Can Neck Strengthening Reduce Migraines? The RCT Says Not Yet

Migraine starts in the nervous system, but the neck may still influence the pain circuit. Cervical nerves from the upper spine converge with trigeminal inputs in the trigeminocervical complex, a region involved in head and face pain. That anatomy gives neck-focused exercise a credible rationale. It does not yet give it a proven clinical result.

The only sham-controlled trial of craniocervical strengthening for migraine found measurable changes in neck-muscle activity after eight weeks, yet no statistically significant improvement in monthly headache days or pain intensity versus sham treatment. The study was too small to settle the question, but it also did not deliver the outcome patients actually want: fewer migraines.

What the migraine strengthening trial found

In a 2022 randomized controlled trial by Benatto et al., 42 people with episodic migraine were assigned to either craniocervical muscle-strengthening exercise or disconnected ultrasound used as a sham treatment. Each group had 21 participants. The intervention lasted eight weeks, and researchers followed participants for three months afterward.

The exercise program targeted the craniocervical flexor and extensor muscles. The sham group also received home stretching, which matters: stretching may have mild analgesic effects, so this was not a fully inert comparison.

The co-primary outcomes were headache frequency, measured as days per month, and pain intensity on a 0-to-10 numerical rating scale. The trial found no statistically significant between-group difference in either outcome. The publication does not provide a confidence interval in the evidence summarized here, so it would be wrong to reconstruct an effect estimate or claim a near-miss.

It did find a secondary physiological change. During the craniocervical flexion test, participants in the strengthening group showed lower electromyographic activity in the anterior scalene and upper trapezius muscles. That result indicates that the exercise altered neuromuscular behavior. It is a surrogate outcome, not evidence that migraine attacks fell.

This split result is the honest headline: the muscles adapted, while the clinical endpoints did not.

A null result that cannot close the case

The trial's small size limits what its null result can tell us. If a conventional two-group superiority trial aimed to detect a reduction of two migraine days per month, with a typical standard deviation of three to four days per month, it would need roughly 36 to 72 participants per group for 80% power at a 0.05 significance threshold. Benatto and colleagues enrolled 21 per group.

That does not make the trial positive by implication. Underpowered studies can miss real effects, but they can also produce unstable estimates in either direction. The correct conclusion is narrower: eight weeks of this program did not show a statistically significant advantage over this sham condition in this small trial.

Several design features could have diluted a treatment effect. The researchers did not monitor home-exercise adherence, so the actual training dose remains unknown. Eight weeks may have been too short for a peripheral exercise program to influence a sensitized migraine system. Recruitment through social media and university radio may also have produced a group that differs from the broader migraine population. None of those limitations proves benefit; they explain why one small trial cannot provide a final answer.

Why researchers took the neck hypothesis seriously

The therapeutic evidence is thin, but the cervical findings themselves are consistent. A 2021 systematic review and meta-analysis by Anarte-Lazo et al. found that physical-examination measures, including craniocervical flexion test performance and pressure pain thresholds, distinguish people with migraine from asymptomatic controls and from people with cervicogenic headache. Migraine patients commonly show measurable cervical impairment.

Those findings do not establish that weak or poorly coordinated neck muscles cause migraine. They could be consequences of recurrent attacks, expressions of central sensitization, or both. The same trigeminocervical convergence that makes neck input biologically plausible also permits a two-way relationship: migraine can affect the neck, and neck pain signals may influence head pain.

A 2022 study by Di Antonio et al. supports that dynamic view. In people with episodic migraine, cervical impairments such as craniocervical flexion test performance and pressure pain thresholds varied across the interictal, preictal, ictal, and postictal phases of the migraine cycle. A single neck assessment may therefore capture part of an attack-related state rather than a fixed structural defect.

Measurement quality also matters. A 2024 reliability and validity study by Rodrigues et al. found acceptable test-retest reliability and construct validity for the craniocervical flexion test in people with migraine. That supports its use for measuring cervical function, including the trial's electromyographic secondary outcome. It does not show that improving the test changes migraine burden.

The association extends beyond one adult sample. A 2025 pediatric cross-sectional study reported cervical musculoskeletal dysfunction in children with migraine, while a 2022 study in women with migraine found reduced mobility in neck and back muscle chains. Together, these studies make an age-related degenerative explanation less satisfying. They still remain observational studies.

Which patients might benefit remains unknown

The strengthening trial did not enroll or analyze a subgroup defined by documented cervical dysfunction. That omission matters because any benefit could be diluted if many participants had little cervical involvement to change. Future trials should test whether abnormal craniocervical flexion performance, reduced pressure pain thresholds, segmental hypomobility, or a related profile predicts response.

Aura status does not appear to solve that selection problem. A 2021 study by Carvalho et al. found no significant difference in cervical performance or mobility between migraine with aura and migraine without aura. Aura is clinically important, but this study does not support using it as a shortcut for identifying people with a more cervical-driven phenotype.

Evidence from neck pain offers context, not an answer. A 16-year prospective study by Multanen et al. linked neck muscle strength and cervical mobility with future neck pain and disability in a general population. It did not measure migraine outcomes. The finding supports the idea that cervical function has long-term clinical relevance, but it cannot tell us whether strengthening prevents migraine attacks.

How this compares with other non-drug migraine approaches

Non-pharmacological treatment can reduce migraine frequency under the right conditions. In the EMIKETO randomized trial, a very-low-calorie ketogenic diet reduced migraine frequency more than a hypocaloric balanced diet in people with high-frequency episodic migraine and a body mass index above 27 kg/m². That result does not validate neck exercise, because the population and intervention differed. It does show that a non-drug intervention can move a patient-centered migraine endpoint in a controlled trial.

For craniocervical strengthening, the next study needs more participants, longer treatment, monitored adherence, and preplanned enrollment of people with objectively documented cervical dysfunction. It should preserve headache days and pain intensity as patient-centered endpoints while measuring cervical function and pain sensitivity as supporting outcomes.

The practical verdict

Neck-focused strengthening is a reasonable adjunct to discuss with a qualified clinician when migraine comes with neck pain, stiffness, or tenderness. It should not replace established acute or preventive migraine care, and the current evidence does not support presenting it as a standalone migraine treatment.

The evidence supports a restrained claim: people with migraine often have measurable cervical impairments, and targeted exercise can change neuromuscular measures. One small sham-controlled trial did not show fewer headache days or lower pain intensity. The hypothesis remains open, not confirmed.