PRISM: Exploring Heterogeneous Pretrained EEG Foundation Model Transfer to Clinical Differential Diagnosis

2026-02-28

AI Summary

This study addresses the limited generalizability of existing EEG foundation models, which are typically pretrained on single-source clinical data and struggle to disentangle neurophysiological signals from device- or site-specific artifacts. To overcome this, the authors construct a multi-center EEG pretraining corpus with deliberate geographic and device diversity and employ a unified masked autoencoder (MAE) architecture with consistent preprocessing. They systematically investigate how data heterogeneity influences transfer performance on downstream clinical tasks, revealing a critical trade-off between data diversity and model transferability. Their findings demonstrate that strategically curated diverse data outperforms indiscriminate scale expansion, achieving a 12.3-percentage-point gain in balanced accuracy on an unseen seizure-versus-mimics discrimination task—matching or surpassing the REVE model trained on 92 datasets—and identify six key non-additive bias factors that critically impact evaluation.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems
• Humans and AI: Other Foundations of Human Computation & AI

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: Fairness-aware retrieval and ranking

Abstract

EEG foundation models are typically pretrained on narrow-source clinical archives and evaluated on benchmarks from the same ecosystem, leaving unclear whether representations encode neural physiology or recording-distribution artifacts. We introduce PRISM (Population Representative Invariant Signal Model), a masked autoencoder ablated along two axes -- pretraining population and downstream adaptation -- with architecture and preprocessing fixed. We compare a narrow-source EU/US corpus (TUH + PhysioNet) against a geographically diverse pool augmented with multi-center South Asian clinical recordings across multiple EEG systems. Three findings emerge. First, narrow-source pretraining yields stronger linear probes on distribution-matched benchmarks, while diverse pretraining produces more adaptable representations under fine-tuning -- a trade-off invisible under single-protocol evaluation. Trained on three source corpora, PRISM matches or outperforms REVE (92 datasets, 60,000+ hours) on the majority of tasks, demonstrating that targeted diversity can substitute for indiscriminate scale and that dataset count is a confounding variable in model comparison. Second, on a clinically challenging and previously untested task -- distinguishing epilepsy from diagnostic mimickers via interictal EEG -- the diverse checkpoint outperforms the narrow-source checkpoint by +12.3 pp balanced accuracy, the largest gap across all evaluations. Third, systematic inconsistencies between EEG-Bench and EEG-FM-Bench reverse model rankings on identical datasets by up to 24 pp; we identify six concrete sources including split construction, checkpoint selection, segment length, and normalization, showing these factors compound non-additively.

Problem

• EEG foundation model
• clinical differential diagnosis
• distribution shift
• benchmark inconsistency
• epilepsy mimickers

Innovation

• EEG foundation model
• pretraining diversity
• clinical transfer learning
• benchmark inconsistency
• masked autoencoder