BrainNext: A General-Purpose Self-Supervised Foundation Model for Brain MRI Analysis
2026-07-20 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors created BrainNext, a new AI model that learns from a large number of 3D brain MRI scans without needing labels. It uses a special training method called masked autoencoder combined with a unique 3D neural network design to understand brain anatomy well. After this general learning, BrainNext can be fine-tuned for specific tasks like identifying tumors or estimating brain age. Tested on a medical imaging challenge, it performed very well, showing it can handle different brain MRI jobs effectively. This work suggests that training on lots of unlabeled brain scans helps build versatile models for neuroimaging.
foundation modelsself-supervised learningbrain MRImasked autoencoder3D Bi-Directional xLSTM-UNettransfer learningbrain segmentationbrain-age estimationFOMO 2025 challengevolumetric representations
Authors
Moona Mazher, Abdul Qayyum, Steven A. Niederer, Daniel C. Alexander
Abstract
Foundation models pretrained using self-supervised learning have transformed computer vision by learning transferable representations from large-scale unlabeled data. However, existing foundation models for neuroimaging remain limited by task-specific training, slice-based learning strategies, or relatively small pretraining datasets, restricting their generalizability across diverse brain MRI applications. In this work, we present BrainNext, a general-purpose self-supervised foundation model for volumetric brain MRI analysis. BrainNext combines masked autoencoder (MAE) pretraining with a native three-dimensional Bi-Directional xLSTM-UNet architecture to learn rich anatomical representations from 60,551 unlabeled brain MRI examinations spanning multiple MRI modalities. The pretrained model is subsequently adapted to downstream tasks through lightweight task-specific fine-tuning. We evaluate BrainNext on the Foundation Models for Medical Imaging (FOMO) 2025 Method Track, encompassing classification, segmentation, and brain-age estimation, where it achieved second place overall and ranked first in the meningioma segmentation task on the official FOMO 2025 challenge leaderboard, demonstrating strong transferability across heterogeneous neuroimaging tasks. These results highlight the potential of large-scale self-supervised pretraining to learn robust and transferable volumetric representations, establishing BrainNext as a scalable foundation model for diverse brain MRI applications.