The new effort, called the SEU-Allen Joint Center for Neuron Morphology or the SEU-Allen Center, is using virtual reality headsets that allow researchers to manually trace the cells’ shapes, or morphologies, in 3D as well as computational analyses to help automate the reconstruction of those morphologies.

“Capturing complete neuron morphologies is very complicated. It takes people a long time to reconstruct each cell,” said Hanchuan Peng, Ph.D., Associate Investigator at the Allen Institute for Brain Science and Director of the SEU-Allen Center. “We want to get to a point where our reconstruction is fully automated and then double-checked by a person.”

Picking out one neuron’s complete shape, with all its delicate and microscopically thin axon fibers entangled with thousands of other neurons in the brain, is a laborious and painstaking process. But working with the entire brain is important in understanding the neurons’ true shapes, Peng said, as the cells are so intertangled and their axons so long that any cut in the brain may sever important pieces of the exact neuron a researcher is trying to reconstruct.

This project is an essential part of understanding how all the cells in the brain are different from each other — and ultimately, how they work to carry out all the brain’s different functions — said Hongkui Zeng, Ph.D., Executive Director of Structured Science at the Allen Institute for Brain Science and one of the leaders of the SEU-Allen Center. The collaboration fits in with the Allen Institute’s efforts to create a comprehensive “periodic table” of all the different types of cells in the mouse and human brains, known as the Allen Cell Types Database.

Complete neuron shapes captured using VR technology developed by the SEU-Allen Center.

“Understanding the cells’ complete shapes is an important part of characterizing brain cells,” Zeng said. “And we need to carry out this work in whole brains because many neurons talk to other neurons in really distant brain areas, meaning they send long axon connections throughout the entire brain. It’s very important to understand the wiring diagram of the brain, that is, who is talking to whom in the brain — this is an essential problem for the field of neuroscience.”

The SEU-Allen team is not the only research group attempting to reconstruct neurons in their entirety. But there are a few aspects to their approach that are unique, Peng said. They’re using VR headsets to help with the manual reconstruction — this lets the researchers see the neuron in 3D rather than flat on a computer screen. And they’re working in teams to pair software code development with manual reconstructions in an iterative process that allows real-time improvement of their technique. The SEU-Allen Center’s workspace at Southeast University includes a large screen visible from all the researchers’ desks that has current data on the team’s progress, updated every 30 minutes.

Right now, the Center comprises about a dozen staff, with more researchers who rotate through temporarily to help, Peng said. They’re producing between 50 to 100 new complete neuron reconstructions a month, with the goal of speeding up their work in the near future by further automating the process. The researchers are also hoping other collaborators may be interested in helping the effort. Because the data is openly available, anyone who has a VR headset can participate in the reconstruction remotely.

“This collaboration is a great match for those of us interested in producing high quality, open-access data for neuroscience,” said Gang Wu, Ph.D., Vice President for Research at Southeast University. “We believe such data and novel technologies produced in this collaboration will not only help the entire field understand better the complexity and beauty of neurodata, but will also attract more scholars into the brain science field.”

Get the latest news from the Allen Institute.