The Informatics Backbone of the Brain Genomics Superstruct Project Open Data Release
Timothy O'Keefe (Harvard University, Neuroinformatics Research Group), Victor Petrov (Harvard University, Neuroinformatics Research Group), Gabriele Fariello (Harvard University, Neuroinformatics Research Group), Avram Holmes (Harvard University, Center for Brain Science), Randy Buckner (Harvard University, Center for Brain Science)
Large scale imaging data sets are necessary to address complex questions regarding the relationship between brain and behavior. Generating, storing and analyzing the required data are a daunting enterprise for many independent research groups. In 2007, the Open Access Series of Imaging Studies (OASIS) sought to remove these obstacles by developing a distribution model for free and reusable magnetic resonance imaging data sets (Marcus et al., 2007). The community has benefited from these and other open data initiatives including the 1000 Functional Connectomes Project (Biswal et al., 2010) and the upcoming NIH Human Connectome Project data release (Van Essen et al., 2012). The availability of open data creates opportunities for researchers to contribute scientifically while spending less time and resources gathering independent, and often redundant, data. In the spirit of these initiatives, the Brain Genomics Superstruct Project Open Data Release presented here reflects the public release and informatics behind a uniform, high-quality collection of neuroimaging, cognitive, behavioral and derived data for 1,500 human participants. These data sets will be available from a hosted or downloadable installation of the eXtensible Neuroimaging Archive Toolkit (XNAT; Marcus et al., 2007). Each data set will contain T1-weighted and bandwidth-matched T2-weighted structural data, low-resolution DTI, resting state BOLD acquisitions and, for a subset of subjects, DSI data amenable to tractography. Demographic, cognitive (e.g., WAIS III, WMS III), personality (e.g., STAI-T, NEO) and lifestyle metrics will also be provided. Each data set will be accompanied by a fully-automated quality assessment of functional acquisitions, manual quality assessments of anatomical acquisitions, and pre-computed analyses of intrinsic connectivity (Van Dijk et al., 2010) and morphometrics (Fischl et al., 2000; 2004). We will present details regarding the underlying informatics needed to capture, vet and publicly expose these data. We expect that this Brain Genomics Superstruct Project Open Data Release will prove a valuable resource, fueling discoveries, particularly within the NIH Human Connectome Project.
Preferred presentation format: