Skip to content

Classification and diagnosis

TL;DR — Migraine is a clinical diagnosis defined by recurring symptom patterns, not by an imaging, blood, electrophysiological or genetic test. ICHD-3 separates migraine without aura, migraine with aura and chronic migraine, permits concurrent medication-overuse headache, and deliberately classifies some clinically important phenotypes in an appendix while evidence accumulates (IHS 2018, PMID 29368949; Olesen 2024, PMID 38166472). The attack is polyphasic: prodromal and postdromal symptoms matter even though formal criteria concentrate on headache and aura (Paemeleire 2023, PMID 38043959). Aura usually evolves over minutes and contains positive neurological phenomena, but neither gradual spread nor positive symptoms excludes ischemia; a first, atypical or abruptly maximal deficit must not be labelled migraine by pattern recognition alone (Lebedeva 2018, PMID 28994605; Scutelnic 2024, PMID 39680237). Routine neuroimaging is not indicated for stable migraine with a normal examination, while a changed pattern or red flag changes that calculation (Evans 2020, PMID 31891197).

What is being classified

ICHD-3 is a phenomenological classification: diagnoses are assigned from attack number, duration, pain qualities, associated symptoms, frequency and exclusion of a better explanation. It standardizes research cohorts but is not a biological taxonomy; increasing migraine polygenic burden across non-headache, probable-migraine and definite-migraine categories suggests that the clinical boundary cuts through a continuum (IHS 2018, PMID 29368949; Häppölä 2022, PMID 34648375).

Dimension Operational distinction Consequence
Aura Fully reversible focal neurological symptoms meeting temporal and symptom criteria Subtypes the disorder and changes the vascular differential
Frequency Chronic migraine requires ≥15 headache days/month for >3 months, with migraine features or migraine-specific acute response on ≥8 days/month Determines trial eligibility and several treatment pathways
Medication exposure MOH is diagnosed concurrently when both sets of criteria are met Avoids forcing an either/or diagnosis
Special phenotype Hemiplegic, brainstem, retinal, vestibular and menstrual presentations use additional criteria Often widens the exclusion work-up
Certainty “Probable” migraine misses one required feature Preserves clinically migraine-like cases without diluting definite cohorts

The classification has evolved because overly restrictive earlier criteria excluded recognizable chronic disease and required improvement after medication withdrawal before MOH could be confirmed. ICHD-3 instead allows prospective diagnosis and dual coding (Ashina 2023, PMID 36732518). Historical revisions improved comparability, but changes mean prevalence and trial cohorts created under different editions are not automatically interchangeable (Olesen 2024, PMID 38166472).

Migraine without aura

ICHD-3 requires at least five attacks lasting 4–72 hours when untreated or unsuccessfully treated; at least two of unilateral location, pulsating quality, moderate/severe intensity, or aggravation by routine activity; and either nausea/vomiting or both photophobia and phonophobia (IHS 2018, PMID 29368949). Children may have 2–72-hour attacks, and bilateral pain is more common earlier in life.

The checklist should not be mistaken for a complete attack description. Prodrome, headache and postdrome can include fatigue, cognitive change, neck discomfort, yawning and sensory sensitivities; individual attacks vary within the same person (Paemeleire 2023, PMID 38043959). “Sinus headache,” neck pain and non-pulsatile or bilateral pain do not by themselves exclude migraine; the full recurring syndrome and absence of a better cause carry more diagnostic weight.

Feature Diagnostic use Limitation
Activity aggravation One of four pain characteristics Patients may avoid activity, obscuring the feature
Nausea/vomiting One associated-symptom route Absent in many attacks and less consistently reported by children
Photo- and phonophobia Must both be present if nausea/vomiting is absent Often inferred from behaviour in young children
Attack duration 4–72 h in adults; shorter lower bound in children Early effective treatment truncates observed duration
Laterality/pulsation Supportive, not individually mandatory Bilateral and pressure-like migraine occurs

Migraine with aura

Aura comprises fully reversible visual, sensory, speech/language, motor, brainstem or retinal symptoms. The diagnosis requires at least two attacks and at least three of six temporal/phenotypic features: gradual spread over ≥5 minutes, succession of symptoms, each symptom lasting 5–60 minutes, unilateral symptom, positive symptom, and headache within 60 minutes (IHS 2018, PMID 29368949). Motor aura may last longer than 60 minutes.

Visual aura is the most common form; systematic review documents substantial diversity in shape, colour, field position and evolution, making a stereotyped “zig-zag line” an inadequate screening description (Joppeková 2025, PMID 40597581). Field testing found the revised criteria more sensitive than earlier criteria for typical aura, while brainstem-aura criteria risked over-inclusion (Li 2015, PMID 25424707). Earlier validation work showed that criteria combining gradual development, duration and symptom quality can achieve high sensitivity and specificity, but their performance depends on the comparison group (Eriksen 2005, PMID 15654035).

