The single most helpful test is magnetic resonance imaging with diffusion-weighted imaging (MRI-DWI), because it shows acute infarction within minutes of onset and separates true stroke from its mimics at a level CT cannot match. The boundary matters: DWI's advantage depends on when the scan is done and on pretest probability, so the test is most decisive when a patient with sudden focal deficits is scanned promptly and the clinical picture is genuinely ambiguous. This article is general educational information only and is not a diagnosis or treatment plan.

Why DWI Outperforms Non-Contrast CT When the Clock Is Still Young

Non-contrast CT is the first scan in most stroke pathways because it rules out hemorrhage fast, but it misses early ischemia at a rate that makes it a weak discriminator for mimics. According to the American Heart Association/American Stroke Association 2019 guideline update for the early management of acute ischemic stroke, CT sensitivity for acute ischemia in the first hours is low, and small or brainstem infarcts are frequently invisible. That matters here because mimics and small strokes look identical on examination.

DWI changes that picture. Restricted diffusion appears as hyperintensity on DWI with corresponding hypointensity on the apparent diffusion coefficient map, and it develops within minutes of cytotoxic edema. A 2021 consensus statement from the American College of Radiology Appropriateness Criteria rated MRI without and with contrast as the most appropriate imaging study for acute stroke when MRI is available, with CT remaining appropriate when speed or availability dominates.

The practical consequence is a rule you can apply at the scanner: a negative DWI in a patient with a convincing syndrome should push you toward a mimic, while a positive DWI in a low-probability patient should push you to revisit the examination. Neither scan substitutes for the history. Roughly a third of patients referred to stroke centers with a suspected stroke turn out to have something else, a proportion reported repeatedly in stroke-mimic cohorts and consistent with the 2019 AHA/ASA emphasis on accurate triage before thrombolysis decisions.

Cost and time enter the decision honestly. A meta-analysis by the Cochrane Stroke Group in 2013 comparing MRI and CT for acute stroke concluded MRI was more sensitive but added roughly 15 to 25 minutes of door-to-imaging time in typical workflows, a figure that varies by whether the scanner sits in the emergency department. That delay is the price of the better discriminator, and it's why many centers run CT first for thrombolysis candidates and MRI immediately after, or use MRI directly when the diagnosis is in question.

The Timing Rule That Actually Decides Whether MRI Helps

Timing is the variable that most changes the answer, and it's the one the top-ranking pages usually compress into "MRI is better." Start with the accepted window: DWI remains positive for days to weeks after infarction, so it retains value well past the thrombolysis window, unlike perfusion imaging, which reflects current hemodynamics. If the patient woke with symptoms or cannot give a last-known-well time, DWI is the test that still tells you whether an infarct exists and roughly how old it is.

The reverse failure case is real. In very small brainstem or lacunar infarcts, DWI can be falsely negative in the first 24 hours, particularly with slices thicker than 3 mm or motion artifact. A 2009 study in Stroke by Brazzelli and colleagues found DWI sensitivity above 90% for supratentorial infarcts but noticeably lower for infratentorial ones. If suspicion stays high after a negative scan, repeat imaging after 24 hours rather than accepting the negative.

Scan after treatment changes the reading too. Once thrombolysis or thrombectomy succeeds, DWI may show only a small residual lesion while the examination has already improved, so the scan reflects tissue fate, not the presenting deficit. Order the test to answer the question you actually have: is there an infarct, how large, and where. Don't order it hoping to explain a deficit that has already resolved.

A workable timing sequence looks like this: CT within the first 20 minutes to exclude hemorrhage; MRI-DWI as soon as the patient is stable if the syndrome is atypical, if the deficit is small, or if a mimic is on the list; and repeat DWI at 24 to 72 hours when the first scan is negative but the story won't let go. Run this sequence in that order rather than scanning twice in the first hour.

First Steps at the Bedside Before You Order Any Imaging

The most useful diagnostic step in a mimic workup costs nothing and takes about 60 seconds: a capillary glucose. Hypoglycemia is the classic stroke mimic, and the American Diabetes Association Standards of Care in Diabetes (2024) set the level requiring immediate treatment at below 70 mg/dL (3.9 mmol/L) for a person with diabetes, with neuroglycopenic symptoms possible at that threshold. Treat and re-examine before you image.

