High performance computing advances neuroscience research

1287

Around the world, researchers and medical professionals are hard at work developing methods and processes to find cures, improve diagnoses and develop cost-effective treatments for the diseases and disorders that plague our population.

Closer to home, in Ontario, academic and industry researchers are working to improve the quality of life for those living with ailments, creating better patient experiences and aiming to reduce spending within the health care system by working with a set of technology-leading tools.

These tools focus on making sense of the mountains of “big data” that have already been collected to make advances that were not previously possible. For example, Dr. Mark Daley, an associate professor of Computer Science and Biology at Western University, is using high performance computing (HPC) at the IBM Canada Research and Development Centre (CRDC) to further his research in neuroscience and modeling the brain.

“To have this leg-up at this critical juncture is really exciting,” says Dr. Daley, “this will help to make [Canadians] world leaders in many disciplines by enabling us to make sense of the vast amounts of data we are generating.”

The CRDC is offering Canadian researchers such as Dr. Daley the opportunity to access an advanced HPC infrastructure that includes the most powerful supercomputer in Canada and some of the most advanced analytics software available today. Dr. Daley’s brain research is one of fifteen health care projects using The Southern Ontario Smart Computing Innovation Platform (SOSCIP) capabilities.

Real-time analysis of human brain networks

An important tool in the diagnosis of neurological disorders is functional magnetic resonance imaging, or fMRI. The challenge with this process is that the current fMRI results often take more than a week to be analyzed and can require additional tests, which incur further expenses to the health care provider. Dr. Daley and his research team believe there are better methods to the review process and as such, have developed a working prototype that could enable faster, more accurate diagnoses in near-real time.

In addition, the working prototype uses dynamically adapted brain scanning. This allows the test to be stopped, tuned and adjusted while it’s underway. This allows for review of neural functional connectivity networks. What’s more, these neural functional connectivity networks have shown potential as diagnostic indicators for several brain disorders including autism, schizophrenia, Alzheimer’s and ADHD.

Dr. Daley and his colleague Rhodri Cusack are also looking at neonates flagged as having potential neurological problems.  Dr. Daley explains:  “With premature babies it’s very difficult to do behavioral testing.  For example you can’t say ‘how many fingers am I holding up?’ In this case we need some sort of quantitative metric such as brain scanning.  In the case of infants who are at high risk, we’re hoping that we can identify simple metrics where the graph theory tool can tell us if an infant is at risk and needs to see a pediatric neurologist straight away.”

“The tools IBM is providing allow us to take these colossal data sets and make sense of them in an automated way,” Dr. Daley notes. “In neuroscience, we can often generate terabytes of data from images detailing both the structure and the function of the brain. The new computational resources will allow us to automate that analysis so we can boil that down into simple models.”

From a real-world application and commercialization perspective, this tool Dr. Daley and his team have developed is expected to create a product that would be installable with existing MR facilities so that a doctor would then have access to a stock tool set.

The SOSCIP project portfolio includes close to 40 projects and the results of these projects are expected to hold much potential and deliver significant advancements in their respective fields.