Free Data Worth Sharing
NeuroJSON.io is an NIH-funded open data portal for neuroimaging datasets, designed to make scientific data:
- Findable: Fully searchable metadata and datasets
- Accessible: Open, lightweight JSON format
- Interoperable: Compatible across platforms and programming languages
- Reusable: Rich metadata, visualizations, and long-term viability
NeuroJSON leverages modern web technologies and scalable NoSQL databases and the JSON standard to distribute large-scale, complex imaging data in a human- and machine-readable form.
- Search — Browse across multiple databases, datasets, and modalities
- Preview — Interact with JSON metadata and visualize imaging data (2D/3D) in browser
- Download — Retrieve datasets in JSON format, ready for use in Python, MATLAB/Octave, C++, and more
- Upload - Contribute your own datasets to NeuroJSON.io (Steps to contribute)
- REST API — Automate your workflows with lightweight endpoints, designed for smooth integration into both local analyses and large-scale pipelines
- Visit https://neurojson.io
- Use the search page to find datasets or subjects of interest
- Click any dataset to preview or download data
- For automation, use the REST API
We welcome your datasets!
- NeuroJSON prefers BIDS-compliant data
- Convert datasets to JSON using
NeuroJSON Client(neuroj):- Install Docker (skip this step if you already have it): Get Docker
- Download neuroj via docker:
docker pull openjdata/neuroj:v2025 - Example (convert a single dataset to JSON via neuroj):
docker run openjdata/neuroj:v2025 neuroj -i /path/to/database/rootfolder -o /path/to/output/json/folder -db openneuro -ds ds000001 --convert - See the full list of available NeuroJSON Client commands
- Watch our
tutorial video - convert data
- Download NeuroJSON Client (neuroj)
- Convert your dataset to JSON
- Validate metadata
- Open a ticket to upload your dataset
Contributions ensure long-term public availability and reusability.
- Lightweight endpoints for download
- JSON responses designed for integration with cloud and local workflows
Example (Load by URL with REST-API in Python):
pip install jdata bjdata numpy
import jdata as jd
data = jd.loadurl('https://neurojson.io:7777/openneuro/ds000001')
# List all externally linked files
links = jd.jsonpath(data, '$.._DataLink_')
# Download & cache anatomical nii.gz data for sub-01/sub-02
jd.jdlink(links, {'regex': 'anat/sub-0[12]_.*.nii'})
| Metric | Value |
|---|---|
| Databases | 22 |
| Datasets | 1,529 |
| Subjects | 58,026 |
| Links | 580,857 |
| Data Size | 38 TB |
- NIH-funded data dissemination service
- Maintained by the COTI Lab, Northeastern University
- Contact: [email protected]
We’re continuing to grow NeuroJSON.io to better serve the community. Some of the upcoming directions include:
- 🔜 Expanding databases and datasets — broadening coverage to include more sources and subjects
- 🔜 Enhancing visualization — improving 2D/3D previews for richer and more intuitive exploration of data
- 🔜 Streamlining uploads — introducing new features to make dataset contributions more automatic and user-friendly