The Blue Brain Project is based at the École Polytechnique Fédérale de Lausanne (EPFL), a research university in Geneva, Switzerland. It is an attempt:
... to build biologically detailed digital reconstructions and simulations of the rodent, and ultimately the human brain.
(3)
The project is highly computational and uses what is currently (2019) the world's 6th fastest supercomputer, Piz Daint, in Lugano, Switzerland. (4) I started an interest in neuroscience by hanging out with scientists from Arizona State University and the University of California San Diego while working for Science Approach here in Tucson. I also have an interest in Computational Biology taking two graduate courses at the University of Arizona. Last year I took a MOOC from EPFL on Edx. (5) It was very difficult but I did learn something about the project.
At a basic level, digital reconstruction is a Big Data problem that gathers data from 100's of thousands of neuroscience papers in all sorts of formats including images, laboratory notes, and even hand drawings from experiments using wildly different methodologies of data gathering in a dozen languages. It seems like a ridiculously daunting task that requires all the latest developments in machine learning and natural language processing. There has been some progress made with the rat visual cortex. (2) Neurons can be classified into morphological types (m-types), electrical types (e-types), and synaptic types (s-types). M-types are classified by how the neuron actually looks, and the branching pattern of the dendrites and axons. In the rat somatosensory cortex, they have identified 55 different m-types. E-types are classified by the shape of the electrical firing pattern. There are 11 e-types, 1 excitatory type, and 10 inhibitory. These two categories combine into 207 morpho-electric types (me-types). There are thousands of synapse types. They are distributed according to how the different neurons are connected. (1)
Regions of the brain are made up of layers that look like blobs of dough nestled on top of one another. A neuron circuit is the smallest hexagonal cylinder of tens of thousands of connected neurons linking down through all the different layers. A minimum of 300 of these cylinders seem to be needed for so-called emergent patterns to form. Emergent patterns are patterns that emerge from the complex non-linear electrical structure of these regions containing layers of neurons. These patterns form the higher-level cognitive outputs of the brain. (6)
It is necessary to note that this process is a computational process from beginning to end. Data is collected and generalized, and morphology is digitized. Connectivity and type distribution is calculated. Missing data is filled in statistically. Micro circuits are formed and simulated. Results are published giving recommendations for further research. Bad data is trimmed. As new data is entered into the model it gradually gets better. The short-term goal is to simulate the mouse brain. The long-term goal is human brain simulation.
Markram, Henry, Eilif Muller, Srikanth Ramaswamy, Michael W. Reimann, Marwan Abdellah, Carlos Aguado Sanchez, Anastasia Ailamaki, et al. “Reconstruction and Simulation of Neocortical Microcircuitry.” Cell 163, no. 2 (October 8, 2015): 456–92. https://doi.org/10.1016/j.cell.2015.09.029.
Ramaswamy, Srikanth, Jean-Denis Courcol, Marwan Abdellah, Stanislaw R. Adaszewski, Nicolas Antille, Selim Arsever, Guy Atenekeng, et al. “The Neocortical Microcircuit Collaboration Portal: A Resource for Rat Somatosensory Cortex.” Frontiers in Neural Circuits 9 (October 8, 2015). https://doi.org/10.3389/fncir.2015.00044.
Blue Brain Portal. “Blue Brain Portal.” Accessed May 22, 2020. https://portal.bluebrain.epfl.ch/.
Greene, Tim. “10 of the World’s Fastest Supercomputers.” Network World, November 18, 2019. https://www.networkworld.com/article/3236875/embargo-10-of-the-worlds-fastest-supercomputers.html.
edX. “EdX.” Accessed May 29, 2020. https://www.edx.org/.
Blue Brain Portal. “Blue Brain Cell Atlas.” Accessed May 29, 2020. https://portal.bluebrain.epfl.ch/resources/models/cell-atlas/.
Commentaires