Then work the structured list. The most common mimics in cohort studies are seizure with postictal Todd paresis, migraine with aura, hypoglycemia, toxic-metabolic encephalopathy, functional or conversion disorder, Bell palsy, and peripheral vertigo. Each has a discriminating bedside feature, and the features do more work than any single scan.

1: Draw a point-of-care glucose and correct anything below 70 mg/dL, then reassess the deficit. 2: Ask about the first symptom and its spread: a march from hand to face suggests seizure or migraine, not a vascular territory. 3: Examine for non-vascular patterns such as isolated global aphasia with no weakness, or deficits that cross vascular boundaries. 4: Check medications and tox screens for sedatives, anticholinergics, and sympathomimetics. 5: Verify the NIH Stroke Scale score yourself rather than accepting a field score, because Todd paresis and functional deficits both distort it. 6: Only then order imaging, with the sequence set by what the bedside exam left uncertain.

Step 4 deserves emphasis because it's where mimics hide. A patient on lithium with a level of 2.5 mmol/L presents with dysarthria, ataxia, and confusion that reads as posterior circulation stroke until the level returns. Similarly, a first seizure in an older adult can produce a dense hemiparesis lasting hours.

If the glucose is normal and the story is vascular, proceed to imaging without delay. If the story is clearly non-vascular, image anyway when the patient is old or has vascular risk factors, since the two conditions coexist more often than the mimic literature implies. The bedside exam narrows the differential; it does not close it.

Building the Mimic Differential in Vascular Territory Order

Organizing mimics by the vascular syndrome they imitate keeps the workup focused and prevents the shotgun panel that delays imaging. Use the table below as a working frame; the "best discriminator" column names the test or finding that separates the mimic from infarct most reliably.

Vascular syndrome Most faithful mimic Best discriminator What MRI-DWI shows
Middle cerebral artery Postictal Todd paresis Witnessed seizure, postictal confusion clearing over hours Negative in uncomplicated Todd paresis
Posterior circulation Peripheral vertigo, vestibular neuritis HINTS exam, direction-changing nystagmus Small brainstem infarct, sometimes falsely negative early
Lacunar Hypoglycemia, functional disorder Glucose, inconsistency on distraction Small deep infarct or normal
Global aphasia without hemiparesis Alzheimer disease progression, delirium Onset speed, fluctuating course Cortical infarct in two territories
Bilateral signs Toxic-metabolic encephalopathy Drug levels, ammonia, renal function Usually normal or old small-vessel disease

The middle cerebral artery row is the one clinicians get wrong most often. Todd paresis after a focal seizure can persist for hours and resolves without thrombolysis, so a witnessed seizure or a postictal phase with gradual clearing should pause the thrombolysis decision long enough for MRI. The posterior circulation row is the other trap: HINTS testing (head impulse, nystagmus, test of skew) outperforms early MRI for distinguishing central from peripheral vertigo in the hands of a trained examiner, a finding established in the 2009 Stroke study by Kattah and colleagues, while DWI can be negative in the first 48 hours for small brainstem lesions.

Use the table to ask a single question per row: does the evidence in the discriminator column fit? If the answer is no, the mimic is less likely and the imaging threshold drops. Don't order the full mimic panel for every patient; order the test the table points to.

One structural caution: mimics cluster by age. In patients under 50, migraine, seizure, and functional disorder dominate; over 70, hypoglycemia, toxic-metabolic states, and dementia-related presentations dominate. Adjust the differential to age before you adjust the imaging order.

CT Angiography, Perfusion, and Lumbar Puncture: When the Answer Isn't DWI

Sometimes the helpful test isn't the infarct scan at all. CT angiography of the head and neck answers a different question — is there a large vessel occlusion or dissection — and in the 2015 endovascular trials summarized in the New England Journal of Medicine, CTA defined the treatment-eligible population as much as the clinical exam did. Order CTA when the syndrome suggests large vessel occlusion even if the DWI is equivocal, because thrombectomy eligibility turns on vessel status.

