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

TL;DR

This work introduces the Attention-based Temporal Learner with Dynamic Graph Neural Network (AT-DGNN), a novel framework for EEG-based emotion recognition that outperforms existing SOTA methods and advances graph-based learning methodology in brain-computer interfaces (BCI), significantly improving the accuracy of EEG-based emotion recognition.

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 https://github.com/xmh1011/AT-DGNN.

Figures (2)

• Figure 1: Structure of AT-DGNN. The AT-DGNN model comprises two core modules: a feature extraction module (a) and a dynamic graph neural network learning module (b). The feature extraction module consists of a temporal learner, a multi-head attention mechanism, and a temporal convolution module. These components effectively leverage local features of EEG signals through a sliding window technique, thereby enhancing the model's capacity to dynamically extract complex temporal patterns in EEG signals. In the graph-based learning module, the model initially employs local filtering layers to segment and filter features from specific brain regions. Subsequently, the architecture employs three layers of stacked dynamic graph convolutions to capture complex interactions among different brain regions. This structure enhances the AT-DGNN's capacity for integrating temporal features effectively.
• Figure 2: Local-global graph definitions ding2023lggnet. (a) General definition $G_g$. (b) Frontal definition $G_f$. (c) Hemispheric definition $G_h$.