Atoms of Thought: Universal EEG Representation Learning with Microstates

TL;DR

This paper introduces a universal EEG microstate tokenizer that improves representation learning across various neuroinformatics tasks, outperforming traditional features and enhancing interpretability.

Contribution

The authors develop a universal microstate tokenizer from large EEG datasets and demonstrate its effectiveness across multiple downstream tasks.

Findings

Microstate-based EEG representations outperform traditional features.

The approach enhances interpretability and scalability of EEG analysis.

Universal microstate tokenizer benefits diverse neuroinformatics applications.

Abstract

Learning universal representations from electroencephalogram (EEG) signals is a cutting-edge approach in the field of neuroinformatics and brain-computer interfaces (BCIs). Conventionally, EEG is treated as a multivariate temporal signal, where time- or frequency-domain features are extracted for representation learning. This paper investigates a simple yet effective EEG representation, i.e., microstates. Microstates represent the building blocks of brain activity patterns at a microscopic time scale. We build a universal microstate tokenizer from a large medical EEG dataset by clustering continuous EEG signals into sequences of discrete microstates. The microstate tokenizer is then adopted universally across a series of downstream tasks, including sleep staging, emotion recognition, and motor imagery classification. Experimental results show that EEG representation learning with microstates outperforms traditional time-domain and frequency-domain features under different models and across different tasks. Further analysis shows that microstates offer greater interpretability and scalability, thereby opening up applications in both cognitive neuroscience and clinical research.