It has been a year since the launch of a research and development initiative between IBM Canada, the governments of Canada and Ontario, and a consortium of seven universities, backed by a total investment of $210 million.
Mark Daley, associate professor at the University of Western Ontario (UWO), credits this initiative to the accelerated pace of his lab’s research on brain graphs, a mathematical abstraction of relationships between brain regions. Two regions of the brain are deemed to be connected if they show similar patterns of activity over a short period of time, explained Daley.
Daley showed a live demo (see video) at this initiative’s one-year anniversary event hosted at the IBM Innovations Center in Toronto on April 10. For the live demo, Daley videoconferenced with two team members, Jinhui Qin, a postdoc, and Rhodri Cusack, associate professor at UWO. The neural activity of a postdoc undergoing functional magnetic resonance imaging played out in a dynamic display of dots interconnected with lines – brain graphs. Brain graphs from a scan of Daley’s brain are available online.
“Without this initiative, a live demo wouldn’t have been possible today. We’d have a prototype two years from now,” Daley said after the live demo. No wonder. This research is supported by an IBM technology backbone that includes Canada’s fastest supercomputer, generating a near real-time analysis of fMRI data.
Researchers have been applying graph theory to human neurological data for several years, Daley said. However, to his and his colleagues’ knowledge, they’re the first group to apply graph theory to a brain in real-time.
“Most researchers scan for long periods of time to generate one graph that shows the average connectivity. We’re interested in how that connectivity changes over time. Most importantly, we’re doing it in real-time,” Daley said.
Daley and his colleagues believe that their technology could reduce the cost of fMRI in the clinical setting. This technology turns functional neuroimaging from a linear process to one that’s akin to a feedback loop, explained Jingyun Chen, a postdoc in Daley’s lab.
Thanks to real-time feedback, clinicians will be able to adjust experimental protocols in real-time, decreasing the need to bring back patients for additional scans. In addition to decreasing repeat visits, this technology will decrease the time a patient spends in a scanner, accelerate diagnoses, and decrease wait times.
Assessing cognitive impairment in non-communicative patients like neonates works especially well with this technology because very little behavioural testing can be conducted with them and they move around a lot. However, the benefits of this technology are not only limited to non-communicative patients.
The technology summarizes complex networks into “scores” that may help clinicians identify potential pathologies. A lot of literature shows how certain “scores” indicate specific neuropsychiatric disorders.
For example, when somebody presents with a depressed phenotype, it’s hard to identify the particular type of depression with traditional methods. However, with this technology clinicians will be more likely to identify the specific type of depression.
On a broader scale, this technology becomes an enabling technology, Daley said, because it allows us to look at two brains and ask questions, like: “How is this person’s brain quantitatively different from this person’s brain?”
Daley and his colleagues are set on making this technology a “push-button system” as much as possible. Daley cautions, however, that this technology is not intended to make a diagnosis but rather to help raise red flags. The technology can be accessed via a Web browser on a laptop, a tablet, or a smartphone.
