Structured Prototype-Guided Adaptation for EEG Foundation Models

2026-02-19

AI Summary

This work addresses the limited generalization of existing EEG foundation models under scarce subject-specific supervision, which stems from a structural mismatch between supervisory signals and the high-dimensional parameter space. To overcome this, we propose SCOPE, a novel framework that first constructs geometrically regularized task priors and confidence-aware pseudo-labels, then introduces a lightweight Structured Prototype-based Conditional Adapter (ProAdapter) to enable efficient cross-subject transfer while keeping the foundation model frozen. By innovatively integrating structured prototypes with a confidence-aware mechanism, SCOPE consistently achieves significant improvements in both generalization performance and adaptation efficiency across three EEG tasks and five foundation models in label-scarce scenarios.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Computer Vision: Other Foundations of Computer Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Electroencephalography (EEG) foundation models (EFMs) have achieved strong performance under full fine-tuning but exhibit poor generalization when subject-level supervision is limited, a common constraint in real-world clinical settings. We show that this failure stems not merely from limited supervision, but from a structural mismatch between noisy, limited supervision and the highly plastic parameter space of EFMs. To address this challenge, we propose SCOPE, a Structured COnfidence-aware Prototype-guided adaptation framework for EFM fine-tuning. SCOPE follows a two-stage pipeline. In the first stage, we construct reliable external supervision by learning geometry-regularized task priors, constructing balanced class-level prototypes over the resulting embeddings, and producing confidence-aware pseudo-labels from their agreement to filter unreliable signals on unlabeled data. In the second stage, we introduce ProAdapter, which adapts frozen EEG foundation models via a lightweight adapter conditioned on the structured prototypes. Experiments across three EEG tasks and five foundation model backbones demonstrate that SCOPE consistently achieves strong performance and efficiency under label-limited cross-subject settings.

Problem

• EEG foundation models
• limited supervision
• structural mismatch
• cross-subject generalization
• label scarcity

Innovation

• prototype-guided adaptation
• EEG foundation models
• confidence-aware pseudo-labeling
• ProAdapter
• label-limited learning