Towards Generalisable Foundation Models for Brain MRI

TL;DR

BrainFound is a self-supervised foundation model for brain MRI that extends vision transformers to 3D, supporting diverse tasks and modalities, outperforming existing methods especially with limited labels.

Contribution

We introduce BrainFound, a novel 3D self-supervised foundation model for brain MRI that adapts DINO-v2 for volumetric data and multimodal inputs, enhancing generalization and downstream performance.

Findings

Outperforms existing self-supervised and supervised methods

Effective across multiple MRI modalities and clinical scenarios

Reduces need for extensive annotations

Abstract

Foundation models in artificial intelligence (AI) are transforming medical imaging by enabling general-purpose feature learning from large-scale, unlabeled datasets. In this work, we introduce BrainFound, a self-supervised foundation model for brain MRI, built by extending DINO-v2, a vision transformer originally designed for 2D natural images. BrainFound adapts DINO-v2 to model full 3D brain anatomy by incorporating volumetric information from sequential MRI slices, moving beyond conventional single-slice paradigms. It supports both single- and multimodal inputs, enabling a broad range of downstream tasks, including disease detection and image segmentation, while generalising across varied imaging protocols and clinical scenarios. We show that BrainFound consistently outperforms existing self-supervised pretraining strategies and supervised baselines, particularly in label-scarce and multi-contrast settings. By integrating information from diverse 3D MRI modalities (e.g., T1, T2, FLAIR), it enhances diagnostic accuracy and reduces dependency on extensive expert annotations. This flexibility makes BrainFound a scalable and practical solution for 3D neuroimaging pipelines, with significant potential for clinical deployment and research innovation.