Psychiatry is restricted by the absence of objective biological measures able to detect and discriminate between psychopathologies. Instead, psychiatric nosology remains exclusively syndromic, relying on clusters of signs and symptoms. Attention-Deficit Hyperactivity Disorder (ADHD) provides a salient example of where the imprecise nature of current nosological distinctions and absence of defining biological metrics represent fundamentally confounding factors limiting a better understanding of the disease. New structural and functional imaging tools may render these issues tractable for the first time; however, they require large-scale datasets and techniques capable of characterizing individuals.
Toward this end, the ADHD-200 Consortium, a self-organized, grassroots global initiative has emerged with the expressed mission of accelerating the implementation of imaging-based discovery science for ADHD. On March 1, 2011, the effort provided the world with its first bolus of data - 776 resting-state fMRI and anatomical datasets aggregated across 8 independent imaging sites, 491 of which were obtained from typically developing individuals and 285 in children and adolescents with ADHD (ages: 7-21 years old). Publicly available for unrestricted research use, the database will continue to grow.
To highlight the potential of this effort the coordinators of the ADHD-200 organized a global competition, aimed at identifying algorithms/approaches capable of classifying individual participants based on their functional imaging data. Participants were instructed to develop and train a diagnostic classifier (ADHD-Combined type, ADHD-Inattentive type, typically developing control) based on the ADHD-200. From there, the participants were invited to assess the quality of their classifier on a test set comprised of previously unreleased datasets without diagnostic labels. Winners were chosen based on their classification accuracy of the unlabeled test set. 50 teams expressed their intent to participate, with 21 of them completing the challenge - nearly all teams performed well above chance.
In this Research Topic of Frontiers in Neuroscience, we highlight the methods and approaches used by the teams for this competition. We believe the dissemination of the data and the methods from this work will accelerate the rate at which neuroscience can inform clinical characterization of many developmental neuropsychiatric disorders.
Psychiatry is restricted by the absence of objective biological measures able to detect and discriminate between psychopathologies. Instead, psychiatric nosology remains exclusively syndromic, relying on clusters of signs and symptoms. Attention-Deficit Hyperactivity Disorder (ADHD) provides a salient example of where the imprecise nature of current nosological distinctions and absence of defining biological metrics represent fundamentally confounding factors limiting a better understanding of the disease. New structural and functional imaging tools may render these issues tractable for the first time; however, they require large-scale datasets and techniques capable of characterizing individuals.
Toward this end, the ADHD-200 Consortium, a self-organized, grassroots global initiative has emerged with the expressed mission of accelerating the implementation of imaging-based discovery science for ADHD. On March 1, 2011, the effort provided the world with its first bolus of data - 776 resting-state fMRI and anatomical datasets aggregated across 8 independent imaging sites, 491 of which were obtained from typically developing individuals and 285 in children and adolescents with ADHD (ages: 7-21 years old). Publicly available for unrestricted research use, the database will continue to grow.
To highlight the potential of this effort the coordinators of the ADHD-200 organized a global competition, aimed at identifying algorithms/approaches capable of classifying individual participants based on their functional imaging data. Participants were instructed to develop and train a diagnostic classifier (ADHD-Combined type, ADHD-Inattentive type, typically developing control) based on the ADHD-200. From there, the participants were invited to assess the quality of their classifier on a test set comprised of previously unreleased datasets without diagnostic labels. Winners were chosen based on their classification accuracy of the unlabeled test set. 50 teams expressed their intent to participate, with 21 of them completing the challenge - nearly all teams performed well above chance.
In this Research Topic of Frontiers in Neuroscience, we highlight the methods and approaches used by the teams for this competition. We believe the dissemination of the data and the methods from this work will accelerate the rate at which neuroscience can inform clinical characterization of many developmental neuropsychiatric disorders.
