🧠 fNIRS Open Dataset Contest 2026

Contest Overview

Type: Data sharing challenge

Goal: Encourage the release of high-quality, well-documented fNIRS datasets that others can conveniently reuse.

Rationale: Open data accelerates scientific discovery. By sharing your fNIRS dataset(s), you contribute to reproducible research, enable extended and advanced analyses, and help train the next generation of neuroscientists. This contest recognizes researchers who go above and beyond in making their data well-documented and easily accessible.

Scoring: Committee votes: 30% — Panel of ~5 expert members
Community votes: 70% — Open voting for all SfNIRS members and broader community

Timeline

Jan 2026: Submission window is open

Sep 1, 2026: Submission deadline

Sep–mid Oct 2026: Open community voting period

Oct 16–19, 2026: Winners announced at SfNIRS meeting in Macau, China

        Recognition and Awards:

1st Prize USD $2,000 or flight + conference fee + hotel in Macau (up to $2,000 total)
2nd Prize USD $1,000 or conference fee + hotel (up to $1,000 total)
3rd Prize USD $500 or conference fee waiver

Rating Criteria

Standard Compliance: Whether the dataset follows best practices and up-to-date standards, including BIDS, 10-20 system standardized localization, or other established data standards. Datasets not currently supported by these standards are also acceptable, but all files should use standardized formats such as SNIRF. For multi-modal datasets, all included data files are expected to be stored using data formats with widely supported open-standards, such as NIfTI, HDF5, or NeuroJSON/JSON formats.

Quality: Clear documentation with detailed annotations. Encouraged elements include block design diagrams, probe layout, and data quality assessments.

Quantity: Datasets with large numbers of subjects, data channels, trials, measurement duration, and/or modalities. Rich metadata including biological factors (participants.tsv/participants.json) and behavioral measurements.

Reproducibility: All relevant tasks, metadata, scripts, and stimuli should be provided to extend data acquisition and replicate analyses. Sufficient metadata and acquisition details for meaningful secondary analyses.

BIDS Requirement: All datasets following the BIDS standard must pass the BIDS validator. Submitters are also recommended to use the dataset_description.json creator to ensure format compatibility.

Dataset Eligibility

Datasets that have not been previously published — including new datasets and datasets associated with prior publications where the data was not publicly released
New datasets are encouraged to submit as a Data Paper for the "fNIRS Commons: Datasets and Tools for Open, Reproducible Science" Special Issue of Neurophotonics (no publication fees for Data Paper)
Can include measurements from humans, animals, or synthetic datasets (e.g., simulated fNIRS signals for developing signal processing methods)
Can be submitted from any fNIRS research lab
Must follow ethical and legal requirements, such as de-identification of human data
Submissions can include any type of fNIRS data: task-based, resting-state, or naturalistic recordings
Dataset must be submitted to one or multiple public data hosting websites (OpenNeuro.org, NeuroJSON.io, DANDI Archive, etc.) and provide public URLs in the submission form

Dataset Hosting and Conversion Resources

BIDS-ifyfing fNIRS Hands-On Workshop:

NeuroJSON.io

For this data contest, the NIH-funded NeuroJSON project offers free data hosting services at our data portal, NeuroJSON.io, for all participating datasets. In addition, NeuroJSON developers will also commit resources to help participants convert their datasets to the standardized BIDS format (or NeuroJSON format if the dataset does not fit within the scope of BIDS). Please reach out to the project maintainer, Dr. Qianqian Fang (q.fang@neu.edu), to request help with data conversion.

NeuroJSON.io is an NIH-funded neuroimaging dataset dissemination portal with a focus on accelerating fNIRS dataset sharing. By utilizing the universally supported JSON format and high-performance NoSQL databases, the NeuroJSON platform can host large amounts of complex datasets with high scalability, easy searchability, and programmable access in most programming environments via RESTful APIs.

SfNIRS Standardization and Open Science Committee