Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Representation Learning of EEG Data with Masked Autoencoders

Yifei Zhou, Sitong Liu

TL;DR

This study designs a masked autoencoder (MAE) to guide deep learning models to learn electroencephalography (EEG) signal representation and finds that the MAE is an effective brain signal learner and significantly improves learning efficiency.

Abstract

Self-supervised learning has been a powerful training paradigm to facilitate representation learning. In this study, we design a masked autoencoder (MAE) to guide deep learning models to learn electroencephalography (EEG) signal representation. Our MAE includes an encoder and a decoder. A certain proportion of input EEG signals are randomly masked and sent to our MAE. The goal is to recover these masked signals. After this self-supervised pre-training, the encoder is fine-tuned on downstream tasks. We evaluate our MAE on EEGEyeNet gaze estimation task. We find that the MAE is an effective brain signal learner. It also significantly improves learning efficiency. Compared to the model without MAE pre-training, the pre-trained one achieves equal performance with 1/3 the time of training and outperforms it in half the training time. Our study shows that self-supervised learning is a promising research direction for EEG-based applications as other fields (natural language processing, computer vision, robotics, etc.), and thus we expect foundation models to be successful in EEG domain.

Enhancing Representation Learning of EEG Data with Masked Autoencoders

TL;DR

This study designs a masked autoencoder (MAE) to guide deep learning models to learn electroencephalography (EEG) signal representation and finds that the MAE is an effective brain signal learner and significantly improves learning efficiency.

Abstract

Self-supervised learning has been a powerful training paradigm to facilitate representation learning. In this study, we design a masked autoencoder (MAE) to guide deep learning models to learn electroencephalography (EEG) signal representation. Our MAE includes an encoder and a decoder. A certain proportion of input EEG signals are randomly masked and sent to our MAE. The goal is to recover these masked signals. After this self-supervised pre-training, the encoder is fine-tuned on downstream tasks. We evaluate our MAE on EEGEyeNet gaze estimation task. We find that the MAE is an effective brain signal learner. It also significantly improves learning efficiency. Compared to the model without MAE pre-training, the pre-trained one achieves equal performance with 1/3 the time of training and outperforms it in half the training time. Our study shows that self-supervised learning is a promising research direction for EEG-based applications as other fields (natural language processing, computer vision, robotics, etc.), and thus we expect foundation models to be successful in EEG domain.
Paper Structure (17 sections, 1 equation, 3 figures, 2 tables)

This paper contains 17 sections, 1 equation, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Masked modeling in language and vision
  4. Masked autoencoder for EEG data
  5. EEG-based gaze estimation
  6. Methods
  7. Masking mechanism
  8. Encoder design
  9. Decoder design
  10. Reconstruction task
  11. Experiment Setting
  12. EEG data
  13. Training
  14. Results
  15. Encoder's performance
  16. ...and 2 more sections

Figures (3)

  • Figure 1: Pre-training and fine-tuning model architectures. EEG signals collected from multiple channels are arranged into a matrix. (a) We mask random elements from the input EEG signal matrix. Our MAE learns to recover these missing signals. (b) Our main purpose is to measure the encoder's performance change after MAE pre-training, so we remove the decoder and fine-tune the encoder to predict gaze positions.
  • Figure 2: Fine-tuning results under different settings.
  • Figure 3: Fine-tuning loss curves. For each decoder setting, top two results among all the masking ratios ($r$) are presented.