BrainWave: A Brain Signal Foundation Model for Clinical Applications

2024-02-15

AI Summary

Significant inter-subject variability in EEG signals, coupled with heterogeneous recording modalities and diverse neurological disorders, severely limits model generalizability. Method: We propose the first unified foundation model compatible with both invasive and non-invasive EEG. Trained on 16,000 patients and 40,000 hours (13.79 TB) of multi-center, heterogeneous clinical EEG data, it integrates temporal self-supervised pretraining, cross-center EEG alignment, hierarchical spatiotemporal encoding, and contrastive representation disentanglement to achieve unified EEG representation learning. Contribution/Results: The model enables zero-shot generalization across recording modalities and disease categories, and performs few-shot recognition without fine-tuning. It achieves state-of-the-art performance across multiple neurological diagnosis tasks: 82.3% average accuracy under zero-shot transfer and >89.1% accuracy for 5-shot classification per class.

Technology Category

• Machine Learning: Multimodal Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Computer Vision: Multi-modal Vision

Application Category

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

Abstract

Neural electrical activity is fundamental to brain function, underlying a range of cognitive and behavioral processes, including movement, perception, decision-making, and consciousness. Abnormal patterns of neural signaling often indicate the presence of underlying brain diseases. The variability among individuals, the diverse array of clinical symptoms from various brain disorders, and the limited availability of diagnostic classifications, have posed significant barriers to formulating reliable model of neural signals for diverse application contexts. Here, we present BrainWave, the first foundation model for both invasive and non-invasive neural recordings, pretrained on more than 40,000 hours of electrical brain recordings (13.79 TB of data) from approximately 16,000 individuals. Our analysis show that BrainWave outperforms all other competing models and consistently achieves state-of-the-art performance in the diagnosis and identification of neurological disorders. We also demonstrate robust capabilities of BrainWave in enabling zero-shot transfer learning across varying recording conditions and brain diseases, as well as few-shot classification without fine-tuning, suggesting that BrainWave learns highly generalizable representations of neural signals. We hence believe that open-sourcing BrainWave will facilitate a wide range of clinical applications in medicine, paving the way for AI-driven approaches to investigate brain disorders and advance neuroscience research.

Problem

• Developing reliable neural signal models for clinical diagnosis
• Addressing variability in individuals and brain disorder symptoms
• Overcoming limited diagnostic classifications for neurological diseases

Innovation

• Foundation model for invasive and non-invasive neural recordings
• Pretrained on 40,000 hours of brain signal data
• Enables zero-shot transfer learning across brain diseases