The top 30 migraine triggers, ranked by how often they actually show up in clinical data.

If you came here looking for a single villain, I have bad news: there isn’t one. The top migraine trigger by clinical frequency is stress (reported by 70–80% of patients), followed by sleep deprivation, hormonal shifts, weather changes (especially barometric pressure), and dietary triggers like alcohol and skipped meals. That ordering is consistent across the biggest studies — Kelman’s 1,207-patient survey, Pavlovic’s literature summary, Marmura’s clinical review.

But the ranking is statistical. Your triggers are personal, and they almost certainly aren’t a single item from the list — they’re a combination that stacks until you cross your threshold. That’s the most important sentence on this page, and I’ll defend it below.

TL;DR — the top 10

Rank	Trigger	% of patients reporting	Typical lag
1	Stress (acute or “let-down”)	70–80%	0–24 hours
2	Sleep disturbance (too little or too much)	50–65%	6–24 hours
3	Hormonal fluctuation (menstrual, perimenopause)	50–60% (women)	1–3 days
4	Skipped meals / fasting	40–60%	4–12 hours
5	Weather changes (pressure, humidity, temperature)	40–55%	6–24 hours
6	Bright or flickering lights	35–50%	Minutes–hours
7	Alcohol (especially red wine)	30–45%	3–12 hours
8	Dehydration	30–40%	4–12 hours
9	Strong odors / perfumes	25–40%	Minutes–hours
10	Caffeine excess or withdrawal	25–40%	6–24 hours

Numbers vary by study. These ranges combine Kelman (2007), Pavlovic (2014), and the AMF top-10 list. Everything past #10 gets harder to rank reliably — the studies stop agreeing.

The threshold model — why “avoid your triggers” is bad advice

Before the list, the concept that makes the list usable.

The old framing of migraine triggers was binary: “X causes migraines for you, so avoid X.” It doesn’t work. People avoid chocolate religiously and still get attacks. People drink red wine and don’t always react. So they conclude their triggers are mysterious.

The modern framing is the threshold model, sometimes called trigger stacking:

• Every person with migraine has an attack threshold.
• Each trigger contributes a “load” toward that threshold.
• One trigger usually isn’t enough. Multiple triggers, close together, push you over.
• The threshold itself moves — it’s lower when you’re depleted (poor sleep, illness, period) and higher when you’re rested.

This is why the same trigger sometimes causes an attack and sometimes doesn’t. A glass of red wine on a Friday after a relaxing week is fine. The same glass of red wine after three nights of bad sleep, on day 1 of your cycle, during a storm front — devastating.

The practical consequence: don’t chase the single trigger. Reduce the cumulative load on high-risk days. If you wake up after 5 hours of sleep on a low-pressure day, that’s not the day to also skip lunch and drink wine.

With that framing in place — the list.

Stress and lifestyle (1–6)

1. Acute stress

The top trigger in every major study. The mechanism is thought to involve cortisol, sympathetic nervous system activation, and central sensitization. Acute stress alone is bad; stress combined with anything else on this list is worse. Track: subjective stress 0–10 daily, plus major events (deadlines, conflicts, travel). Look for attacks 0–24 hours after spikes.

2. Stress let-down (“weekend migraines”)

The drop in cortisol after sustained stress. Classic Saturday-morning migraines after a hard work week. The let-down trigger is sneaky because it doesn’t feel like stress — by the time the headache starts, you’re relaxed. Track: attacks within 24–48 hours of finishing a high-pressure stretch.

3. Sustained anxiety / depression

Distinct from acute stress. Chronic mood disorders are both a trigger and a comorbidity of migraine. CBT and SNRI-class antidepressants reduce frequency for many sufferers. Track: weekly mood baseline alongside attack frequency.

4. Overexertion (especially sudden exertion)

Heavy lifting, intense workouts, even sex can trigger attacks via vascular changes. “Primary exertional headache” is its own ICHD category. Gradual exercise is protective; sudden spikes aren’t. Track: attacks within 1–6 hours of intense physical effort.

5. Posture and neck tension

Hours hunched over a laptop or phone activate the trigeminocervical complex, which can feed into and amplify a migraine attack via convergent sensory pathways. Common in remote workers. Track: screen-hour totals and neck pain in the 24 hours before attacks.

