Abstract:We present the MEEG dataset, a multi-modal collection of music-induced electroencephalogram (EEG) recordings designed to capture emotional responses to various musical stimuli across different valence and arousal levels. This public dataset facilitates an in-depth examination of brainwave patterns within musical contexts, providing a robust foundation for studying brain network topology during emotional processing. Leveraging the MEEG dataset, we introduce the Attention-based Temporal Learner with Dynamic Graph Neural Network (AT-DGNN), a novel framework for EEG-based emotion recognition. This model combines an attention mechanism with a dynamic graph neural network (DGNN) to capture intricate EEG dynamics. The AT-DGNN achieves state-of-the-art (SOTA) performance with an accuracy of 83.74% in arousal recognition and 86.01% in valence recognition, outperforming existing SOTA methods. Comparative analysis with traditional datasets, such as DEAP, further validates the model's effectiveness and underscores the potency of music as an emotional stimulus. This study advances graph-based learning methodology in brain-computer interfaces (BCI), significantly improving the accuracy of EEG-based emotion recognition. The MEEG dataset and source code are publicly available at <a class="link-external link-https" href="https://github.com/xmh1011/AT-DGNN" rel="external noopener nofollow">this https URL</a>.

What problem does this paper attempt to address?

This paper attempts to address the issue of improving the accuracy of emotion recognition based on electroencephalogram (EEG) in emotion recognition tasks. Specifically, the paper proposes a new framework—Attention Mechanism Dynamic Graph Neural Network (AT-DGNN)—by introducing music as an emotion-inducing stimulus and combining Graph Neural Networks (GNN) with attention mechanisms to capture complex EEG dynamic features. ### Main Issues and Solutions 1. **Insufficient Diversity of Emotion Induction**: - Existing EEG emotion datasets (such as DEAP and SEED) mainly rely on visual stimuli, which limits the exploration of the relationship between music, emotions, and EEG signals. - **Solution**: Developed a new multimodal dataset MEEG, utilizing music to induce emotional states, improving the effectiveness of emotion induction and the accuracy of the model. 2. **Poor Performance of Existing Methods in Complex Emotion Recognition Tasks**: - Existing deep learning methods mainly treat EEG signals as 2D time series data, making it difficult to capture complex dynamic features. - **Solution**: Proposed the AT-DGNN framework, combining sliding window technology, attention mechanisms, and multi-layer dynamic graph neural networks, significantly improving the accuracy of emotion recognition. 3. **Limitations of Existing Datasets**: - Existing datasets (such as DEAP) have experimental biases that affect the comprehensive evaluation of model performance. - **Solution**: The MEEG dataset reduces the influence of subjective factors by strictly selecting participants and music stimuli, improving the effectiveness of emotion induction. ### Main Contributions of the Paper 1. **MEEG Dataset**: - Provides a multimodal EEG emotion dataset containing various music-induced emotional states, superior to existing DEAP datasets, improving the accuracy and F1 score of the model. 2. **AT-DGNN Framework**: - Combines sliding window technology, attention mechanisms, and multi-layer dynamic graph neural networks to effectively capture the dynamic features of EEG signals, significantly enhancing emotion recognition performance. - Achieved the highest accuracy of 83.74% and 86.01% in arousal and valence dimensions, respectively, surpassing existing state-of-the-art methods (LGGNet series). 3. **Performance Evaluation**: - Validated the effectiveness of the AT-DGNN framework by comparing it with existing CNN, TCN, and GNN methods, and revealed the contributions of each component through ablation experiments. ### Summary This paper significantly improves the accuracy of EEG-based emotion recognition by introducing the music-induced emotion dataset MEEG and the innovative AT-DGNN framework, providing new ideas and methods for the development of Brain-Computer Interface (BCI) technology.

MEEG and AT-DGNN: Improving EEG Emotion Recognition with Music Introducing and Graph-based Learning

EEG Emotion Recognition Based on Self-attention Dynamic Graph Neural Networks

A Efficient Multimodal Framework for Large Scale Emotion Recognition by Fusing Music and Electrodermal Activity Signals

Attention-based Temporal Graph Representation Learning for EEG-based Emotion Recognition

Music emotion recognition based on temporal convolutional attention network using EEG

Using electroencephalogram for emotion recognition based on filter-bank long short-term memory networks

Recording Brain Activity While Listening to Music Using Wearable EEG Devices Combined with Bidirectional Long Short-Term Memory Networks

Electroencephalogram Emotion Recognition Based on Empirical Mode Decomposition and Optimal Feature Selection.

MAS-DGAT-Net: A dynamic graph attention network with multibranch feature extraction and staged fusion for EEG emotion recognition

The PMEmo Dataset for Music Emotion Recognition

[An improved electroencephalogram feature extraction algorithm and its application in emotion recognition]

Enhancing Emotion Recognition Using Region-Specific Electroencephalogram Data and Dynamic Functional Connectivity

Exploring non-Euclidean relationships in EEG emotion recognition: A dynamic graph attention network

A Multimodal Framework for Large-Scale Emotion Recognition by Fusing Music and Electrodermal Activity Signals

Spatial-temporal features-based EEG emotion recognition using graph convolution network and long short-term memory

Emotion Detection from EEG Signals Using Machine Deep Learning Models

EEG Emotion Recognition Using Dynamical Graph Convolutional Neural Networks

A comparative analysis of machine learning methods for emotion recognition using EEG and peripheral physiological signals

Multi-source Selective Graph Domain Adaptation Network for cross-subject EEG emotion recognition

A Feature‐based on Potential and Differential Entropy Information for Electroencephalogram Emotion Recognition

An Efficient LSTM Network for Emotion Recognition from Multichannel EEG Signals