Written and medically reviewed by Tae Ho Guak, MD, board-certified neurologist · Reviewed July 2026 · Editorial policy

A normal MRI is not a diagnosis. It is the absence of one particular kind of problem — a tumor, an aneurysm, a bleed, a mass. Migraine, cluster headache, tension-type headache, medication-overuse headache: none of them show up on an MRI. They never have. So if you are holding a report that says “no acute intracranial abnormality” and you still have a headache four days a week, the scan did its job. It just didn’t answer your question.
I’ve had this conversation several thousand times twenty-five years. A patient sits down, slides the disc or the printed report across the desk, and says some version of: “They told me it’s clean, so I don’t know why I’m here.” Almost always they are relieved and quietly humiliated at the same time — relieved there’s no tumor, humiliated because “clean scan” got translated somewhere along the way into “it’s stress” or “it’s in your head.” Both of those readings are wrong.
What an MRI is actually looking for
Brain imaging is a structural test. It shows anatomy — tissue, blood, fluid, bone. It is very good at finding things that take up space or don’t belong: tumors, subdural collections, hydrocephalus, large strokes, malformations. It is also reasonably good, with the right sequences, at showing things like sinus vein thrombosis or the sagging brain and dural enhancement of spontaneous intracranial hypotension — though those require someone to specifically ask for them.
What it cannot show is function. Migraine is a disorder of how the trigeminovascular system and the brainstem behave, not of what the brain is made of. There is nothing to photograph. The same is true of cluster headache, of tension-type headache, of trigeminal neuralgia in most cases, and of medication-overuse headache. These are diagnosed by history — by the pattern, the timing, the associated symptoms, the response to treatment. The International Classification of Headache Disorders (ICHD-3), which is the standard the entire field uses, defines them almost entirely on clinical criteria. Imaging appears in those criteria mainly as a way to exclude something else.
So the honest translation of your report is: “We looked for the dangerous structural causes, and we didn’t find any.” That is genuinely good news. It is not the end of the workup.
Why the scan often gets ordered in the first place
In an emergency department or a busy primary care clinic, the first job is to not miss the catastrophe. That’s correct triage. A scan is fast, objective, and defensible. Taking a proper headache history — how the pain started, whether it throbs, whether light bothers you, how many days a month you reach for a painkiller, what you took last week — takes twenty minutes and a specific set of questions. Under time pressure, imaging often substitutes for that conversation.
The result is a common and frustrating pattern: the patient gets the expensive test and not the cheap one that would have actually named the problem. I’ve written more about how this happens in Headache Not Getting Better? What Doctors Miss — the short version is that most “undiagnosable” headaches are simply undiagnosed, not undiagnosable.
The three things I look for after a normal scan
First: how many days a month are you taking something for pain? This is the single highest-yield question in headache medicine and almost nobody asks it. If the answer is ten or more days a month for triptans, combination analgesics, or opioids — or fifteen or more for plain NSAIDs or acetaminophen — you may have layered medication-overuse headache on top of whatever you started with. The scan will be normal. The headache will be daily. And it will not improve until the overuse is addressed. I cover the mechanics of this in Medication-Overuse Headache: Painkillers as Cause.
Second: does the headache actually fit migraine, even if nobody has called it that? Migraine is chronically under-diagnosed because people picture it as a one-sided, vomiting, dark-room event. It often isn’t. Bilateral pressure, neck pain, nausea you’ve stopped noticing, a mild dislike of bright screens, worse with movement — that is migraine, and it is routinely labeled “tension headache” or “sinus headache” instead. If the label is wrong, the treatment will be wrong, and the headache will look treatment-resistant when it is merely mistreated.
Third: is there a positional, temporal, or trigger pattern the scan was never designed to capture? Headache that is clearly worse upright and better lying flat points toward a CSF leak — and a routine MRI without contrast can easily miss it. Headache that wakes you at the same hour nightly, headache that appears only with exertion, headache that started on one identifiable day and never stopped — each of these is a specific diagnosis with a specific workup. A normal scan doesn’t rule any of them out. It just means the next step is a better history, not a bigger machine.
When a normal scan should still worry you
Imaging is reassuring, but it is not permanent reassurance, and it doesn’t override certain symptoms. Get re-evaluated — promptly, and in person — if any of the following are true, regardless of what the last scan said: the headache changed character or became abruptly severe; you have new neurological symptoms such as weakness, numbness, double vision, or difficulty speaking; you’re over 50 with a genuinely new headache; you have fever with neck stiffness; you have persistent vomiting; you have a history of cancer or immune suppression; or your vision is dimming or greying out in episodes.
That last one matters more than people realize. Idiopathic intracranial hypertension can produce an MRI a general radiologist calls normal while pressure is quietly damaging the optic nerve. It needs an eye exam and often a lumbar puncture — not another scan. The full red-flag list is here: Headache Red Flags: When to Go to the ER.
What to do with the report you’re holding
Keep it. It has real value — it means a future doctor doesn’t need to re-image you, and it means the conversation can move on to the part that actually leads to relief. Then bring three things to your next appointment: a four-week record of headache days, a complete list of every medication you took on those days including over-the-counter ones, and a plain description of what a bad day actually feels like from the first hour to the last.
In my clinic, that packet is worth more than the MRI. It’s usually what tells me the diagnosis. Repeating the scan almost never does.
Next steps
- Free red-flag checklist — the warning signs that mean imaging or an ER visit is genuinely warranted.
- Talk to a specialist — if your scan is normal and your headache isn’t, a proper headache history is the missing step.
Written by Dr. Tae-Ho Guak, MD — board-certified neurologist, 25 years and over 100,000 patients seen. This article is for education and does not replace an in-person medical evaluation. If you have a sudden, worst-ever headache, new neurological symptoms, fever with neck stiffness, or vision loss, go to an emergency department immediately.
From the textbook. Chapter 1, “The Normal Scan: Where Headache Medicine Begins,” of my clinical textbook Headache Medicine covers this topic in full for doctors, with the evidence and decision pathways behind it. About the book →
