Towards a general-purpose foundation model for fMRI analysis

2025-06-11

AI Summary

Functional MRI (fMRI) analysis suffers from poor reproducibility and limited cross-task generalizability due to complex, task-specific preprocessing pipelines and modeling approaches—hindering its broader application in neuroscience and clinical diagnostics. To address this, we propose NeuroSTORM, the first general-purpose foundation model for fMRI, directly operating on 4D functional volumes. NeuroSTORM innovatively integrates the Mamba architecture with a voxel-wise shifted scanning strategy to enable spatiotemporal-aware sequence modeling. We further introduce a spatiotemporal contrastive pretraining paradigm and a lightweight prompt-based fine-tuning mechanism to enhance transferability and adaptability. Trained on large-scale, multicenter fMRI data, NeuroSTORM achieves state-of-the-art performance across five diverse downstream tasks—including age/sex prediction, phenotypic inference, and neurological disorder diagnosis—and demonstrates superior generalization on clinical datasets from the United States, South Korea, and Australia.

Technology Category

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

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Functional Magnetic Resonance Imaging (fMRI) is essential for studying brain function and diagnosing neurological disorders, but current analysis methods face reproducibility and transferability issues due to complex pre-processing and task-specific models. We introduce NeuroSTORM (Neuroimaging Foundation Model with Spatial-Temporal Optimized Representation Modeling), a generalizable framework that directly learns from 4D fMRI volumes and enables efficient knowledge transfer across diverse applications. NeuroSTORM is pre-trained on 28.65 million fMRI frames (>9,000 hours) from over 50,000 subjects across multiple centers and ages 5 to 100. Using a Mamba backbone and a shifted scanning strategy, it efficiently processes full 4D volumes. We also propose a spatial-temporal optimized pre-training approach and task-specific prompt tuning to improve transferability. NeuroSTORM outperforms existing methods across five tasks: age/gender prediction, phenotype prediction, disease diagnosis, fMRI-to-image retrieval, and task-based fMRI classification. It demonstrates strong clinical utility on datasets from hospitals in the U.S., South Korea, and Australia, achieving top performance in disease diagnosis and cognitive phenotype prediction. NeuroSTORM provides a standardized, open-source foundation model to improve reproducibility and transferability in fMRI-based clinical research.

Problem

• Developing a general-purpose fMRI model for improved reproducibility and transferability
• Addressing limitations of task-specific models in fMRI analysis
• Enhancing clinical utility of fMRI across diverse applications

Innovation

• Mamba backbone processes full 4D fMRI volumes
• Spatial-temporal optimized pre-training improves transferability
• Task-specific prompt tuning enhances diverse applications