Alzheimer’s disease is traditionally characterized by the buildup of amyloid plaques and tau protein tangles, but brain inflammation appears to play an early role in this process. A new study provides evidence that two separate inflammatory proteins in the blood signal two distinct paths of brain damage, both of which eventually lead to memory problems. The findings were published in Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring.
Alzheimer’s disease involves a complex interplay of biological events that begin long before memory loss becomes noticeable. Researchers suspect that brain inflammation acts as an early trigger in this disease process, rather than just a reaction to it. A study covered by PsyPost in 2025 indicated that markers of brain inflammation increase years before symptoms appear, right alongside the initial formation of amyloid plaques.
Building on this timeline, a 2021 cross-sectional study found that measuring an inflammatory protein called GFAP in the blood can reliably detect early amyloid accumulation in the brain before cognitive issues arise. Similarly, a 2021 study of cognitively unimpaired and impaired adults found that blood levels of GFAP specifically reflect amyloid plaque buildup rather than tau protein tangles.
The new study builds on this progression by exploring how GFAP and a second inflammatory marker, YKL-40, fit into a larger sequence of biological events.
“There is growing evidence that inflammation and the brain’s immune response play an important role in Alzheimer’s disease, potentially well before symptoms emerge. But the literature has been somewhat inconsistent,” said study co-authors Michael A. Yassa and Batool Rizvi. Yassa is a professor of neurobiology and behavior, the James L. McGaugh Endowed Chair, and director of the Center for the Neurobiology of Learning and Memory at the University of California, Irvine; Rizvi is a postdoctoral scholar in the Department of Public Health Sciences at the University of California, Davis.
“That made us wonder whether we were trying to fit several different biological processes into a single pathway,” the researchers added. “We hypothesized that different forms of neuroinflammatory activity might be associated with different aspects of disease.”
To map these intersecting biological pathways, the team—whose work is detailed on the Yassa Lab website—recruited 126 older adults from the local community. The participants had an average age of 70 and were cognitively unimpaired, meaning they showed no signs of dementia or mild cognitive impairment and performed normally on standard memory tests.
To assess the participants’ brain health, the scientists collected a variety of medical data. They drew blood to measure the levels of three specific proteins: YKL-40 and GFAP, which are markers of nervous system inflammation, and p-tau217, a marker of tau protein tangles. The participants also underwent two types of brain scans. Magnetic resonance imaging (MRI) was used to measure the volume of the hippocampus and the thickness of the medial temporal lobe, brain regions essential for memory.
The MRI scans also allowed the team to quantify white matter hyperintensities. These are bright spots on a brain scan that indicate small blood vessel disease and structural damage in the brain’s white matter. Additionally, positron emission tomography (PET) scans were used to measure the accumulation of amyloid beta plaques in the brain.
Finally, the participants completed a series of neuropsychological assessments, including the Rey Auditory Verbal Learning Test. This test measures a person’s ability to learn a list of words and remember them after a delay or after being exposed to distracting information. The researchers specifically focused on a measure of memory called retroactive interference, which shows how well a person can retain original information when new information is introduced.
To make sense of all these measurements, the team used a statistical technique called structural equation modeling. This approach allows scientists to test proposed chains of cause and effect by examining how multiple variables interact simultaneously.
The modeling revealed two separate pathways linking inflammation to brain damage. First, higher levels of the inflammatory protein YKL-40 in the blood were associated with more white matter damage in the brain. However, YKL-40 levels did not relate to amyloid plaque buildup.
Second, higher levels of the inflammatory protein GFAP were associated with increased amyloid plaque buildup. But unlike YKL-40, GFAP levels did not relate to white matter damage. This provides evidence that YKL-40 and GFAP represent two parallel, independent tracks of neuroinflammation, one tied to blood vessel damage and the other tied to Alzheimer’s-related plaque.
