How do we model brain activity?
Learn transferable representations across spatial resolutions and long recordings.
Research
Models and methods for
understanding the brain
Nine projects on modeling brain activity, applying foundation models, and using data and compute well.
Research map
Three questions
A guide to how these projects relate across modeling, applications, and data strategy.
What can foundation models enable?
Use learned representations to create individualized brain atlases and generate anatomy-conditioned fMRI dynamics.
How should data and compute be used?
Study different representations of fMRI data, choose pretraining and transfer tasks, and scale data, models, and training time.
Featured research
View all researchGenerative modeling
NeurIPS 2026 · OralBrainWorld
A structural prior brings anatomy and four-dimensional brain activity into one generative framework.
Explore BrainWorld
Representation learning
ICML 2026Omni-fMRI
Content-adaptive voxel patches preserve spatial detail while making large-scale pretraining more efficient.
Explore Omni-fMRI
Diffusion pretraining
MICCAI 2026 · Best Paper ShortlistBrain-DiT
Metadata-conditioned diffusion connects resting, task, and naturalistic fMRI in a shared representation framework.
Explore Brain-DiT
Native-space representation
ECCV 2026FlexiBrain
Learn from native-space fMRI across different resolutions with physical-space patching and a Mamba-JEPA encoder.
Explore FlexiBrain
Select a project
Explore the collection
Models & methods09
Explore nine projects by modality
and research purpose.
9 research projectsOriginal papers, code and methods
Explore research Omni-fMRI
Atlas-free brain representations with content-adaptive voxel patches. Preserve spatial detail while making large-scale pretraining more efficient.
Learn moreExplore research Brain-DiT
Metadata-conditioned diffusion pretraining brings resting-state, task, naturalistic and other brain states into a shared representation framework.
Learn moreExplore research BrainWorld
Use individual brain anatomy as a structural prior to generate whole-brain fMRI dynamics across space and time.
Learn more
Explore research BrainTaskonomy
Organize pretraining domains and downstream transfer with measured task relationships, curriculum learning and budget-aware source selection.
Learn more
Explore research fMRI Scaling Study
A controlled study of how data, model size and training duration affect generalization in fMRI foundation models.
Learn more
Explore research FlatClip
Turn cortical activity into flatmap sequences and reuse a frozen vision encoder for a lightweight surface-based learning baseline.
Learn more
Explore research SLIM-Brain
Route informative temporal windows to a hierarchical encoder to reduce the cost of voxel-level fMRI representation learning.
Learn moreFlexiBrain
Learn from native-space fMRI across different resolutions with physical-space patching and a Mamba-JEPA encoder.
Learn more
DCA
Build personalized voxel-level brain atlases with pretrained fMRI embeddings and spatially guided clustering.
Learn moreNo research found
Try another modality, purpose or keyword.
The foundations of discovery
Data collections
Explore 45 data collections and paper-specific cohorts, with source links, access conditions, and reported scale.
Explore the collectionsAbout fMRI Atlas
A directory of research in functional brain imaging. Each project presents its source document and method overview, with code links where available. The data collection index includes public resources and paper-specific cohorts; consult each paper for its complete data inventory and evaluation protocol.
A research directory connecting original projects and sources.