Deciphering the secrets of the brain

Unlocking the secrets of the brain, especially its architecture and wiring, is one of the big challenges in modern life sciences. That is why the National Institutes of Health (NIH) in the USA, one of the world’s largest research agencies, has included this in its programme. As part of the NIH BRAIN Initiative, a Swiss researcher has now been awarded a major grant of up to 2.6 million US dollars. The neurobiologist Adrian Wanner, a group leader at the Paul Scherrer Institute PSI, is the project’s principal investigator. Andreas Schaefer from the Francis Crick Institute in London is also closely involved.

The NIH’s decision to invest such a large sum in a project at a Swiss institute demonstrates the exceptional competitiveness of Swiss researchers and confirms PSI’s position as a centre for world-class research. “For a young research group leader to receive such a large grant, especially from another country, is by no means commonplace; it testifies to his great scientific talent and the confidence that the international community has in Switzerland as a research location,” says Gebhard Schertler, Head of the Department of Biology and Chemistry, who is delighted with the good news from the United States. Schaefer adds, “This funding will further strengthen the existing collaboration between our groups and institutes.”

Exploring one of nature’s most complex structures

The brain is one of the most complex structures found in nature. The human brain, for example, contains about 100 billion brain cells and many times that number of connections, called synapses. Scientists believe that in order to explain how the brain processes complex information from the environment and transforms it into thoughts, decisions and actions, it is necessary to know the patterns in which the brain cells are wired. They refer to this wiring diagram of the brain as the connectome. It can be used to extract detailed information showing how different types of cells in the brain are connected; this is fundamental particularly to understanding brain diseases such as Alzheimer’s.

The path to this goal is arduous, as demonstrated by the process of decoding the connectome of the nematode Caenorhabditis elegans which was completed in 1986 after years of research. Even though the roundworm’s brain contains just 302 nerve cells and 5000 synapses, mapping it is nevertheless considered a milestone in brain research. To map the connectome of a brain or brain sample, researchers cut the tissue into tiny slices. These ultra-thin slices, 30 to 40 nanometres thick, are then stained and analysed in a high-resolution electron microscope. The resulting images can be used to reconstruct the nerve cells and their connections, the synapses, and to identify the different types of cells. In the case of the nematode, it was possible to reconstruct the nerve cells by hand. However, this is no longer feasible for larger brains. Reconstructing the connectome of a fly’s brain, for example, would take some 2000 person-years. In other words, 50 people doing nothing else would have to spend their entire 40-year working lives to accomplish the task.

Using artificial intelligence to decipher the brain

In recent years, however, researchers have succeeded in automating the reconstruction process. Advances in imaging and the use of artificial intelligence (AI) have reduced the human input needed by a factor of 50. “It’s almost crazy how well AI works,” Wanner says happily. His colleagues at Princeton University, for example, recently succeeded in reconstructing the connectome of a fly, almost entirely automatically – it took “only” some 50 person-years to perform the manual corrections by humans. If this process can be automated further, even the connectome of small mammals such as mice will be within reach. This is the reasoning behind the NIH’s BRAIN Initiative strategy and funding.

For now, there is still a catch to the procedure outlined. Ultra-thin slices in the range of 30-40 nanometres are difficult to handle. Mistakes can be made when positioning the slide, individual slices can be lost, they can break or crinkle, or the knife can leave notches. More than 50 percent of the errors in the AI’s analysis algorithms are attributable to flaws in the slicing. This is why, for the moment, the AI’s results still have to be checked by hand. On average, this takes about one working week for a single nerve cell in a mouse’s brain – far too long.

A microchip fabrication process can help

The PSI and Crick team are taking a different approach, one that allows more robust imaging and could therefore be crucial in studying the mouse’s connectome. The initial slices are 250 to 500 nanometres thick, because thicker sections are much easier to handle. A multi-beam scanning transmission electron microscope is then used to record images of these sections. Next, a broadband ion beam removes a layer of the sample just a few nanometres thick, and the entire process is repeated. “We copied this procedure from microchip manufacturers,” explains Wanner. After 25 to 50 polishing steps, the entire section has been analysed and the computer can generate a high-resolution 3D image from the differences between the individual images.

Fewer slicing errors and the higher resolution provided by broadband ion beam polishing means that the AI algorithms can draw on better information when reconstructing the nerve cells. The sections are scanned not just by a single electron beam, but by 64 parallel beams, which also makes imaging very fast and reliable. Preliminary trials included in the research application to the NIH have already demonstrated that the new method works. The next three years will show whether this can lead to significant savings and speed up research into the mouse connectome. During this time, the PSI-Crick project will be funded by the NIH.

Text: Werner Siefer

© PSI provides image and/or video material free of charge for media coverage of the content of the above text. Use of this material for other purposes is not permitted. This also  includes the transfer of the image and video material into databases as well as sale by third parties.

About PSI

The Paul Scherrer Institute PSI develops, builds and operates large, complex research facilities and makes them available to the national and international research community. The institute's own key research priorities are in the fields of future technologies, energy and climate, health innovation and fundamentals of nature. PSI is committed to the training of future generations. Therefore about one quarter of our staff are post-docs, post-graduates or apprentices. Altogether PSI employs 2200 people, thus being the largest research institute in Switzerland. The annual budget amounts to approximately CHF 420 million. PSI is part of the ETH Domain, with the other members being the two Swiss Federal Institutes of Technology, ETH Zurich and EPFL Lausanne, as well as Eawag (Swiss Federal Institute of Aquatic Science and Technology), Empa (Swiss Federal Laboratories for Materials Science and Technology) and WSL (Swiss Federal Institute for Forest, Snow and Landscape Research). (Last updated in May 2023)