Neuromodulation and behavioral care¶
TL;DR — Behavioral and device treatments can reduce migraine burden without systemic drug exposure, but effects are intervention-specific and sham/attention controls matter. CBT meta-analysis supports reductions in headache frequency and disability, with variable study size and methods (Bae 2021, PMID 35056352). For acute attacks, evidence reviews find benefit for remote electrical neuromodulation (REN), non-invasive vagus nerve stimulation (nVNS), external trigeminal stimulation and single-pulse transcranial magnetic stimulation (sTMS), with device-specific effects and no significant overall excess of adverse events versus sham in the 2021 review (VanderPluym 2021, PMID 34128998; Clark 2022, PMID 34698941). Device trials face distinctive bias: participants may infer allocation from sensation, visible muscle effects or absent paresthesia, and sham devices can themselves be active (Huang 2024, PMID 38245660). Sleep, regular meals, hydration, exercise, diary use and stress skills are best framed as a feasible stability plan, not an exhaustive trigger-avoidance regime (Robblee 2019, PMID 31710587). Implanted occipital stimulation has surgical/device complications and uncertain benefit, so it is not equivalent to non-invasive options (Schwedt 2011, PMID 20847083).
Why these interventions belong together—and where they differ¶
| Category | Primary target | Typical role | Main evidence problem |
|---|---|---|---|
| CBT/relaxation | Cognition, coping, arousal, behaviour | Prevention/add-on | Therapist time and attention control |
| Biofeedback | Physiological self-regulation | Prevention | Protocol heterogeneity |
| Mindfulness | Attention and response to symptoms | Prevention/disability | Mixed headache diagnoses and small trials |
| Lifestyle regularity | Sleep, meals, exercise, hydration, stress | Foundation/risk management | Multi-component attribution |
| Non-invasive device | Peripheral/central neural circuit | Acute and/or preventive by device | Sham credibility and device-specific dosing |
| Implanted stimulation | Occipital/peripheral nerve | Refractory chronic disease | Surgery, hardware events and enrichment |
“Non-pharmacologic” is not synonymous with low effort, low cost or no adverse effects. Access to trained therapy, device purchase, daily adherence and time burden can be substantial.
Cognitive behavioral therapy¶
CBT uses education, cognitive reappraisal, relaxation, pacing, problem solving and behavioural experiments to reduce disability and improve attack management. It is not predicated on migraine being psychogenic.
The 2021 meta-analysis found CBT reduced headache frequency and disability versus controls, but studies varied in diagnostic mix, therapist contact, comparator and follow-up (Bae 2021, PMID 35056352). A broader psychological-treatment meta-analysis similarly reported benefit with heterogeneity and risk-of-bias limitations (Lee 2019, PMID 30764752).
| Outcome | Plausible CBT pathway | Measurement caution |
|---|---|---|
| Headache days | Sleep/stress regulation and reduced escalation | Diary adherence can itself change reporting |
| Disability | Less avoidance and improved activity planning | Can improve despite unchanged pain |
| Distress | Reduced catastrophizing/anxiety | Not the same as migraine-frequency efficacy |
| Acute-medication use | Earlier plan and less panic redosing | Under-treatment should not be rewarded |
| Self-efficacy | Skills and predictable response plan | Often self-report and unblinded |
Implementation review found a gap between efficacy studies and routine chronic-headache care, with workforce, reimbursement, referral and fidelity barriers (Perlini 2020, PMID 32450871). Digital delivery may expand reach but can dilute therapeutic support and selectively retain engaged users.
Acute-care behavioral evidence is still largely feasibility-oriented: a narrative review identified relaxation, mindfulness and hypnosis approaches but also major workflow, staffing and comparative-efficacy gaps in emergency and inpatient settings (Rosenberg 2020, PMID 32979092).
