Migraine and Sleep: Insomnia, Sleep Quality, CBT-I and the Bidirectional Evidence

Bottom line: Migraine and sleep are connected in both directions, but “poor sleep causes migraine” is too simple. Migraine attacks can disrupt sleep, insomnia can increase migraine burden, irregular schedules can change vulnerability, and several sleep disorders can coexist with migraine. Subjective sleep complaints are consistently worse in people with migraine, while objective sleep studies show a much more heterogeneous picture. Treating insomnia is worthwhile on its own and may also reduce headache burden, but sleep treatment should complement—not replace—standard migraine care.[1-9]

Migraine and sleep form a feedback loop

Migraine is often described as being “triggered by lack of sleep.”

That can be true for some people, but the relationship is broader.

A person with frequent migraine can experience:

• pain that delays sleep onset;
• nausea or sensory sensitivity that fragments sleep;
• anxiety about the next attack;
• medication effects;
• disrupted daytime activity;
• irregular sleep after an attack; and
• compensatory oversleeping after sleep loss.

At the same time, poor sleep can increase pain sensitivity, emotional reactivity and physiologic stress and can destabilize routines that matter for migraine control.[1,9]

The most useful model is therefore bidirectional rather than one-way.

People with migraine report substantially worse sleep

A 2021 meta-analysis found adults with migraine had significantly worse Pittsburgh Sleep Quality Index scores than healthy controls, with a pooled standardized effect of about g = 0.75.[4]

The subjective sleep difference was larger in chronic migraine than episodic migraine in that analysis.[4]

That is a meaningful and fairly consistent signal:

People with migraine tend to feel that their sleep is worse.

But subjective sleep quality and objective sleep architecture are not interchangeable outcomes.

Objective sleep is disrupted too—but not in one universal way

A 2026 systematic review examined 30 studies using objective measures such as polysomnography and actigraphy in migraine and cluster headache.[2]

The major finding was heterogeneity.

Among adult migraine studies:

• about half of studies that assessed sleep efficiency reported a significant reduction;
• most studies reporting number of awakenings found more awakenings, suggesting fragmentation;
• sleep microstructure and arousal profiles differed across studies; and
• there was no single consistent pattern of hyperarousal or hypoarousal.[2]

A smaller number of studies even suggested arousal patterns may change across the migraine cycle, including the night before an attack.[2]

That is a much stronger scientific position than claiming “migraine reduces deep sleep” or “migraine destroys REM.”

REM findings are not a universal migraine biomarker

The earlier 2021 meta-analysis found a small reduction in percentage of REM sleep among adults with migraine compared with controls.[4]

That is real pooled evidence.

But the newer objective-sleep review makes clear that the broader architecture literature is heterogeneous and dependent on study design, migraine phase and population.[2]

So the defensible statement is:

• some pooled analyses detect REM differences;
• objective sleep disruption is plausible and measurable;
• no single sleep-stage signature currently diagnoses migraine or explains every attack.

This is exactly why the sleep cluster separates Subjective vs Objective Sleep.

Insomnia deserves treatment even if migraine remains

Insomnia is one of the most common sleep disorders reported in migraine populations.[9]

That matters because insomnia is not simply “a symptom to tolerate until the headaches improve.”

If someone has chronic migraine and chronic insomnia, both conditions can be treated as real clinical problems.

The evidence for CBT-I is especially useful here.

CBT-I improves sleep in chronic migraine

A randomized pilot trial assigned adults with chronic migraine and comorbid insomnia to brief CBT-I or a behavioral control.[5]

CBT-I produced significantly larger improvements in:

• actigraphy-measured total sleep time;
• actigraphy-measured sleep efficiency; and
• self-reported insomnia severity.[5]

That is important because the sleep benefit was not merely a questionnaire effect.

The study also suggested fewer headaches at follow-up, but the headache result did not survive the study's Bonferroni-adjusted significance threshold.[5]

So this trial strongly supports CBT-I for the insomnia and provides promising, not definitive, migraine evidence.

