Brain Simulation, Big Data, And A New Computing Paradigm
February 17, 2014

Research To Unlock The Brain’s Mysteries Detailed At AAAS Conference

redOrbit Staff & Wire Reports - Your Universe Online

Experts from the Ecole Polytechnique Fédérale de Lausanne (EPFL) in Switzerland are combining experimental data with the core principles of brain organization in order to create a detailed computer model which will allow them to conduct supercomputer-based simulations of the brain’s inner workings.

“Brain simulation allows measurements and manipulations impossible in the lab, opening the road to a new kind of in silico experimentation,” said the EPFL’s Henry Markram. The amount of neuroscience data is resulting in a tremendous deluge of brain-related information, with more constantly on the way – making it a challenge to search, access and analyze all this data.

The project is one of several being pursued by scientists affiliated with the Human Brain Project, a 10-year, $1.3 billion project designed to construct a new information and communications technology infrastructure for neuroscience and brain-related research in medicine.

The project was launched in October 2013 and involved over 80 top universities and research institutions from Japan, China, the US and 22 European nations. The group plans to announce over a dozen new research partnerships prior to the end of the month, and gave an update on their ongoing projects Sunday during “Inventing New Ways to Understand the Human Brain” at the 2014 AAAS Annual Meeting in Chicago.

“Understanding the human brain is one of the greatest challenges facing 21st century science,” officials from the EPFL said in a statement Sunday. “If we can rise to this challenge, we will gain profound insights into what makes us human, develop new treatments for brain diseases, and build revolutionary new computing technologies that will have far reaching effects, not only in neuroscience.”

In the panel discussion, EPFL’s Sean Hill, leader of the Human Brain Project’s Neuroinformatics Platform, explained that the newly-developed platform will provide scientists with the tools they need to manage, navigate and annotate spatially-referenced atlases of the human brain. These tools will be a fundamental part of the project’s brain-modeling effort, he added.

The Human Brain Project’s Neuroinformatics Platform will combine several different types of information, the researchers explained. For example, Seth Grant of the University of Edinburgh Centre for Neuroregeneration explained that he and his colleagues were developing new techniques to decode the underlying molecular principles of the brain’s organization.

One aspect of that is how individual proteins assemble into larger complexes, and as Grant explained during the AAAS meeting, this research could be applied to treat the genetic mutations behind disorders such as autism and schizophrenia. Likewise, understanding more about how the brain computers can assist the development of new technology, combining realistic brain models with hardware for an entirely new paradigm of computing.

Furthermore, Dr. Christof Koch explained at the conference that he and colleagues from the Allen Institute for Brain Science were working on a 10-year plan to understand the brain’s structure and function by mapping mice and human cell types with computer simulations.

Their goal is to figure out the cell’s network, as well as how they encode, relay and process information. “The project… promises massive, multimodal, and open-access datasets and methodology that will be reproducible and scalable,” the researchers explained.

Harvard University scientists are also participating in the Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative, which is designed to map all of the brain’s neurons using advancing technology. At the meeting, genetics professor George Church gave a progress report on new tools for measuring brain cell development and connectivity in both rodent and human samples.