The largest 3D reconstruction of a fragment of the human brain has been created

Кубический миллиметр мозговой ткани может показаться незначительным. Но если учесть, что этот маленький квадрат содержит 57 000 клеток, 230 миллиметров кровеносных сосудов и 150 миллионов синапсов, что в сумме составляет 1 400 терабайт данных, то исследователи из Гарварда и Google добились огромного прогресса.

Команда Гарварда под руководством Джеффа Лихтмана, профессора молекулярной и клеточной биологии Джереми Р. Ноулза и недавно назначенного декана научного факультета, совместно с исследователями Google создали самую крупную на сегодняшний день трехмерную реконструкцию мозга человека на синаптическом уровне, показывающую в ярких деталях каждую клетку и ее сеть нейронных связей в части человеческого височного кортекса, размером примерно с половину зернышка риса.

The achievement, published in Science, is the latest in a nearly decade-long collaboration with Google Research scientists who combine Lichtmann electron microscopy with AI algorithms for color coding and reconstruction the extremely complex neural wiring of mammals. The paper's three co-first authors were former Harvard postdoc Alexander Shapson-Ko, Google Research's Michal Januszewski, and Harvard postdoc Daniel Berger.

The ultimate goal of the collaboration, supported by the National Institutes of Health's BRAIN Initiative, is to create a high-resolution map of the neural connectivity of the entire mouse brain, which will require about 1,000 times more data than they just obtained from one cubic millimeter of human cortex.

The word "fragment" sounds ironic. A terabyte is a huge amount to most people, but a fragment of a human brain - just a tiny, tiny piece of a human brain - is still thousands of terabytes."

Jeff Lichtman, Jeremy R. Knowles Professor of Molecular and Cellular Biology

The latest map in Science reveals details of brain structure that have not previously been observed, including a sparse but powerful network of axons connected by up to 50 synapses. The team also noted some features in the tissue, such as a small number of axons forming extensive spirals. Because their sample was taken from a patient with epilepsy, they are unsure whether such unusual formations are pathological or simply rare.

Lichtman's area of research is "connectomics," which, similar to genomics, seeks to create complete catalogs of the brain's structure down to individual cells and connections. Such completed maps will light the way to new understandings of brain function and diseases about which scientists still know very little.

Modern Google artificial intelligence algorithms make it possible to reconstruct and map brain tissue in three dimensions. The team has also developed a set of publicly available tools that researchers can use to explore and annotate the connectome.

“Given the huge investment that went into this project, it was important to present the results in a way that others can now benefit from,” said Google Research Fellow Viren Jain.

The team will next target a region of the mouse hippocampus, which is important to neuroscience for its role in memory and neurological disease.