This week (Oct 17-19th), I attended the BIDS-Stats-Models meeting in Stanford, a collaborative workshop devoted to the extension of the 'Brain Imaging Data Structure' (BIDS) to statistical models.
BIDS defines a folder structure to organize brain images. Since its creation in 2015, many tools and databases have adopted this standard making it easier to exchange datasets within and between labs. Beyond the original specification –which covered anatomical, functional and diffusion MRI–, extensions are underway for PET, MEG, Arterial Spin Labeling and more.
BIDS covers basics metadata for functional MRI data (fMRI) which makes it possible to build automatic pipelines for preprocessing (e.g. fmriprep). But, to date, there is no support to describe the necessary pieces for fMRI statistical analyses. The BIDS-Stats-Models extension aims at adding this support. The goal of this meeting was to solidify the current draft of the BIDS-Stats-Models proposal. Here is a brief overview of what happened.
A review of what was accomplished and plans for future steps are at: bids-standard/bids-specification#60. Interested in this effort? Please consider adding your own contributions: https://bids.neuroimaging.io!
I'd like to thank Franklin Feingold, Chris Gorgolewski, Tal Yarkoni and Russ Poldrack for this great meeting.
The event was covered on Twitter with #bidsstatmodels.
The text was updated successfully, but these errors were encountered: