cs.CVOct 27, 2026

Systematic Morphometric Brain Autoencoding for Representation Learning in Neurodegenerative Diseases

Authors: Nuria Balbás-Esteban, Albert Belenguer-Llorens, Elaheh Moradi, Jussi Tohka, Vanessa Gómez-Verdejo, Carlos Sevilla-Salcedo

Organizations: Department of Signal Processing and Communications, Universidad Carlos III de Madrid, Leganés, 28911, Madrid, Spain · Department of Health Sciences and Technology (DHEST), ETH Zurich, 8092, Zürich, Switzerland · SIB Swiss Institute of Bioinformatics, Lausanne, Switzerland · A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, 70211, Finland · 

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Early identification of neurodegenerative disorders using structural T1-weighted MRI is essential for timely clinical intervention, yet supervised deep learning models often struggle to generalize due to limited annotations and scanner-induced distribution shift. Self-supervised learning offers a promising alternative; however, existing approaches are largely adapted from natural image paradigms, relying on contrastive objectives or generic masked reconstruction with voxel-wise losses, which may not fully capture the distributed morphometric patterns associated with neurodegeneration. We introduce SYMBA (SYstematic Morphometric Brain Autoencoding), a self-supervised pretraining framework specifically designed for structural brain MRI. SYMBA integrates anatomically constrained masking, a multi-scale structural similarity (MS-SSIM) reconstruction objective, and a progressive curriculum that transitions from local anatomical detail to global brain morphology. These components explicitly incorporate neuroanatomical structure into the self-supervised objective rather than relying on generic image priors. Pretrained on 6,953 T1-weighted MRI scans from nine heterogeneous cohorts, SYMBA is evaluated using frozen linear probing on two independent unseen datasets, enabling a strict assessment of out-of-distribution generalization. Across multiple clinical tasks, SYMBA consistently outperforms recent neuroimaging foundation models, achieving 0.930 AUROC for Alzheimer's disease classification, 0.772 for mild cognitive impairment conversion prediction, and strong performance in amyloid status prediction. In low-label regimes, SYMBA demonstrates high sample efficiency, reaching approximately 0.90 AUROC for AD/CN classification using only six labeled subjects.

Figures & tables

Figure 1: Overview of SYMBA. The model learns latent representations from masked structural MRI volumes using anatomically constrained region-based masking, followed by reconstruction. The latent representations are then used to conduct linear probing on downstream tasks .
Figure 2: Overview of the anatomically constrained masking operator used in SYMBA. The masking geometric regions are restricted to brain tissue, sampled within the brain support, and mirrored across the sagittal plane. Then, all region supports are combined into a binary mask. Masking complexity is progressively increased throughout training via curriculum learning.
Database# Scans# ParticipantsCountry of OriginPopulationAge Ranges
Wayne600300USAHealthy(18 - 83 years)
SALD494494ChinaHealthy(19 - 80 years)
IXI550550UKHealthy(19 - 86 years)
DLBS315315USAHealthy(21 - 89 years)
OASIS28421378USAHealthy Aging & Alzheimer’s/Dementia(42 - 95 years)
CamCAN652652UKHealthy(18 - 88 years)
Table 1: Number of samples, country of origin, nature of the diagnosis and metadata of the different databases used
TaskLabel# Samples
ADNIAIBL
AD vs CNAD39569
CN858436
sMCI vs pMCIsMCI235-
pMCI263-
3-ClassAD39569
Table 2: Data split and sample distribution per task for low-data regime evaluation test.
Figure 3: Downstream classification performance (AUROC) as a function of the proportion of labelled training data available, shown on a logarithmic x-axis. (a) Classification tasks A) AD vs CN, B) sMCI vs pMCI, C) Amyloid, D) Sex. (b) for the AD vs CN and Sex tasks. In both panels, SYMBA corresponds to a frozen encoder pre-trained using the proposed self-supervised framework with a linear classifier; No Pretraining denotes the same architecture trained from random initialization under full supervision; and AnatCL, BrainIAC, BrainMorph, and BrainSegFounder are publicly available pre-trained neuroimaging foundation models evaluated under the same frozen-encoder protocol as SYMBA.
Figure 4: Representative reconstructions produced ResNet under the masking ratio (75%) and patch size 16 for the axial and coronal view. Columns correspond to the original image, masked input, reconstructed image, total absolute difference and masked difference respectively.
More in this topic

Medical Image Representation Learning

Medical image representation learning builds reusable visual features for diagnosis, segmentation, retrieval, and other clinical imaging tasks.

