In This Story
Abolfazl Safikhani, an assistant professor in the George Mason University Department of Statistics, is a principal investigator on a new National Institutes of Health (NIH) grant of over $3 million, exploring new approaches for early detection of Alzheimer’s disease.
“We propose using resting-state functional MRI (fMRI) to look at the functional connectivity changes in people who are going through the disease progression, versus healthy people,” says Safikhani. “By comparing functional connectivity changes among healthy individuals with those suffering from the disease, we aim to identify individuals at risk. The approach shows great promise—when we compute the functional connectivity in a novel way, we can classify Alzheimer’s patients with more than 98 percent accuracy.”
There are early biomarkers of potentially developing Alzheimer’s, such as abnormal levels of proteins in cerebrospinal fluid, but they have limitations. Safikhani and colleagues suggest that readily available and noninvasive detection methods like fMRI can work at a preclinical stage, when disease-modifying therapies are most effective. For this project, they will use existing data sets without scanning new subjects.
“We focus on fMRI because it's noninvasive and it could capture functional connectivity as opposed to the brain’s structural connectivity,” he said. “Because if the patient already has some atrophy, i.e., changes in the structure of the brain, it may be too late to start treatment.” He noted that the disease goes through certain steps as it progresses in a person’s brain, but the order and steps are still rather unclear to neurologists.
Given the small number of available samples in any single study, the proposal develops a transfer learning framework to utilize data from multiple sources. This approach is not without challenges itself, given patient data from various sources will be used: different machines with their own variations are used in imaging, there are confidentiality issues, and different data sources represent different subsets of the population. In addition, Safikhani said, “Software development is an important aspect of this project. We want practitioners to take advantage of the methodological developments that we are going to hopefully make in the next five years.”
Getting an fMRI for routine checks on Alzheimer’s progression—akin to annual mammograms or colonoscopes as people age—is unlikely to be standard practice soon. But Safikhani said that those who are at increased risk, such as people who have suffered a traumatic brain injury or veterans who have head injuries, would be potential candidates.
The grant is from NIH’s National Institute on Aging and Safikhani is working with another PI, Ali Shojaie, a professor at the University of Washington, who he collaborated with while he was a visiting scholar there in 2017. The NIH grant review process involves giving each proposal an average score and placing it in a quantile relative to the others, only funding those that fall above a particular quantile. Safikhani and colleagues’ proposal was in the top one percent.
Their work, at the intersection of statistics, time-series research, and neuroscience, could help chart a clearer path toward detecting Alzheimer’s before its effects take hold, turning early insight into early intervention.