Doctors diagnose dementia through a combination of medical history, cognitive testing, and brain imaging. There is no single test that definitively proves dementia—instead, physicians evaluate how memory, thinking, and reasoning have changed over time compared to a person’s baseline abilities. A doctor might start by asking detailed questions about when symptoms first appeared, what specific tasks have become difficult (like paying bills or managing medications), and whether a family member or close friend has noticed changes in judgment or language.
The diagnostic process typically involves an office visit where the doctor performs a brief cognitive screening, reviews medical and family history, and discusses any medications the person takes. For example, a 68-year-old woman might come in because her daughter noticed she’s repeating the same story in one conversation—something she never did before. The doctor would then administer memory tests, order blood work to rule out treatable conditions like low B12 or thyroid dysfunction, and possibly order brain scans like an MRI to check for structural changes or other causes of cognitive decline. A full dementia evaluation can take several hours across multiple appointments, and it often involves a specialist called a neuropsychologist or geriatrician who administers longer, more detailed testing than a primary care doctor can provide in a routine visit.
Table of Contents
- What Cognitive Screening Tests Do Doctors Use?
- Neuroimaging and Why Brain Scans Are Ordered
- Detailed Memory and Cognitive Testing
- Blood Tests and Emerging Biomarkers
- Distinguishing Types of Dementia
- The Role of Informant History
- Lumbar Puncture and Cerebrospinal Fluid Analysis
- Frequently Asked Questions
What Cognitive Screening Tests Do Doctors Use?
The most commonly used cognitive screening tool is the Mini-Cog, a three-minute test that combines a clock-drawing exercise with a three-word recall task. A person is asked to draw a clock showing a specific time and remember three unrelated words; the doctor then asks them to recall those words a few minutes later. The Mini-Cog is quick enough to use in a busy medical practice, making it a practical first-line screening tool that flags whether more extensive testing is needed. Another standard test is the Montreal Cognitive Assessment (MoCA), which takes 10 to 15 minutes and evaluates multiple areas: memory, attention, language, and visual-spatial skills.
The MoCA is more comprehensive than the Mini-Cog but still brief enough for an office setting. A 75-year-old man taking the MoCA might struggle with naming objects or drawing a 3D cube—scores below 26 typically warrant further investigation. The Mini-Mental State Exam (MMSE) is an older, 30-point screening tool that’s less commonly used now because it’s somewhat dependent on education level; someone who did not finish high school may score lower even without cognitive impairment. One important limitation is that these screening tools are not diagnostic—they are preliminary filters. A borderline or abnormal score does not mean someone has dementia; it means more testing is justified to determine the actual cause and severity of cognitive change.
Neuroimaging and Why Brain Scans Are Ordered
An MRI (magnetic resonance imaging) of the brain is often the first imaging study ordered because it provides detailed pictures of brain structure without radiation exposure and can reveal shrinkage (atrophy) in specific regions associated with different types of dementia. In Alzheimer’s disease, the hippocampus—a region critical for memory—typically shrinks; in frontotemporal dementia, the frontal lobes show preferential shrinkage. An MRI also rules out other causes of cognitive decline, such as brain tumors, previous strokes, or normal-pressure hydrocephalus, a condition where fluid accumulates in the brain and can mimic dementia but is sometimes reversible.
A CT (computed tomography) scan is faster and less expensive than MRI, uses X-rays, and is helpful for detecting recent strokes or bleeding, but it provides less detail of brain tissue. A PET (positron emission tomography) scan can show areas of the brain with reduced metabolism or abnormal protein buildup—amyloid-PET and tau-PET scans are increasingly used in specialized dementia centers to identify Alzheimer’s pathology—but these scans are expensive, require an injection of radioactive tracer, and are not available in all communities. The limitation is that brain imaging cannot definitively diagnose dementia types; an MRI showing hippocampal shrinkage suggests Alzheimer’s disease, but the diagnosis is confirmed through the full clinical picture combined with cognitive testing and symptoms. Additionally, many older adults have brain imaging abnormalities (like small strokes or mild atrophy) that cause no symptoms—finding them on a scan can create unnecessary worry and does not always change treatment.
