Subscribe to eTOC

Low Brain Network Connectivity and a Positive Amyloid PET Scan Found to Predict Cognitive Decline in Healthy Adults

ARTICLE IN BRIEF

Figure

MODEL ESTIMATES of preclinical Alzheimer cognitive composite (PACC) score decline according to rs-fcMRI networks.Slopes represent PACC trajectories according to connectivity in all functional networks. Blue line indicates baseline connectivity one standard deviation (SD) above the group mean (high connectivity); purple line indicates mean baseline connectivity; red line indicates baseline connectivity one SD below the group mean (low connectivity).

Among healthy adults, those with the lowest default mode network (DMN) connectivity — the network of interacting brain regions — and a positive amyloid scan shown via the PET radiotracer scan, Pittsburgh compound PiB, had poorer cognitive outcomes (at a group level) and were more likely to progress from preclinical Alzheimer's disease to the symptomatic phase of the disease, according to a new study.

In healthy older adults, lower functional brain connectivity coupled with evidence of amyloid from a positive positron emission tomography (PET) scan predicted cognitive decline three years down the road.

Those with the lowest default mode network (DMN) connectivity — the network of interacting brain regions — and a positive amyloid scan shown via the PET radiotracer scan, Pittsburgh compound P (PiB), had poorer cognitive outcomes (at a group level) and were more likely to progress from preclinical Alzheimer's disease (AD) to the symptomatic phase of the disease, a team of researchers reported in the June 7 online edition of Neurology.

The study author Reisa A. Sperling, MD, professor of neurology and director of the neuroimaging program at Massachusetts General Hospital and director of the Center for Alzheimer's Research and Treatment at Brigham and Women's Hospital, said that she doesn't think that measures of functional connectivity are yet ready to be used for diagnosing Alzheimer's disease, but she said they will be valuable in research studies. The signal is still too noisy and hard to interpret, she said.

Figure

DR. REISA A. SPERLING said she doesnt think that the measures of functional connectivity are yet ready to be used for diagnosing Alzheimers disease, but she said they will be valuable in research studies.

“It can't tell you who has AD and who doesn't,” Dr. Sperling said. “But it is a useful marker with amyloid imaging. We still don't know why some people with a head full of amyloid develop AD symptoms and others don't, and we may not for years to come,” she said. “This may give us a new marker to see who is at greater risk for cognitive decline. It's likely to tell us something about resilience to pathology. We have struggled to predict who will be in trouble over the next four years.”

Functional MRI has also been hard to standardize across centers, Dr. Sperling noted. “It's a good measure of what your brain is doing at that moment. And that's why it gets us closer to cognition. Functional connectivity is much more dynamic.”

Dr. Sperling is the principal investigator of the A4 prevention trial, which is using fMRI to measure networks at four time points during the study.

STUDY DESIGN

The researchers set out to study the utility of resting-state functional connectivity (rs-fcMRI) as a predictor of future cognitive decline in 237 people between 63 and 90 years old with preclinical AD who are enrolled in the Harvard Aging Brain Study.

The brain networks, which function independently but also talk to other networks to process information and act on it, comprise four areas that relate to cognitive domains (DMN, salience, dorsal attention, and frontoparietal control regions) and three non-cognitive domains (primary visual, extrastriate visual, and motor areas).

The researchers used the baseline connectivity of these networks to see if they could predict cognitive changes in the preclinical Alzheimer cognitive composite (PACC) an average of three to four years after they collected the initial rs-fcMRI and the PiB scan. The PACC battery measures a broad range of cognitive tasks.

Rachel Buckley, PhD, a post-doctoral fellow who worked on the study design and analysis, said that the baseline recordings of either the DMN or the salience network, along with high amyloid-beta (Abeta) burden, were powerful enough to predict a decline on the PACC scores at the three-year time interval. Those with the lowest DMN network connectivity and the highest Abeta levels scored lower on the PACC test than those with high connectivity and high Abeta.

