Subject Representation Learning from EEG using Graph Convolutional Variational Autoencoders

Aditya Mishra, Ahnaf Mozib Samin, Ali Etemad, Javad Hashemi
1/13/2025
eess.SPcs.LG

No official code found. Best implementation: adityamishra-ml/GC-VASE (2 stars, last commit 2025-08-11).

Abstract

We propose GC-VASE, a graph convolutional-based variational autoencoder that leverages contrastive learning for subject representation learning from EEG data. Our method successfully learns robust subject-specific latent representations using the split-latent space architecture tailored for subject identification. To enhance the model's adaptability to unseen subjects without extensive retraining, we introduce an attention-based adapter network for fine-tuning, which reduces the computational cost of adapting the model to new subjects. Our method significantly outperforms other deep learning approaches, achieving state-of-the-art results with a subject balanced accuracy of 89.81% on the ERP-Core dataset and 70.85% on the SleepEDFx-20 dataset. After subject adaptive fine-tuning using adapters and attention layers, GC-VASE further improves the subject balanced accuracy to 90.31% on ERP-Core. Additionally, we perform a detailed ablation study to highlight the impact of the key components of our method.

View on arXivView PDF

Code Implementations(1)

adityamishra-ml/GC-VASECommunity76%

Subject Representation Learning from EEG using Graph Convolutional Variational Autoencoders

20PythonDec 25, 20241 years ago

Similar Papers

CoSupFormer : A Contrastive Supervised learning approach for EEG signal Classification

83% match2025

Source-free Subject Adaptation for EEG-based Visual Recognition

83% match1 repo4172023

Variational Autoencoder Learns Better Feature Representations for EEG-based Obesity Classification

83% match2023

A Contrastive Learning Based Convolutional Neural Network for ERP Brain-Computer Interfaces

82% match2024

Pre-Training Graph Contrastive Masked Autoencoders are Strong Distillers for EEG

82% match1 repo132024

CSLP-AE: A Contrastive Split-Latent Permutation Autoencoder Framework for Zero-Shot Electroencephalography Signal Conversion

82% match2 repos152026

Cite this paper

@article{mishra2025subject,
  title  = {Subject Representation Learning from EEG using Graph Convolutional Variational Autoencoders},
  author = {Aditya Mishra and Ahnaf Mozib Samin and Ali Etemad and Javad Hashemi},
  year   = {2025},
  eprint = {2501.16626},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2501.16626v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2501.16626.svg)](https://codeofpaper.com/paper/2501.16626)

Discussion