How this instrument works
The scale examines eleven numbered categories, but three of them split further into separate sub-scores because the underlying finding genuinely differs by side or by component: level of consciousness breaks into three parts (1a alertness, 1b answering questions, 1c following commands), and motor strength is scored separately for each limb — arm and leg, left and right — because one-sided weakness is exactly the clue that points toward which side of the brain a stroke has damaged. Gaze, visual fields, facial palsy, ataxia, sensation, language, dysarthria, and extinction or inattention do not have that same natural split, so each stands as a single score. Fifteen scored components in total, summing to a range of 0 to 42.
Thomas Brott and colleagues published the scale in 1989 as a way to quantify stroke severity reproducibly at the bedside, replacing looser, less standardised descriptions with a number that could be tracked over time and compared between examiners. It has since become the standard initial severity measure in stroke care worldwide, feeding directly into time-sensitive treatment decisions — eligibility for clot-dissolving medication and for mechanical clot retrieval both weigh the NIHSS score alongside how long ago symptoms began.
None of that makes this page a substitute for the exam itself. The NIH Stroke Scale requires a trained, certified examiner physically testing a patient — checking eye movement, asking them to name objects, testing strength against resistance — and scoring what is actually observed; it cannot be filled in from a description or a guess. This instrument only totals scores an examiner has already assigned and shows how the sum was built; it assumes the exam happened and was performed correctly, and it is not a way to diagnose or evaluate a stroke without one.
- Score each of the fifteen components from the standardised exam, items 1a through 11, using the official NIHSS definitions.
- Select the observed value for each item; items without a natural left/right split have one field, motor and consciousness items have several.
- Read the total — the sum of all fifteen sub-scores, from 0 (no deficit) to 42 (maximum severity).
Worked example — a moderate mixed deficit totals 17
Add the fifteen sub-scores from a moderate, mixed presentation: 1 (drowsy but arousable) + 1 (answers one question correctly) + 1 (follows one command) + 1 (partial gaze palsy) + 1 (partial visual field cut) + 2 (partial facial palsy) + 2 (left arm drifts to the bed within 10 seconds) + 0 (right arm holds against gravity) + 2 (left leg drifts) + 0 (right leg holds) + 1 (mild ataxia) + 1 (mild sensory loss) + 2 (severe aphasia) + 1 (mild dysarthria) + 1 (mild extinction). Sum the fifteen values and the total comes to 17, in the moderate-to-severe range where treatment decisions move quickly.
The two ends of the scale bracket that number cleanly. Score the best possible finding — 0 — on every one of the fifteen items and the total is 0, meaning no measurable deficit on this exam. Score the worst possible finding on every item instead (3, 2, 2, 2, 3, 3, 4, 4, 4, 4, 2, 2, 3, 2, 2) and the fifteen values sum to the maximum of 42, a devastating presentation typically associated with a comatose patient and a very large stroke.
Questions
Why are some items split into sub-scores and others aren't?
Consciousness and limb strength carry information that a single number would flatten: 1a, 1b, and 1c separately capture alertness, orientation, and the ability to follow commands, while 5a/5b and 6a/6b score each arm and leg on its own because one side weakening more than the other is a localising sign — it points toward the opposite side of the brain. Gaze, language, and the rest are assessed as a whole, so one score each is enough to capture what the exam is checking.
Can a total score alone tell me how severe a stroke is?
Broadly, yes, by convention: scores of 0 usually mean no deficit, 1 to 4 a minor stroke, 5 to 15 moderate, 16 to 20 moderate-to-severe, and above 20 severe. But two patients can reach the same total through very different combinations of findings — a language deficit weighted heavily in one, pure motor weakness in another — so the itemised breakdown often matters as much as the sum for planning care.
Who is qualified to perform this exam?
A clinician trained and certified in the NIH Stroke Scale — commonly a neurologist, emergency physician, stroke nurse, or paramedic who has completed the standardised training and passed the certification video modules. The exam involves hands-on testing: checking eye movements, asking specific questions, testing strength against resistance, and more. This calculator totals scores that examiner has already assigned; it does not perform or replace the exam.
How does the NIHSS affect treatment decisions?
It is one of the inputs, alongside time since symptom onset and imaging findings, that guides eligibility for clot-dissolving medication and for mechanical clot retrieval in large-vessel strokes. Very low scores may mean a treatment's risk outweighs a small measurable deficit, while very high scores can shift the urgency and the choice of intervention. The scale informs that judgment; it does not make the decision by itself.
Is this a substitute for calling emergency services?
No, and it was never meant to be one. Anyone with sudden weakness, slurred speech, vision loss, or confusion needs emergency evaluation immediately — stroke treatment is highly time-sensitive, and every minute of delay costs brain tissue. This page exists for people who already know how to perform the standardised exam and simply want the arithmetic totaled correctly and shown.
Does a score of 0 mean there was no stroke?
Not necessarily. A score of 0 means no deficit was detected on this particular exam, at this particular moment — some strokes, especially smaller ones affecting areas the scale does not test well, such as parts of the cerebellum, can produce real symptoms without moving the NIHSS total. A normal score does not rule out a stroke on its own; imaging and clinical judgment still matter.
References
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.