Healthy Brain Network
- Sponsored Open Neuroscience Intiatives
Sponsorship of Major Open Science Initiatives In the Imaging Field
The Child Mind Institute has become the proud sponsor of the 1000 Functional Connectomes Project and its International Neuroimaging Data Initiative. As described below, these efforts have distributed thousands of datasets to scientists around the world, who just a few years ago viewed a few dozen datasets as a major resource.
The Functional Connectomes Project
The era of discovery science for human brain function was inaugurated by the collaborative launch of the 1000 Functional Connectomes Project (FCP) on December 11, 2009, by leading members of the functional magnetic resonance imaging (fMRI) community. Following the precedent of full unrestricted data sharing, which has become the norm in molecular genetics, the FCP entailed the aggregation and public release (via www.nitrc.org) of over 1200 resting state fMRI (R-fMRI) datasets collected from 33 sites around the world. The data release has led to more than 40,000 downloads from around the world (1,223 cities in 78 countries) and two dozen publications in just over two years. The FCP was featured in Nature-Medicine and Nature-Methods and in the NIMH director's weekly blog. Also indicative of the FCP's scientific impact is the success of the initial publication demonstrating the feasibility of data-pooling and discovery science with the datasets (Biswal et al., 2010). Data processing steps employed to carry out the feasibility analyses were made available on 3/3/10, yielding almost 600 downloads from NITRC in the first three months. The first FCP-based paper by independent investigators has already been published in PNAS (Tomasi & Volkow, 2010) and many additional papers are in progress.
The International Neuroimaging Data-Sharing Initiative (INDI)
Having provided the first large-scale demonstration of the feasibility and scientific value of open sharing of R-fMRI data, the next major challenge is to make the aggregation and sharing of well-phenotyped datasets a cultural norm for the imaging community. In order to accomplish this, two major paradigm shifts will be required. First, comprehensive phenotypic information must be made available with imaging datasets to facilitate sophisticated data-mining—a process by which novel relationships between phenotypic and imaging data can be revealed. Scientists are often reluctant to release their phenotypic data, because of the concern that competitors will gain an advantage in answering specific scientific questions. Unfortunately, this parochial attitude leads to the perpetuation of highly under-powered studies and a plethora of false positive and false negative results, while potentially massive amounts of data remain locked-up in individual labs.
A second paradigm shift would be from retrospective to prospective data sharing. That is, in contrast to the FCP release, which primarily comprised datasets that had already been published, prospective data sharing involves regularly scheduled (e.g., weekly, monthly, or quarterly) sharing of data collected at contributing sites, as the data is being collected. The notion of sharing newly acquired data, rather than waiting until those data have been published, is novel in the imaging community, but is common practice in fields such as genetics where discovery science has been successfully implemented. In order for such a shift in practice to occur, one or more imaging groups must take the lead, and set an example for the field.
Based on the continued need for cultural advancement and modeling in order to sustain and accelerate shifts towards data-sharing, the INDI was formed as a next generation FCP effort. Comprised of two components, INDI-Pro and INDI-Retro, INDI aims to provide a model for the broader imaging community while simultaneously creating a public dataset capable of dwarfing those that most groups could obtain individually.
- ABIDE: Despite significant advances in understanding the neurobiological basis of autism, the sheer variety of clinical presentations and relatively sparse nature of inquiry have hampered the creation of a more total picture. The Autism Brain Imaging Data Exchange (ABIDE) aims to accelerate understanding of autism using the tenets of open neuroscience and the sharing tools developed by INDI and other initiatives.
- ADHD 200 Consortium: The ADHD-200 Sample is a grassroots initiative, dedicated to accelerating the scientific community's understanding of the neural basis of ADHD through open-data sharing and discovery-based science. The related ADHD-200 Global Competition puts the data in the Sample to the test in hopes of eventually demonstrating the clinical efficacy of imaging-based approaches.
- C-PAC: Open neuroscience is not only about sharing data or strategy—it is also about sharing tools and the knowledge to use them. The Configurable Pipeline for the Analysis of Connectomes is a software package designed to be used by anyone whose interests turn to what neuroimaging can tell us about the brain's function. In short, it is designed for the interdisciplinary group who make up the new open science.