Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Jingyi Wang, Zhiqun Wang, Guiran Liu

TL;DR

The paper tackles emotion recognition from EEG recorded during music listening using wearable devices, addressing the challenge of high-dimensional neural data. It introduces a Bi-LSTM with an attention mechanism fed by differential entropy ($DE$) features and a learnable 3D adjacency graph to capture spatial-temporal EEG patterns. Validated on the SEED and DEAP datasets, the approach achieves $98.28\%$ and $92.46\%$ accuracy in multiclass emotion recognition, respectively, outperforming SVM and EEG-Net baselines. This work demonstrates robust, real-time emotion decoding suitable for brain-computer interfaces and affective computing with portable EEG, and points to future gains from multimodal data integration and hardware improvements.

Abstract

Electroencephalography (EEG) signals are crucial for investigating brain function and cognitive processes. This study aims to address the challenges of efficiently recording and analyzing high-dimensional EEG signals while listening to music to recognize emotional states. We propose a method combining Bidirectional Long Short-Term Memory (Bi-LSTM) networks with attention mechanisms for EEG signal processing. Using wearable EEG devices, we collected brain activity data from participants listening to music. The data was preprocessed, segmented, and Differential Entropy (DE) features were extracted. We then constructed and trained a Bi-LSTM model to enhance key feature extraction and improve emotion recognition accuracy. Experiments were conducted on the SEED and DEAP datasets. The Bi-LSTM-AttGW model achieved 98.28% accuracy on the SEED dataset and 92.46% on the DEAP dataset in multi-class emotion recognition tasks, significantly outperforming traditional models such as SVM and EEG-Net. This study demonstrates the effectiveness of combining Bi-LSTM with attention mechanisms, providing robust technical support for applications in brain-computer interfaces (BCI) and affective computing. Future work will focus on improving device design, incorporating multimodal data, and further enhancing emotion recognition accuracy, aiming to achieve practical applications in real-world scenarios.

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

TL;DR

The paper tackles emotion recognition from EEG recorded during music listening using wearable devices, addressing the challenge of high-dimensional neural data. It introduces a Bi-LSTM with an attention mechanism fed by differential entropy () features and a learnable 3D adjacency graph to capture spatial-temporal EEG patterns. Validated on the SEED and DEAP datasets, the approach achieves and accuracy in multiclass emotion recognition, respectively, outperforming SVM and EEG-Net baselines. This work demonstrates robust, real-time emotion decoding suitable for brain-computer interfaces and affective computing with portable EEG, and points to future gains from multimodal data integration and hardware improvements.

Abstract

Electroencephalography (EEG) signals are crucial for investigating brain function and cognitive processes. This study aims to address the challenges of efficiently recording and analyzing high-dimensional EEG signals while listening to music to recognize emotional states. We propose a method combining Bidirectional Long Short-Term Memory (Bi-LSTM) networks with attention mechanisms for EEG signal processing. Using wearable EEG devices, we collected brain activity data from participants listening to music. The data was preprocessed, segmented, and Differential Entropy (DE) features were extracted. We then constructed and trained a Bi-LSTM model to enhance key feature extraction and improve emotion recognition accuracy. Experiments were conducted on the SEED and DEAP datasets. The Bi-LSTM-AttGW model achieved 98.28% accuracy on the SEED dataset and 92.46% on the DEAP dataset in multi-class emotion recognition tasks, significantly outperforming traditional models such as SVM and EEG-Net. This study demonstrates the effectiveness of combining Bi-LSTM with attention mechanisms, providing robust technical support for applications in brain-computer interfaces (BCI) and affective computing. Future work will focus on improving device design, incorporating multimodal data, and further enhancing emotion recognition accuracy, aiming to achieve practical applications in real-world scenarios.
Paper Structure (17 sections, 7 equations, 8 figures, 5 tables)

This paper contains 17 sections, 7 equations, 8 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Wearable EEG Signal Monitoring Devices
  4. EEG Recording and MRI Neuroimaging
  5. Applications of Recurrent Neural Networks in EEG Analysis
  6. Auditory Neural Stimulation on the Brain
  7. Method
  8. EEG Feature Extraction
  9. Construction of 3D Adjacency Matrix of Graph Convolutional Neural Networks
  10. Bi-LSTM constructs EEG signal recognition
  11. Experiment
  12. Datasets
  13. Data Preprocessing and Feature Extraction
  14. Emotion Evoked EEG Signal Generation and Acquisition
  15. Bi-LSTM Construction and Simulation Parameters
  16. ...and 2 more sections

Figures (8)

  • Figure 1: Schematic Diagram of the Proposed Bi-LSTM Framework for Recording Brain Activity While Listening to Music. EEG Signals Are Extracted from the SEED Dataset, and ERP Signals Are Extracted from the DEAP Dataset. The Two Datasets Are Processed Independently and Do Not Interfere with Each Other.
  • Figure 2: 3D location map of EEG channels and location channel connectivity map. Left: 3D location map of EEG channels; Right: Initialize the global channel connectivity graph.
  • Figure 3: Bidirectional Long Short-Term Memory Network Network
  • Figure 4: Single LSTM Cell Diagram. Left: Original LSTM Cell Diagram; Right: LSTM Cell Diagram with Attention Gate.
  • Figure 5: EEG electrode 130 system placement method. The blue labels indicate the 31 EEG channels selected for the second phase of the experiment, with 13 channels in the prefrontal cortex and 18 channels in the occipital lobe.
  • ...and 3 more figures