Migraine and your nervous system: what HRV and heart rate can and can't tell you

Short answer

HRV and migraine are linked at the group level, not the individual level. Between attacks, people with migraine show, on average, weaker autonomic test responses and, in small studies, lower heart rate variability than people without it. But HRV cannot predict your next attack: across 1,615 people who log migraines in Welltory, heart rate near a logged migraine was only about 1 bpm lower than on their own ordinary days.

Before you blame yourself, check what your body already recorded. If you have felt your whole system shift around an attack — the yawning, the heavy fatigue, the pulse that seems off — you are not imagining it: autonomic symptoms are part of how migraine is described in the clinical literature, and an attack is not your fault. What a tracker adds is a personal record, not a verdict. If you want somewhere to start, try a guided migraine check-in in Welltory.

Note: this article explains research and our own aggregated, anonymised data; it is not medical advice. Migraine is diagnosed and treated by a clinician. A sudden, severe or new kind of headache needs urgent care — see the red flags further down.

What is HRV, and why would it matter for migraine?

Heart rate variability (HRV) is the variation in time between one heartbeat and the next. A healthy heart does not tick like a metronome: the gap between beats stretches and shortens from moment to moment, and much of that fine adjustment comes from the autonomic nervous system — the part of the nervous system that runs things you do not consciously control. Its sympathetic branch works like an accelerator; its parasympathetic branch, carried largely by the vagus nerve, works like a brake. Measures such as RMSSD mostly reflect that vagal brake, which is why researchers call them "vagally mediated" HRV.

Migraine is a neurological disorder, not "a bad headache." The International Classification of Headache Disorders (ICHD-3) describes attacks lasting 4–72 hours when untreated or unsuccessfully treated, with at least two of these features: one-sided pain, pulsating quality, moderate or severe intensity, and pain made worse by routine activity like climbing stairs — plus nausea or sensitivity to both light and sound. ICHD-3 also describes a prodrome that can start hours or days before the headache, with symptoms such as fatigue, repetitive yawning, cravings for particular foods, low mood, being unusually active or sluggish, and neck stiffness or pain.

Several of those symptoms belong to the autonomic nervous system. A 2024 review by Pavelić and colleagues lists nausea, vomiting, sweating, pallor, palpitations and light-headedness among the autonomic symptoms that precede, accompany and outlast attacks. So the question is reasonable: if migraine involves the nervous system that steers your heart rhythm, does your heart rhythm carry a trace of it?

The answer comes in three layers, and they point in different directions:

• The group layer — do people with migraine, on average, differ from people without it? Yes, modestly.

• The attack layer — does something change in the hours around an attack? Sometimes, in small and inconsistent ways.

• Your layer — can a reading tell you that an attack is coming, or that a headache is a migraine? No, not with any method available today.

Keeping those three apart is the whole point of this guide. Most confusion about HRV and migraine comes from reading a group-level finding as if it were a personal prediction.

What does research show about HRV and the nervous system in migraine?

Four pieces of work cover most of what is known, and they do not agree perfectly.

A 2016 meta-analysis of vagal HRV. Koenig and colleagues pooled seven studies comparing people with headache disorders against healthy controls. For RMSSD, the pooled comparison covered 122 patients and 125 controls, and HRV was lower in the headache group with a medium effect size (Hedges' g −0.63). For high-frequency HRV, the effect was smaller (g −0.30) and its confidence interval crossed zero. The conclusion: vagally mediated HRV is reduced in people with headache disorders, but the total number of people studied was small.

A 2024 meta-analysis of standard autonomic tests. Pavelić, Zebenholzer and Wöber pooled seven studies that used the classic bedside autonomic tests — deep breathing, the Valsalva manoeuvre, standing up (orthostatic challenge) and a sustained hand grip (isometric challenge). Measured between attacks, the 424 people with migraine had lower results than 268 healthy controls on every test: deep breathing g −0.32, orthostatic g −0.28, isometric g −0.55, and a Valsalva ratio 0.17 lower. The authors read this as the autonomic system operating "at a lower threshold of homeostasis" between attacks.

