This past June I presented our work “ReproIn: automatic generation of shareable, version-controlled BIDS datasets from MR scanners” at the Organization for Human Brain Mapping meeting in Singapore. In recent years, the international neuroimaging community pushed forward initiatives to create standard data formats to enhance reproducibility and transparency, with the goal of increasing sharing of datasets across researchers. Adoption of an open and common standard also allowed the creation of analysis routines that can be tested and shared across centers, increasing the reliability of research in Cognitive Neuroscience.
At the Dartmouth Brain Imaging Center (http://dbic.dartmouth.edu/) we developed an automated pipeline to standardize Magnetic Resonance (MR) datasets into the Brain Imaging Data Structure (BIDS) format directly from acquisition. In the past, researchers had to manually convert these datasets—a painstaking task that took away much of precious time needed to address important scientific questions. With our pipeline, PIs and graduate students in our center can start using automated open tools the same day they collect their data. In the spirit of open science, we released all our material freely (http://reproin.repronim.org), so that other researchers around the world can also save time and improve their data curation practices.
I presented this work at the Informatics oral session during OHBM. The talk was very well received and a large audience attended, comprising international leaders in the field and young investigators. The large attendance showed the importance and need for such automated tools by the neuroimaging community. Because our work is based on open tools, my talk also highlighted the importance of contributing to open source software to avoid “reinventing the wheel” and provide the community with stable and reliable means to advance the science in our field. After the talk and during the poster presentation, I had many insightful discussions that will guide the future development of our pipeline.
I was able to present in Singapore thanks to the support of the Graduate Student Council, the Neukom Institute, the Department of Psychological & Brain Sciences, and my advisor, professor M. Ida Gobbini. This work was co-authored by researchers from three different institutions, to whom I’m deeply grateful for their help. At Dartmouth, James Dobson, Terry Sackett, Chandana Kodiweera, James Haxby, and Yaroslav Halchenko; at MIT, Mathias Goncalves and Satrajit Ghosh; and at the University of Magdeburg, Germany, Michael Hanke. Support for the development also came from grants NIH #1P41EB019936-01A1, NSF #1429999. Slides from my talk are available here and a copy of my poster can be found here.