BigBrain poised to make a big impact

Discovery
by Tim Hornyak, BA’95
Montreal Neurological Institute Alan Evans was one of the driving forces behind the creation of BigBrain (Photo: Christinne Muschi)

The Montreal Neurological Institute’s Alan Evans was one of the driving forces behind the creation of BigBrain, the first high-resolution 3D digital model of a human brain  (Photo: Christinne Muschi)

For a woman whose brain spent seven years under the proverbial microscope, she’s rather mysterious. The 65-year-old is anonymous, her background is a secret, but her brain is now intimately known to science. It’s the focus of an international endeavour called BigBrain that is helping redefine the limits of neurology.

Part of the $1.3 billion Human Brain Project that is aimed at producing a supercomputer simulating the brain, BigBrain is the first brain atlas offering microscopic detail. It allows researchers to view brain structures in 3D at a resolution of a mere 20 micrometers, which is far thinner than a strand of hair. That’s 50 times the resolution of previous digital brain maps of 1 millimeter resolution, and allows the viewing of the largest neurons and their connections.

This is a significant effort on the road to fully understanding the workings of the 86 billion neuron human brain. The Montreal Neurological Institute (the Neuro) is a key player in the project, which involved slicing the donor’s brain into thousands of cellophane-thin sections.

“Our group did all of the computational analysis involved in spatial registration of the 7,404 sections, which was a messy challenge,” says Alan Evans, a professor of neurology and neurosurgery at the Neuro and a co-creator of BigBrain. “We also developed the software for image processing and 3D visualization of the data.”

The Neuro team worked with Katrin Amunts and others from Germany’s Forschungszentrum Jülich, one of Europe’s major research centres, and published a paper about the BigBrain project in a recent edition of Science. The paper is the fruit of a decade of work on the donor brain and testament to an intensive production process. The brain was fixed in formalin and preserved in wax before an instrument called a microtome (Evans describes it as “a giant deli slicer”) was used to divide it into extremely thin sections. They were then stained for cellular structures and then digitally scanned. The preparation and scanning of each brain slice required about 1,000 hours of labour.

But handling these shavings inevitably produced wrinkles and tears in the tissue that showed up in the scans. Evans and his team were responsible for detecting these errors when they cropped up, using corrective software to ensure that the scans added up to a coherent whole. The data involved was so large it required a computing grid that was distributed across Canada.

“It’s about 100,000 processors across the country – this is the entire supercomputing grid of Compute Canada,” says Evans.

The finished data set, however, has a volume of 1 terabyte (1,000 gigabytes) and is already available to researchers for free through the BigBrain website; over 10,000 accounts to access it were opened in the first week. They are able to easily download the sections and use them when analyzing MRI and PET scans of living patients’ brains. The data is also expected to help scientists map features like neural activity more precisely and aid in the treatment of brain illnesses like epilepsy.

Evans says professional reaction to the project varies. “There’s a tremendous response from the people who do systems neuroscience, who are interested in how different parts of the brain communicate with each other,” he says. “At the same time, some of the researchers who work at more of the basic molecular level said it’s just a technical achievement and it’s not science. I think the truth lies somewhere in between.” In an interview with BBC, Amunts compared BigBrain to “using Google Earth. You can see details that are not visible before we had this 3D reconstruction.”

While the BigBrain data form a new frontier in brain mapping, they do not account for variability in individual brains. The donor had no known neurological diseases, but how would her brain compare to that of a man or a child? But those may be in the offing: Amunts wants to create BigBrain atlases for such brain types, and even those with developmental disorders such as autism, to show developmental differences.

“We’re basically using BigBrain as a gold standard for analysis of individual brains done with MRI,” says Evans. “The technology for acquisition and for computational analysis is now in place. So everything that we’re done we can do a lot faster.”

The nameless woman whose brain took up space on computers all across the country may have started a revolutionary trend.

 

Share this article:

Comments are closed.