LAION-fMRI is a deeply-sampled 7T fMRI dataset of brain responses to visual images, built to uncover how the human brain sees and understands the world. Five participants viewed 25,052 unique natural images across the launch release’s 150 main image-viewing sessions, capturing hundreds of thousands of brain responses at 1.8 mm resolution with an ultra-high-field 7T MRI scanner. Across all acquired fMRI protocols, the dataset contains 165 fMRI sessions; the supplemental image sessions and their additional shared images will be released later.
The images span everything from everyday photographs - drawn from a 120M image-text corpus (Roth & Hebart, 2025) - to abstract shapes and visual illusions, ensuring the dataset covers the full breadth of human visual experience. Every image was measured multiple times, delivering exceptional signal quality and setting new standards for the field.
Beyond functional brain scans, the dataset includes rich complementary data: retinotopic mapping, functional localizers, precision diffusion MRI, and behavioral responses - making it one of the most deeply characterized neuroimaging resources assembled to date.
Scale - thousands of unique images per participant (including 1,492 shared images in the launch release), 30 main image-viewing sessions each, up to 12 repeats for shared images
Acquisition - multi-echo 7T fMRI at 1.8 mm isotropic, 1.9 s TR
Broad sampling - natural photographs, prior benchmark images (NSD, THINGS), plus out-of-distribution test stimuli
Single-trial betas - GLMsingle-derived response estimates with strong noise ceilings
Complementary data - retinotopy, functional localizers, diffusion MRI, behavioral responses
Open - freely available for research
Getting Started¶
Get started quickly with basic examples
Overview of all data, spaces, and ROIs
Download and access instructions
Interactive 3D brain explorer
Citation¶
Until the dataset paper is available, please cite the VSS 2026 conference presentation:
Zerbe, J., Roth, J., Mell, M. M., Herholz, P., Knapen, T., & Hebart, M. N. (2026). LAION-fMRI: A densely sampled 7T-fMRI dataset providing broad coverage of natural image diversity. Talk 25.11, Scene Perception Talk Session, Vision Sciences Society Annual Meeting, May 16, 2026. VSS abstract.
If you use the provided GLMsingle beta estimates, also cite GLMsingle. See Example Methods Text for BibTeX and suggested methods wording.