“What surprised us was how clearly the two inflammatory markers separated,” the researchers noted. “That suggests that calling something simply ‘neuroinflammation’ may obscure important biological differences. Different inflammatory or glial responses may be associated with different aspects of Alzheimer’s disease and brain aging.” This refers to reactions from glial cells, which are non-neuronal cells that support and protect the brain.
Despite their separate tracks, both white matter damage and amyloid plaque accumulation independently linked to higher blood levels of p-tau217. In other words, both vascular damage and amyloid pathology appear to converge, promoting the accumulation of toxic tau tangles.
Following this point of convergence, the pathway led directly to structural brain changes and cognitive deficits. Higher levels of p-tau217 were associated with a thinner medial temporal lobe cortex and a smaller hippocampal volume. In turn, a smaller hippocampus was associated with worse memory performance on the word-recall test.
“So rather than seeing Alzheimer’s disease as a single chain of events, our findings suggest that several biological processes may be occurring in parallel and ultimately contributing to the same downstream brain changes,” the authors said.
The results align with research covered by PsyPost in 2026, which indicated that elevated blood levels of tau and hippocampal shrinkage are tied to memory impairment in older adults without dementia. One difference in measurement is that the earlier study tracked memory decline trajectories over several years, whereas the new study modeled a single cross-sectional snapshot in time that included markers of brain inflammation.
The new findings are also consistent with a study covered by PsyPost in 2024, which found that tau accumulation and hippocampal shrinkage relate to cognitive decline even in individuals with low levels of amyloid plaque. Together, these studies suggest that tau and brain shrinkage are predictors of memory issues, regardless of which initial pathway triggered the damage.
As with all research, there are some caveats. The study relied on observational data collected at a single point in time, which restricts the ability to confirm a strict cause-and-effect sequence among the biological markers. “The most important limitation is that this was a cross-sectional study, meaning that the biomarkers were measured at approximately the same stage rather than tracked over many years,” the researchers explained. “Our statistical model allowed us to test a biologically motivated sequence of relationships, but it cannot establish that one biomarker caused the next or prove the order in which these changes occur.”
Another limitation is that blood-based markers of brain inflammation can sometimes reflect systemic inflammation from other parts of the body. Furthermore, the researchers noted that the participant sample lacked diversity. “The study also involved a relatively modest sample of cognitively unimpaired older adults recruited from a single research cohort, and the participants were predominantly White,” they said. “That limits how broadly we can generalize the findings.”
Future research will need to track individuals over many years to verify the temporal sequence of these two inflammatory pathways. “If YKL-40 and GFAP are identifying different biological processes, we want to know when each process begins, how it changes as Alzheimer’s pathology develops, and whether it predicts who is most likely to experience subsequent neurodegeneration or cognitive decline,” the authors explained.
The researchers hope to eventually use combinations of blood biomarkers and brain scans to identify distinct biological profiles of Alzheimer’s risk. “The broader point is that Alzheimer’s disease is biologically heterogeneous,” they noted. “Our findings suggest that two people could potentially arrive at similar downstream signs of Alzheimer’s disease through somewhat different combinations of biological processes.”
Because different pathways may be active in different individuals, a single intervention might not work for everyone. “Identifying distinct biological pathways is an important step toward eventually developing a more individualized approach to prevention and treatment,” the researchers concluded.
The study, “Parallel neuroinflammatory pathways to cerebrovascular burden and amyloid beta in Alzheimer’s disease,” was authored by Batool Rizvi, Jenna N. Adams, Alison Bamford, Soyun Kim, Mithra Sathishkumar, Nicholas J. Tustison, Lisa Taylor, Nandita Tuteja, Liv McMillan, Bin Nan, Hengrui Cai, Yuritza Y. Escalante, Novelle J. Meza, Alyssa L. Harris, Rond Malhas, Adam M. Brickman, Mark Mapstone, Elizabeth A. Thomas, and Michael A. Yassa.