Biofeedback is among the better-supported approaches for chronic adult headache: several randomised trials show reduced attack frequency and intensity, usually alongside medication rather than instead of it. Neurofeedback, by contrast, has barely been studied for this indication.
Level 1. AAPB contains no controlled neurofeedback study for adult headache: two case reports only (Linden 2015, 22 then 40 sessions of qEEG-guided neurofeedback in an athlete with post-concussion headache; Martic-Biocina 2017, 25 sessions combining EEG neurofeedback, respiration and vascular training in one chronic migraine patient: attacks reduced by 50 %, contribution of each component not separable). The 2022-2026 search adds no clinical amplitude neurofeedback trial. Golshan 2025 (n = 101 randomised, 61 analysed: 8 weeks of 10-minute daily Muse-headband 'neurofeedback mindfulness' vs audiobook) found improvement in both arms, inconclusive between-group differences and a significant advantage only on headache-management self-efficacy; a consumer headset used without a clinician, it is not clinical neurofeedback and is not counted. Infra-low frequency training has its own row.
Other neurofeedback methods
Infra-low frequency (ILF) / infra-slow (ISF)
Ages
Adult (18+)
Techniques
Infra-low frequency (ILF)
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Two controlled studies, neither replicated. Arina 2022: randomised cross-over, 10 sessions of infra-low frequency neurofeedback and 10 sham sessions in tension-type headache, with a psychoeducation session and emotional support throughout; modelled headache probability fell with genuine sessions and not with sham (sample size, blinding and effect sizes not given in the abstract). Carlson 2025: 87 combat veterans with post-concussive headache, sleep and attention symptoms randomised to 20 half-hour ILF sessions or 8 weekly 15-minute health discussions, treatment as usual continued in both arms, intention-to-treat analysis; headache improved significantly from baseline to end of the intervention (p < 0.0001), but the abstract reports within-group change rather than the between-arm contrast, and the population is post-traumatic rather than primary headache. One adequately sized trial in a neighbouring indication and one small cross-over study: possibly efficacious, without independent replication in primary headache.
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 maintained. The AAPB rating rests on three meta-analyses (Nestoriuc & Martin 2007, migraine: medium average effect vs no treatment, weak specificity; Nestoriuc 2008, tension-type headache: 53 studies, medium to large effect on frequency stable over 15 months, medium vs placebo and small vs relaxation; Lee 2019) and on independent randomised trials: Kolbe 2020 (n = 70, waitlist, medium to strong effects on frequency, duration and intensity), Li 2016 (n = 100, fewer and shorter headaches and less medication than drug therapy at 3-12 months), Tehrani 2021 (n = 86, biofeedback plus medication superior to medication alone), Rausa 2016 (n = 27, EMG in medication-overuse headache: frequency reduced only in the biofeedback arm, no group difference at one year) and Odawara 2015 (n = 27, waitlist). The 2022-2026 window adds Paudel 2025 (9 RCTs, n = 558: frequency and severity reduced vs waitlist; no significant difference in frequency vs pharmacotherapy or CBT in 3 trials, n = 278), Goel 2026 (n = 100, four arms: biofeedback, CBT and their combination all reduced MIDAS disability vs treatment as usual) and Poole 2026 (n = 279, app-based self-administered biofeedback vs waitlist: 0.9 fewer monthly migraine days, 95 % CI 0.3-1.5). Randomised superiority over waitlist and over pharmacological care in independent settings: Level 4 criteria met. Level 5 is not assigned: superiority over a credible placebo or a reference treatment is not replicated (Martino Cinnera 2023, EMG: pooled intensity ES 0.21, 95 % CI −0.02 to 0.44, no effect on frequency; Treadwell 2025 judged the adult biofeedback evidence insufficient).
In short
Clinical reading
AAPB Level 4 confirmed by NeuroLogic for EMG and thermal biofeedback: independent RCTs (Kolbe 2020, Li 2016, Tehrani 2021, Rausa 2016, Goel 2026, Poole 2026) and meta-analyses (Nestoriuc 2007, 2008; Paudel 2025). Benefit demonstrated against waitlist and as an add-on to pharmacological care; no replicated superiority over a credible placebo or an active treatment. Amplitude neurofeedback: Level 1 (case reports); ILF: Level 2 (Arina 2022, Carlson 2025).
Protocols
Frontal or cervical EMG and digital thermal biofeedback, typically 8-12 sessions with home practice; HRV increasingly used.
Limits
Gains are not uniformly maintained at one year; biofeedback is nearly always delivered with relaxation, making its specific contribution hard to isolate; the HRV app trials (Minen 2021, Cuneo 2023) are null on their primary outcome; the new self-administered digital trials do not inform clinic practice; no controlled trial of clinical neurofeedback. Paediatric evidence is covered under 'Pediatric headache'.
Study base
Three AAPB-indexed meta-analyses and six RCTs; 2022-2026 base: 14 publications indexed in the archive (10 biofeedback/HRV, 3 neurofeedback, 1 mixed), including three meta-analyses and five RCTs.
Brendan's perspective
Level 4 for EMG and thermal biofeedback is right, and it is one of the older honest successes in this field: independent groups, meta-analyses, benefit against waitlist and on top of medication. Level 5 stays out of reach because superiority over a credible placebo or a reference treatment has not been replicated — Nestoriuc 2007 already found the specificity weak, and Paudel 2025 finds no advantage over pharmacotherapy or CBT. The row that should stop a reader is neurofeedback at level 1. Two case reports. For an indication where neurofeedback is marketed without restraint, the adult literature is essentially empty, and the infra-low row sits a level higher not because ILF is better but because someone bothered to run two small controlled studies. In clinic: frontal or cervical EMG, thermal training, HRV, eight to twelve sessions with real home practice. Tell the client the gains are not reliably held at one year.
Nestoriuc & Martin (2008) Meta-analysis of biofeedback for tension-type headache: Efficacy, specificity, and treatment moderators doi:10.1037/0022-006X.76.3.379
Lee et al. (2019) Efficacy of psychological treatment for headache disorder: A systematic review and meta-analysis doi:10.1186/s10194-019-0965-4
Kolbe et al. (2020) Effectiveness of biofeedback for primary headache: A randomized controlled trial doi:10.1055/a-1059-9356
Paudel et al. (2025) Efficacy of biofeedback for migraine: A systematic review and meta-analysis doi:10.1016/j.ctim.2025.103153
Poole et al. (2026) App-based self-administered biofeedback for the prevention of episodic migraine: A randomized controlled trial doi:10.1177/03331024261480724
Treadwell et al. (2025) Behavioral interventions for migraine prevention: A systematic review and meta-analysis doi:10.1111/head.14914
Arina et al. (2022) Infra-Low Frequency Neurofeedback in Tension-Type Headache: A Cross-Over Sham-Controlled Study doi:10.3389/fnhum.2022.891323
