A neural mapping approach that pegs results from more than two dozen previous Alzheimer's studies found that reproducibility improves when trying to isolate symptoms to a brain network rather than a single area of the brain.
This is the first study to use "meta analysis network mapping" to essentially show results from prior studies that can't be reproduced to a single location in the brain are mappable to a distributed network, said Ryan Darby, MD an assistant professor of neurology at Vanderbilt University Medical Center, who is first author of the paper that will run in Brain in January.
"The study changes the assumption for where symptoms or cognitive processes should localize - away from a specific region and toward localizing to a brain network. Changing that assumption improves reproducibility," said Darby.
Neural networks include regions in different parts of the brain that are connected together (akin to distant cities connected by highways), so the study zoomed out its examination to think about symptom-specific circuitry rather than a single spot.
Of the 26 Alzheimer's studies examined, reproducibility to a specific region came to 20 to 30 percent - a low rate that highlights one of the persistent challenges in studies. But when looking at reproducibility to a network, the rate surged to 100 percent.
Darby called the results powerful.
The inquiry started with Alzheimer's disease and dementia because the theory in those disorders suggests it spreads across connected brain networks.
The study also found that studies of Alzheimer's disease patients with delusions localized to the same brain network that Darby found associated with brain lesions causing delusions. In the future he wants these findings and method to shed light on a path to therapy by identifying the symptom-specific brain network.
"We're hoping to look at individual patients with dementia and then eventually individual psychiatric patients. What we hope to see is that there actually is a consistent localization for the same symptom across different diseases, which is hard because different diseases disrupt brain function in different ways- but if we unite things at a network level then we can ask, 'is this the common way to understand this symptom across different diseases?' That's what I hope," said Darby.