NIH Blueprint: The Human Connectome Project

Overview of the Human Connectome Project

The HCP is mapping the human connectome as accurately as possible in a large number of normal adults and is making this data freely available to the scientific community using a powerful, user-friendly informatics platform. 

The project is being carried out in two phases by a consortium of 36 investigators at 11 institutions. In Phase I (years 1–2, Fall 2010–Spring 2012), data acquisition and analysis methods were optimized for 16 major project components. In Phase II (years 3 – 5, Summer 2012–Summer 2015), data is being acquired from a cohort of 1,200 healthy adults, with primary data acquisition at three of these institutions (see Phase II: Logistics of Data Acquisition). The resultant datasets are being made publicly available at regular intervals, thereby enabling many explorations and analyses of brain circuitry even as data collection continues.

To obtain brain connectivity maps of the highest quality, HCP is using cutting-edge MR hardware, including new 3T and 7T MR scanners and customized head coils. MR data acquisition has been optimized through refinements to our pulse sequences and key pre-processing steps. 

Information about brain connectivity is being obtained using two powerful and complementary MR imaging modalities: diffusion imaging and resting-state fMRI. 

  • Diffusion imaging is used to chart the trajectories of fiber bundles coursing throughout the brain’s white matter. This is being done using HARDI (High Angular Resolution Diffusion Imaging) to acquire the data and probabilistic tractography to estimate fiber trajectories and generate maps of structural connectivity between gray matter regions.
  • Resting-state fMRI (R-fMRI) is providing comprehensive descriptions of functional connectivity between different gray matter regions, based on correlations in the fMRI BOLD signal among functionally interacting brain regions.

Additional information about brain function is being obtained using Task-fMRI, in which subjects carry out a variety of behavioral tasks in the MR scanner. In addition, some subjects are being studied using magnetoencephalography (MEG) combined with electroencephalography (EEG), yielding information about brain function on a millisecond time scale. Behavioral testing using a battery of tests to assess sensory, motor, and cognitive function is enabling assessment of brain circuits associated with particular behavioral features or traits.

The subjects being studied in Phase II include twins and their non-twin siblings (see Recruitment), thereby enabling exploration of the heritability of various brain circuits. Genotyping of all subjects will enable genome-wide association studies (GWAS) to evaluate genetic influences on brain circuitry.

The massive amounts of experimental data of many different types obtained by the HCP will be integrated (see: multi-modal integration) using a variety of analysis and visualization tools. Sophisticated network modeling tools will be made freely available, enabling analyses of key characteristics of brain connectivity in individual subjects and in populations distinguished by behavioral or other attributes. 

We have a robust and reliable informatics platform to provide a stable, well-structured database for storing vast amounts of HCP data. A powerful supercomputer with dedicated time for the HCP is enabling complex analyses of connectivity data and other analyses to be carried out efficiently and incorporated into this database. A user-friendly platform for data mining, analysis, and visualization is being built to enable investigators around the world to capitalize on these enormously rich datasets. Extensive outreach efforts are being made to inform the scientific community about the availability and use of these datasets. We will also provide a variety of training opportunities to promote utilization of the data and the associated tools.

The Human Connectome Project is providing a treasure trove of information at an unprecedented level of detail. Evaluation of this data must respect the profound anatomical and functional complexity of the human brain (see A Neurobiologically Grounded Connectome).