Open Resting-State fMRI & Structural MRI for ADHD Research
data-indi/adhd200

The ADHD-200 Sample is one of the earliest and most influential open neuroimaging resources for studying Attention Deficit Hyperactivity Disorder (ADHD). Released as part of the 1000 Functional Connectomes Project / INDI, the dataset provides unrestricted, non-commercial research access to multi-site MRI and phenotypic data from children and adolescents with and without ADHD.

This initiative was created to accelerate discovery-based science and to support the development of objective biomarkers for ADHD.

Dataset Overview

• Total subjects: 776
  • Typically developing (TD): 491
  • ADHD (all subtypes): 285
• Age range: 7–21 years
• Sites: 8 independent imaging centers
• Modalities:
  • Resting-state fMRI
  • Anatomical MRI (T1-weighted)
• Phenotypic data:
  • Diagnosis
  • ADHD symptom scores
  • Age, sex
  • IQ measures
  • Lifetime medication status
• QC information:
  • Preliminary rs-fMRI quality ratings (usable vs. questionable), based on visual inspection

All data have been fully de-identified in accordance with HIPAA and 1000 Functional Connectomes Project protocols.

Scientific Motivation

ADHD affects 5–10% of school-age children and results in substantial lifelong impairment and economic burden (> $36B/year in the US). Despite decades of research:

• There is no comprehensive pathophysiological model of ADHD.
• Clinical diagnosis lacks objective biological tools.

The ADHD-200 Sample was designed to remove barriers to data access, enabling the global scientific community to:

• Build predictive biomarkers
• Improve diagnostic classifiers
• Investigate the neural basis of ADHD
• Evaluate developmental trajectories and heterogeneity

Multi-Site Structure

Participating imaging sites include:

• Bradley Hospital / Brown University
• Kennedy Krieger Institute
• NeuroIMAGE sample
• New York University Child Study Center
• Oregon Health & Science University
• Peking University
• University of Pittsburgh
• Washington University in St. Louis

Data quality, scanning protocols, and phenotypic variables vary by site; a phenotypic key is provided by the original project to ensure consistent interpretation.

Known Data Fixes (Historical Notes)

The ADHD-200 team issued corrections shortly after release:

April 11, 2011 – Structural Voxel Size Fix

• 73 subjects had incorrectly oriented structural images (x/z dimension swap).
• Affected sites: Peking_1, NYU_part2, OHSU.
• Corrected structural images and a list of affected subjects were re-released.

July 1, 2011 – Phenotypic Correction

• 49 subjects had errors in gender or IQ fields.
• Affected sites: Peking_1, KKI, NYU_part1, OHSU.
• A quick-fix phenotypic CSV was released.

These corrections are already incorporated into most modern mirrors of ADHD-200.

The ADHD-200 Global Competition (2011)

To stimulate methodological innovation, the project launched a global open competition with two tracks:

1. Best Diagnostic Classification Algorithm

Participants used ADHD-200 imaging + phenotypics to classify:

• ADHD-Combined type
• ADHD-Inattentive type
• Typically developing controls

A held-out test set (unlabeled) was released on July 1, 2011. Submissions were evaluated by accuracy.

2. Most Innovative Neuroscientific Analysis

Participants submitted novel ADHD pathophysiology manuscripts using the March 1 release.

Winners were invited to present at the NIDA pre-conference symposium before the Society for Neuroscience Annual Meeting.

Licensing & Access

• Usage: Free for non-commercial research (in line with the 1000 Functional Connectomes Project).
• Requirement: Register/login at NITRC to download.
• Source:
  • NITRC ADHD-200 page
  • 1000 Functional Connectomes Project portal

When publishing, specify the sites and data subsets used and acknowledge funding sources listed by the ADHD-200 consortium.

Citation

If you use the ADHD-200 Sample, please cite the original INDI / ADHD-200 publications and acknowledge the participating institutions.

Notes for This Repository

This release (data-indi/adhd200) provides:

• Standardized NIfTI derivatives (when permitted)
• DSI Studio–ready diffusion/tractography files when applicable
• Harmonized phenotypic tables (phenotypic.csv)
• QC summaries (qc.tsv)
• Consistent BIDS-style organization across sites

Users wishing to reconstruct full pipelines should refer to the original ADHD-200 documentation hosted on NITRC.

Resting-State fMRI, Structural MRI, and Extended Phenotyping for Autism Research
data-indi/abide2

The Autism Brain Imaging Data Exchange II (ABIDE II) expands upon the success of ABIDE I by assembling an even larger and more richly characterized neuroimaging sample for Autism Spectrum Disorder (ASD). Released in June 2016, ABIDE II was supported by the National Institute of Mental Health (R21MH107045) and brings together over 1,100 MRI datasets contributed by 19 research sites worldwide.

