NIH Blueprint: The Human Connectome Project

News and Updates

Press Releases,Project News,Recommended Reading,The Science of Connectome | July 20, 2016

Nature article: Cortical brain maps at the highest resolution to date

Screen Shot 2016-07-20 at 11.21.20 AM

Matthew Glasser, Ph.D. of the Van Essen lab at Washington University in St. Louis.

Major new work based on Human Connectome Project data and methods published in July 20 issue of Nature promises to be a boon to neuroanalysis research for years to come.

The study, A multi-modal parcellation of human cerebral cortex, led by Matthew Glasser and David Van Essen of Washington University, used information derived from structural and functional MRI data collected on 210 HCP subjects to create a new 180 region per hemisphere map of the cerebral cortex of the human brain. Further improvements to previous maps were achieved by using the multimodal surface matching algorithm pioneered by HCP investigators at Oxford U to precisely align the individual brains before analysis. Results were validated and the maps applied to individuals from an independent set of 210 HCP subjects.

Although the new map will be very important in raising the accuracy of work to delineate the connections between brain regions by the HCP and others, it will continue to improve as more, higher resolution data is added to the analyses.

As Glasser told the Washington University Record:

“We ended up with 180 areas in each hemisphere, but we don’t expect that to be the final number,” Glasser said. “In some cases, we identified a patch of cortex that probably could be subdivided, but we couldn’t confidently draw borders with our current data and techniques. In the future, researchers with better methods will subdivide that area. We focused on borders we are confident will stand the test of time.”

and added in Nature:

“We’re thinking of this as version 1.0,” says Glasser. “That doesn’t mean it’s the final version, but it’s a far better map than the ones we’ve had before.”

The parcellation and Connectome Workbench scenes for each of the main article and supplemental figures are being shared in the new Brain Analysis Library of Spatial maps and Atlases (BALSA) database being developed by the Van Essen lab at Washington University.

The parcellation for use as a reference is most easily accessed in the Glasser_et_al_2016_HCP_MMP1.0_5_StudyDataset.scene study dataset.

Nature produced a video highlighting the work in the context of previous brain mapping efforts:

In addition to the article itself, Nature is distributing a wealth of supplemental information, as David Van Essen told the Washington University Record:

“We were able to persuade Nature to put online almost 200 extra pages of detailed information on each of the 180 regions as well as all of the algorithms we used to align the brains and create the map,” Van Essen said. “We think it will serve the scientific community best if they can dive down and get these maps onto their computer screens and explore as they see fit.”

The seminal work has also garnered much press:

B.T. Thomas Yeo & Simon B. Eickhoff authored a Nature News and Views article.

Nature News: Human Brain Mapped in Unprecedented Detail

Washington University Record: Map provides detailed picture of how the brain is organized

NIH News: Connectome map more than doubles human cortex’s known regions

New York Times: Updated Brain Map Identifies Nearly 100 New Regions

Wall Street Journal: Brain Mappers Create a Detailed Atlas of the Human Cortex

The Scientist: Mapping the Human Connectome

Popular Science: This New, Ultra-Detailed Map Of The Brain Could Change Medicine

Wired: A New Map of the Brain Redraws the Boundaries of Neuroscience

 

Press Releases,Project News | September 2, 2014

Beta Release of Group-ICA maps, Node timeseries, and Network matrices

The Human Connectome Project (HCP) WU-Minn consortium is pleased to announce a beta-release of Resting-state fMRI (rfMRI) analyses on the HCP 500 Subjects data released in June 2014. This rfMRI data was processed, using data from all 468 subjects having four rfMRI runs, yielding the following outputs:Netmat_Fig

  • A group-average “parcellation”, obtained by means of group-ICA (independent component analysis).
  • Subject-specific sets of “node timeseries” – for each subject, a representative time series per ICA component (“parcel” node).
  • A subject-specific “parcellated connectome”or “netmat” – for each subject, a nodes x nodes matrix – the functional connectivity between node timeseries.

As a shorthand, we are calling this rfMRI analysis dataset the HCP 500 Subjects Parcellation, Timeseries, Netmats Release, or “HCP 500 Subjects PTN Release”, for short.

These rfMRI analyses are being “beta” released because the analysis methods used are still works-in-progress and may be improved in the future. We encourage the community to look at the released data and discuss the methods and results.

What’s in the HCP 500 Subjects PTN Release? The data is available as a single package release containing:

  • Group-ICA “parcellations” at five different dimensionalities (levels of detail), including spatial-ICA maps for each distinct ICA decomposition dimensionality and volumetric MNI152 3D-space projected versions of the spatial-ICA maps, primarily for display purposes.
  • Node-timeseries (individual subjects), estimated by 2 distinct timeseries estimation approaches for each group-ICA dimensionality.
  • Netmats (parcellated connectomes). Full and partial correlation netmats for each subject and at the group-level (all subjects) for each choice of group-ICA dimensionality and node timeseries estimation approach.

A summary document for the HCP 500 Subjects PTN Release is included in the released package. It contains details about methods used in these analyses, descriptions of released files, example MATLAB/FSLNets code, and references.

Access the HCP 500 Subjects PTN Release data on the HCP website. Download the dataset on the WU-Minn HCP 500 Subjects project page on ConnectomeDB under the HCP500 Parcellation + Timeseries + Netmats (PTN) heading.

Have more questions?  We encourage you send your questions to hcp-users@humanconnectome.org and join the hcp-users discussion group (http://www.humanconnectome.org/contact/#subscribe), so that you can tune in to technical discussions on issues that may be of interest.

The Science of Connectome | August 1, 2010

Visualizing the Networked Brain: A Parcellation Scheme for Human Left Lateral Parietal Cortex

This network image comes from the latest thesis work by Steve Nelson, working under the auspices of Steve Petersen, PhD, the James S McDonnell Professor of Cognitive Neuroscience at Washington University in Saint Louis. Nelson’s research is using fcMRI and fMRI to parcell the parietal cortex, yielding a fascinating diagram of function and structure.

The left half of the figure is a graphical layout of the functional correlations between regions. The different colors represent different subnetworks defined by modularity optimization. The right half plots the different subnetworks on the brain, showing how each subnetwork is distributed across the cortex.

Parcellation Scheme for human left lateral parietal cortex

Parcellation Scheme for human left lateral parietal cortex

Here is the abstract from the paper, which was published in Neuron, Volume 67 Issue 1 on July 15th, 2010.

The parietal lobe has long been viewed as a collection of architectonic and functional subdivisions. Though much parietal research has focused on mechanisms of visuospatial attention and control-related processes, more recent functional neuroimaging studies of memory retrieval have reported greater activity in left lateral parietal cortex (LLPC) when items are correctly identified as previously studied (“old”) versus unstudied (“new”). These studies have suggested functional divisions within LLPC that may provide distinct contributions toward recognition memory judgments. Here, we define regions within LLPC by developing a parcellation scheme that integrates data from resting-state functional connectivity MRI and functional MRI. This combined approach results in a 6-fold parcellation of LLPC based on the presence (or absence) of memory-retrieval-related activity, dissociations in the profile of task-evoked time courses, and membership in large-scale brain networks. This parcellation should serve as a roadmap for future investigations aimed at understanding LLPC function.

Neuron. 2010 Jul 15;67(1):156-70.