The other day I received an email from Stanford psychologist Russell Poldrack, PhD, letting me know about a new study, about which he wrote, “A format for how people should arrange their files probably sounds like the most boring thing on earth.”
I really couldn’t have agreed with him more. File storage? Yawn.
But he went on to describe the paper, which appeared recently in Nature Scientific Data, as an example of how creating a community standard – in this case laying out how to organize files containing brain images – could make it easier for scientists to share data and collaborate.
Getting science done is something that I do care about, so I decided to give Poldrack a chance.
In the article, Poldrack and his co-authors point out that imaging data from MRIs are stored in many different formats and can’t always be shared between labs. What this means is that two scientists who are both interested in a common problem – autism, for example – might not be able to share imaging data, and that slows the pace of research.
The group introduced something called the Brain Imaging Data Structure, aka "BIDS", a standard way of organizing and storing MRI data. It’s kind of like telling all your colleagues to use compatible word processing software so you can share documents. Without some compatibility you would never be able to read or learn from their work.
“BIDS will let people take shared datasets from any lab and immediately know how to process them,” Poldrack told me.
The first author on the paper, Chris Gorgolewski, PhD, is a postdoctoral fellow in Poldrack’s lab and together they co-lead the Center for Reproducible Neuroscience, which has the goal of helping neuroscientists share data to make science more transparent and reproducible.
It’s true that data storage remains somewhat tedious to this non-imaging expert, but as someone who follows brain research, it’s good to know that scientists will be better able to compare findings and perhaps learn more about what makes us tick.