Can MRI Detect Amyloid Plaques in the Brain?

A standard brain MRI can’t show amyloid plaques, the protein clumps linked to Alzheimer’s disease. What it can show is the damage and other changes that tend to come with them, which is why it’s still one of the first scans in a memory workup.¹ Other tests show the amyloid itself, while research scanners get far closer to seeing plaques than the MRI at your local hospital.

Why a Regular MRI Misses Plaques

MRI builds its pictures from how tissues respond to a strong magnet. The amyloid protein in plaques gives off almost no magnetic signal of its own.² When plaques do show up on research scans, it’s mostly because of tiny iron deposits sitting alongside them.³ Even then, in brain tissue only the dense, fiber-like kind of amyloid changed the MRI signal, and looser, diffuse plaques left no trace.⁴

A routine scan is tuned to tell tissues apart by their water content and structure, which is why it shows shrinkage, strokes and tumors so well. Plaques are a tiny amount of protein embedded in tissue that is otherwise very similar, so they create no contrast against the brain around them.

Size is a second barrier. Individual plaques measure in fractions of a millimeter, far below the detail a routine hospital scan resolves. So even where iron is present, a routine scan can’t pick out single plaques.

That shapes what a radiologist’s report on a memory MRI will say. It won’t mention amyloid or plaques, and a normal report doesn’t rule amyloid out. Words like “atrophy” or “volume loss” describe shrinkage, not plaques. If you want to know whether amyloid is present, ask the doctor which test would actually check for it.

Plaques can also build up for years before memory symptoms start, and a scan taken in that early stage can look entirely unremarkable. A clear MRI and a positive amyloid test are therefore not contradictory.

What an MRI Does Show in a Memory Workup

People with more amyloid tend to lose brain volume faster in memory and thinking areas.⁵ People with high amyloid, whatever their symptoms, often have a smaller basal forebrain, a small region involved in memory and attention.⁶ But shrinkage has many causes, so it hints at amyloid without proving it.

In Alzheimer’s-type changes, shrinkage often shows first in memory-related regions such as the hippocampus, which is why radiologists look at these areas closely. Some shrinkage is expected with ordinary aging, so a report often compares the findings with what is typical for your age. Vascular disease, other dementias, heavy alcohol use and past head injury can all shrink the brain too. Shrinkage supports a diagnosis only alongside symptoms and other tests.

Amyloid can also build up in the walls of brain blood vessels. Those vessels can leak, leaving microbleeds, tiny spots of old bleeding.⁷ Special MRI settings that are sensitive to iron in blood are good at finding them.⁸

On those settings, microbleeds appear as small dark dots. They may not be visible at all on a standard sequence, so a scan without those settings can miss them. A report listing tiny bleeds near the brain’s surface can point toward amyloid in vessel walls, while bleeds deep in the brain more often relate to long-standing high blood pressure. That location is a clue, not proof, and a doctor weighs it with the rest of the picture.

Doctors use a clinical evaluation plus MRI first to set the context, before blood or PET tests.⁹ The MRI can reveal strokes, tumors or fluid buildup that can also cause memory problems. A prior stroke, a slow-growing tumor, or fluid pressing on the brain can each explain symptoms, and some are treatable. Finding one changes the plan, but it doesn’t necessarily rule out Alzheimer’s, because more than one condition can be present in the same person.

A clear MRI still matters. It can exclude treatable causes, it gives a baseline to compare against later scans, and it is needed before some Alzheimer’s drugs.

Tests That Do Show Amyloid

An amyloid PET scan uses a small injected radioactive tracer that sticks to plaques and lights them up. An MRI shows the structure of the brain and its damage, while amyloid PET shows whether a tracer has attached to amyloid, so it answers a different question. The result is usually reported as positive or negative for amyloid rather than as a detailed description of the brain’s shape.

In telling Alzheimer’s from normal aging, PET caught about 93 in 100 cases and MRI about 85.¹⁰ In separating Alzheimer’s from mild memory problems, both did only fair, catching roughly 6 to 7 in 10.¹¹ PET is also much more expensive than MRI because of tracer production and specialized equipment.¹²

The appointment itself is fairly simple. A technician places an IV in your arm and injects a small amount of tracer, a quick step. You then wait while the tracer travels to the brain and attaches to any plaques, and the length of the wait depends on the tracer used. In the scanner you lie still with your head in a ring-shaped device, and holding still matters because movement blurs the images. The tracer is radioactive, so the clinic will give instructions for afterward, and the amount is small. Insurance coverage varies, so check first.

