Network (TSGCN), for enhanced EEG-based emotion recognition.
- To address the challenges of information sparsity and cross-subject variations in emotion recognition.
Main Methods:
- TSGCN integrates neural oscillation changes across time windows and topological brain region structures.
- Minimum Category Confusion (MCC) loss reduces inconsistencies between subjective and predicted labels.
- Deep and Shallow feature Dynamic Adversarial Learning (DSDAL) enhances cross-subject generalization.
Main Results:
- TSGCN significantly outperforms state-of-the-art methods on public EEG emotion recognition datasets.
- Ablation studies confirm the performance and robustness contributions of TSGCN's components.
- Investigations validate TSGCN's effectiveness in overcoming emotion recognition challenges.
Conclusions:
- TSGCN offers a robust and effective solution for EEG-based emotion recognition.
- The proposed model demonstrates superior performance by effectively capturing complex neural dynamics.
- TSGCN shows strong generalization capabilities, addressing critical limitations in current approaches.
