“High Throughput Connectomics: The Building of a Brainscope”
By Professor Nir Shavit, MIT (Massachusetts Institute of Technology, USA
Nir Shavit is a professor of computer science at MIT. He received B.Sc. and M.Sc. degrees in Computer Science from the Technion – Israel Institute of Technology in 1984 and 1986, and a Ph.D. in Computer Science from the Hebrew University of Jerusalem in 1990. Shavit is a co-author of the book The Art of Multiprocessor Programming, is a winner of the 2004 Gödel Prize in theoretical computer science for his work on applying tools from algebraic topology to model shared memory computability, and a winner of the 2012 Dijkstra Prize for the introduction and first implementation of software transactional memory. He has recently become interested in computational neurobiology, and in particular is involved in developing new ways of using high performance computing to analyze data in order to uncover the microscopic structure and function of neural tissue.
Connectomics is an emerging area of neurobiology that uses cutting edge machine learning and image processing techniques to extract brain connectivity graphs from electron microscopy images. This talk will describe how a team of researchers from MIT and Harvard plan to extract the complete connectivity graph of a cubic millimeter of brain tissue. Though it is the size of a grain of salt, the dataset is 2 Petabytes in size, and will contain about a hundred thousand neuron bodies and a billion synapses.
It has long been assumed that the processing of such large connectomics datasets will require mass storage and farms of CPUs and GPUs and will take years. I will discuss the feasibility of designing a high-throughput connectomics-on-demand system that runs on a multicore machine with less than 100 cores and extracts connectomes at the terabyte per hour pace of modern electron microscopes.