Their own caveats matter as much as the result. The seven studies came from only three research groups; all but one carried a high risk of bias; none followed the latest autonomic testing guidelines; and each person was tested once, so a single day stood in for a whole month.

A small 24-hour study during and between attacks. Zhang and colleagues recorded 24-hour heart rhythm in 18 people with episodic migraine and 18 matched controls. During an attack, SDNN (another HRV measure) averaged about 57 ms, against about 116 ms in the same people between attacks and about 136 ms in controls. Between attacks, however, HRV did not differ significantly from controls — which does not fit neatly with the between-attack differences above.

A pattern that changes direction. Summarising the field in 2026, Faisal and colleagues describe autonomic dysfunction in migraine as bidirectional — lower sympathetic activity between attacks and higher during them — which, in their words, makes stable markers difficult to define.

What this adds up to. At the group level, the autonomic nervous system in migraine looks slightly different, most consistently in tests of how the heart responds to a challenge. The effects are small to medium, which means the readings of people with and without migraine overlap heavily. That is why none of these tests is used to diagnose migraine, and why a single low HRV number says nothing about whether you have it.

One thing to raise separately. If you also get dizzy, shaky or a racing heart when you stand up, mention it to your doctor as its own topic. A heart rate that jumps on standing can have causes of its own, such as postural orthostatic tachycardia syndrome (POTS), and that deserves its own evaluation rather than being filed under migraine.

What happens in your body in the hours and days before a migraine?

The prodrome is the part of migraine people most want their data to show, so it helps to know how solid the evidence is.

How common is a prodrome? Less certain than many articles suggest. A 2022 meta-analysis by Eigenbrodt and colleagues estimated that 29% of people with migraine in population studies, and 66% of those seen in clinics, report at least one premonitory symptom. Heterogeneity was extreme (I² 99%), risk of bias was high in 20 of the 29 included studies, and the authors concluded the evidence was insufficient to estimate prevalence reliably.

What does it feel like when people do recognise it? The largest prospective data come from the screening period of the PRODROME trial, reported by Schwedt and colleagues in 2025. 920 adults with two to eight attacks a month, chosen because they said they could recognise their prodrome, logged events in an electronic diary for 60 days — 4,802 events in all. The most common symptoms were sensitivity to light (57.2%), fatigue (50.1%), neck pain (41.9%), sensitivity to sound (33.9%), difficulty thinking or concentrating (30.0%) and dizziness (27.8%). 81.5% of the logged prodromes were followed by headache within one to six hours. Two caveats: the participants were selected for already recognising their prodrome, and the trial was funded by a drug manufacturer.

Some "triggers" may be the attack already starting. In a 2021 study of 53 people with migraine, Karsan and colleagues found that several self-reported triggers matched early attack symptoms: bright light as a trigger agreed with premonitory light sensitivity, and skipping meals or specific foods as triggers agreed with premonitory food cravings. Stress as a trigger showed good agreement with mood change in the early phase of attacks that were triggered in the lab. The practical point: craving chocolate and then getting a migraine does not prove the chocolate did it. We cover this in more depth in your migraine trigger might be the migraine already starting.

What a wearable sees on a headache day. In a 12-week study of people with tension-type headache or migraine who wore Fitbit trackers, reported by Cerrada and colleagues, 350 participants had enough data for the main analysis. On headache days, people slept more, moved less and had a lower maximum heart rate — and the more a headache affected daily life, the lower activity and maximum heart rate went. That is a picture of what an attack does to a day (lying down, cancelling plans), not a marker of what comes before it.

What this means for your own numbers. A dip in your morning HRV the day before an attack could be part of a prodrome — or a short night, a stressful week, an infection, alcohol, or ordinary day-to-day noise. None of the research above lets you tell which from the number alone. What it does support is looking at your data alongside what you felt and did, over weeks rather than one morning.

Can HRV predict a migraine?

