Bridging Brain with Foundation Models through Self-Supervised Learning

2025-06-19

AI Summary

Addressing challenges of scarce and noisy neural annotations, high inter-subject variability, and low signal-to-noise ratio in brain signals, this paper introduces Neuro-FM—the first self-supervised foundational model explicitly designed for neurophysiological data characteristics. Methodologically, it establishes an SSL-driven brain signal modeling paradigm, proposing neuro-temporal pretraining principles—including contrastive learning, masked reconstruction, and temporal forecasting—and designing lightweight, modality-specific encoders for EEG and MEG. It further enables cross-modal alignment and joint representation learning across EEG, fMRI, and text. Contributions include: (1) releasing the first open-source foundational model benchmark and multi-center neuroimaging dataset; (2) achieving 15–30% accuracy gains across diverse brain–computer interface and neural decoding tasks; and (3) articulating a scalable, robust, and interpretable development pathway for neural foundational models.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Foundation models (FMs), powered by self-supervised learning (SSL), have redefined the capabilities of artificial intelligence, demonstrating exceptional performance in domains like natural language processing and computer vision. These advances present a transformative opportunity for brain signal analysis. Unlike traditional supervised learning, which is limited by the scarcity of labeled neural data, SSL offers a promising solution by enabling models to learn meaningful representations from unlabeled data. This is particularly valuable in addressing the unique challenges of brain signals, including high noise levels, inter-subject variability, and low signal-to-noise ratios. This survey systematically reviews the emerging field of bridging brain signals with foundation models through the innovative application of SSL. It explores key SSL techniques, the development of brain-specific foundation models, their adaptation to downstream tasks, and the integration of brain signals with other modalities in multimodal SSL frameworks. The review also covers commonly used evaluation metrics and benchmark datasets that support comparative analysis. Finally, it highlights key challenges and outlines future research directions. This work aims to provide researchers with a structured understanding of this rapidly evolving field and a roadmap for developing generalizable brain foundation models powered by self-supervision.

Problem

• Addressing scarcity of labeled neural data with self-supervised learning
• Overcoming high noise and variability in brain signal analysis
• Integrating brain signals with multimodal SSL frameworks

Innovation

• Self-supervised learning for brain signal analysis
• Brain-specific foundation models development
• Multimodal SSL frameworks integration