Written and maintained by CASRAI Editorial Board
Last updated
The Human Connectome Project (HCP) is an open-science neuroimaging initiative established by the National Institutes of Health (NIH) Blueprint for Neuroscience Research to map neural pathways in the healthy adult brain. Supported across 16 NIH institutes under grant 1U54MH091657, the project combined high-field magnetic resonance imaging (MRI), magnetoencephalography (MEG), and behavioral testing across 1,200+ healthy participants. The resulting data assets serve as a reference benchmark for studying brain circuitry, developmental trajectories, and neurological conditions in accordance with FAIR data principles.
What It Is
The Human Connectome Project is a systematic initiative characterizing anatomical and functional brain connectivity across the healthy human lifespan. Launched in 2010 under the NIH Blueprint for Neuroscience Research, the project addressed limitations in neuroimaging resolution, sample size, and cross-site protocol variation. Rather than compiling disparate scans across heterogeneous scanners, the HCP established standardized acquisition sequences and public preprocessing pipelines to examine macroscale neural circuits at high spatial and temporal detail.
Beyond mapping brain architecture, the HCP pioneered open multimodal dataset dissemination. By hosting deeply phenotyped collections in a centralized neuroimaging data repository, the initiative enabled reproducible secondary analysis across structural MRI (sMRI), resting-state functional MRI (rfMRI), task fMRI (tfMRI), diffusion MRI (dMRI), and MEG, alongside cognitive and sensory batteries.
Who Runs It and What It Covers
The flagship phase of the Human Connectome Project was conducted by two academic consortia funded by the NIH:
- The Washington University–University of Minnesota (WU-Minn) Consortium: Led by David Van Essen and Kamil Ugurbil, mapping healthy young adults on customized 3-Tesla (3T) and 7-Tesla (7T) scanners paired with behavioral testing.
- The Massachusetts General Hospital–UCLA Consortium: Led by Bruce Rosen and Arthur Toga, optimizing diffusion hardware, including 300 mT/m gradient coils to image white matter microstructure.
Subsequent extensions expanded coverage across the lifespan and clinical cohorts, including HCP-Development (ages 5–21), HCP-Aging (ages 36–100+), the Adult Aging Brain Connectome (AABC), and disease-focused connectomes.
Dataset Scope and Cohort Numbers
Official release documentation records the following cohort sizes and scan configurations:
- HCP Young Adult (S1200 Release): Includes behavioral and 3T MRI data from 1,206 healthy participants (ages 22–35; collected 2012–2015), with 1,113 completing structural MRI scans. Subcohorts include 184 participants with 7T multimodal MRI, 95 with resting-state/task MEG, and 46 completing full 3T repeat scans.
- Adult Aging Brain Connectome (AABC Release 2, updated January 2026): Encompasses cross-sectional and longitudinal 3T MRI and phenotypic data from 1,396 aging adults (ages 36–90+), comprising 2,878 total sessions, 2,789 imaging visits across 1,390 participants, and 7T magnetic resonance spectroscopy (MRS) for 282 individuals.
- Image Preprocessing Pipelines: Datasets include unprocessed archives and standardized minimally preprocessed outputs. Diffusion pipelines incorporate FMRIB Software Library (FSL) EDDY tools for slice outlier detection and motion correction. Functional pipelines feature ICA-FIX artifact removal and MSM-All surface registration.
How to Search or Use It
Researchers access HCP datasets through ConnectomeDB powered by BALSA (db.humanconnectome.org) or cloud mirrors via eight verified procedural steps:
- Navigate to ConnectomeDB: Visit the database homepage at
db.humanconnectome.organd select active collections, including HCP-Young Adult, HCP Retest, or AABC. - Register a User Account: Locate the sign-in area and create a profile using an institutional email address, confirming registration through the automated email link.
- Accept Open Access Data Use Terms: Complete the electronic agreement for the WU-Minn HCP Consortium Open Access Data Use Terms to generate an institutional compliance record.
- Query and Filter Cohorts: Use ConnectomeDB search filters to isolate participant subsets by modality (3T, 7T, MEG), behavioral measures, or quality control flags (such as the
QC_Issueattribute). - Select Data Packages: Choose between unprocessed NIfTI files, minimally preprocessed volumetric and surface packages, or group-average dense connectomes.
- Submit Restricted Data Application (If Required): If research requires sensitive participant measures—specifically exact chronological age, twin/sibling zygosity (
ZygosityGT), or handedness—the institutional principal investigator must submit an Electronic Restricted Access Application. - Download via Direct Transfer or Cloud Bucket: Download packages through the browser, synchronize cohorts from the Amazon Web Services (AWS) Open Data S3 bucket (
hcp-openaccess) using AWS CLI, or order pre-loaded hard drives via “Connectome in a Box”. - Visualize with Connectome Workbench: Open extracted CIFTI-2 and GIFTI surface/volume files in Connectome Workbench (
wb_view) for 3D visualization or process them via command-line utilities (wb_command).
Strengths and Limits
Strengths
- Acquisition Quality: Sub-millimeter anatomical resolution (0.7 mm isotropic T1w/T2w) and fast temporal sampling provide high fidelity for structural and functional mapping.
- Multimodal Breadth: Integrated structural, functional, diffusion, and MEG data paired with standardized cognitive testing enable cross-modal brain-behavior modeling.
