Grid-based Google for Brain Imaging, Neurological Disease Research
Grid computing, long used by physicists and astronomers to crunch masses of data quickly and efficiently, is making the leap into the world of biomedicine. Supported by EU-funding, researchers have networked hundreds of computers to help find treatments for neurological diseases such as Alzheimer’s. They are calling their system the “Google for Brain Imaging.”
Through the Neugrid project, the pan-European grid computing infrastructure has opened up new channels of research into degenerative neurological disorders and other illnesses, while also holding the promise of quicker and more accurate clinical diagnoses of individual patients.
The infrastructure, set up with the support of EUR 2.8 million in funding from the European Commission, was developed over three years by researchers in seven countries. Their aim, primarily, was to give neuroscientists the ability to quickly and efficiently analyze magnetic resonance imaging (MRI) scans of the brains of patients suffering from Alzheimer’s disease. But, their work also has helped open the door to the use of grid computing for research into other neurological disorders, and many other areas of medicine.
“Neugrid was launched to address a very real need. Neurology departments in most hospitals do not have quick and easy access to sophisticated MRI analysis resources. They would have to send researchers to other labs every time they needed to process a scan. So, we thought, why not bring the resources to the researchers rather than sending the researchers to the resources,” explains Giovanni Frisoni, a neurologist and the deputy scientific director of IRCCS Fatebenefratelli, the Italian National Centre for Alzheimer’s and Mental Diseases, in Brescia.
The Neugrid team, led by David Manset from MaatG in France and Richard McClatchey from the University of the West of England in Bristol, laid the foundations for the grid infrastructure, starting with five distributed nodes of 100 cores (CPUs) each, interconnected with grid middleware and accessible via the Internet with an easy-to-use Web browser interface. To test the infrastructure, the team used datasets of images from the Alzheimer’s Disease Neuroimaging Initiative in the United States, the largest public database of MRI scans of patients with Alzheimer’s disease and a lesser condition termed mild cognitive impairment.
“In Neugrid, we have been able to complete the largest computational challenge ever attempted in neuroscience: we extracted 6,500 MRI scans of patients with different degrees of cognitive impairment and analyzed them in two weeks,” Dr. Frisoni, the lead researcher on the project, says, “on an ordinary computer it would have taken five years!”
Though Alzheimer’s disease affects about half of all people aged 85 and older, its causes and progression remain poorly understood. Worldwide, more than 35 million people suffer from Alzheimer’s, a figure that is projected to rise to over 115 million by 2050 as the world’s population ages.
Patients with early symptoms have difficulty recalling the names of people and places, remembering recent events and solving simple math problems. As the brain degenerates, patients in advanced stages of the disease lose mental and physical functions and require round-the-clock care.
The analysis of MRI scans conducted as part of the Neugrid project should help researchers gain important insights into some of the big questions surrounding the disease, such as which areas of the brain deteriorate first, what changes occur in the brain that can be identified as biomarkers for the disease and what sort of drugs might work to slow or prevent progression.
Neugrid built on research conducted by two prior EU-funded projects: Mammogrid, which set up a grid infrastructure to analyze mammography data, and AddNeuroMed, which sought biomarkers for Alzheimer’s. The team is now continuing their work in a series of follow-up projects. An expanded grid and a new paradigm Neugrid for You (N4U), a direct continuation of Neugrid, will build upon the grid infrastructure, integrating it with high performance computing (HPC) and cloud computing resources. Using EUR 3.5 million in European Commission funding, it also will expand the user services, algorithm pipelines and datasets to establish a virtual laboratory for neuroscientists.
“In Neugrid we built the grid infrastructure, addressing technical challenges, such as the interoperability of core computing resources and ensuring the scalability of the architecture. In N4U, we will focus on the user-facing side of the infrastructure, particularly the services and tools available to researchers,” Frisoni says. “We want to try to make using the infrastructure for research as simple and easy as possible,” he continues, “the learning curve should not be much more difficult than learning to use an iPhone!”
N4U also will expand the grid infrastructure from the initial five computing clusters through connections with CPU nodes at new sites, including 2,500 CPUs recently added in Paris in collaboration with the French Alternative Energies and Atomic Energy Commission (CEA), and in partnership with Enabling grids for e-science Biomed VO, a biomedical virtual organization.
Another follow-up initiative, outGRID, will federate the Neugrid infrastructure, linking it with similar grid computing resources set up in the United States by the Laboratory of Neuro Imaging at the University of California, Los Angeles, and the CBRAIN brain imaging research platform developed by McGill University in Montreal, Canada. A workshop was recently held at the International Telecommunication Union, an agency of the United Nations, to foster this effort.
Frisoni is also the scientific coordinator of the DECIDE project, which will work on developing clinical diagnostic tools for doctors built upon the Neugrid grid infrastructure.
“There are a couple of important differences between using brain imaging datasets for research and for diagnosis,” he explains. “Researchers compare many images to many others, whereas doctors are interested in comparing images from a single patient against a wider set of data to help diagnose a disease. On top of that, datasets used by researchers are anonymous, whereas images from a single patient are not, and protecting patient data becomes an issue.”
The DECIDE project will address these questions in order to use the grid infrastructure to help doctors treat patients. Though the main focus of all these new projects is on using grid computing for neuroscience, Frisoni emphasizes that the same infrastructure, architecture and technology could be used to enable new research — and new, more efficient diagnostic tools — in other fields of medicine.
“We are helping to lay the foundations for a new paradigm in grid-enabled medical research,” he says.
Neugrid received research funding under the European Union’s Seventh Framework Programme (FP7).