Ten years ago, Dr. Jeff Lichtman, a molecular and cellular biology professor at Harvard, received a small brain sample in his lab. Despite its tiny size, the 1-cubic-millimeter tissue contained 57,000 cells, 230 millimeters of blood vessels, and 150 million synapses. Lichtman and his team amassed 1,400 terabytes of data from this sample, equivalent to over a billion books. Collaborating with Google scientists for a decade, they transformed this data into the most detailed human brain map to date.
A rendering shows all of the excitatory (pyramidal) neurons in a part of the brain sample, at varying degrees of magnification and tilt. They are colored by size; the cell body (central core) of the cells ranges from 15 to 30 micrometers across. Berger/Harvard University
The brain sample came from an epilepsy patient, and the team used diamond blades to slice it into ultra-thin sections for electron imaging. Google utilized AI-based processing to analyze the 300 million images, identifying cell types and connections. The result is an interactive 3D model of brain tissue, available online via "Neuroglancer." This breakthrough offers unprecedented insights into the human brain's structure and could illuminate various medical conditions.
The study, co-authored by Lichtman and Google’s Viren Jain, was published in Science. The collaboration underscores the potential of detailed brain mapping in advancing neuroscience. By providing a comprehensive view of the brain’s microstructure, this project may help researchers understand and address unresolved medical issues, offering valuable comparisons between normal and disordered brains.