Relaxation, biofeedback and mindfulness¶
Relaxation training aims to reduce autonomic arousal and muscle tension; biofeedback provides real-time physiological signals such as thermal or electromyographic feedback. Autogenic training meta-analysis across conditions reported medium-to-large pre–post effects, but only a subset concerned headache and older study methods limit migraine-specific inference (Stetter 2002, PMID 12001885).
Mindfulness meta-analysis for primary headache found possible pain benefit but small, heterogeneous studies and unstable conclusions (Gu 2018, PMID 29578127). Chronic-pain mindfulness-based cognitive therapy improves some psychological outcomes, but disease-general evidence should not be relabelled migraine efficacy (Pei 2021, PMID 33241941).
The most defensible use is preference-sensitive add-on care with explicit outcomes, not a claim that stress causes all attacks.
Lifestyle: regularity over prohibition¶
SEEDS summarizes sleep, exercise, eating, diary and stress management (Robblee 2019, PMID 31710587).
| Domain | Operational target | Harm from overreach |
|---|---|---|
| Sleep | Consistent timing, adequate duration, evaluate sleep disorder | Rigid sleep anxiety and blaming unavoidable disruption |
| Exercise | Gradual regular aerobic/resistance activity as tolerated | Post-exertional attacks if dose escalates abruptly |
| Eating/hydration | Regular meals and adequate fluid | Restrictive “migraine diets” without individual evidence |
| Diary | Track days, medicines, function and cycles | Hypervigilance and excessive trigger attribution |
| Stress skills | Recovery, pacing, relaxation and problem solving | Implying psychological weakness |
Trigger studies are vulnerable to recall and expectation. A factor preceding an attack can be a premonitory symptom rather than a cause—for example, food craving may be produced by the early attack. N-of-1 testing with repeated exposure and minimal restriction is more informative than universal avoidance lists.
Non-invasive device evidence¶
The 2021 acute-treatment review graded REN and nVNS evidence moderate and sTMS/eTNS evidence low, with no significant excess adverse events versus sham overall (VanderPluym 2021, PMID 34128998). A device-specific meta-analysis also found acute benefit but emphasized small numbers and heterogeneity (Clark 2022, PMID 34698941).
| Device | Stimulation site/mechanism | Acute/preventive evidence anchor | Characteristic limitation |
|---|---|---|---|
| sTMS | Magnetic pulse over occipital cortex | Acute migraine-with-aura sham RCT (Lipton 2010, PMID 20206581) | Aura-enriched population; device availability |
| REN | Upper-arm peripheral nerves; conditioned pain modulation | Acute multicentre sham RCT (Yarnitsky 2019, PMID 31074005) | Sensation may unblind; app/device adherence |
| nVNS | Cervical vagus stimulation | PRESTO acute sham RCT (Tassorelli 2018, PMID 29907608) | Primary endpoint/timepoint nuances; training |
| eTNS | Supraorbital/trigeminal stimulation | Preventive sham RCT (Schoenen 2013, PMID 23390177); acute ACME (Chou 2019, PMID 30449151) | Forehead sensation and sham activity |
| ONS | Implanted occipital leads | Refractory chronic-migraine trials | Surgical/hardware burden and paresthesia unblinding |
Single-pulse transcranial magnetic stimulation¶
In a randomized sham-controlled study of migraine with aura, portable sTMS applied at aura onset improved pain-free response at two hours and sustained outcomes without major safety signals (Lipton 2010, PMID 20206581). The aura-specific enrollment means evidence cannot automatically be generalized to every migraine phenotype.
TMS has contraindications related to implanted electronic/metal devices and seizure context that require device-specific screening. A low adverse-event rate in a migraine trial does not replace electromagnetic safety procedures.
Remote electrical neuromodulation¶
REN stimulates upper-arm peripheral nerves below a painful threshold to engage conditioned pain modulation. The multicentre randomized trial found improved 2-hour pain relief/freedom and MBS freedom versus sham with favourable tolerability (Yarnitsky 2019, PMID 31074005).