- Open-Science Standard: Rigorous governance and public preprocessing routines provide an authoritative open-data benchmark used to validate computational tools like fMRIPrep preprocessing pipelines.
- Test-Retest Validation: A dedicated 46-subject 3T retest dataset allows empirical quantification of scan-rescan reproducibility.
Limits
- Storage Demands: Complete multimodal packages span hundreds of gigabytes per subject, requiring substantial local high-performance storage.
- Healthy Adult Demographics: The flagship cohort excluded neurological and psychiatric conditions, limiting direct generalization to clinical populations without disease controls.
- Privacy Governance Constraints: Because datasets include twins and family structures, the data cannot be treated as completely anonymous; researchers must strictly avoid attempts at re-identification.
Alternatives
The table below summarizes key differences between the Human Connectome Project and other major neuroimaging initiatives.
| Initiative / Repository | Target Population and Sample Size | Primary Modalities | Data Access Model | Distribution Platform |
|---|---|---|---|---|
| Human Connectome Project (HCP) | Healthy young adults (n=1,206; ages 22–35); Lifespan cohorts (ages 5–90+) | 3T and 7T MRI (sMRI, dMRI, rfMRI, tfMRI), MEG | Two-tier: Open Access (registration) and Restricted (PI application) | ConnectomeDB / BALSA, AWS Open Data S3 |
| UK Biobank Imaging | Adults (target n=100,000; ages 40–69 at baseline) | 3T brain MRI, cardiac MRI, abdominal MRI, DXA | Restricted access via project application and fee | UK Biobank Research Analysis Platform |
| Adolescent Brain Cognitive Development (ABCD) | Adolescents (n=11,880; recruited at ages 9–10) | 3T MRI (structural, diffusion, functional), biospecimens | Controlled access through NIMH Data Archive (NDA) | NIMH Data Archive (NDA) |
| Alzheimer’s Disease Neuroimaging Initiative (ADNI) | Older adults: normal, MCI, and Alzheimer’s disease (n=2,000+) | 1.5T and 3T MRI, PET (amyloid, tau, FDG), fluid biomarkers | Application-based research access with formal DUA | LONI Image and Data Archive (IDA) |
| Human Brain Project / EBRAINS | Multi-species atlas data, computational simulations | Histology, high-field MRI, intracranial EEG | Open and controlled tiers under European data standards | EBRAINS Infrastructure |
Frequently Asked Questions
What is the difference between HCP Open Access Data and Restricted Data?
Open Access Data includes all raw and preprocessed imaging files and standard behavioral scores, accessible upon registering at ConnectomeDB and electronically signing the Open Access terms. Restricted Data covers sensitive variables—specifically exact age in single years, twin/sibling zygosity, and handedness—requiring an institutional PI application to prevent participant re-identification.
Do I need Institutional Review Board (IRB) approval to use Human Connectome Project data?
The HCP Data Use Terms require researchers to comply with home institutional policies. Because the dataset contains family pedigrees and twins, it is not classified as fully de-identified where combining variables could risk re-identification. Investigators must consult their institutional IRB or Ethics Committee to confirm exemption or approval prior to initiating secondary analyses.
How should Human Connectome Project data be cited in academic publications?
Publications must include the following acknowledgment: “Data were provided [in part] by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University.” Authors must also cite the primary overview paper (Van Essen et al., 2013, NeuroImage) and preprocessing papers (Glasser et al., 2013, NeuroImage). The consortium must not be listed as a manuscript co-author.
Can researchers redistribute modified or derived HCP datasets?
Researchers may redistribute original or derived data (such as parcellation schemes or statistical maps) only under the exact same Data Use Terms. Downstream recipients must register with ConnectomeDB and accept these terms. Commercial distribution or unauthorized monetization is prohibited.
What software is required to open and visualize HCP neuroimaging files?
HCP data use CIFTI-2 and GIFTI formats alongside standard NIfTI files. Connectome Workbench (wb_view for interactive visualization and wb_command for command-line scripting) is the primary native tool. Common neuroimaging software including FSL, FreeSurfer, AFNI, and Python libraries (Nibabel, Nilearn) also provide native support.
Sources
- Human Connectome Project. HCP Young Adult Data Use Terms. Washington University in St. Louis. URL: https://www.humanconnectome.org/study/hcp-young-adult/data-use-terms. (Checked: 2026-11-08).
- Human Connectome Project. 1200 Subjects Data Release: Reference Manual and Documentation. WU-Minn HCP Consortium. URL: https://www.humanconnectome.org/study/hcp-young-adult/document/1200-subjects-data-release. (Checked: 2026-11-08).
- ConnectomeDB. BALSA Neuroimaging Database and Repository. Washington University in St. Louis. URL: https://db.humanconnectome.org. (Checked: 2026-11-08).
- National Institutes of Health. NIH Blueprint for Neuroscience Research: Human Connectome Project. URL: https://neuroscienceblueprint.nih.gov/. (Checked: 2026-11-08).
- Van Essen, D. C., et al. (2013). The WU-Minn Human Connectome Project: An overview. NeuroImage, 80, 62–79. DOI: 10.1016/j.neuroimage.2013.05.041.
- Glasser, M. F., et al. (2013). The minimal preprocessing pipelines for the Human Connectome Project. NeuroImage, 80, 105–124. DOI: 10.1016/j.neuroimage.2013.04.127.
