Brain imaging scans can reveal specific structural and functional changes that distinguish normal aging from dementia, but the differences are often subtle and require expert interpretation. A person with normal cognitive aging may show mild brain shrinkage and changes in white matter, while someone with Alzheimer’s disease typically displays atrophy concentrated in the hippocampus and temporal lobes—areas critical for memory formation. For example, an 80-year-old with normal cognition might have slight ventricular enlargement on an MRI, which is expected with age, whereas an 80-year-old with dementia would show hippocampal atrophy along with other regional changes that correlate with their cognitive decline.
Neuroimaging isn’t a single “dementia test”—it’s one tool among many. Different scan types reveal different pathology: MRI shows structural changes, PET imaging detects abnormal protein accumulation, and CT primarily rules out stroke or tumor. A scan might show changes consistent with dementia, but the clinical diagnosis depends on cognitive testing, medical history, and sometimes cerebrospinal fluid or blood biomarkers. No scan alone can diagnose Alzheimer’s disease with certainty during life, though emerging blood tests for tau and amyloid proteins are changing this landscape.
Table of Contents
- What Do MRI Scans Show in Normal Aging Versus Dementia?
- PET Imaging and Protein Buildup—The Hallmark of Alzheimer’s
- Structural Signatures of Specific Dementias
- What CT Imaging Can and Cannot Tell You
- Limitations and Why Imaging Alone Is Not Diagnostic
- Blood Biomarkers and the Shift Away from Structural Imaging
- Choosing the Right Imaging in Clinical Practice
What Do MRI Scans Show in Normal Aging Versus Dementia?
MRI imaging is the workhorse of brain imaging for dementia evaluation because it shows anatomical detail and is non-radioactive. In normal aging, the brain gradually shrinks—a process called atrophy—particularly in frontal and temporal regions. The fluid-filled spaces around the brain (ventricles and sulci) widen correspondingly. This normal atrophy can begin in the 30s and accelerates after age 60, but it follows a predictable pattern and does not correlate with cognitive decline in healthy older adults. In contrast, Alzheimer’s disease produces selective, asymmetrical atrophy. The hippocampus—a seahorse-shaped structure buried deep in the temporal lobe—shrinks disproportionately early in the disease.
Radiologists measure hippocampal volume and look for loss relative to normal reference ranges. A 65-year-old with memory complaints and bilateral hippocampal atrophy on MRI has a much higher risk of progressing to dementia than a 65-year-old with normal hippocampal volume. Other dementias show different patterns: frontotemporal dementia causes frontal lobe atrophy, vascular dementia shows multiple small infarcts, and Lewy body dementia may appear relatively preserved on MRI despite significant cognitive symptoms. white matter—the brain’s “wiring” connecting different regions—also changes differently with normal aging versus dementia. Normal aging produces white matter hyperintensities, visible as bright spots on certain MRI sequences, particularly in the deep brain. These are common after age 70 and are associated with cardiovascular risk factors. However, extensive white matter disease can accelerate cognitive decline and is seen more severely in vascular dementia.
PET Imaging and Protein Buildup—The Hallmark of Alzheimer’s
Positron emission tomography (PET) imaging uses radioactive tracers to visualize metabolic activity and pathological protein accumulation. amyloid PET scans detect amyloid-beta plaques, and tau PET scans detect neurofibrillary tangles—the two pathological hallmarks of Alzheimer’s disease. In people with normal cognition, amyloid can be present in the brain for years or decades before symptoms appear; amyloid positivity alone does not equal dementia. However, when amyloid accumulation combines with tau pathology and cognitive symptoms, the diagnosis is more certain. A critical limitation of PET imaging is cost and availability. PET scans are expensive (often $3,000–$5,000), require specialized facilities with cyclotrons or generator units, and are not widely covered by insurance for symptomatic patients. Insurance companies often classify amyloid and tau PET as research tools rather than diagnostic procedures.
Many community hospitals and rural areas lack PET capacity entirely. For these reasons, PET is primarily used in research settings and specialized dementia centers, not in routine primary care evaluations. Fluorodeoxyglucose (FDG) PET, a different tracer, shows glucose metabolism. The Alzheimer’s disease pattern includes reduced metabolism in the temporoparietal and posterior cingulate regions—areas that light up normally in healthy brains. Frontotemporal dementia shows a characteristic frontal and temporal hypometabolism. Lewy body dementia often shows relative sparing of the posterior cingulate (unlike Alzheimer’s), which can help differentiate it. However, FDG-PET abnormalities overlap significantly with normal aging in some regions, and interpretation requires expert neuroradiologists.
Structural Signatures of Specific Dementias
Different dementia types have distinct imaging fingerprints. Frontotemporal dementia (FTD) produces pronounced atrophy of the frontal and anterior temporal lobes, sometimes asymmetrical. Patients with FTD often have behavioral changes—apathy, disinhibition, or loss of empathy—before memory loss, and their MRI reflects this anatomy. A 55-year-old with personality change and behavioral disinhibition showing selective frontal atrophy on MRI is a classic presentation for behavioral variant FTD. Vascular dementia is characterized by multiple lacunar infarcts (small, deep brain strokes), territorial infarcts, or diffuse white matter disease, all visible on MRI or CT. The cognitive decline correlates with stroke location and burden.