Combined trial evidence suggests headache benefit—but the evidence base is small

A 2018 sequential Bayesian analysis combined completer data from two randomized behavioral-insomnia trials.[6]

The analysis estimated that the CBT-I groups reduced monthly headache frequency by about 6.2 days more than control, with a 95% credible interval from -9.7 to -2.7 days.[6]

That looks impressive.

But several caveats matter:

• the analysis used completers rather than a simple full intent-to-treat pool;
• the underlying trials were small;
• behavioral controls can themselves affect headache behavior; and
• the result should not be treated as proof that CBT-I is a primary migraine preventive equivalent to established migraine therapies.

A more recent randomized trial of a multidisciplinary program that included CBT for pain and insomnia also reported sustained improvements in both insomnia and migraine outcomes over follow-up.[7]

The direction is encouraging, but the evidence base is still much smaller than for CBT-I's effect on insomnia itself.

The newest systematic review calls the CBT-I evidence promising, not settled

A 2026 systematic review examined 23 intervention studies involving 1,941 participants and explicitly asked both directions of the migraine-sleep question.[1]

It found:

• several migraine-targeted treatments reduced migraine burden, with variable effects on sleep;
• digital CBT-I significantly reduced headache days and improved sleep parameters in included evidence;
• standard CBT-I evidence for migraine outcomes was mixed; and
• small samples, heterogeneous interventions and inconsistent outcome measures limited generalizability.[1]

That is a useful editorial guardrail.

“CBT-I helps insomnia in migraine” is stronger than “CBT-I is proven migraine prevention.”

Treating migraine itself can improve subjective sleep

The relationship also runs the other way.

A 2025 meta-analysis evaluated seven studies with 989 participants receiving oral or non-oral migraine preventive treatment.[3]

Six of seven studies reported improvement in monthly migraine days, and five of six studies reporting sleep outcomes found relevant improvement in subjective sleep quality.[3]

The pooled effect on self-reported sleep quality was large: Cohen's d = 1.09 (95% CI 0.57–1.62).[3]

That sounds decisive, but several limitations matter:

• sleep outcomes were questionnaire-based;
• treatments differed substantially;
• some treatments were no longer aligned with current international migraine guidance; and
• improving pain can make sleep feel better without proving normalization of sleep architecture.[3]

The conclusion should therefore be:

Effective migraine prevention can improve perceived sleep in people who also have sleep problems, but the effect is not interchangeable with treating a separate insomnia disorder.

Melatonin is a particularly easy place to overclaim

Melatonin is attractive because migraine and sleep both involve circadian biology.

But the 2026 intervention systematic review found no significant melatonin impact within the migraine/sleep intervention evidence it included.[1]

That should not be expanded into “melatonin never helps migraine” or “melatonin never helps sleep.”

Different melatonin questions involve:

• migraine prevention;
• circadian phase shifting;
• general insomnia;
• dose;
• clock timing; and
• population.

Those are separate evidence questions.

See Melatonin Timing vs Dose for why timing matters.

Regularity may matter as much as chasing a perfect bedtime

Migraine care often emphasizes consistency in sleep, meals, hydration and activity.

From a sleep perspective, that makes mechanistic sense: large swings in bedtime, wake time and total sleep can repeatedly alter circadian timing and sleep pressure.

But “consistent sleep” should not become a rigid rule that a person must sleep exactly the same number of minutes every night.

The stronger principle is reducing large, repeated schedule swings where practical.

See Sleep Regularity and Weekend Catch-Up Sleep.

Both too little and too much sleep can be reported as triggers

Migraine diaries often identify both sleep deprivation and unusually long sleep as attack-associated events.

This can seem contradictory until timing is considered.

Long sleep may occur because:

• the person is recovering from prior sleep debt;
• a migraine prodrome already increased fatigue;
• an attack disrupted the previous night;
• illness or medication increased sleepiness; or
• a large schedule shift produced circadian misalignment.

So an attack after “sleeping too much” does not prove that extra sleep itself caused the attack.

Reverse causation is possible.

Morning headache does not automatically mean sleep apnea

Morning headache is often listed as a clue for obstructive sleep apnea.

It can be one clue—but it is not specific.