Aura versus transient ischemia or seizure

Pattern More typical of aura More typical of ischemia/seizure Why it is not decisive
Onset Gradual spread across ≥5 min Sudden deficit for TIA; seconds for many seizures Imaging-confirmed ischemia can evolve or produce positive symptoms
Symptom type Positive visual/sensory phenomenon followed by negative symptom Negative loss of function in TIA; stereotyped positive motor/sensory event in seizure All three can mix positive and negative phenomena
Succession Visual → sensory → language symptoms over minutes Simultaneous deficits in TIA Migraine symptoms can be simultaneous
Duration Usually 5–60 min per non-motor symptom Variable Duration distributions overlap
Recurrence Similar attacks over years TIA often new; seizure often highly stereotyped Recurrent vascular events and changing aura both occur

Explicit TIA criteria differentiated TIA from aura in the derivation study, but subsequent prospective work found that neither ICHD aura criteria nor proposed TIA criteria produce perfect separation (Lebedeva 2018, PMID 28994605; Lebedeva 2018, PMID 28372496). A 2024 diagnostic-accuracy study found migraine-like symptoms unexpectedly frequent in imaging-confirmed infarction, directly challenging the assumption that positive or spreading symptoms are non-ischemic (Scutelnic 2024, PMID 39680237). The safe inference is asymmetric: a classic recurrent aura pattern supports migraine, but an unfamiliar focal episode still warrants vascular or seizure assessment.

Hemiplegic migraine requires reversible motor weakness plus another reversible aura symptom. Familial and sporadic forms share criteria; interview, examination during an attack when possible, and exclusion of stroke, epilepsy, encephalitis and secondary syndromes remain central (de Boer 2024, PMID 38307656). Common migraine is polygenic, whereas rare pathogenic variants in CACNA1A, ATP1A2 and SCN1A explain subsets of hemiplegic migraine; absence of a variant does not exclude the clinical diagnosis (de Boer 2024, PMID 38307656).

Episodic, high-frequency and chronic migraine

“Episodic migraine” is operationally anything below the chronic threshold rather than a separate ICHD-3 diagnosis. The sharp boundary at 15 headache days/month is administratively useful but biologically imperfect: symptom burden rises continuously, and high-frequency episodic migraine can resemble chronic migraine more than low-frequency disease.

For chronic migraine, a diary should distinguish total headache days, migraine-feature days and acute-medication days. The ≥8 migraine days may be days meeting migraine criteria, believed at onset to be migraine and relieved by a triptan or ergot, or fulfilling aura criteria (IHS 2018, PMID 29368949). Earlier chronic-headache cohorts often failed strict older criteria, illustrating how classification rules create or erase study populations (Papetti 2019, PMID 30890994).

MOH is a secondary headache diagnosis in a person with a pre-existing headache disorder who has headache on ≥15 days/month and regular overuse for >3 months. The exposure threshold is generally ≥10 days/month for triptans, opioids, ergots or combination analgesics and ≥15 days/month for simple analgesics; the underlying migraine and MOH are coded together (Ashina 2023, PMID 36732518). This is an exposure-defined diagnosis, not proof that medication alone caused chronification.

Diaries and measurement error

A prospective diary reduces recall error and makes frequency, duration, medication exposure and menstrual timing visible. It is especially important near the 15-day chronic threshold, the medication-overuse thresholds and for menstrual migraine.

Validation problem Quantitative result Interpretation
Child/adolescent self-labelled migraine day vs ICHD diary classification κ=0.50; PPV 0.66; NPV 0.85 in 438 detailed entries “Migraine day” is not self-evident even within diagnosed cohorts (Kellier 2023, PMID 37140013)
Self-reported menstrual migraine vs diary ICHD-3 diagnosis Only about two thirds agreed; statistical-vs-diary κ=0.28 Retrospective menstrual labels are unreliable (Verhagen 2022, PMID 35514214)
Population menstrual-migraine classification 1,532 of 9,184 women with migraine met ICHD-3 MM criteria in a Danish questionnaire study Definition and observation window materially affect prevalence (Chalmer 2023, PMID 37184838)
Computerized ICHD-3-beta support in a specialist cohort Correctly recognized 159/160 migraine without aura, 36/36 with aura and 20/21 chronic migraine Rules can be implemented consistently, but specialist diagnosis was the reference (Dong 2014, PMID 24934331)

Screening is not diagnosis

ID Migraine asks about nausea, disability and photophobia. In the original primary-care validation, 451 of 563 recruited patients completed expert evaluation; a positive screen was designed to flag likely migraine, not establish a final diagnosis (Lipton 2003, PMID 12913201). A 22-study meta-analysis and Chinese validation confirmed useful diagnostic accuracy but also heterogeneity across settings (Wang 2015, PMID 26244435). Setting matters: a Brazilian tertiary cohort reported sensitivity 92% but specificity 60%, illustrating why positive predictive value rises when migraine prevalence is already high (Mattos 2017, PMID 28746431). A North-Indian translation study similarly treated neurologist ICHD-3 assessment as the reference standard (Sahu 2023, PMID 37970241).