The preferred path is clinical evaluation and MRI first, then validated blood tests, then PET if uncertainty remains and the result could change care.¹³ A blood test looks for amyloid-related proteins and is less invasive and cheaper than PET, which is why it often comes earlier. A spinal tap is another option: a needle in the lower back collects a small sample of the fluid around the spinal cord, and the lab measures amyloid and related proteins in it. Which tests are available depends on the clinic, so ask which it offers and which it trusts, and whether a blood test is possible before a PET.

A positive amyloid result doesn’t always mean dementia. Many older adults have amyloid without symptoms, so results are read alongside memory testing. A doctor should explain what a positive result changes for you.

Why MRI Matters on Anti-Amyloid Drugs

Lecanemab, a drug that clears amyloid, can cause swelling or small bleeds in the brain that show up on MRI, which doctors call ARIA.¹⁴ In one clinic, about 1 in 5 treated patients developed ARIA.¹⁵ People with more microbleeds on their starting MRI were more likely to develop it, so the baseline scan helps decide how risky treatment is.¹⁶

The scan before the first infusion shows the starting number of tiny bleeds and any other changes. Without it, a doctor couldn’t tell whether a spot on a later scan is new or was there already. Findings at baseline can lead the team to change the plan, or to decide the risks are too high.

MRIs are required before the 5th, 7th and 14th infusions.¹⁷ Another is recommended around week 52, especially for APOE4 carriers or those with earlier ARIA.¹⁸ New symptoms like headache or confusion call for an extra scan.¹⁹ If drug-related swelling or bleeding appears, MRI is repeated monthly until it settles.²⁰ Portable low-field scanners may one day make these checks quicker.²¹

The swelling or bleeding can occur with no symptoms at all, so a person can feel fine while a scan shows a change. Between infusions, report a new headache, confusion or vision changes, or dizziness and unsteadiness. If a scan does show a change, the team looks at how extensive it is and whether symptoms are present when deciding whether to continue.

People who carry the APOE4 gene variant face a higher risk of ARIA and may need more scans, so it’s worth asking about genetic testing first. Doctors may offer the test before starting, and the result shapes the monitoring plan.

Will MRI Ever See Plaques Directly

In mice bred to grow plaques, MRI without dye has spotted plaques as small as 50 thousandths of a millimeter.²² In mouse brain tissue, plaques seen on MRI matched those confirmed under a microscope.²³ The limits are that this tissue wasn’t living and the scanning conditions used can’t be copied in a patient.

The fiber-like plaques most like human ones carry little iron and are harder to see.²⁴ The mice are also bred to grow plaques quickly, which makes them denser and easier to find than in a human brain. A plaque seen in a mouse shows the principle is possible, not that a patient scan will work the same way.

In people, 7-tesla research scanners, more than twice as strong as most hospital machines, reveal detail once seen only after death.²⁵ One found more than twice as many microbleeds as a conventional scanner.²⁶ These machines are found mostly at research centers, and a person usually reaches one only by joining a study. Better detail on bleeds and structure isn’t the same as seeing amyloid, which is still mostly inferred from iron.

Iron is both help and hurdle. Some plaques collect iron, and iron is easy for MRI to detect, so it makes those plaques stand out. Iron mapping on MRI closely tracked amyloid on PET in people with mild memory problems.²⁷ But iron also builds up in the brain with normal aging and in other conditions, so iron on a scan could have other sources. Researchers therefore compare it with a PET scan to check it.

AI reading ordinary MRIs can guess amyloid status, but is often wrong, in one case right only about 6 in 10 times.²⁸ Another AI model did better but still fell short of a direct test.²⁹ Researchers are also designing contrast dyes that latch onto amyloid, so far tested in mice.³⁰ The brain’s protective barrier keeps most dyes out, one reason these remain experimental.³¹

None of these methods is routine care, and a hospital won’t offer them for diagnosis. Research studies sometimes recruit volunteers, so it’s worth asking your memory clinic about trials.