Summary: Shady Rahayel, PsyD, PhD, and his team used machine learning to analyze brain MRIs of people with REM sleep behavior disorder, identifying two distinct atrophy patterns that may predict whether an individual will develop Parkinson’s disease or Lewy body dementia.
Key takeaways:
- People with RBD are at high risk of developing Parkinson’s disease or Lewy body dementia, with each linked to different patterns of brain atrophy.
- Brain atrophy in Parkinson’s tends to start deep in the brain and spread outward, while in LBD, it begins in the cortex and spreads inward.
- The AI algorithm estimates disease progression without needing longitudinal scans from the same individuals.
Shady Rahayel, PsyD, PhD, is studying the distinct patterns of atrophy that REM sleep behavior disorder leaves in the brain. How does this atrophy begin and progress in people with REM sleep behavior disorder (RBD), and how can it predict the development of Parkinson’s disease and Lewy body dementia (LBD)?
Hidden Warning Sign
His research team’s preliminary findings are published in The Lancet eBioMedicine.
“The people with RBD who come to see us are in good health, but we know that something is wrong,” says Rahayel, a Université de Montréal medical professor, in a release. “Of those who subsequently develop a disease, half will have Parkinson’s and the other half LBD.”
Having the second most common form of dementia after Alzheimer’s, “LBD patients are no longer able to function in everyday life,” Rahayel says. “As well as dementia, they will have Parkinson’s-like symptoms, vivid visual hallucinations, fluctuating attention, and other symptoms.”
Building a Global Brain Scan Database
At the Centre for Advanced Research in Sleep Medicine, at Montreal’s Sacré-Coeur Hospital, researchers have access to a large database of patients with REM sleep behavior disorder. But to find a biomarker with the help of artificial intelligence, Rahayel’s computational neuroscience team needed even more MRI brain scans.
Tracking down the information, they contacted 11 sleep centers around the world to assemble a database of 1,276 MRI scans of people at risk of, or with, Parkinson’s disease or Lewy body dementia, and also of healthy people.
Machine Learning Spots Diverging Disease Patterns
Using machine learning and computational models, the researchers identified two trajectories of brain atrophy progression.
“Normally, in a study of this type, you have to monitor the subjects annually and check the progression of the disease using brain scans,” Rahayel says. “But this powerful algorithm can estimate the stage of progression of the disease based on scans of different people.”
Lewy body dementia seems to be associated with brain atrophy that begins in the cortex and then spreads to the interior of the brain, while in Parkinson’s the atrophy progresses from the interior to the exterior of the brain.
Why the Atrophy Paths Differ
“The question now is why,” Rahayel says. The next steps will be to investigate the factors that lead to this deterioration in the cortex, such as vascular lesions, the effects of drugs, and the impacts of lifestyle choices.
“Now that we have identified these new progression patterns, our goal is to be able to determine from an MRI whether a person has one of them so that we can provide the best possible care,” Rahayel says.
He would like to be able to apply the model in a clinical setting and define biological markers that are more reliable than clinical assessment, which can be influenced by external factors.
Toward Clinical Use of MRI Biomarkers
“If a person has just started a new medication, for example, this could influence their memory score on a cognitive test,” he says. “Biomarkers are more objective.”
Rahayel hopes that a better understanding of how these diseases develop in the brain will support the search for more effective treatments.
We recommend for you:
ID 140391471 © Lightpoet | Dreamstime.com