A person with a single large territorial stroke affecting the thalamus or strategic white matter can have sudden cognitive change, while multiple small lacunes accumulate over years. Unlike Alzheimer’s, vascular dementia does not have a specific pathological protein; imaging shows the actual ischemic damage. Lewy body dementia shows relatively preserved hippocampal volume and normal-appearing temporal lobes compared to Alzheimer’s disease—a critical difference that can reduce diagnostic confusion. The key on DaTscan (dopamine transporter imaging) is reduced dopamine in the basal ganglia striatum, reflecting loss of dopaminergic neurons. this scan is not widely available but is helpful in specialized centers. Patients with Lewy body dementia often have Parkinsonian features, visual hallucinations, and fluctuating cognition, and their imaging reflects the subcortical pathology.
What CT Imaging Can and Cannot Tell You
Computed tomography (CT) is fast, widely available, and much cheaper than MRI (typically $200–$1,500). It is excellent for ruling out acute problems: stroke, intracranial hemorrhage, subdural hematoma, or brain tumor. For an 80-year-old with sudden memory loss and confusion, a CT is often the first imaging to exclude stroke or bleed. However, CT is poor at showing subtle atrophy or white matter changes because it has lower tissue contrast than MRI.
An MRI might reveal hippocampal atrophy that a CT could miss entirely. CT also exposes patients to ionizing radiation and is less precise for assessing dementia-related brain changes. CT should not be the first choice for evaluating a patient with progressive memory loss in a stable clinical context—MRI is superior. CT is better reserved for acute presentations or when MRI is contraindicated (e.g., some pacemakers, though newer models are MRI-compatible). Newer CT perfusion imaging, which shows blood flow patterns, can detect hypoperfusion in Alzheimer’s disease and other dementias, but this remains primarily a research tool and is not standard in most centers.
Limitations and Why Imaging Alone Is Not Diagnostic
A fundamental limitation of all structural imaging is that brain changes do not always correlate with symptoms. Some people have significant hippocampal atrophy on MRI but normal cognition; others have minimal imaging changes despite cognitive decline. Conversely, dementia can progress rapidly with minimal visible structural changes early on. The lag between pathology (visible on amyloid PET) and cognitive symptoms can be years, making imaging in asymptomatic people ethically and clinically problematic. Aging and dementia pathology overlap. A 75-year-old with normal cognition might have white matter hyperintensities, mild brain atrophy, and even amyloid in the brain—all considered “normal” for that age.
Distinguishing normal aging from preclinical dementia requires serial imaging, biomarkers, and cognitive testing, not a single scan. Additionally, vascular disease, cerebral amyloid angiopathy (amyloid in blood vessel walls), and neuroinflammation all contribute to cognitive decline but are not easily visible on standard imaging. A warning: patients and families sometimes expect imaging to definitively diagnose dementia. Clinicians must explain that imaging provides supporting evidence—increased likelihood of dementia—but cannot rule it in or out alone. A normal MRI does not exclude dementia (cognitive impairment can reflect other causes like depression, thyroid disease, or medication effects). Conversely, abnormal imaging in an asymptomatic person does not diagnose dementia.
Blood Biomarkers and the Shift Away from Structural Imaging
Recent advances in blood biomarkers—phosphorylated tau, phosphorylated tau variants, and plasma amyloid-beta—are shifting diagnostic practice away from reliance on imaging. These biomarkers can be detected in blood with a simple test, reducing the need for expensive PET or lumbar puncture (cerebrospinal fluid collection). A person with cognitive symptoms and elevated phosphorylated tau in blood is at high risk for Alzheimer’s pathology, with or without imaging confirmation.
The advantage is accessibility: a blood test costs a few hundred dollars and can be done in any clinic, whereas PET requires specialized facilities. The disadvantage is that blood biomarkers detect pathology but do not show the anatomical extent of brain damage the way MRI or PET do. An MRI provides information about how much atrophy has already occurred and where; a blood biomarker indicates pathological process but not anatomical burden.
Choosing the Right Imaging in Clinical Practice
For a patient with cognitive complaints, the choice of imaging depends on clinical context. If the presentation is acute (sudden confusion, stroke-like symptoms), CT head is first-line to exclude hemorrhage or stroke. If the presentation is progressive memory loss in an older adult with normal neurological exam, MRI brain is standard to look for hippocampal atrophy, white matter disease, and rule out subdural hematoma or other structural lesions.
Amyloid PET or tau PET is reserved for specialist evaluation—typically memory disorder centers or research studies—and is rarely ordered in primary care. In the future, as amyloid-targeting therapies (like aducanumab or lecanemab) require proof of amyloid pathology, amyloid PET or plasma biomarkers may become more common in routine practice. Currently, however, plasma phosphorylated tau and amyloid-beta ratios in blood are increasingly used as gatekeepers: if blood biomarkers are negative in a symptomatic patient, advanced imaging may be deferred. If blood biomarkers are positive, MRI is typically obtained to assess structural change, and specialist evaluation is recommended.
- —
