Harvard and Google Map 1 Cubic Millimeter of Human Brain, Revealing Vast Data Potential
Harvard and Google spent 10 years mapping a 1 cubic millimeter section of the human brain, resulting in 1.4 petabytes of data, equivalent to 1.4 million gigabytes. This project involved slicing the tissue into 5,000 wafers and using a $6 million electron microscope, along with Google's machine learning models for 3D reconstruction. The full human brain is estimated to yield 1.4 zettabytes of data, requiring a 140-acre data center at a cost of $50 billion. The next phase will involve mapping a mouse hippocampus over the next five years.

Harvard and Google have mapped a 1 cubic millimeter section of the human brain, producing 1.4 petabytes of data from 57,000 cells and 150 million synapses. The mapping involved slicing brain tissue into 5,000 wafers, using a high-cost electron microscope, and employing Google's machine learning for data reconstruction.
A full human brain would generate roughly 1.4 zettabytes of data, necessitating a 140-acre data center costing an estimated $50 billion. The next research phase will focus on mapping a mouse hippocampus over the next five years, which is necessary before attempting to map the entire human brain.




Comments