ABIDE II strengthens the foundation for discovery-based ASD neuroscience by:

• increasing sample size,
• improving phenotypic depth,
• broadening age ranges and symptom characterization, and
• providing rare longitudinal data for a subset of participants.

All datasets follow HIPAA and 1000 Functional Connectomes Project / INDI anonymization guidelines.

Dataset Overview

• Total MRI datasets: 1,114
  • ASD: 521
  • Typically developing controls (TD): 593
• Age range: 5–64 years
• Sites: 19 institutions (10 charter + 7 new contributors)
• Phenotyping: Expanded clinical detail for ASD and comorbid symptoms
• Longitudinal samples:
  • 38 individuals scanned twice, 1–4 years apart (two collections)

Modalities available:

• Resting-state fMRI (R-fMRI)
• T1-weighted anatomical MRI
• Quality metrics for MRI acquisition
• Extended phenotypic profiles

The improved phenotyping in ABIDE II enables more precise analyses of ASD heterogeneity, symptom dimensions, and developmental biomarkers.

Scientific Motivation

ABIDE I showed that multi-site R-fMRI aggregation is feasible and scientifically valuable.
However, ASD presents deep heterogeneity across genetics, behavior, and neurobiology.

ABIDE II was created to:

• Build larger, more diverse, and better-characterized cohorts
• Support robust machine learning and biomarker discovery
• Enable finer-grained connectome analyses across development
• Provide longitudinal views of ASD brain development
• Promote open, reproducible neuroscience

The expanded dataset addresses the complexity of ASD by integrating broader symptom measures, refined behavioral assessments, and larger sample sizes, fostering discovery across biological, clinical, and computational domains.

Data Access & Licensing

Consistent with the 1000 Functional Connectomes Project / INDI, ABIDE II data are:

• Free for non-commercial research use
• Distributed under Creative Commons Attribution–NonCommercial–ShareAlike License (CC BY-NC-SA)
• Downloadable through NITRC, requiring:
  • A registered NITRC account
  • Registration with the 1000 FCP / INDI site
  • Authentication before download (required to avoid permission errors)

Available Files

• ABIDE II Composite Phenotypic File
• Longitudinal Phenotypic File
• Phenotypic Data Legend
• Imaging Data
  • Site-specific R-fMRI and T1w datasets
  • MRI data quality metrics
• QC resources

For technical issues, refer to the INDI forums.

Participating Sites (19 Institutions)

• Barrow Neurological Institute
• Erasmus University Medical Center Rotterdam
• ETH Zürich
• Georgetown University
• Indiana University
• Institut Pasteur & Robert Debré Hospital
• Katholieke Universiteit Leuven
• Kennedy Krieger Institute
• NYU Langone Medical Center — Sample 1
• NYU Langone Medical Center — Sample 2
• Olin Neuropsychiatry Research Center (Hartford Hospital)
• Oregon Health & Science University
• Trinity Centre for Health Sciences
• San Diego State University
• Stanford University
• University of California Davis
• University of California Los Angeles
• University of Miami
• University of Utah School of Medicine

Longitudinal Cohorts

• University of California Los Angeles (Longitudinal Sample)
• University of Pittsburgh School of Medicine (Longitudinal Sample)

Notes for This Repository

The data-indi/abide2 release includes:

• Harmonized phenotypic tables
• Optional QC tables (qc.tsv)
• DSI Studio–ready derivatives (*.sz, *.gqi.fz, *.qsdr.fz) when DWI is available
• Standardized folder layout matching ABIDE I and other INDI datasets in the Fiber Data Hub

Raw archives, DICOM collections, and original site-level structures are available via NITRC for users who need complete imaging packages.

Acknowledgements

ABIDE II was supported by:

• National Institute of Mental Health — NIMH R21MH107045
• Additional contributions from the Nathan S. Kline Institute and the Child Mind Institute
• Philanthropic gifts from Joseph P. Healey, Phyllis Green, and Randolph Cowen

We acknowledge the ABIDE II directors (Adriana Di Martino, Michael P. Milham), the INDI team, the participating institutions, and all researchers who contributed data.

Citation

Please cite ABIDE II and all relevant site-level funding sources when using these data.
All ABIDE publications and updates are available on the INDI / ABIDE website.