6. Travel (jet lag, time zone shift)

Combines disrupted sleep, disrupted meals, cabin pressure changes, and stress. Of the top 30, travel may be the single most reliable multi-trigger stack. Track: attacks within 24–72 hours of arrival at a new time zone.

Sleep (7–9)

7. Sleep deprivation

The #2 trigger overall. Less than 6 hours sleep correlates strongly with next-day attacks in prospective diary studies. Mechanism: reduced glymphatic clearance, altered serotonin, increased central sensitization. Track: sleep duration nightly and look at the 24-hour window after short nights.

8. Oversleeping

Lying in on Saturday is a classic trigger. The mechanism isn’t fully known but probably involves shifted circadian rhythm and missed meals/caffeine. Track: attacks after nights >9 hours, especially if normal is 7.

9. Irregular sleep schedule / shift work

Even if total sleep is adequate, variable bedtimes destabilize the trigger threshold. Shift workers have elevated migraine rates compared to fixed-schedule workers. Track: standard deviation of your bedtime, not just the average.

Hormonal (10–13)

10. Menstrual cycle (estrogen withdrawal)

The drop in estrogen in the 2 days before menstruation triggers attacks in 50–60% of women with migraine. “Menstrual migraine” is a recognized ICHD subtype. Track: day of cycle for every attack. The pattern usually surfaces within 2–3 cycles.

11. Ovulation

A smaller estrogen spike mid-cycle triggers attacks in a subset of women. Less studied than menstrual migraine, but real. Track: attacks around day 13–15 of a typical cycle.

12. Perimenopause

The chaotic hormonal years before menopause (typically 40s) often increase migraine frequency before they decrease it. Post-menopause, frequency usually drops. Track: if you’re in your 40s and frequency is rising, this is likely contributing.

13. Hormonal contraceptives

Combined oral contraceptives can either trigger or prevent migraine — depends on the formulation and on whether you have migraine with aura (in which case combined OCs are usually contraindicated due to stroke risk). Track: before vs. after starting any new hormonal medication.

Weather (14–17)

14. Barometric pressure drops

The most-studied weather trigger. A drop of 6–10 hPa within 24 hours is the most-cited threshold. ~75% of pressure-sensitive sufferers feel storms coming. I’ve written a whole piece on barometric pressure thresholds — short version: rate of change matters more than absolute pressure. Track: local pressure history + attack timestamps.

15. Temperature swings

Rapid temperature changes (especially heat waves) trigger attacks for ~20% of weather-sensitive sufferers. Likely involves vasodilation. Track: day-over-day temperature deltas, not just absolute heat.

16. High humidity

Often pairs with low pressure to compound effect. Humidity alone is a weaker trigger than pressure or temperature, but the combination is potent. Track: humidity alongside pressure.

17. Wind patterns (Foehn, Santa Ana, Chinook)

Specific dry-wind events correlate with migraine spikes in regions where they occur. The mechanism is debated — likely involves rapid pressure and humidity changes plus ion balance shifts. Track: if you live in a Foehn/Santa Ana/Chinook region, log wind events as their own field.

Dietary (18–23)

18. Alcohol (especially red wine)

The most reliable dietary trigger — holds up in prospective diary studies, unlike most others. Mechanism involves vasoactive compounds (tyramine, histamine, sulfites) plus dehydration. Red wine is the worst offender; clear spirits cause attacks for fewer people. Track: type and quantity of alcohol vs. 12-hour attack window.

19. Skipped meals

Hypoglycemia is a documented trigger. Attacks typically start 4–12 hours after a missed meal. Often combines with stress (working through lunch). Track: time between meals, especially on busy days.

20. Dehydration

Often paired with skipped meals and exertion. Direct mechanism: reduced plasma volume → altered cerebral perfusion. Track: rough daily water intake; look for attacks on low-intake days.

21. Aged cheeses and cured meats (tyramine)

Tyramine is a vasoactive amine that triggers attacks in a subset of sufferers. Foods include aged cheddar, parmesan, salami, prosciutto. The effect is real but smaller than alcohol. Track: specific food items, not just “cheese.”

22. Chocolate

Frequently blamed, often unfairly. Some studies find chocolate is more often a premonitory craving — your brain, already entering the prodromal phase, craves carbs and chocolate before pain begins. The mechanism, if real as a trigger, may involve phenylethylamine or caffeine content. Track: time of consumption vs. attack — and whether premonitory symptoms preceded the craving.

