This Digital Sliver of Rat Brain Took Researchers 20 Years to Reconstruct

This hypnotic mass of wiring isn't interactive art. It's a digitally reconstructed sliver of a rat's brain, and it's taken researchers 20 years to craft.

In a study published Thursday in the journal Cell, researchers from the Blue Brain Project present a first draft of their digitally reconstructed slice of a young rat neocortical microcircuit—using over 30 software tools. It contains a whopping 31,000 neurons, 55 layers of cells, and 207 different neuron subtypes, and is a visual culmination of the group's efforts over the years.

The aim of the Blue Brain Project, which started in 2005, is to build detailed digital reconstructions and simulations of a rat's brain (and eventually a human's brain). In the process, the researchers want to better understand the data measured in neuroscience, certain diseases that affect the brain, and to apply their knowledge to building new types of computing devices.

"It's a tremendous relief that we got to this point," Sean Hill, co-director of the Blue Brain Project at the Ecole Polytechnique de Lausanne, told me. "We've learned so many principles about how this core microcircuit is built, so in the future we can build additional ones more quickly. […] The brain circuitry is also performing an enormous amount of computation with very low power. There's clearly a lot to learn there that would apply to computing technology."

Video: YouTube

Neuroscientists the world over have been working for decades to understand the brain. Thus far, they've mapped out its circuitry, investigated different neurons, and measured how they fire. But digitally reconstructing the entire brain still remains a holy grail as the organ is so complex.

"The brain is extremely complicated. We're very far from being able to systematically and completely measure all of the content and all of the properties in the brain just because of the scale," said Hill. It's an enormous amount of data to just go in there and measure it all using electron microscopy [a microscope using a beam of electron as a source of light]."

Hill joined the Brain Project in 2006. At the initial stages, the researchers dovetailed off of the work from director of both the Blue Brain and the Human Brain Project, Henry Markram's lab. Markram had measured all of the characteristics of a single circuit within the brain (a rat's cortical microcircuit), but wasn't satisfied with keeping that knowledge within his lab's four walls.

"In 2005, Henry had a vision that we should somehow be able to take that data and put it together in order to better understand how all these measurements help us to understand this brain circuit," said Hill. "That's when the Blue Brain Project was launched."

Over the years, the researchers have been using a supercomputer from IBM to build a reconstruction of the neocortical microcircuit based on the first set of brain data that had been measured. They then ran simulations of the electrical activity so they they could understand how everything interacted.

"We wanted to know how do these neurons interact, how do they communicate through the synapses, and how does that produce this emergent network behavior," said Hill.

The current study and visualizations could eventually allow researchers to predict the location, number, and amount of calcium ions, which influence different brain states, that flow through the 40 million synapses within the digital piece of brain. It also allows them to simulate how neurons behave under different conditions, or as Markram said in a press statement, "a spectrum of states."

"This is a very powerful tool to preserve, but to also manipulate the data and to try and understand the emergent phenomena of brain states and diseases," Idan Segev, co-author of the study and professor of computational neuroscience at the Hebrew University of Jerusalem, told me. "It's a historic moment, and a very powerful kind of new approach for brain research."

The results of the study have been released to both the public and the scientific community. The idea is for researchers around the world to make use of the data collected so far, and to encourage them to collaborate by contributing their own information, and by asking questions, so that the digital reconstruction of the rodent's brain can continue to evolve.

The visualization took so long to reconstruct as the project required multiple experiments, with researchers iteratively building the digital brain slice through both their own data, and data from around the world.

"The point is this had never been done before. There had never really been an effort to say let's take all of these facts that we have measured about this particular brain, and integrate it," said Hill. "It's a very small brain circuit, but it really represents the prototypical circuit that occurs through the rat's cortex. […] We have a huge amount of work to do to get to the human brain, but it represents an initial prototype to understanding this core mammalian circuit."

Next up, the researchers will be scaling up their work on the rodent brain, and aiming to build larger brain areas. While the ramifications of the group's research could touch research fields beyond neuroscience, Hill asserted their aim was primarily to advance neuroscience, as opposed to help progress the development of artificial intelligence.

"There's a whole movement now in deep learning, and there is potential for the knowledge that we're gaining about the real biological system to be brought to bear on those efforts. But our focus is on creating a tool to accelerate neuroscience," said Hill. "But that can certainly have impacts on other things."