CT perfusion enters when the question is salvageable tissue rather than whether an infarct exists. The DAWN and DEFUSE 3 trials, published in the New England Journal of Medicine in 2018, used perfusion thresholds to select patients with wake-up or late-presenting stroke for thrombectomy up to 24 hours from last known well. If your patient with a mimic differential has a proximal occlusion on CTA, perfusion decides what to do; if the CTA is clean, perfusion adds little.

Lumbar puncture is the test that MRI cannot replace for subarachnoid hemorrhage when CT is negative and suspicion persists. The standard rule from the 2012 AHA/ASA scientific statement on subarachnoid hemorrhage management holds that CT within six hours by a modern scanner has very high sensitivity, and beyond that a negative CT should be followed by lumbar puncture looking for xanthochromia. In a patient whose "stroke" is actually a thunderclap headache with neck stiffness, the LP is the answer and MRI may only show subtle sulcal signal.

Electroencephalography belongs in the list when the deficit fluctuates or consciousness is impaired without a structural lesion. Non-convulsive status epilepticus can present as a persistent aphasia or hemineglect, and a 20-minute EEG changes management in a meaningful share of such patients, as reported in ICU EEG series in Neurology. Order it after imaging, not instead of it.

The Stroke Mimic Score and Other Decision Aids: What They Add

Clinical decision instruments exist, and their value is in flagging when to test, not in replacing the test. The most cited is the Telestroke Mimic Score, published by Ali and colleagues in Stroke in 2018 and derived to predict non-stroke diagnoses in a telestroke cohort. Its components — absence of facial droop, negative visual field cut, and a history of seizure or migraine — are bedside features you already collect; the score just forces you to weigh them before thrombolysis.

A second instrument, the FABS score, uses six items in a similar role and was validated in emergency department cohorts. Neither reaches the discrimination of DWI, and neither is a stand-alone exclusion rule: a score favoring mimic does not substitute for imaging in a patient with ongoing deficits, and a score favoring stroke does not justify thrombolysis without hemorrhage exclusion.

The practical use is triage. If the mimic score is high and the deficit resolves or is inconsistent, MRI replaces the thrombolysis clock as the priority. If the score is low and the deficit is fixed, CT and treatment proceed in parallel with MRI. That sequencing keeps the decision aid where it belongs, upstream of the scan.

Don't over-read the scores in patients with prior stroke. A patient with old deficits and a new event scores oddly on several instruments because the items assume a first presentation. In that case, clinical judgment plus DWI outperforms any score you can plug in.

A Worked Sequence for an Ambiguous Hemiparesis

Consider a 58-year-old man brought in at hour two with right face and arm weakness and mild expressive aphasia. He has hypertension, and his wife says he "looked odd" starting after a brief episode of jerking in the right arm. On arrival the weakness is fixed, glucose is 112 mg/dL, and the field NIH Stroke Scale score is 4.

The most helpful test here is MRI-DWI, not because CT is unavailable but because the witnessed jerking raises the possibility of postictal Todd paresis, and the small deficit makes early CT even less sensitive. Get the CT to exclude hemorrhage, then get the DWI. If DWI shows a small cortical infarct in the middle cerebral artery territory, the seizure was the presentation of the stroke and he is a thrombolysis candidate within the window. If DWI is negative, the story favors Todd paresis, and a repeat DWI at 24 hours plus an EEG is the sequence.

Compute the decision honestly: the MRI adds 15 to 25 minutes based on the Cochrane comparison, and that delay matters only if it pushes past the thrombolysis window. At hour two, it doesn't. At hour four and a half, every minute counts, so the CT-to-needle path should run without waiting for the MRI.

Now change one fact: the same deficit after a glucose of 48 mg/dL. Correct it and re-examine. If the deficit clears, the mimic diagnosis is made without any imaging, and the DWI is no longer the most helpful test because the question is already answered. The sequence is not fixed; the test that helps most is the one that resolves the remaining uncertainty.

Where DWI Fails and What to Do Instead

The first failure case is the hyperacute tiny infarct. DWI can be negative within the first hours for very small brainstem lesions, and the correct response is repeat imaging, not reassurance. A normal scan plus a convincing syndrome is a provisional answer.