No — not for an individual, and not with any method you can use today.

The most serious attempt so far is a 2026 study by Faisal and colleagues. They used data from 146 people in a clinical trial of app-based biofeedback, with wearables recording muscle tension, HRV and skin temperature during daily sessions alongside a headache diary — 21,550 days of data, of which 4,061 were moderate-to-severe headache days. Their best time-series model scored an AUC of 0.84 for the next day. That sounds impressive until you read the detail: its sensitivity was 0.39, meaning it recognised fewer than four in ten of the actual headache days, and did well mostly by correctly calling headache-free days. For a three-day window, the AUC was 0.76 with sensitivity 0.58.

The authors are careful about what this means. The participants were highly motivated trial volunteers doing daily biofeedback, which may itself change the signals. The most useful inputs were headache intensity and duration from the diary, followed by heart-rate scores from the sessions. And they state that the model must be validated in independent samples before any clinical use, and that such tools would likely be regulated as medical devices.

Our own data point the same way, from the other side: the heart-rate difference we see around a logged migraine is real across the group but tells us almost nothing about any one person's day (details in the next section).

In practice HRV is one input among several, not the deciding one. For migraine, the other inputs are your diary, your sleep, your cycle, your week and your own sense of your body — and your neurologist.

If you want the full comparison of what smartwatches claim and what studies show, see can a smartwatch predict a migraine?

What our data shows: heart rate around a logged migraine

We looked at what happens to heart rate and HRV around the moment people write "migraine" in Welltory. This is an internal, aggregated analysis. Nothing here detects or predicts migraine, and it should be read with its caveats, which we put first.

What the data is. Migraine has no preset tag in Welltory, so people type their own — the catalogue holds 2,865 distinct migraine spellings used by 2,710 people, and the analysis matched 301 of them across several languages. The analysis sample is 1,634 people who logged a migraine between 2023 and August 2026 and had heart measurements both near a label and at least 14 days away from any migraine label; 1,615 remain in the main comparisons. Each person's measurements within 24 hours of a label were compared with their own measurements on ordinary days, at least two weeks from any migraine label. Where the device could be identified, about 91% of readings came from Apple Watch, with smaller shares from Galaxy Watch and the phone camera.

Read this first:

• These are self-recorded labels, not a record of anyone's health. Tagging runs one to two orders of magnitude below plausible migraine frequency: clinical episodic migraine means roughly one to four attacks a month, while the 95th percentile here is 12 a year. 48.7% of person-years contain exactly one tagged attack. Absence of a label is not absence of a migraine, so "ordinary days" almost certainly include untagged attacks.

• A label timestamps typing, not onset. We know when someone logged a migraine, not when it began, how long it lasted or how bad it was. Everything we call "timing" below is the timing of reporting.

• Who logs is not who suffers. The cohort is 90.1% female, against roughly three to one in the population with migraine — the logging channel is far more female-skewed than the condition.

• Nothing here is yet shown to be specific to migraine. The comparison days exclude only migraine labels, and no neutral-label check (for example, "meeting" or "workout") has been run through the same pipeline. This is the largest gap in the work.

What we found:

• Heart rate was about 1 bpm lower near a logged migraine — −0.95 bpm with the device held constant — in 60.2% of the 1,615 people. The result held across every way we sliced it: per-person medians, only people with many readings, each time of day separately, only readings the app rates as high quality, and within each device.

• It sits close to the moment of logging. Within one hour of the label the difference was −1.42 bpm; it peaked in the three hours after the report (about −2 bpm) and was mostly after, not before. Readings taken one to four days away from the report showed a difference 3–5 times smaller (about −0.2 to −0.4 bpm) — the effect belongs to the day of the report, not to the week around it.

• It points the "wrong" way. Pain and migraine are usually associated with sympathetic activation and a faster heart. This is slower, and we do not yet know why.

