PhysioSync: Temporal and Cross-Modal Contrastive Learning Inspired by Physiological Synchronization for EEG-based Emotion Recognition

🚀 Introduction

• PhysioSync is a pre-training framework for EEG-based emotion recognition that leverages temporal and cross-modal contrastive learning inspired by physiological synchronization.
• It incorporates Cross-Modal Consistency Alignment (CM-CA) to capture dynamic relationships between EEG and Peripheral Physiological Signals (PPS), and Long- and Short-Term Temporal Contrastive Learning (LS-TCL) to model emotional fluctuations across different time resolutions.
• Experiments on DEAP and DREAMER demonstrate that PhysioSync achieves robust improvements in both uni-modal and cross-modal settings.

📦Create environment

We recommend Python >= 3.8.

Install dependencies:

pip install -r requirements.txt

(Alternatively, if you use conda:)

conda create -n physiosync python=3.8
conda activate physiosync
pip install -r requirements.txt

📊 Dataset

We evaluate PhysioSync on benchmark EEG-based emotion recognition datasets. Please download the datasets from their official sources

• DEAP
• DREAMER

🏃Pre-Training and Fine-Tuning

Take DEAP's "Dependent" as an example

Pre-Training

python main_pretrain_1s.py
python main_pretrain_5s.py

Fine-Tuning

cd Fine_tuning
python fine_tuning.py 

📊 Experimental Results

We report results under Dependent (within-subject) and Cross-subject evaluation protocols for DEAP and DREAMER datasets.

1️⃣ Dependent (Within-Subject) Results

Dependent DEAP results

Dependent DREAMER results

2️⃣ Cross-Subject Results

Cross-subject results

📖Citation

@ARTICLE{11199963,
  author={Cui, Kai and Li, Jia and Liu, Yu and Zhang, Xuesong and Hu, Zhenzhen and Wang, Meng},
  journal={IEEE Transactions on Computational Social Systems}, 
  title={PhysioSync: Temporal and Cross-Modal Contrastive Learning Inspired by Physiological Synchronization for EEG-Based Emotion Recognition}, 
  year={2025},
  volume={},
  number={},
  pages={1-14},
  doi={10.1109/TCSS.2025.3602913}}