New 3D Brain Atlas Takes Neuroimaging to Next Level

Talan, Jamie

doi: 10.1097/01.NT.0000432679.14632.84
Back to Top | Article Outline


As part of an international initiative, neuroscientists and computer mapping experts have created a 3D brain atlas that provides resolution that is 50 times greater than existing brain maps.

A team of scientists sliced hair-thin sections of a perfectly preserved human brain and like masterful enigmatologists assembled them into the most powerful 3-D digital atlas — 7,400 slices lining up with all the intricate folds and sulci to reveal fine cellular detail that has never been seen before.

Neuroimaging experts who were not involved with the new atlas said it was premature to talk about the concrete application of the new 3D brain map, but they speculated that it might be useful for the study of Alzheimer's disease and other neurological disorders to understand the cells and regions involved, and to track changes over time. And, in a more practical context, neurosurgeons may be able to use the fine maps when determining where to implant electrodes.

The first anatomical explorers of the human brain used their eyes and imagination and what they knew about populations of neurons and other brain cells to develop the first two-dimensional maps. That was in the early 1900s and these atlases have been debated and refined ever since. Now, this new atlas — dubbed BigBrain and created by researchers in Germany and Canada — provides resolution that is 50 times greater than the available atlases, which means that scientists can look at small cellular circuitry, even single cortical layers and sub-layers, of cerebral cortex, to help unravel the brain's true inner-workings.

Part of the European Human Brain Project — an international initiative comprising experts in neuroscience, medicine, and computing — the new reference brain is available at no cost to scientists worldwide. The details of their effort, which took a team a thousand hours to slice, stain, reassemble and digitize with a high-resolution flatbed scanner, is published in the June 21st issue of Science.

The scientists say that it will change the field of neuroanatomy.

“This is a powerful tool to facilitate neuroscience research,” said lead author Katrin Amunts, PhD, director of the Cecile and Oskar Vogt Institute for Brain Research at the Heinrich Heine University Dussledorf in Germany. “The famous cytoarchitectural atlases of the early 1900s were simplified drawings of a brain and were based on pure visual analysis of cellular organization patterns,” she added.

“These atlases do not give you the ability to integrate information at the level of cortical layers, columns, microcircuits or cells,” the scientists wrote in their study. Until now.

The challenge was this: the cerebral cortex has a lot of folds and a large number of specialized regions, considerable variability across individuals, and it's big — with billions of brain cells. But modern computing power was now on their side.

The investigators used a tool called a microtome to slice the paraffin-covered brain of a 65-year old woman into 20-micrometer thick sections. Each section was then mounted on slides and stained to identify cell structures. The massive task was then digitized with the scanner to reconstruct it into a high-resolution 3D model. The sections were registered to MRI data for an undistorted frame of reference, they said.

Co-investigator Alan Evans, PhD, a professor at the Montreal Neurological Institute at McGill University, said that the hair-thin slicing of fragile human tissue led to rips and tears that had to be repaired in the digital version.

Dr. Amunts added that neuroscientists can now use the atlas to layer data onto it to better understand how humans think and feel and behavior. It can also be used to study brain changes over the course of neurological diseases and the effects of normal aging. Maps can be linked to data from fMRI, MRI, SPECT, PET, and other brain scanning technologies. “We can now gain a new understanding of the normal structure of different functional areas of the brain, such as the motor cortex or a region that is important for learning and memory, and we can also measure numerous structural properties,” said Dr. Amunts.

The investigators are now looking to other post-mortem brains to expand their effort and factor in structural variability. This atlas reflects one human brain — an older woman with no history of neurological problems. They are also layering on maps of white matter to study the anatomical relationship between cortical microanatomy and the fibers that connect up the cerebral cortex.

“It is a common basis for scientific discussions because everybody can work with this brain model,” added Karl Zilles, PhD, a co-investigator in the project and a senior professor of the Jülich Aachen Research Alliance. Dr. Zilles said that they are also integrating receptor data into the atlas. This will allow neuroscientists to study “the relationship between cortical microanatomy and the key molecules of neurotransmission.”

The German and Canadian researchers have also recently met with scientists at the Allen Institute, headquartered in Seattle, to talk about integrating the institute's gene expression atlas into BigBrain. It could someday also be used as an atlas for neurosurgeons looking for stereotaxic and topological positioning in the brain.

“This is a reference brain,” said Dr. Amunts. “It can be used as a computational model to examine the organization of the human brain. The whole point is that it can simulate how the normal brain develops and also look at the brain undergoing neurodegeneration.”

“It changes the game,” added Dr. Zilles. “We can now distinguish between even smaller areas of the brain.”

Back to Top | Article Outline


“This team has done a heroic task of serially sectioning a human brain,” said Ed Lein, PhD, an investigator at the Allen Institute. He and his colleagues mapped the mouse brain atlas and they are now working on developing anatomical and gene expression atlases of human and non-human primate brains. “It reconstructs the volume of the human brain and scientists will be able to map MRI coordinates and link it to any type of data. It is a great resource.”

Allen Institute scientists said that they will be taking their gene expression data and mapping it on the atlas so that “anyone interested in the function of a given gene can see where it is distributed in the brain,” said Dr. Lein. Pharmaceutical companies also use atlases to understand cells and receptors and brain regions targeted by their compounds.

Michael Hawrylycz, PhD, another Allen Institute scientist, said that the new atlas provides a clearer look at boundaries from one brain region to the next. He said that the challenge will be to annotate the atlas — to assign neuroanatomic labels to the regions of the atlas that have meaning. Although this can be challenging in 3D it can be done by careful delineations of 2D sections. “For now, the first step is to look at this data and annotate it to make sense of the boundaries.”

“This is a tour de force,” added Arthur Toga, PhD, director of the laboratory of neuroimaging at UCLA. “It will bridge the gap between cellular imaging and the whole brain. The atlas will allow us to refine some of these boundaries.” He likens it to flying high in a plane and being able to make out the shapes of cars and tracks of houses. “You see the big picture but none of the fine detail. This brings us close to the ground so you can see things we have not been able to make out,” he explained.

That said, he added that a singular brain does not “allow us to appreciate the level of variability in the human brain.” He said that BigBrain will pose new challenges. How do scientists go across spatial scale and resolution? How do they handle data volume from this type of resource? How do you analyze and distribute the data? And how can one comprehensively take an inherently disruptive sampling approach without introducing some experimental error?

Those questions aside, he will be trying to incorporate the data into his research.

Back to Top | Article Outline


•. Access to the free atlas:
•. Amunts K, Lepage C, Borgeat L, et al. BigBrain: An ultrahigh-resolution 3D human brain model. Science 2013; 340:1472-1475.
•. Neurology Today archive on brain atlases:
© 2013 American Academy of Neurology