LUNA: Efficient and Topology-Agnostic Foundation Model for EEG Signal Analysis

TL;DR

LUNA is a scalable, topology-agnostic EEG foundation model that efficiently processes diverse electrode layouts, achieving state-of-the-art results while significantly reducing computational costs.

Contribution

It introduces a novel latent space approach that reconciles different electrode topologies and scales linearly with channel count, enabling versatile EEG analysis.

Findings

Achieves state-of-the-art AUROC scores on EEG abnormality detection.

Reduces computational costs by up to 300x compared to traditional models.

Demonstrates consistent performance across various electrode configurations.

Abstract

Electroencephalography (EEG) offers a non-invasive lens into human brain activity, but building large-scale models is hampered by topological heterogeneity: each public EEG data defines its own electrode layout, limiting generalization. We introduce LUNA (Latent Unified Network Architecture), a self-supervised foundation model that reconciles disparate electrode geometries while scaling linearly -- not quadratically -- with channel count. LUNA compresses multi-channel EEG into a fixed-size, topology-agnostic latent space via learned queries and cross-attention. Downstream transformer blocks then operate exclusively on this latent representation using patch-wise temporal self-attention, decoupling computation from electrode count. Pre-trained on TUEG and Siena (over 21,000 hours of raw EEG across diverse montages) using a masked-patch reconstruction objective, LUNA transfers effectively to four downstream tasks: abnormality detection, artifact rejection, slowing classification, and emotion recognition. It demonstrates highly competitive performance across several benchmarks, achieving state-of-the-art results on TUAR and TUSL, e.g., 0.921 AUROC on TUAR, while reducing FLOPs by 300x and trimming GPU memory use by up to 10x. Critically, these gains are consistent across all evaluated electrode configurations. Code is available at https://github.com/pulp-bio/BioFoundation