Figure

DR. RACHEL BUCKLEY: “These are still early days. Functional connectivity is predicting cognitive decline but the steepest drops on test scores are in those with a lot of amyloid in their brain. If we could find a way to reduce the cost of screening for Alzheimers, that would be the Holy Grail. This is a first step.”

“These are still early days,” added Dr. Buckley. “Functional connectivity is predicting cognitive decline but the steepest drops on test scores are in those with a lot of amyloid in their brain. If we could find a way to reduce the cost of screening for Alzheimer's, that would be the Holy Grail. This is a first step.”

They are now trying to determine a cutoff for high or low connectivity on an individual level. It is not clear what comes first: the intrinsic network dysfunction that drives Abeta burden or Abeta burden that puts pressure on these networks.

EXPERTS WEIGH IN

Commenting on the study, William Seeley, MD, professor of neurology at the Memory and Aging Center at the University of California, San Francisco, said: “This is a convincing step forward. This is one of the first studies to carefully determine whether network connectivity at baseline predicts cognitive decline in the setting of amyloid positivity. To me, it converges with structural MRI findings that predict decline or conversion to dementia in the setting of a positive amyloid scan.”

Dr. Seeley and his team have been working on developing a method for measuring functional connectivity to help determine prognosis in individual patients.

The question is, would functional connectivity be predictive without knowing whether there is a positive or negative PiB scan?

PiB developer William E. Klunk, MD, PhD, the distinguished professor of psychiatry and neurology and the Levidow-Pittsburgh Foundation chair in Alzheimer's Disease and Dementia Disorders and co-director of the Alzheimer's Disease Research Center at the University of Pittsburgh, said that he suspects that all the study subjects with an amyloid positive scan will eventually show decreased connectivity in these networks and deficits in cognitive testing.

Figure

DR. WILLIAM SEELEY: “This is a convincing step forward. This is one of the first studies to carefully determine whether network connectivity at baseline predicts cognitive decline in the setting of amyloid positivity. To me, it converges with structural MRI findings that predict decline or conversion to dementia in the setting of a positive amyloid scan.”

Figure

DR. WILLIAM E. KLUNK said that he suspects that all the study subjects with an amyloid positive scan will eventually show decreased connectivity in these networks and deficits in cognitive testing. “That time frame will determine the value of measuring functional connectivity. Just knowing whether a PiB scan is positive tells you one thing, and just knowing that people have low network connectivity tells you something else.”

“That time frame will determine the value of measuring functional connectivity,” he said. “Just knowing whether a PiB scan is positive tells you one thing, and just knowing that people have low network connectivity tells you something else. “

He said that it is impressive that the high functional connectivity group had no changes on their cognitive studies — with or without a high amyloid burden. “That is surprising,” he said. The problem is that there is no standard for high and low functional connectivity.

“Do I need to study 237 cognitively normal elderly people, or can I apply their study data to our patients in Pittsburgh? Can I simply do a relatively cheap fMRI and know whether I then need to send someone for a PET scan? Wouldn't that be great? But how do you generalize this across centers?”

Beau M. Ances, MD, PhD, FAAN, professor of neurology at Washington University School of Medicine in St. Louis, said the results “fit nicely into our predictions. This fills in the gap in what is happening in the brain over time in people with low connectivity in the presence of amyloid. It will be important to look at several time points and compare the rate of change over time. It would also be important to look at tau imaging in the future.”

Michael Greicius, MD, MPH, associate professor of neurology at Stanford University, has been doing functional imaging studies since 2002, and co-founded a company, SBGneuro, that does image analysis for clinical trials.

There are challenges with such tests, he said. “It is important to know how to acquire resting state fMRI, clean it correctly and analyze it,” Dr. Greicius said. “There is no question that resting state is noisier than structural MRI and amyloid PET. But it is potentially more dynamic over a short period of time.”

LINK UP FOR MORE INFORMATION:

• Buckley RF, Schultz AP, Hedden T, et al. Functional network integrity presages cognitive decline in preclinical Alzheimer's disease http://www.neurology.org/content/early/2017/06/07/WNL.0000000000004059.short. Neurology 2017; Epub 2017 June 7.