• It is far too small to identify an attack. Within each person, heart rate separated migraine-label days from ordinary days with a median AUC of 0.530, where 0.500 is a coin toss — and that figure is already flattering, since each person's AUC was oriented to the direction their own data happened to show. Fewer than 12% of people reached 0.70 on any measure.

• The early HRV rise did not survive. An early version of this analysis read "heart rate lower and HRV higher." The HRV half disappeared once we used the app's heart-rate-corrected measure, restricted to high-quality readings, and accounted for device: the phone camera reads RMSSD about 12 ms higher than the same person's watch.

• People switch to the phone camera. Near a migraine label, camera readings rose from 3.39% to 10.68% of all measurements; within individual users, the odds of a camera reading were about 1.7 times higher (Mantel-Haenszel OR 1.69). That switch did not explain the heart-rate change.

• Migraine and headache look the same in the body. Compared with a sample of 1,251 people who used the one-tap headache tag, the physiology was indistinguishable — any difference big enough to matter would have shown up, though a very small one could still be hiding. That null is bounded rather than clean: the smallest difference this comparison could have detected was a Cliff's delta of 0.070, and on the subset where we could resolve each person's time zone (1,148 against 638 people) a difference did appear — d −0.06, p=0.039 — so the null does not survive that restriction. The headache group is also a 1% sample whose own case/control response was never established; it works as a comparator, not as a studied group in its own right. The differences were in how and when people reported — the hour of day, the day of the week and the notes attached to a measurement.

• People differ, consistently. Among the 377 people in this analysis, drawn from the 384 who had logged at least four attacks covered by measurements, the same person tended to respond the same way across their own attacks. But nothing we know about a person (age, sex, weight, how often they log) tells us in advance which way they will respond. A stable response is not an attribution either: someone who always logs while lying down would produce the same kind of consistency.

• heart rate about 1 bpm lower near a logged migraine, across 1,615 people — that your heart rate can tell you a migraine is coming

• strongest in the hours after the report — anything about when attacks begin — labels mark typing, not onset

• no reliable HRV change after corrections — that HRV is irrelevant to migraine — only that this data does not show a change around labels

• no physiological difference we could detect between migraine and headache labels — that migraine and headache are the same condition — they are defined by symptoms, not heart rate

• more phone-camera readings near a label — a measurement error explaining the result — it did not

One more finding belongs to our sibling article on stress: migraines were reported about 11% less often on weekends than these same people's weekend app use would lead you to expect (5,904 reports from 1,148 people) — the opposite of the popular "weekend migraine" idea. Again, that is reporting, not onset, and engagement confounds every rate of this kind: people who feel unwell open the app more often, and normalising for someone's own activity only half-solves that. We unpack it in stress, let-down and migraine.

So what can HRV and heart rate tell you about your migraines?

Put the research and our data together and a fairly clear division appears.

The honest middle ground is this: HRV and resting heart rate are useful as context. They show how your body has been coping — short nights, heavy weeks, illness — and let you look at what the days before your attacks looked like, next to your diary. They are not a migraine sensor.

Living with migraine: the questions people ask next

Does a migraine raise your heart rate?

It can feel that way, and palpitations are listed among the autonomic symptoms that can accompany migraine attacks, according to Pavelić and colleagues' 2024 review. But the measured picture is mixed. In a small 24-hour study, HRV fell sharply during migraine attacks, and in a Fitbit study of 350 people, maximum heart rate was lower on headache days, mostly because people moved less. In our own data, heart rate near a logged migraine was about 1 bpm lower, not higher. If a migraine comes with a racing heart plus chest pain, shortness of breath or fainting, treat that as a medical emergency rather than a migraine symptom.

Is migraine bad for my heart in the long run?

This is a fair worry and deserves a real conversation with your doctor. A 2009 BMJ meta-analysis by Schürks and colleagues found migraine linked to a higher risk of ischaemic stroke (pooled relative risk 1.73), concentrated in migraine with aura (2.16); for migraine without aura the increase was not statistically significant. Risk was higher in women, people under 45, smokers and women using oral contraceptives. The same analysis found no overall link between migraine and heart attack or death from cardiovascular disease. These are group-level associations, so make sure your doctor knows if you have aura, and look after blood pressure and smoking — rather than reading meaning into a single HRV reading.