23. MSG, nitrates, artificial sweeteners (aspartame)

Frequently cited, weakly supported. Controlled trials of MSG specifically haven’t reliably reproduced the effect. Aspartame has slightly better evidence in case reports. Track: if you suspect one, do a 30-day elimination and reintroduction with logged attacks.

Sensory (24–27)

24. Bright or flickering lights

Photophobia during an attack is universal; bright lights as a trigger are reported by 35–50%. Fluorescent flicker, screen glare, sunlight on snow or water. Track: environment in the hour before attack onset.

25. Strong odors

Perfumes, cleaning products, gasoline, cigarette smoke. Osmophobia (odor sensitivity) is one of the most migraine-specific symptoms — non-migraineurs rarely report it. Track: specific odors and timing.

26. Loud or repetitive sounds

Concerts, construction, open-plan offices. Often paired with sensory overload generally. Track: sound exposure plus other stack components.

27. Screen time / blue light

Modest direct effect; bigger effect through neck posture, sleep disruption (late-night screens), and eye strain. Track: total screen hours rather than just “device use.”

Medication-related (28–30)

28. Caffeine withdrawal

Skipping your usual coffee triggers attacks for habitual users — typically 12–24 hours after the last dose. The same caffeine that aborts attacks at modest doses triggers them on withdrawal. Track: caffeine intake daily, not just on attack days.

29. Medication overuse headache (MOH)

Using abortive medication (triptans, NSAIDs, combination analgesics) more than 10–15 days per month causes a chronic daily headache that resists treatment. It looks like worsening migraine; it’s actually the rescue meds. Track: abortive medication use days per month — if you’re above 10, talk to a neurologist.

30. Vasodilator medications

Nitrates (for heart conditions), some blood pressure drugs, PDE5 inhibitors (sildenafil), and hormonal medications can trigger attacks via vascular mechanisms. If migraine frequency rose after a new prescription, this is worth investigating. Track: start dates of new medications vs. attack frequency before/after.

What the list doesn’t tell you

A statistical top-30 is a starting menu, not a diagnosis. Three things the ranked list will not give you:

1. Which of these matter for you. Stress is #1 globally; for you, it might be irrelevant. For me, weather (#14) and sleep (#7) are the dominant pair, and stress barely registers. You can’t read that off this page.
2. How your triggers stack. The top-30 lists each trigger as if it acts alone. They don’t. Prospective diary studies (summarized in Pavlovic 2014) suggest multiple-trigger days are several-fold more likely to produce an attack than single-trigger days, even when the single trigger was historically strong.
3. Your threshold. Some people get attacks at any combination of two triggers. Others need three or four to stack. The threshold isn’t on this list. It’s in your own history.

Finding your triggers

The honest answer to “what are my migraine triggers?” is: log everything, for at least 8 weeks, and let correlation surface the real pattern.

Minimum dataset per day:

• Sleep duration and approximate quality
• Subjective stress (0–10)
• Major food/drink (alcohol, caffeine, anything unusual)
• Weather (pressure if you can — most weather apps don’t show it)
• Cycle day if relevant
• Any unusual sensory exposure (perfume, screens, loud environments)
• And of course: any attack, with timestamp, severity, and duration

The first month of this feels pointless. The second month, patterns start to become obvious. By month three, you usually know your top three and you can ignore the rest of the list.

A paper migraine diary template works fine for this. So does Notes app. So does a spreadsheet. The medium doesn’t matter; the consistency does.

If you’d rather not run the analysis manually — and want a comparison of the iOS tracker landscape — that’s the niche Migra fills. It’s the iOS app I built because I was tired of trying to eyeball correlations across screenshots of my Notes app. It pulls real-time barometric pressure from the iPhone’s altimeter, integrates with Apple Health for sleep and cycle data, and after ~20 logged attacks the on-device AI starts surfacing your specific trigger stack — not the population average from the table above. All processing is local; nothing leaves the phone. It’s not magic; it’s just the math that’s tedious to do by hand.

If you want a clinical baseline for how much migraine is actually costing you while you do the tracking work, the MIDAS test is the 5-question disability score most neurologists use. Worth a baseline before you change anything.

The one-line takeaway

The top 30 is a map of where to look. The legend is in your own log. Stress and sleep are the most likely entries on your personal list — but the only way to find out which of the other 28 matter to you is to write things down for two months and let the pattern emerge.

That’s the whole job. Triggers don’t reveal themselves to memory. They reveal themselves to data.