Screeners therefore answer “who needs migraine assessment?” They do not distinguish every aura subtype, identify secondary headache or replace a history and examination.

Examination and testing

The examination is expected to be normal between attacks. Its purpose is less to prove migraine than to detect an alternative: papilledema, meningism, impaired consciousness, a persistent focal deficit, systemic illness or a new ocular finding. Emergency-department data show that red flags are individually imperfect and must be interpreted with the phenotype and examination (Munoz-Ceron 2019, PMID 30615622).

The American Headache Society systematic review found no evidence that routine imaging is necessary for patients whose headaches are consistent with migraine, whose neurological examination is normal and who have no atypical features or red flags. Imaging may be considered for an unusual, prolonged or persistent aura; increasing frequency or changed features; first or worst migraine; brainstem aura; confusion; motor manifestations; late-life accompaniments; side-locked headache; or post-traumatic headache, but much of this guidance is consensus-based (Evans 2020, PMID 31891197).

In children with recurrent headache and a normal neurological examination, routine laboratory testing, EEG and neuroimaging are not supported; seizures, recent severe onset, change in type or abnormal examination alter the evaluation (Lewis 2002, PMID 12196640). Pediatric application needs care because shorter attacks and bilateral pain can leave otherwise typical cases outside adult-shaped criteria (Papetti 2019, PMID 30890994).

Common diagnostic failure modes

Failure Corrective rule
Calling every severe headache “migraine” Apply attack pattern and associated-symptom criteria; exclude a better explanation
Calling every visual disturbance “aura” Record monocular/binocular field, positive/negative phenomena, spread, duration and succession
Treating “probable migraine” as no migraine Preserve the uncertainty; many clinical attacks miss one criterion
Treating chronic migraine and MOH as mutually exclusive Code both when both criteria are met
Using treatment response as sole proof Placebo response and nonspecific analgesia prevent diagnostic certainty
Ordering imaging to confirm uncomplicated migraine Use imaging to investigate an atypical feature or secondary cause, not to visualize migraine
Trusting retrospective frequency estimates near thresholds Use a daily diary before reclassification where feasible
Assuming a familiar migraine history explains a new phenotype Reassess any abrupt, persistent or materially changed event

Boundary cases

Vestibular migraine requires recurrent vestibular episodes, a migraine history and migraine features during at least half the episodes, with exclusion of another vestibular diagnosis. The Bárány Society/IHS update retained the 2012 criteria after literature review, while “probable vestibular migraine” remains useful when only migraine history or episode-associated migraine features are present (Lempert 2022, PMID 34719447).

Menstrual migraine is an appendix diagnosis requiring diary-demonstrated attacks in a defined perimenstrual window in at least two of three cycles. Validation found pure menstrual migraine under 1% in one e-diary cohort and poor accuracy of self-report; the label should not rest on impression alone (Verhagen 2022, PMID 35514214). A population study proposed revisions after finding important misclassification risk when attacks occur frequently by chance (Chalmer 2023, PMID 37184838).

Status migrainosus is currently a debilitating attack lasting >72 hours. The duration boundary is historically rather than empirically anchored, and a 2026 review argues that severity, disability, response and phase should be incorporated in future criteria (Robblee 2026, PMID 41902388).

What diagnosis can and cannot claim

The clinical criteria have face validity, field-testing and measurable genetic gradients, but no biomarker serves as an independent gold standard (Häppölä 2022, PMID 34648375; Tinsley 2018, PMID 30291572). Diagnostic reliability therefore depends on how well the phenotype is elicited and recorded. A diagnosis predicts a family of attack patterns and evidence-based treatment options; it does not identify a single mechanism, guarantee response or eliminate future secondary headache.

Open questions

  • Can outcome- or treatment-response validation replace expert-consensus circularity in future migraine criteria? Current validation usually compares rules with specialist diagnosis or another rule set (Dong 2014, PMID 24934331; Tinsley 2018, PMID 30291572).
  • Where should the episodic/chronic boundary sit if disability and biology vary continuously across monthly headache frequency? (IHS 2018, PMID 29368949)
  • Can a validated event-level classifier distinguish aura, TIA and focal seizure without sacrificing sensitivity to atypical ischemia? (Lebedeva 2018, PMID 28994605; Scutelnic 2024, PMID 39680237)
  • Should menstrual migraine criteria incorporate attack frequency and chance clustering rather than a fixed calendar rule alone? (Verhagen 2022, PMID 35514214; Chalmer 2023, PMID 37184838)
  • Which diary definition of a migraine day best predicts disability and treatment response in children and adults? (Kellier 2023, PMID 37140013)