Detailed Memory and Cognitive Testing
When a primary care doctor’s screening raises concern, the next step is often a comprehensive neuropsychological evaluation, which can take 4 to 6 hours and involves detailed testing of memory (both recent and remote), attention, processing speed, language, visuospatial skills, and executive function (planning, decision-making, problem-solving). A neuropsychologist administers tests like the California Verbal Learning Test (CVLT), which involves learning and recalling a list of words under different conditions to measure memory encoding and retrieval, or the Wisconsin Card Sorting Test, which assesses reasoning and the ability to shift mental sets. For instance, if an 80-year-old man is evaluated with the CVLT and recalls only 3 out of 16 words on the first trial when most people recall 6 to 8, this suggests memory impairment.
However, if he recalls 11 words on subsequent trials (showing he can learn with repetition), this pattern is different from someone with Alzheimer’s disease, who typically shows minimal improvement with repetition. The pattern of test results helps the neuropsychologist identify which brain regions are affected and narrow down the type of dementia. One real challenge is distinguishing normal aging from pathological decline and depression from dementia—older adults often score lower on some cognitive tests due to slower processing speed or reduced attention span with age alone, rather than disease. Neuropsychologists use age- and education-adjusted scores and compare results to published norms to make this distinction, but there is still gray area in mild cases.
Blood Tests and Emerging Biomarkers
Blood tests cannot currently diagnose dementia, but they are essential to rule out treatable causes of cognitive decline. Doctors routinely check vitamin B12 and folate levels (deficiency causes memory problems), thyroid function (hypothyroidism can mimic cognitive decline), and blood glucose control (diabetes increases dementia risk). Other blood work includes checking liver and kidney function (to ensure the brain is not being affected by organ disease) and screening for infections like syphilis or HIV, which can cause cognitive symptoms. In recent years, blood-based biomarkers have emerged that detect amyloid, tau, and other proteins associated with Alzheimer’s disease and can indicate who has brain pathology years before symptoms appear.
Tests like plasma phospho-tau and blood amyloid-beta ratios are becoming available at specialized centers and can help identify people with preclinical Alzheimer’s disease, though these tests are not yet standard in routine practice and are expensive. A person with memory complaints might have a blood biomarker that is positive for Alzheimer’s pathology, yet still have a normal cognitive test if the disease is very early—this creates a dilemma about whether to treat or monitor. The tradeoff with biomarkers is that knowing you have brain pathology before symptoms can guide prevention efforts (lifestyle changes, cognitive training) and enrollment in research trials, but currently no proven disease-modifying treatments exist for asymptomatic people. Additionally, some people with Alzheimer’s pathology on imaging never develop symptoms during their lifetime, suggesting other factors protect the brain or that pathology alone is not sufficient for dementia to emerge.
Distinguishing Types of Dementia
Once cognitive impairment is confirmed, the next challenge is determining the specific type of dementia, because treatment and prognosis differ. Alzheimer’s disease typically starts with memory loss; vascular dementia often presents with difficulty concentrating or slowed thinking; and frontotemporal dementia often starts with personality changes or inappropriate behavior before memory is affected. A 72-year-old woman whose first symptom was becoming uncharacteristically rude and impulsive suggests frontotemporal dementia, whereas gradual forgetting of recent events over months suggests Alzheimer’s disease. The diagnostic criteria for different dementias also consider the pattern on cognitive testing, the areas of brain shrinkage on MRI, and sometimes genetic testing—for example, genetic mutations in presenilin-1 or amyloid precursor protein cause early-onset familial Alzheimer’s disease and are identified through DNA analysis.
