LAION-fMRI: 7-Tesla fMRI responses to 25,000+ natural images across 165 sessions

A dense-sampling 7-Tesla fMRI dataset from the Hebart Lab in which five participants viewed more than 25,000 unique natural photographs over 165 hours and 165 scan sessions, built to study how the human visual system represents and understands natural scenes. The scale is the point: responses are measured often enough per participant that effects can be replicated within an individual rather than only across a group. Organised as BIDS, with sub-01 through sub-07 at the root (five participants, non-contiguous labels), plus derivatives/, stimuli/ and archive/. Measured 2026-10-27: 75,680 objects, 2.89 TiB. File mix is 8,171 .json sidecars, 4,400 .nii.gz, 3,608 FreeSurfer .gii surface geometries, 2,202 .png, 288 .label, 180 .mgz and 174 .h5, so both volumetric and surface-space data are present. The heaviest single files are the GLMsingle plus tedana beta estimates under derivatives/glmsingle-tedana/ at about 1.23 GiB per session. The fMRI data, derivatives and metadata are CC0 1.0; the raw LAION stimulus images are not, and require accepting a separate Data Use Agreement on the project site before download.

Category
Neuroscience
Provider
Hebart Lab
License
CC0-1.0 (fMRI data, derivatives, metadata); raw stimulus images require a Data Use Agreement
Hosted on
Amazon S3, bucket laion-fmri, us-west-2
Access
Free and public. No account with the data provider is needed.
Source
https://registry.opendata.aws/laion-fmri/

Analyze this dataset without downloading it

On Scigantic, LAION-fMRI: 7-Tesla fMRI responses to 25,000+ natural images across 165 sessions opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

Related Neuroscience datasets