Related Experiment Video
Updated: Jun 16, 2025
Applications of EEG Neuroimaging Data: Event-related Potentials, Spectral Power, and Multiscale Entropy
Published on: June 27, 2013
Emotion recognition of EEG signals based on contrastive learning graph convolutional model
Yiling Zhang1, Yuan Liao1, Wei Chen1
1College of electronic and optical engineering & college of flexible electronics (future technology), Nanjing University of Posts and Telecommunications, Jiangsu 210023, People's Republic of China.
This study introduces a novel Contrastive Learning Graph Convolutional Network (CLGCN) to decode emotions from electroencephalogram (EEG) signals, achieving high accuracy by focusing on commonalities across subjects. The method effectively analyzes brain connectivity for improved emotion recognition.
Area of Science:
- Neuroscience
- Machine Learning
- Signal Processing
Background:
- Electroencephalogram (EEG) signals provide insights into brain activity related to emotions.
- Individual variability in EEG signals poses challenges for emotion recognition.
- Existing methods struggle to generalize across different subjects.
Purpose of the Study:
- To develop an innovative approach for emotion recognition from EEG signals that addresses inter-subject variability.
- To identify commonalities in EEG data across distinct subjects for robust emotion decoding.
- To improve the accuracy and generalizability of emotion recognition models.
Main Methods:
- Introduction of the Contrastive Learning Graph Convolutional Network (CLGCN).
- CLGCN integrates contrastive learning (CL) for multisubject data learning and graph convolutional networks (GCN) for analyzing brain connectivity.
- The model generates a standardized brain network learning matrix to capture distinctive features and channel nodes related to emotional states.
Main Results:
- The CLGCN model achieved 97.13% accuracy on the DEAP dataset and over 99% accuracy on the SEED and SEED_IV datasets in a five-fold cross-validation setting.
- In incremental learning experiments, the model achieved 92.8% accuracy for new subjects with only 5% of data fine-tuning.
- These results demonstrate the model's high efficacy and ability to generalize.
Conclusions:
- The combination of CL and GCN in CLGCN significantly enhances the accuracy of decoding emotional states from EEG signals.
- This approach offers valuable insights into the underlying mechanisms of emotional processing in the brain.
- CLGCN presents a promising method for overcoming inter-subject variability in EEG-based emotion recognition.
