What Is a Functional MRI (fMRI) and What Is It Used for?

The short answer: A functional MRI (fMRI) tracks blood flow changes while you do simple tasks, showing which brain areas handle speech, movement, and vision. Its main clinical use is planning brain surgery, per the ACR-ASNR-SPR practice parameter (amended 2023). It cannot diagnose a psychiatric condition in one person.

How does a functional MRI work?

A standard brain MRI shows structure. A functional MRI adds a picture of what the brain is doing. It uses MR imaging to measure the tiny changes in blood flow that happen when an area works harder. Active areas consume more oxygen, and the common method, called BOLD (blood oxygenation level dependent), picks up that change.

This is an indirect signal. The scan measures nerve cell activity only indirectly. Software then compares your brain during a task with your brain at rest and overlays a color map on your anatomy. Like other MRI, fMRI uses no radiation.

What is fMRI used for in patient care?

The ACR practice parameter lists these primary indications:

• Planning surgery or biopsy for a brain tumor, vascular malformation, or other focal lesion
• Finding critical language, motor, sensory, or vision areas near that lesion
• Finding critical areas and language side before epilepsy surgery
• Planning radiation treatment and checking those areas afterward
• Testing how brain blood vessels respond, before a revascularization procedure

Surgeons want to know where function sits before they operate, because mapping cannot be generalized from one brain to another. In epilepsy, a 2017 American Academy of Neurology guideline says fMRI may be considered in place of the invasive Wada test for finding the language side in several epilepsy types. Most of those recommendations carry a modest Level C rating.

The FDA wording needs care. RadiologyInfo says the FDA has approved fMRI only for surgical planning. The FDA record we read is a device clearance: a 2012 510(k) summary for Brainlab iPlan software lists BOLD MRI mapping as a planning module.

What is fMRI used for in research, and what can it not do for one person?

Researchers use fMRI to study how normal, diseased, or injured brains work. Much of this work looks for patterns across many people. A pattern across a group does not automatically apply to one person, and the limits are well documented.

• Psychiatric diagnosis: A 2018 American Psychiatric Association review found that, outside neurocognitive disorders, neuroimaging has yet to have a significant impact on diagnosing or treating individual patients. It applied a suggested standard of at least 80% sensitivity and 80% specificity, and found the literature does not support a diagnostic imaging biomarker. A 2022 Nature analysis of about 50,000 people found brain-behavior links smaller than expected, needing thousands of participants to replicate.
• Repeat reliability: A 2020 meta-analysis of 90 experiments found poor test-retest reliability for common task measures (mean ICC .397). It addressed biomarker and individual-differences research, not surgical maps.
• Pain: A 2017 International Association for the Study of Pain task force concluded that using brain imaging to support or dispute a chronic pain claim "is not warranted."
• Lie detection: It has been marketed for personnel screening and offered in court. In United States v. Semrau (2012), the Sixth Circuit upheld excluding it, noting the method had not been fully examined in real-world settings.

If a clinic offers fMRI for depression, ADHD, or pain, ask for published sensitivity and specificity from independent groups.

What happens at a clinical fMRI appointment, step by step?

Protocols differ by center, so treat this as a typical path. It is built from the ACR parameter and patient guides from Memorial Sloan Kettering and Columbia. For general scanner basics, see our guide to your first MRI.

1. Task planning. Your doctor sends the clinical question, and the team reviews prior brain imaging to choose tasks.
2. Safety screening. Get the exact name and manufacturer of any implant. Without them, you may not be scanned that day. Ask about skin patches, glucose monitors, and insulin pumps.
3. The day of the scan. Try to be awake and alert, and take regular medicines unless told otherwise. Wear contact lenses if you use them, so you can see the tasks.
4. Caffeine. Advice conflicts here. MSK and Columbia say to avoid it that day. RadiologyInfo says to keep your usual habit. Follow your center.
5. Task training. Staff train you just before the scan and check reading level, handedness, senses, and focus. A mock scanner may help children.
6. In the scanner. You lie on your back with arms at your sides, wearing earbuds and goggles. You can speak with the team throughout.
7. The tasks. Expect things like finger-to-thumb movement or hand squeezes or thinking of words that start with a certain letter. In one ACR example, 6 task blocks alternate with 6 rest blocks, each 20 to 30 seconds. That works out to about 4 to 6 minutes per task set. If a task is too hard, tell the technologist.
8. Possible rest scan. Some exams add resting-state imaging, about 6 to 10 minutes of lying quietly.
9. Possible standard MRI. fMRI does not need contrast, but an added MRI may use it.
10. Afterward. You can return to normal activities. A specialist reads the images and sends a report to your doctor.

Timing varies. Columbia says about an hour, and Cleveland Clinic says up to an hour. MSK says about 2 hours, which includes the MRI after the fMRI.

Holding still matters more than usual, because fMRI is especially sensitive to head motion. Our article on moving during an MRI explains why. Anxiety can make stillness harder, and anxiety.md covers anxiety more broadly. Tell your doctor early, because the ACR notes that sedation may impair your ability to do the tasks.

How accurate is fMRI for surgical planning?

It is useful, but it does not replace every other test. A 2018 review in NeuroImage: Clinical reported these points:

• Agreement with direct brain stimulation during surgery ranges from 67% to 100%, depending on the patient and task.
• Language maps tend to agree less well than sensory, motor, and vision maps.
• Stimulation and the Wada test remained the clinical gold standards at the time of the review.
• fMRI shows areas involved in a task, but cannot separate essential areas from participating ones.
• Tumors can change local blood flow. That can hide a working area, a false negative that matters most before surgery.

RadiologyInfo notes that doctors may recommend more tests to confirm results when a critical decision depends on them. A separate question is what your scan shows about structure, and our brain MRI findings article covers common report terms.

What should you ask before a functional MRI?

This checklist comes from the sources above:

• What question will this scan answer, and how will it change my plan?
• Which tasks will I do, and can I practice them first?
• Should I have caffeine that day?
• Will I also have a standard MRI or contrast, and how long will it all take?
• If I cannot do the tasks, is a resting-state scan or a separate session possible?
• Will I still need a Wada test or brain stimulation mapping?
• Will the report state which tasks I did and the data quality?
• Is prior approval needed? Billing often uses CPT codes 70554, 70555, and 96020.

Coverage varies by plan. NC Medicaid covers fMRI effective Sept. 1, 2023. One Pennsylvania health plan policy mirrors the ACR list and says other uses are not supported by current evidence.

The bottom line

A functional MRI maps where your brain handles key tasks, and its accepted clinical role is planning treatment near critical areas. Expect short simple tasks and about an hour or more at the center. Research on psychiatric illness, pain, and lying is real, but it does not yet give a reliable answer for one person.

Last updated: October 2026. This article is for informational purposes only and does not constitute medical advice. Ask your neurosurgeon, neurologist, or imaging center how fMRI applies to you.