Research Signal – Source date 2025-06-21
A dementia diagnosis describes a person’s cognitive and daily-function difficulties; by itself, it does not identify every process affecting the brain. Alzheimer’s disease pathology and injury to blood vessels often occur together, and the combination can raise dementia risk. That overlap makes a practical question difficult: if someone has both, what could a blood test or scan distinguish?
A paper published on 21 June 2025 reports proceedings from a workshop held in January by the Albert Research Institute for White Matter and Cognition. Researchers discussed possible biomarkers—measurable signs in blood or images—for vascular contributions to cognitive impairment and dementia, or VCID. The candidates included blood proteins, cell-derived particles, and several kinds of brain imaging. They measure different things, so a result would need a clear interpretation before it could help explain an individual’s symptoms.
Dementia labels can conceal several brain processes
Dementia is a clinical description, not a guarantee that one specific pathology is responsible. In many people with cognitive decline, more than one dementia-associated pathology is present. A diagnosis of Alzheimer’s disease dementia, for example, does not by itself establish how much vascular injury also contributes.
This matters because vascular changes may be clinically silent on their own yet interact with other brain pathologies. The workshop proceedings cite evidence that coexisting Alzheimer’s pathology and vascular pathology are associated with nearly twice the risk of clinical dementia compared with Alzheimer’s pathology alone. A useful VCID measure would therefore add information about a process contributing to decline, rather than simply assign another broad label.
Blood proteins can connect vascular changes with clinical patterns
One line of work discussed at the workshop examined thousands of proteins circulating in plasma. Researchers related 4,877 proteins to MRI features of cerebral small vessel disease, including white matter hyperintensity volume, and to cognitive or clinical features such as motoric cognitive risk and vascular dementia. Some associations were replicated in external groups. The implicated proteins included ones primarily expressed by oligodendrocytes, as well as proteins involved in metabolic and immune processes.
The breadth of these associations is useful for discovery: plasma may reflect both brain and vascular disease. But a protein associated with an MRI feature or a clinical pattern is not automatically a specific readout of the cause of a person’s cognitive decline. The proceedings also describe analyses that prioritized oligodendrocyte myelin glycoprotein and neuronal pentraxin receptor as candidate markers, and a protein network enriched for complement activity as potentially involved in white matter injury. These are research leads, not established clinical tests.
Inherited disease offers another way to look for markers. CADASIL, caused by mutations in the NOTCH3 gene, is a rare small vessel disease with a known progression from early white matter injury toward stroke and, in some people, cognitive impairment. Studies presented at the workshop compared plasma proteins across people at different clinical stages. Earlier proteomic work had found patterns suggesting changes in fibrosis, blood vessel growth, and metabolism. Because the disease’s course is more defined than the many causes grouped under sporadic VCID, researchers can ask whether protein changes accompany a shift toward cognitive problems. Whether those findings translate to more common forms of vascular injury remains an open question.
Cell-derived particles offer a more targeted blood sample
Proteins are not the only material in blood that may carry information about the brain. Extracellular vesicles are small, membrane-bound particles released by cells; their surfaces and internal contents can offer clues about their cell of origin and state. Researchers presented work using vesicles in mouse stroke models, enriched by markers associated with neutrophils or activated endothelial cells. The approach aims to examine RNA released as brain injury evolves and compare it with changes in brain tissue.
This is a more experimental route than broad plasma protein association studies. If vesicles can be reliably tied to particular cell types and changes, they could help distinguish where a blood-borne signal comes from. The proceedings describe this as a possible “liquid biopsy” approach, not a validated way to diagnose VCID in people.
Scans measure different aspects of vascular health
Brain imaging candidates also answer different questions. Cerebral blood flow estimates how much blood reaches tissue, but it can change with metabolism as well as with vascular function. Oxygen extraction fraction, or OEF, instead reflects the balance between oxygen delivered to the brain and oxygen used. In longitudinal aging studies discussed at the workshop, OEF increased over time, especially among people with higher vascular risk, and its changes were associated with vascular risk progression and white matter lesion growth. The reported pattern differed from Alzheimer’s pathology: OEF rose with vascular pathology and fell with Alzheimer’s pathology, while blood flow could be affected by either.
Another imaging candidate is enlarged perivascular spaces, the visible spaces around small blood vessels. An automated method was used to measure these spaces in post-mortem MRI from 817 older adults whose brains were also examined after death. Their location mattered: frontal and total counts were associated with infarcts, while spaces in temporal and occipital regions were associated with cerebral amyloid angiopathy, a condition involving amyloid in vessel walls. Basal ganglia spaces were independently associated with hypertension. Across the study, more spaces were also associated with lower cognition beyond the effects of measured pathologies and other factors.
These findings make perivascular spaces a possible clue to several processes, not a single-purpose marker. The workshop discussed them in relation to sleep-driven waste clearance, but their associations with infarcts, amyloid vessel disease, hypertension, and cognition show why location and context affect what an image may mean.
Context changes what a candidate marker can tell us
People with Down syndrome illustrate why assumptions about vascular risk need care. Although this population is protected from some common vascular risk factors, people with Down syndrome who develop dementia often have both Alzheimer’s pathology and cerebrovascular disease. An extra copy of chromosome 21 is associated with early Alzheimer’s pathology, and amyloid can also accumulate in brain vessels as cerebral amyloid angiopathy. The proceedings report that this vascular pathology can accompany microbleeds and reduced cerebral blood flow. A marker interpreted only through the lens of typical risk patterns could miss this combination.
A different context comes from preclinical retinal research. In a mouse model of small vessel disease, researchers observed reduced visual sensitivity, changes in retinal blood vessels and glial cells, and altered inflammatory and vascular gene activity. Such findings point to the retina as a possible window worth investigating. They also matter when interpreting vision-dependent cognitive tests in animal models: performance may reflect visual changes as well as cognition. The mouse results do not show that retinal imaging can diagnose vascular contributions in people.
Combining data may sharpen interpretation
Artificial intelligence approaches discussed at the workshop can combine neuroimaging, cognitive assessments, and clinical histories. In principle, seeing these sources together could help distinguish patterns associated with Alzheimer’s disease from those associated with vascular dementia, including earlier cerebrovascular changes. The value would depend on whether the patterns correspond to underlying disease processes, rather than merely sorting people into labels.
That requirement was a central challenge in the workshop discussion. VCID covers a wide range of conditions and mechanisms, so each candidate measure needs a defined relationship to the process it is meant to capture. The evidence presented spans different populations, measurements, and stages of research. No candidate described in these proceedings provides a clinically usable test that separates vascular from Alzheimer’s contributions for an individual patient.
What clearer dementia diagnoses could enable
Reliable detection of a vascular contribution could help trials enroll people whose disease process matches a treatment under study and help clinicians explain mixed causes of decline to patients. That capability depends on candidate measures proving interpretable and useful across the people being assessed. The workshop proceedings describe promising ways to investigate the overlap; they do not show that biomarker-guided treatment selection improves outcomes.
Key Sources
- Primary source: The future of biomarkers for vascular contributions to cognitive impairment and dementia (VCID): proceedings of the 2025 annual workshop of the Albert research institute for white matter and cognition.
DOI or Journal Page · Peer-Reviewed Paper · 2025-06-21
Keep exploring