Automated release for abide1_1

Resting-State fMRI, Structural MRI, and Phenotypics for Autism Research
data-indi/abide1

The Autism Brain Imaging Data Exchange I (ABIDE I) is the first large-scale, grassroots initiative to aggregate and openly share resting-state fMRI, anatomical MRI, and phenotypic data for Autism Spectrum Disorder (ASD). Released in August 2012, ABIDE I combines previously collected data from 17 international imaging sites, enabling researchers to study ASD neurobiology at unprecedented scale.

This effort demonstrated that harmonizing multi-site resting-state fMRI is feasible and highly impactful—ABIDE I has since become one of the most widely used ASD neuroimaging resources.

Dataset Overview

• Total MRI datasets: 1,112
  • ASD: 539
  • Typically developing controls (TD): 573
• Age range: 7–64 years (median ~14.7)
• Sites: 17 contributing institutions
• Modalities:
  • Resting-state fMRI (R-fMRI)
  • T1-weighted anatomical MRI
• Phenotypic data:
  • Diagnosis (ASD/TD)
  • Age, sex
  • Handedness
  • IQ measures
  • ADOS / ADI-R scores when available
  • Medication status
• QC: Preliminary scan usability ratings provided (usable vs. questionable)

All datasets were anonymized according to HIPAA and INDI / 1000 Functional Connectomes protocols.

Scientific Motivation

ASD affects communication, behavior, and social interaction. However, understanding its neural mechanisms remains challenging due to heterogeneity and small sample sizes.

ABIDE I was created to:

• Enable large-scale characterization of ASD functional and structural brain organization
• Promote data-driven discovery science
• Lower barriers to reproducibility and biomarker development
• Provide the community with a reference atlas for ASD functional connectomics

ABIDE I has since powered hundreds of manuscripts across neuroimaging, machine learning, and developmental neuroscience.

Participating Sites (17 Total)

ABIDE I aggregates data from:

• California Institute of Technology
• Carnegie Mellon University
• Kennedy Krieger Institute
• Ludwig Maximilians University Munich
• NYU Langone Medical Center
• Olin / Institute of Living, Hartford Hospital
• Oregon Health & Science University
• San Diego State University
• Social Brain Lab (BCN NIC UMC Groningen & Netherlands Institute for Neurosciences)
• Stanford University
• Trinity Centre for Health Sciences
• UCLA — Sample 1
• UCLA — Sample 2
• University of Leuven — Sample 1
• University of Leuven — Sample 2
• University of Michigan — Sample 1
• University of Michigan — Sample 2
• University of Pittsburgh School of Medicine
• University of Utah School of Medicine
• Yale Child Study Center

(Several sites contribute multiple cohorts.)

Data Access & Usage Agreement

ABIDE I follows the 1000 Functional Connectomes Project policy:

• Usage: Free for non-commercial research
• License: Creative Commons Attribution–NonCommercial–ShareAlike
• Requirement:
  • Register and log in to NITRC
  • Register with the 1000 FCP / INDI portal
  • Download requires being logged in; otherwise, a permission error appears

Downloads Provided:

• Composite phenotypic file
• Phenotypic legend / data dictionary
• R-fMRI and T1-weighted imaging data (site-specific profile pages list available files)

For issues, users should consult the INDI forums.

Supported Platforms

To improve accessibility, ABIDE I is mirrored across multiple informatics platforms:

• COINS
• LORIS
• LONI IDA
• NITRC-IR
• NIDB
• BIRN HID

Each platform hosts datasets prepared by the INDI team, allowing users to explore different data management systems for their own workflows.

Notes for This Repository

The data-indi/abide1 release provides:

• Clean NIfTI organization when permitted
• DSI Studio–ready diffusion derivatives (*.sz, *.gqi.fz, *.qsdr.fz) for subjects with DWI available
• Harmonized phenotypic tables (phenotypic.csv)
• Optional QC tables (qc.tsv)
• Consistent folder structure matching other INDI collections in Fiber Data Hub

Users requiring the full imaging archives (tarballs, DICOM, or site-level directories) should download directly from NITRC.

Acknowledgements

Primary support came from:

• NIMH K23MH087770 (Di Martino)
• Leon Levy Foundation
• Gifts from Joseph P. Healy and the Stavros Niarchos Foundation to the Child Mind Institute
• NIMH R03MH096321 (Milham)

We acknowledge the ABIDE consortium founders, the INDI data aggregation team, and the participating institutions for enabling this landmark open dataset.

Citation

Please cite the ABIDE I initiative and appropriate site-level funding sources when using these data.
A full list of ABIDE manuscripts and publications is available on the ABIDE website.