REN’s separation between arm stimulation and head pain can improve acceptability, but blinding depends on matching a convincing sensation. Pediatric emergency-department pilot work shows feasibility interest but is not definitive comparative efficacy (Orr 2025, PMID 39290050).
Non-invasive vagus nerve stimulation¶
PRESTO randomized 248 adults with episodic migraine to nVNS or sham within 20 minutes of pain onset. nVNS was superior at some early pain-freedom endpoints, while the prespecified 2-hour endpoint interpretation was more nuanced; repeated-use findings require attention to hierarchy (Tassorelli 2018, PMID 29907608).
A systematic review/meta-analysis across migraine and cluster headache found signals for efficacy with heterogeneity by disorder and protocol (Lai 2020, PMID 32166843). In chronic-migraine prevention, the EVENT pilot prioritized feasibility/safety and had a short two-month blinded phase followed by open treatment, insufficient for a definitive preventive effect (Silberstein 2016, PMID 27412146).
External trigeminal nerve stimulation¶
The PREMICE preventive trial used 20 minutes daily for three months and found benefit versus sham on migraine days and responder outcomes, with a small sample (Schoenen 2013, PMID 23390177). A 24-person Italian treatment-naive series added short-term feasibility and safety information but no control group (Russo 2015, PMID 26197977). An acute open-label study suggested pain reduction but lacked a control (Chou 2017, PMID 28580703); ACME subsequently provided randomized sham evidence for acute relief (Chou 2019, PMID 30449151).
Combination with flunarizine showed greater improvement than either alone in one randomized study, but the design does not establish synergy across other drugs or populations (Jiang 2019, PMID 30428122).
Sham integrity and contextual effects¶
Device placebo is not inert by default. Low-intensity stimulation may have biological activity; no sensation can reveal sham; paresthesia intensity can predict guessed assignment. A network meta-analysis found different non-pharmacological placebos produce different preventive outcomes (Huang 2024, PMID 38245660).
Every device trial should report:
- the sham waveform/intensity and whether it could be active;
- allocation guesses by participant and assessor;
- sensation and expectation ratings;
- device-use logs;
- rescue medication;
- prespecified endpoint hierarchy and missing-data handling.
Implanted occipital nerve stimulation¶
ONS requires leads, pulse generator and surgery. Trials in refractory chronic migraine have reported benefit in selected/enriched participants but also lead migration, infection, revision and pain at hardware sites (Serra 2012, PMID 22622909). Subthreshold-stimulation crossover work shows improvement can persist when paresthesia is removed, complicating both mechanism and blinding (Slotty 2015, PMID 24812035).
The ONSTIM feasibility trial’s interpretation was limited by feasibility design and endpoints (Schwedt 2011, PMID 20847083). ONS should remain separate from non-invasive consumer devices in both benefit and risk tables.
Integrating multimodal care¶
| Residual problem | Add-on candidate | Measure after 8–12 weeks |
|---|---|---|
| Attack anxiety/avoidance | CBT | Disability and avoided activities |
| Irregular sleep | Behavioural sleep plan/evaluation | Sleep regularity and headache days |
| Drug contraindication | Device matched to phenotype | Pain freedom/function across attacks |
| High acute-drug days | Preventive device + acute plan | Class-specific medication days |
| Persistent disability despite fewer days | Rehabilitation/CBT | Function, work/school participation |
Behavioral and device care should be layered against a stable baseline so marginal benefit is visible. Adding five simultaneous changes may help but prevents learning what is necessary.
Open questions¶
- Which device is best after a specific drug failure in head-to-head pragmatic trials? (Clark 2022, PMID 34698941)
- Can sham calibration preserve blinding without delivering an active neuromodulatory dose? (Huang 2024, PMID 38245660)
- Which CBT components drive headache-day change versus disability change? (Bae 2021, PMID 35056352)
- Do digitally delivered behavioural programs retain efficacy and reduce access disparities outside highly engaged trial samples? (Perlini 2020, PMID 32450871)
- Can an adaptive N-of-1 platform match individuals to devices and behavioural components using multi-attack consistency? (Yarnitsky 2019, PMID 31074005)
Related pages¶
- Acute treatment — drug options against which acute devices are compared.
