UC Berkeley joined a collaboration announced Monday to make data gathered for neuroscience research easier to share and interpret.
Data in the field of neuroscience are not widely shared due to their magnitude and the differing data-collection languages labs use. Some scientists aim to develop a more standard framework that would make data more accessible to other researchers.
A small group of scientists from UC Berkeley, California Institute of Technology and New York University School of Medicine, along with other medical research institutes have teamed up with the goal of creating a standard sharing language.
Called “Neurodata without Borders,” the collaboration will create the language with the hope that it will become an international standard.
UC Berkeley neuroscience adjunct professor Friedrich Sommer hopes the project will “change the culture of neuroscience,” which does not have a history of data-sharing.
Other academic fields have more data-sharing efforts than neuroscience, where data frequently stay within one lab, said Jeff Teeters, a researcher and programmer analyst at UC Berkeley. He said one explanation could be the competitive spirit among neuroscientists and the lack of sharing incentives.
Gyorgy Buzsaki, a professor of neuroscience at NYU’s medical school, said he believes data are not more widely shared due to differences in the way they are gathered, which requires arduous efforts to interpret and hours of work.
“You would not look at a cute picture I promised would make you laugh if you had to download applications and go through a lot of work to see it,” he said as an analogy. “It’s just too much.”
As technology improves and researchers become better at interpreting data, Sommer said the need to share data is increasing because researchers may need others to examine their work. The collaborative project aims to create a framework that will be utilized among many different labs.
One problem in creating a standard language among researchers is the use of “metadata.” Metadata are data describing the context under which other data were gathered, such as temperature or a subject’s behavior.
Sommer and Teeters, who both work for the campus’s Redwood Center for Theoretical Neuroscience, have been working on a website that collects data from various researchers. The data collected on the site will be used to find commonalities among data languages to create a loose framework for data-sharing.
“We are finding we are still far from understanding what this very complex structure in our heads is doing,” Sommer said. “We feel that to get to the next stage of understanding in the field, that data-sharing is a very major part.”