The second is the artifact problem: susceptibility artifact near the skull base, motion in a delirious patient, and poor ADC maps can create false positives. Read DWI with the ADC map every time; a hyperintensity without ADC hypointensity is T2 shine-through, not restricted diffusion, and is a common overcall in white matter disease.

The third failure is diagnostic momentum. A positive DWI in a patient with an old infarct can be read as the cause of new symptoms that are actually functional or metabolic; in the 2018 Stroke mimic score derivation, a substantial share of mimic patients had prior stroke on imaging. Always ask whether the lesion maps to the deficit, not just whether a lesion exists.

The fourth is cost and access in under-resourced settings, where MRI is unavailable after hours and CT is the only option. In that situation, CT plus CTA plus clinical probability is the best available test, and transfer decisions should be made on that basis rather than waiting for MRI. Name the limitation plainly: the answer depends on what the local system can deliver within the window.

FAQ

What is the single best test for distinguishing stroke from a mimic?

MRI with diffusion-weighted imaging is the best single test, and its sensitivity for acute infarction exceeds 90% for supratentorial lesions while CT sensitivity in the first hours is substantially lower, per the 2019 AHA/ASA guideline review. The catch is that DWI's edge shrinks for small brainstem infarcts and vanishes if the scan is delayed past the point where the question has already been answered clinically.

Can a normal MRI rule out a stroke?

No. A normal DWI within the first 24 hours does not exclude a small brainstem or lacunar infarct, and Brazzelli's 2009 Stroke analysis found sensitivity in the infratentorial compartment meaningfully below the supratentorial figure. Repeat imaging after 24 hours and treat the clinical syndrome as the primary evidence when the first scan is negative and suspicion remains high.

How much time does MRI add compared with CT in an acute stroke pathway?

The Cochrane Stroke Group's 2013 meta-analysis put the added door-to-imaging time at roughly 15 to 25 minutes in typical workflows, with wide variation depending on whether the scanner is in the emergency department. That delay is acceptable when the patient is outside the thrombolysis window or the diagnosis is genuinely ambiguous, and it is not acceptable when it pushes a treatable patient past the window.

Is CT angiography more useful than MRI for suspected stroke?

They answer different questions, and the 2015 endovascular trial publications in the New England Journal of Medicine used CTA to define large vessel occlusion before treatment. CTA is the better test for vessel status and dissection; DWI is the better test for whether an infarct exists and how old it is. Most comprehensive centers order both.

What is the most common stroke mimic?

Seizure with postictal Todd paresis, hypoglycemia, and migraine with aura are the most frequently reported mimics in cohort studies, with functional and toxic-metabolic presentations close behind. The order shifts with age: seizure and migraine dominate under 50, while hypoglycemia and toxic-metabolic states dominate in older patients.

Do mimic scores replace imaging?

No. The Telestroke Mimic Score and the FABS score are triage aids derived in Stroke and emergency department cohorts, and they flag when a mimic is likely enough to prioritize MRI over the thrombolysis clock. A score favoring mimic does not exclude stroke, and a score favoring stroke does not exclude hemorrhage; imaging remains mandatory in patients with fixed deficits.

When is lumbar puncture the helpful test instead of MRI?

Lumbar puncture is the answer when subarachnoid hemorrhage is suspected and CT is negative beyond six hours from headache onset, following the 2012 AHA/ASA statement on subarachnoid hemorrhage management. Xanthochromia on the sample is the finding that changes management, and MRI may only show subtle sulcal signal in that window.

Does DWI stay positive after the stroke window closes?

Yes. Restricted diffusion persists for days to weeks after infarction, which is why DWI remains useful in patients who wake with symptoms or cannot give a last-known-well time. That persistence makes it the right test for dating an infarct and confirming it exists, even when perfusion imaging no longer shows salvageable tissue. For immediate danger or a medical emergency, contact local emergency services rather than relying on this information.

This article is general educational information and is not a diagnosis, treatment plan, or substitute for clinical judgment. Imaging and treatment decisions for suspected stroke depend on the individual patient, local protocols, and the resources available; consult a qualified local physician or stroke specialist. If you or someone else has sudden facial droop, arm weakness, or speech difficulty, call local emergency services immediately.