References

  1. Headache Classification Committee of the International Headache Society. The International Classification of Headache Disorders, 3rd edition. Cephalalgia. 2018;38:1-211. PMID 29368949
  2. Olesen J. The International Classification of Headache Disorders: History and future perspectives. Cephalalgia. 2024;44:3331024231214731. PMID 38166472
  3. Paemeleire K, et al. Migraine without aura. Handb Clin Neurol. 2023;198:151-167. PMID 38043959
  4. Häppölä P, et al. Polygenic risk provides biological validity for the ICHD-3 criteria among Finnish migraine families. Cephalalgia. 2022;42:345-356. PMID 34648375
  5. Eriksen MK, et al. Sensitivity and specificity of the new international diagnostic criteria for migraine with aura. J Neurol Neurosurg Psychiatry. 2005;76:212-217. PMID 15654035
  6. Li D, et al. Field-testing of the ICHD-3 beta/proposed ICD-11 diagnostic criteria for migraine with aura. Cephalalgia. 2015;35:748-756. PMID 25424707
  7. Joppeková Ľ, et al. What does a migraine aura look like? A systematic review. J Headache Pain. 2025;26:149. PMID 40597581
  8. Lebedeva ER, et al. Explicit diagnostic criteria for transient ischemic attacks to differentiate it from migraine with aura. Cephalalgia. 2018;38:1463-1470. PMID 28994605
  9. Lebedeva ER, et al. Prospective testing of ICHD-3 beta diagnostic criteria for migraine with aura and migraine with typical aura in patients with transient ischemic attacks. Cephalalgia. 2018;38:561-567. PMID 28372496
  10. Scutelnic A, et al. Field-testing the explicit diagnostic criteria for transient ischemic attack: a diagnostic accuracy study. J Neurol. 2024. PMID 39680237
  11. de Boer I, et al. Hemiplegic migraine. Handb Clin Neurol. 2024;199:353-365. PMID 38307656
  12. Ashina S, et al. Medication overuse headache. Nat Rev Dis Primers. 2023;9:5. PMID 36732518
  13. Papetti L, et al. Features of primary chronic headache in children and adolescents and validity of ICHD-3 criteria. Front Neurol. 2019;10:92. PMID 30890994
  14. Kellier DJ, et al. Validity of self-reported migraine in adolescents and children. Headache. 2023;63:634-641. PMID 37140013
  15. Verhagen IE, et al. Validation of diagnostic ICHD-3 criteria for menstrual migraine. Cephalalgia. 2022;42:1184-1193. PMID 35514214
  16. Chalmer MA, et al. Population-based characterization of menstrual migraine and proposed diagnostic criteria. JAMA Netw Open. 2023;6:e2313235. PMID 37184838
  17. Dong Z, et al. Validation of a guideline-based decision support system for diagnosis of primary headache disorders based on ICHD-3 beta. J Headache Pain. 2014;15:40. PMID 24934331
  18. Lipton RB, et al. A self-administered screener for migraine in primary care: the ID Migraine validation study. Neurology. 2003;61:375-382. PMID 12913201
  19. Wang X, et al. Validation of the Chinese version of ID-Migraine and systematic review with meta-analysis concerning its diagnostic accuracy. J Oral Facial Pain Headache. 2015;29:265-278. PMID 26244435
  20. Mattos ACMT, et al. ID-Migraine questionnaire and accurate diagnosis of migraine. Arq Neuropsiquiatr. 2017. PMID 28746431
  21. Sahu P, et al. Translation and validation of ID-Migraine questionnaire to North-Indian vernacular languages. Ann Indian Acad Neurol. 2023;26:543-548. PMID 37970241
  22. Evans RW, et al. Neuroimaging for migraine: American Headache Society systematic review and evidence-based guideline. Headache. 2020;60:318-336. PMID 31891197
  23. Munoz-Ceron J, et al. Headache at the emergency room: etiologies, diagnostic usefulness of ICHD-3 criteria, red and green flags. PLoS One. 2019;14:e0208728. PMID 30615622
  24. Lewis DW, et al. Practice parameter: evaluation of children and adolescents with recurrent headaches. Neurology. 2002;59:490-498. PMID 12196640
  25. Lempert T, et al. Vestibular migraine: diagnostic criteria. J Vestib Res. 2022;32:1-6. PMID 34719447
  26. Tinsley A, Rothrock JF. What are we missing in the diagnostic criteria for migraine? Curr Pain Headache Rep. 2018;22:84. PMID 30291572
  27. Robblee J, et al. Redefining status migrainosus: a narrative review. Headache. 2026. PMID 41902388