fNIRS Open Dataset Contest

Contest Overview

Type: Data sharing challenge Goal: Encourage the release of high-quality, well-documented fNIRS datasets that others can conveniently reuse. Timeline: Submissions open following the official announcement (end of Nov.) and close on Sep 1, 2026; open for votes between Sep and Oct, and announce the winners during the SfNIRS meeting in Macau (Oct 16-19, 2026). Recognition: The top three (or five) contributors will be recognized during the conference, possibly within existing award categories (no budget changes required, or could be company sponsored. Alex and I will look into it). May have the dataset shown on the cover of Neurophotonics (will ask Anna), featured on SfNIRS website, may send out a special newsletter highlighting the submitted and winning datasets. Scoring:

Committee votes: 30% - need to invite a panel of ~5 members Community votes: 70% - create website for receiving community votes, asking SfNIRS to send out invitations to all members and advertise broadly

Rating criteria

• Standard Compliance: Whether the dataset follows best practices and up-to-date standards, including but not limited to BIDS, NWB, 10-20 system standardized localization, or other established data standards. Datasets that are not currently supported by these standards are also acceptable, but all individual data files should be clearly organized using standardized file formats such as SNIRF, NIfTI, NeuroJSON/JSON formats.
• Quality: Clear documentation with detailed annotations that are easy to understand. Encouraged elements include block design diagrams, probe layout, and data quality assessments wherever applicable.
• Quantity: Datasets featuring a large number of subjects, data channels (trials, measurement duration), and/or modalities will be considered favorably. Detailed and rich metadata are expected, including biological factors (participants.tsv/participants.json) and behavioral measurements when available.
• Reproducibility: All relevant tasks, metadata, scripts, and stimuli should be provided to ensure other groups can easily reproduce results with minimal effort. The dataset should include sufficient metadata and acquisition details to allow meaningful secondary analyses and future extensions.

Dataset eligibility

• Datasets that have not been previously published - including new datasets, and datasets associated with a previously publication, but the dataset was not publicly available/released
• New datasets are encouraged to submit as a Data Paper for the “fNIRS Commons” special issue of Neurophotonics (free of charge)
• Can include measurements from human, animals, may include synthetic datasets (exp: simulated fNIRS signals for developing signal processing methods)
• Can be submitted from any fNIRS research labs
• Dataset must follow ethical and legal requirements, such as de-deidentification of human data
• Submissions can include any type of fNIRS data, such as task-based, resting-state, or naturalistic recordings.

Dataset must be submitted to one or multiple (encouraged) public data hosting websites such as OpenNeuro.org, NeuroJSON.io, DANDI Archive, etc and provide the public URLs in the submission form.

• All datasets formatted following the BIDS standard, the dataset must be able to pass the BIDS validator; submitters are also recommended to use our “dataset_description.json” creator to ensure the format is compatible with the BIDS standard.