The tool is under active development—features and bugs are tracked openly on GitHub. If you have ideas for new functionality or encounter any issues, just let me know and we can work through it together.

The package is on PyPI, the docs are hosted here and the source is on GitHub I’m keen to see applications in diverse domains—let me know if you try it in your own research workflows.

This raises a crucial meta-question: how do we assess the assessor? This platform enables exactly that—providing a structured way to quantify AI’s own biases when analysing academic literature. But that’s a topic for the next thread….

Human judgement remains the gold standard in RoB assessment. This tool is designed to augment that process—acting as a tireless additional reviewer on completed studies, or as a pre-emptive check on protocols.

Risk-of-bias accommodates different workflows—offering both a web interface and command line interface (no coding required). Outputs include standard formats (JSON) for reproducibility and robvis-compatible summaries for visualisation:

The tool’s architecture is modular, allowing systematic implementation of any formal risk-of-bias framework. While it ships with Cochrane RoB2, users can define, load, and share custom frameworks as needed.

I’ve been fortunate to work across a range of fields in recent years, and one practice I’ve come to appreciate—though absent earlier in my basic science career—is the formal evaluation of study bias. This tool uses AI to make that process broadly and systematically accessible.

Risk of bias assessment is a systematic method to evaluate how a study’s design, conduct, or analysis might introduce bias. It’s essential for credible evidence synthesis, yet remains underutilised outside fields like evidence-based medicine.

Introducing risk-of-bias: an open-source AI tool designed to assist with Risk of Bias (RoB) assessment. Built to augment human expertise, it helps researchers systematically evaluate potential flaws in scientific studies.

RT @jrbiesiekierski: We're expanding our Human Nutrition team @UniMelb @SciMelb!. Strong foundation in nutrition science & passionate about….

This group includes: @HanBnrd, @BoasDavid, @artinis_jorn, Sreekanth kura, , @eikemiddell and Sam Powell. We will also announce this at the meeting. Virtual attendance is possible.

The SNIRF Working Group for the fNIRS data standard will be meeting at the @SfNIRS Conference in Birmingham on Friday Sept 13, 2024 during the lunch time break. Please reach out to anyone on the Governance Board or one of the Maintainers if you wish to join.

@BoasDavid @artinis_jorn I couldn't fit the link in the original post :) .

The SNIRF steering committee (@BoasDavid, S. Powell, @artinis_jorn, me) seeks maintainers to join us in maintaining and updating the spec + working with the community. Great opportunity to get involved in fNIRS & open science/data. Interested or have questions? Comment below!.

My friends at Neurode are looking for great software and ML engineers to join their team. It would be a great place to work on an exciting neurotech product.

I'm excited to see the amazing experiments and tools people will build with @aestudiola's Neural Data Simulator. Its modular design enables integration into existing workflows, and it can be easily extended. We'll be presenting the work at the upcoming @BciSociety #BCIMeeting.

Add your comments to the document. And if you’re attending in person, come over and say hi 👋.

RT @fNIRSglossary: Hello team! While we're preparing the final documents, we would like to encourage everyone to join the FRESH study as we….

RT @ibs_bifold: @rob__luke, @mesquitarc, Meryem and I are very happy about all the submissions coming in for the.fNIRS REproducibility Stud….