Release the 'Krakencoder': New algorithm gives scientists unprecedented insight into neural function
Researchers are hopeful that a new algorithm named after a tentacled folklore sea monster can help scientists better understand neural connectivity as it relates to the anatomy of the brain.
The “Krakencoder” is a “joint connectome mapping tool that simultaneously bidirectionally translates between structural and functional connectivity, and between different atlases and processing choices via a common latent representation.” It was designed to spot structural and functional connectivity patterns from multiple imaging exams and combine them to offer providers a more thorough depiction of an individual’s neural health. Essentially, it gives experts multiple views of activity occurring in a single region, much like the instant replays used in televised sports.
Amy Kuceyeski, PhD, senior author of a new paper describing the algorithm, and a professor of mathematics in radiology and neuroscience at Weill Cornell Medicine, explained that Krakencoder will help to effectively combine the all the data acquired in different neurology studies to help scientists better understand the underlying mechanisms of neural activity.
“Everybody uses different methods to take pictures of the brain's networks,” Kuceyeski said. “We’re still just scratching the surface of how brain networks relate to the tasks of everyday life, like solving a math problem, having a conversation with a friend or driving a car.”
“Our fundamental assumption is that each set of choices in the imaging and processing pipelines provides a different view of the same underlying system,” first author Keith Jamison added. “In my head, I saw this as some sort of monster with multiple arms that could reach out and grab different brain representations and digest and congeal them into one unified connectome.”
The Krakencoder was trained on data from more than 700 people included in the National Institutes of Health’s Human Connectome Project, which includes detailed structural and functional imaging of all participants. When tested, researchers involved in the algorithm’s development found that it could analyze a person’s structural connectome to predict their functional connectome with significant accuracy—20 times more accurate than previously developed models, to be exact.
What’s more, Krakencoder also could predict other identifying factors, such an a person’s age, gender and cognitive function that had been determined during external testing. This is something that has been historically challenging using imaging alone, Kuceyeski noted.
Experts involved in the research suggested they are just getting started. Future plans for Krakencode involve combining it with a network modification tool known as NeMo. Researchers are hopeful that by combining the two, they will be able to predict functional outcomes of individuals with neurological injuries or diseases, thereby allowing providers to better treat and manage patients’ neural health.
Learn more about the work here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.