See all 260 papers

Explore similar work

CardsList
Jul 20, 2026

BrainNext: A General-Purpose Self-Supervised Foundation Model for Brain MRI Analysis

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.

Moona Mazher, Abdul Qayyum, Steven A. Niederer +1

UCL Hawkes Institute, Department of Computer Science, University College London, London, United Kingdom · National Heart & Lung Institute, Imperial College London, London, United Kingdom · Division of Cardiovascular Medicine, Stanford Cardiovascular Institute, Stanford University, California, USA

Medical Imaging Foundation ModelsSelf-Supervised Pre-TrainingMedical Image Representation Learning
AskPDF
Apr 30, 2026

BrainDINO: A Brain MRI Foundation Model for Generalizable Clinical Representation Learning

Brain MRI underpins a wide range of neuroscientific and clinical applications, yet most learning-based methods remain task-specific and require substantial labeled data. Here we show that a single self-supervised representation can generalize across heterogeneous brain MRI endpoints. We trained BrainDINO, a self-distilled foundation model, on approximately 6.6 million unlabeled axial slices from 20 datasets encompassing broad variation in population, disease, and acquisition setting. Using a frozen encoder with lightweight task heads, BrainDINO supported transfer across tumor segmentation, neurodegenerative and neurodevelopmental conditions classification, brain age estimation, post-stroke temporal prediction, molecular status prediction, MRI sequence classification, and survival modeling. Across tasks and supervision regimes, BrainDINO consistently equaled or exceeded natural-image and MRI-specific self-supervised baselines, with particularly strong advantages under label scarcity. Representation analyses further showed anatomically organized and pathology-sensitive feature structure in the absence of task-specific supervision. Our findings indicate that large-scale slice-wise self-supervised learning can yield a unified brain MRI representation that supports diverse neuroimaging tasks without volumetric pretraining or full-network fine-tuning, establishing a scalable foundation for robust and data-efficient brain imaging analysis. Code is available at https://github.com/mclwu22/BrainDINO

Yizhou Wu, Shansong Wang, Yuheng Li +5

Department of Radiation Oncology and Winship Cancer Institute, Emory University, Atlanta, GA, USA. · Department of Radiation and Cellular Oncology, The University of Chicago, Chicago, IL, USA. · Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, USA. +2

Medical Imaging Foundation ModelsSelf-Supervised Pre-TrainingMedical Image Representation LearningBrain MRI
AskPDF
Sep 25, 2026

Combining General and Domain-Specific Pretext Tasks for Brain MR Image Segmentation

A key challenge in medical image analysis is the scarcity of large annotated datasets for specific populations and diseases. As deep learning models rely heavily on labeled data, effective transfer learning strategies are needed to reduce the dependence on manual annotations. Self-supervised learning has emerged as a promising approach for developing foundation models by enabling the learning of transferable feature representations from large-scale unlabeled medical imaging datasets. In this study, we investigate voxel-level brain age prediction as a domain-specific self-supervised pretext task and compare it with image inpainting, a widely used non-domain-specific alternative. We further propose a multitask self-supervised pretraining framework that jointly optimizes both objectives to learn complementary neuroimaging representations. The pretrained models are evaluated on three downstream magnetic resonance image segmentation tasks: multiple sclerosis lesion segmentation, ischemic stroke lesion segmentation, and cortical brain structure segmentation. Overall, the proposed multitask pretraining framework consistently outperformed the single-task pretrained models and training from scratch across most experimental settings, demonstrating the benefit of combining domain-specific and general self-supervised learning pretext tasks for the development of generalizable neuroimaging foundation models.\ Code Availability: The source code used in this study is publicly available at https://github.com/TasneemN/Combining-General-and-Domain-Specific-Pretext-Tasks-for-Brain-MR-Image-Segmentation/

Tasneem Nasser, Susanne Schmid, Roberto Souza +1

Biomedical Engineering University of Calgary Calgary AB Canada · Hotchkiss Brain Institute University of Calgary Calgary AB Canada · Electrical and Software Engineering University of Calgary Calgary AB Canada +1

Medical Image Representation LearningSelf-Supervised Pre-TrainingMedical Imaging Foundation ModelsMedical Image SegmentationBrain MRI
AskPDF