However, the types of dementia often overlap; a person can have both Alzheimer’s pathology and vascular changes, making the picture clinically complex. One significant warning is that misdiagnosis occurs, even in specialty settings. Some cases initially diagnosed as Alzheimer’s disease are later found to be primary age-related tauopathy (a tau-based disease) or other conditions. Furthermore, no blood test or imaging study can definitively distinguish all dementia types, so the diagnosis is informed by the whole clinical picture, and some people remain in a diagnostic gray zone until brain autopsy (which is the gold standard but obviously only possible posthumously).
The Role of Informant History
A crucial part of the diagnostic evaluation is detailed history from someone who knows the person well—a spouse, adult child, or close friend who can describe when cognitive changes began and how they have progressed. Doctors ask whether the person has gotten lost in familiar places, whether they’ve had difficulty managing finances or medications, or whether there have been personality or mood changes. This informant history is more reliable than patient self-report, because people in early dementia often lack insight into their own changes and may minimize or deny symptoms.
For example, a 76-year-old man might say his memory is “fine” when his wife reports that he asks the same question five times in an hour and recently got lost driving to a place he’s visited hundreds of times. The wife’s detailed account carries more weight in the diagnostic process. The timeline provided by an informant is also critical; if cognitive decline occurred suddenly over weeks, it suggests a stroke or other acute cause; if it unfolded gradually over years, it suggests a neurodegenerative disease.
Lumbar Puncture and Cerebrospinal Fluid Analysis
In some cases, particularly when dementia begins unusually early (before age 65) or the diagnosis is unclear after standard testing, doctors may recommend a lumbar puncture (spinal tap) to analyze cerebrospinal fluid (CSF) that surrounds the brain and spinal cord. CSF can be tested for amyloid-beta, tau, and phosphorylated tau levels; a low amyloid-beta combined with elevated tau in the CSF is consistent with Alzheimer’s disease.
A 58-year-old woman with rapidly progressive cognitive decline might have CSF analysis that reveals patterns typical of Creutzfeldt-Jakob disease (a rare but serious rapidly progressive dementia), which changes how her doctor approaches treatment and family counseling. Lumbar puncture is generally safe but carries small risks of headache, infection, and spinal cord trauma; it is therefore reserved for cases where the information would meaningfully change diagnosis or treatment. Most people with suspected dementia do not need a lumbar puncture—standard cognitive testing, imaging, and blood work provide sufficient information—but for atypical presentations or young patients, CSF analysis can be clarifying and occasionally identifies a reversible condition.
Frequently Asked Questions
How long does it take to get a dementia diagnosis?
A basic evaluation in primary care may take one or two office visits over several weeks. A full neuropsychological evaluation can take weeks to months, especially if imaging and specialist appointments are needed. Some diagnostic uncertainty persists even after all standard testing, and a diagnosis may be refined as symptoms progress over time.
Can dementia be confused with normal aging or depression?
Yes, frequently. Normal aging involves slower processing and occasional forgetfulness, but not progressive functional decline. Depression can cause poor concentration and memory complaints that mimic dementia; doctors screen for depression as part of the evaluation. Sometimes depression and dementia occur together, complicating the picture.
Are there any risks to the tests doctors use to diagnose dementia?
Cognitive testing and standard office exams carry no medical risk. MRI is safe but contraindicated if someone has metal implants; CT uses radiation; and lumbar puncture carries a small risk of headache or infection. Blood tests involve only a needle stick. The benefits of accurate diagnosis typically outweigh these minor risks.
Can dementia be prevented if caught early?
Current evidence suggests that managing cardiovascular risk factors (blood pressure, diabetes, cholesterol), staying cognitively and socially active, exercising, and eating a heart-healthy diet may reduce dementia risk or slow progression. However, no proven medication prevents dementia in people without symptoms, even those with abnormal biomarkers.
Should I get tested if I’m worried about my memory?
If you or someone close to you notices persistent changes in memory, thinking, or ability to manage daily tasks, a conversation with a primary care doctor is reasonable. Most memory concerns turn out to be normal aging or treatable causes (like low B12), but early evaluation of actual cognitive decline may provide useful information and access to supportive resources.