Can I exercise with migraine, or will it make things worse?

For many people, regular exercise helps with migraine rather than hurting. A 2022 network meta-analysis by Woldeamanuel and Oliveira pooled 21 clinical trials with 1,195 people with migraine: strength training reduced monthly migraine frequency by about 3.5 days, high-intensity aerobic exercise by about 3.1, and moderate aerobic exercise by about 2.2, compared with placebo. During an attack is different — ICHD-3 lists pain made worse by routine activity as a feature of migraine itself. Start gently if exercise has triggered attacks for you before, and ask your doctor about a plan that fits your pattern.

Which doctor should I see for migraine, and will they run tests?

Most people start with a primary care doctor, who can diagnose migraine and refer you to a neurologist or headache specialist if attacks are frequent, change character, or do not respond to treatment. Migraine is diagnosed from your history, using ICHD-3 criteria such as attack length, pain features and associated symptoms — there is no blood test, scan or HRV reading that confirms it. Scans are for ruling out other causes when red flags are present. Autonomic tests like tilt-table testing are used in research and when a doctor suspects a separate condition such as POTS; the 2024 meta-analysis notes there are no standard reference values for using them in migraine care.

Does poor sleep make migraines worse, and does it show in HRV?

Sleep and migraine are closely linked. A 2021 meta-analysis by Stanyer and colleagues found that adults with migraine report clearly worse sleep quality than people without it (Hedges' g 0.75), with a larger gap in chronic migraine, and sleep studies showed less REM sleep. A short night also moves HRV and resting heart rate the next morning, so a low reading before an attack may simply reflect a bad night rather than the migraine itself. The practical use of tracking is to see whether your attacks tend to cluster after short or irregular nights — a pattern to bring to your doctor, not a cause you can prove.

Can migraines follow my menstrual cycle?

Yes, for some people. ICHD-3 describes menstrual migraine as attacks that occur on day 1 of bleeding plus or minus two days — from two days before to the third day of the period — in at least two out of three cycles; when attacks happen only then, it is called pure menstrual migraine. ICHD-3 recommends a prospective diary for research purposes, and many clinicians ask for one. If your phone syncs cycle data from Apple Health, Samsung Health or Health Connect, you can see your cycle next to your logged attacks. Our guide to menstrual migraine covers what to ask about.

Will my migraines get better, and what about long-term medicines?

Migraine often changes over a lifetime, and there are more treatment options than a decade ago. The American Headache Society's 2021 consensus statement describes acute treatments — including newer classes such as gepants (CGRP receptor antagonists) and a 5-HT1F receptor agonist — and preventive options including anti-CGRP monoclonal antibodies, plus neuromodulation devices and behavioural therapies such as CBT, biofeedback and relaxation training. Which of these suits you, and for how long, is your doctor's decision. One thing worth knowing: frequent use of acute headache medicines can itself make headaches more frequent — ICHD-3 calls this medication-overuse headache — so count your medication days. Our migraine treatment guide goes further.

Could my headaches be a sign of something worse?

Most migraines are exactly what they seem, but some headaches need urgent care. Call 911 or go to an emergency department for a sudden, explosive headache ("the worst headache of my life"); a headache with new weakness, numbness, confusion, trouble speaking or vision loss; a headache with fever and a stiff neck; or a headache after a head injury. See a doctor promptly for a headache pattern that is new or has changed — especially a new headache after 50 — a headache that is steadily getting worse, or one that changes with posture or is set off by coughing or straining. These warning signs are drawn from the SNNOOP10 red-flag list published in Neurology, which uses onset after 65 as its formal age cutoff.

How to bring this up with your doctor

Most appointments are short. A little preparation makes them count.

Lead with the pattern, in one sentence. For example: "I have about six migraine days a month, most with nausea and light sensitivity, and they have become more frequent since spring." Frequency, main features and change over time are what a clinician needs first.