A 2024 systematic review and meta-analysis included 23 studies with 15,402 patients with OSA.[8]

The pooled prevalence estimates included:

• any headache: 33%;
• morning headache: 33%;
• migraine: 16%.[8]

Yet the pooled relative risk for headache in people with OSA compared with people without OSA was 1.43 with a 95% CI of 0.92–2.25, meaning the overall increase was not statistically significant.[8]

That is a perfect reason not to diagnose OSA from headache alone.

When should sleep apnea still be considered?

OSA evaluation becomes more relevant when headache occurs alongside features such as:

• loud habitual snoring;
• witnessed breathing pauses;
• gasping or choking during sleep;
• unexplained excessive daytime sleepiness;
• resistant hypertension; or
• other established OSA risk factors.

Migraine itself is not an OSA test.

Likewise, treating suspected apnea with a sedative does not address airway obstruction.

See Sleep Apnea vs Insomnia.

Restless legs and circadian disorders can also coexist with migraine

Migraine populations show increased overlap with several sleep disorders, not only insomnia.[9]

Those can include:

• restless legs syndrome;
• circadian rhythm disorders;
• snoring/OSA; and
• other sleep-wake disorders.

That matters because a person with migraine who says “my sleep is terrible” may need more than a generic sleep-quality intervention.

A symptom-level sleep score cannot identify the mechanism by itself.

Migraine medication can improve or worsen sleep depending on the drug and person

Migraine preventive and acute medications are not one sleep category.

Some can be sedating, some activating, and some relatively neutral. A drug that reduces attack frequency may indirectly improve sleep even if it has no direct hypnotic effect.[1,3]

This is why the site should avoid statements such as:

• “migraine preventives improve deep sleep”; or
• “a sedating preventive is better if you have insomnia.”

Medication choice depends on the migraine phenotype, comorbidities, adverse-effect profile and other clinical goals.

Sleep treatment should complement migraine treatment, not replace it

The 2018 clinical review makes a useful point: identifying and treating sleep disorders should be complementary to standard headache care rather than delaying or superseding it.[9]

That principle prevents two opposite mistakes:

1. ignoring insomnia because “the migraine is the real problem”; and
2. treating sleep as though it explains every migraine attack.

Both conditions deserve their own evidence pathway.

What the evidence does not support

Current evidence does not justify claims that:

• poor sleep is the sole cause of migraine;
• every migraine patient has abnormal objective sleep architecture;
• a small pooled REM difference is a diagnostic migraine biomarker;
• CBT-I is already proven to replace standard migraine prevention;
• better subjective sleep after migraine treatment proves normalized polysomnography;
• melatonin is universally effective for migraine because migraine has circadian biology;
• morning headache automatically means OSA;
• OSA is clearly more common among all people with migraine; or
• oversleeping after a migraine necessarily caused the attack rather than reflecting prodrome or recovery.

A better migraine-and-sleep decision tree

When migraine and poor sleep occur together, ask:

1. Is there enough sleep opportunity?
2. Is chronic insomnia present even on nights without migraine?
3. Are bed and wake times highly irregular?
4. Does poor sleep precede attacks, follow attacks, or both?
5. Are caffeine, nicotine, alcohol, cannabis or medication timing affecting sleep?
6. Are there signs of apnea, restless legs or circadian delay?
7. Is migraine itself adequately treated?

That separates a modifiable sleep comorbidity from the broader neurologic disease.

Bottom line

Migraine and sleep influence each other, but the relationship is not one simple trigger pathway.

People with migraine consistently report worse sleep, while objective sleep studies show heterogeneous changes in efficiency, fragmentation, REM and microstructure. CBT-I is evidence-based for comorbid insomnia and may also reduce headache burden, although the migraine evidence is less definitive. Migraine preventive treatment can improve subjective sleep, but that does not automatically treat a separate insomnia disorder.[1-7]

The strongest principle is:

Treat the sleep disorder if it is present, treat the migraine if it is present, and do not force one to explain all of the other.

Related reading

• CBT-I vs Sleep Supplements
• Subjective vs Objective Sleep
• Sleep Apnea vs Insomnia
• Sleep Regularity
• Weekend Catch-Up Sleep
• Melatonin Timing vs Dose
• Caffeine and Sleep Timing