- Preventive treatment — pharmacologic prevention and combination questions.
- Chronic migraine and chronification — multimodal care in high-frequency disease.
- Migraine in children and adolescents — pediatric behavioural/device evidence.
- Patient experience and advocacy — access, effort and patient-valued outcomes.
References¶
- VanderPluym JH, et al. Acute treatments for episodic migraine in adults. JAMA. 2021. PMID 34128998
- Clark O, et al. Non-invasive neuromodulation in acute migraine: systematic review and meta-analysis. Neurol Sci. 2022. PMID 34698941
- Bae JY, et al. Cognitive behavioral therapy for migraine: systematic review and meta-analysis. Medicina. 2021. PMID 35056352
- Lee HJ, et al. Psychological treatment for headache disorder: systematic review and meta-analysis. J Headache Pain. 2019. PMID 30764752
- Perlini C, et al. Implementation of psychological interventions for chronic headache. BMC Health Serv Res. 2020. PMID 32450871
- Gu Q, et al. Mindfulness meditation for primary headache pain: meta-analysis. Chin Med J. 2018. PMID 29578127
- Pei JH, et al. Mindfulness-based cognitive therapy for chronic pain. Psychol Health Med. 2021. PMID 33241941
- Stetter F, Kupper S. Autogenic training: meta-analysis of clinical outcome studies. Appl Psychophysiol Biofeedback. 2002. PMID 12001885
- Robblee J, Starling AJ. SEEDS for success: lifestyle management in migraine. Cleve Clin J Med. 2019. PMID 31710587
- Huang Y, et al. Effects of non-pharmacologic placebo treatments on migraine prevention. Acta Neurol Belg. 2024. PMID 38245660
- Lipton RB, et al. Single-pulse TMS for acute migraine with aura: sham-controlled trial. Lancet Neurol. 2010. PMID 20206581
- Yarnitsky D, et al. Remote electrical neuromodulation relieves acute migraine. Headache. 2019. PMID 31074005
- Tassorelli C, et al. Non-invasive vagus nerve stimulation as acute migraine therapy: PRESTO. Neurology. 2018. PMID 29907608
- Lai YH, et al. Cervical non-invasive vagus stimulation for migraine and cluster headache. Neuromodulation. 2020. PMID 32166843
- Silberstein SD, et al. Chronic migraine prevention with non-invasive vagus stimulation: EVENT. Neurology. 2016. PMID 27412146
- Schoenen J, et al. Migraine prevention with a supraorbital transcutaneous stimulator. Neurology. 2013. PMID 23390177
- Chou DE, et al. External trigeminal nerve stimulation for acute migraine: open-label trial. Neuromodulation. 2017. PMID 28580703
- Chou DE, et al. Acute migraine therapy with external trigeminal neurostimulation: ACME. Cephalalgia. 2019. PMID 30449151
- Jiang L, et al. Flunarizine plus transcutaneous supraorbital neurostimulation. Acta Neurol Scand. 2019. PMID 30428122
- Russo A, et al. Transcutaneous supraorbital neurostimulation in de novo migraine. J Headache Pain. 2015. PMID 26197977
- Schwedt TJ. Occipital nerve stimulation for chronic migraine: interpreting ONSTIM. Cephalalgia. 2011. PMID 20847083
- Serra G, Marchioretto F. Occipital nerve stimulation for chronic migraine: randomized trial. Pain Physician. 2012. PMID 22622909
- Slotty PJ, et al. Occipital nerve stimulation: randomized trial of subthreshold stimulation. Cephalalgia. 2015. PMID 24812035
- Orr SL, et al. REN for pediatric migraine in the emergency department: randomized pilot. Headache. 2025. PMID 39290050
- Vekhter D, et al. Behavioral treatments for migraine and pain in acute care. Curr Pain Headache Rep. 2020. PMID 32979092