Bring a diary, not just an impression. Dates, how long each attack lasted, how bad it was, aura if any, what you took and on how many days. If you want a template and a sense of how long to keep one, see our headache diary guide. Medication days matter because of medication-overuse headache.

Mention the early symptoms. Yawning, fatigue, neck pain, food cravings or light sensitivity hours before the pain are useful clinical information, and they change how some treatments can be timed. Your doctor decides whether that applies to you.

Mention autonomic symptoms separately. Dizziness or a racing heart on standing, fainting, or a pulse that feels wrong outside of attacks are worth raising as their own topic — they may point to something that deserves its own evaluation, such as POTS.

Bring data as context, and say so. "Here are my morning readings and sleep for the last month, next to my diary" is useful. "My HRV shows I'm about to get migraines" is not something the evidence supports, and a good clinician will say so. Present trends, not conclusions.

Ask these directly: Is this episodic or chronic migraine? Would preventive treatment make sense given my number of migraine days? Am I at risk of medication-overuse headache? Does my aura change anything about my other health risks?

If you feel dismissed, ask what would change their assessment, and whether a referral to a headache specialist is possible. Migraine is common — and being common does not make it minor.

How Welltory helps — and what it cannot do

The limits first. Welltory is a general wellness product, not a medical device. It does not diagnose, predict, monitor, prevent, treat or mitigate migraine or any other condition, and nothing in it is a warning system for attacks. It does not replace a headache diary or your neurologist.

What it can add is the physiological layer next to your diary. For a headache diary, there are many good options. If you want to see how your nervous system — HRV and heart rate — looked in the days around your attacks, Welltory shows these as personal trends. You can take a short spot reading with your phone camera (fingertip over the camera and flash), or Welltory can read measurements from a compatible chest strap, Apple Watch or Samsung Watch. A camera reading is one snapshot: you need to sit still, and it cannot replace continuous heart-rate monitoring or an ECG. How accurate is it? See how accurate HRV from a phone camera is.

A routine that fits migraine:

1. Log the episode and what came before it. If you use an iPhone with an Apple Watch or Oura ring, the Today screen flags stress stretches during the day and asks "What happened?" — tap a suggested tag, type a few words, or just talk: "migraine," "aura," "short night," "long call," "period day 1." You can also add a note any time with the plus icon ("Share your thoughts…"); it goes into your Journal.

2. Look at the days before, calmly. Your morning HRV reading and resting heart rate are shown as your own trends; with an Apple Watch or Oura on iPhone, so are sleep, stress minutes and Battery. Read them next to how you felt — and when the number and your body disagree, believe your body.

3. Give My Patterns a few weeks. After roughly two to three weeks of tagging — at least 7 tagged occurrences in the current month, with some history from the month before — My Patterns shows which tagged situations tend to come before stressful stretches, tags that don't happen often but hit your body hard, day-of-week trends, heart rate during those stretches, and a full list of every time a tag appeared. Tags in My Patterns come only from these stress and rest stretches, which rely on heart-rate data through the day — so for now My Patterns needs an iPhone with an Apple Watch or Oura, and notes added on their own do not feed it.

4. Build a personal log for your doctor. Journal shows your HRV measurements, tags, mood and how you feel physically, notes, workouts and menstrual cycles synced from Apple Health, Samsung Health or Health Connect. From the web app you can export a CSV (Dashboard → choose a chart → Export) and bring it with your own diary to your neurologist appointment.

If you'd like a structured way in, try a guided migraine check-in in Welltory.

Whatever shows up, treat it for what it is: patterns are associations to discuss with your doctor, not proof of a trigger and not a warning system.

How we made it

Made with AI tools, then edited and fact-checked by the Welltory team. See our Editorial & AI policy.

Data analysis by Jane Smorodnikova, co-founder of Welltory and the person who built the methodology behind how we read physiological data.

Written by Tatsiana Yashyna.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data