Biofeedback and neurofeedback reduce pain and fatigue in several randomised trials, but since no symptom is specific to fibromyalgia, evaluation remains difficult. Amplitude neurofeedback is the best-supported modality: two independent randomised trials, against an antidepressant and against an attention control, found it superior. HRV biofeedback, often offered, was not shown to help in the one adequately controlled study.
Updated :
What the research shows
EEG neurofeedback — standard amplitude training
SMR and alpha uptraining with theta suppression; theta/SMR ratio
Ages
Adult (18+)
Techniques
Sensorimotor rhythm (SMR), Alpha training, Theta/beta ratio
AAPB3Probably efficaciousNeuroLogic4Efficacioushigher vs AAPB
Level 4 (AAPB level assigned jointly to biofeedback and neurofeedback.) AAPB states the basis of its level: "a final rating of Level 3 – Probably Efficacious applies here because of two successful RCTs using NFB and one using SEMG feedback" (AAPB ch. 20); two of the three trials are neurofeedback, which is what justifies disaggregating upwards. Kayiran 2010: randomised, rater-blind trial of theta/SMR-ratio neurofeedback versus escitalopram over four weeks — both groups improved on the FIQ, pain, fatigue, depression and anxiety, but the neurofeedback group improved significantly more by the fourth week, with the intended reduction in the theta/SMR ratio. Wu 2021: n = 80 randomised (60 neurofeedback, 20 control), 20 thirty-minute sessions over eight weeks of SMR and alpha uptraining with theta suppression, against weekly phone support as an attention control, assessors blinded — significantly greater improvement in pain severity, pain interference, fibromyalgia symptom severity, sleep-onset latency and sustained attention, with the largest effect sizes for pain (BPI) and symptoms (FIQR). Supporting: Caro & Winter 2011 and Barbosa-Torres & Cubo-Delgado 2021; Ribeiro 2023, whose fibromyalgia subset favoured neurofeedback (SMD −0.73, 95 % CI −1.22 to −0.24, p = 0.001).
Level 3. Or-Borichev 2025: double-blind, randomised, dual-control trial, n = 47 (genuine neurofeedback 21, sham neurofeedback 13, treatment as usual 13), 10 sessions of training on a probe representing amygdala activity. Genuine training produced a significant, immediate and lasting reduction in the Symptom Severity Score (F(2,40) = 7.32, p = 0.00, ηp² = 0.27) and the Fibromyalgia Impact Questionnaire (F(2,40) = 9.85, p = 0.00, ηp² = 0.33), sustained at 10-12-month follow-up; the sham arm improved pain in the long term only and affected no other disease measure; the treatment-as-usual group did not change. Pre-registered (NCT02146495). This is the strongest control design in either AAPB rating. Level 3 rather than 4 solely because the control arms hold 13 participants each and the result is unreplicated outside the originating group — Gurevitch 2024 is a mechanistic companion analysis, not an independent replication.
Source-localised effective-connectivity neurofeedback (alpha, pgACC to S1)
Ages
Adult (18+)
Techniques
Coherence / connectivity
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Anderson 2025: randomised placebo-controlled safety and feasibility trial, n = 30 (15 active, 15 placebo), 12 sessions targeting alpha-band effective connectivity from the pregenual anterior cingulate cortex to primary somatosensory cortex. Training was feasible (recruitment of 6 participants per month, mean adherence 80.5 %, dropout 20 %), safe (no adverse events) and highly acceptable (8.0/10), but "active and placebo groups were comparable in their decrease in pain and functional impact"; the authors call for a fully powered efficacy trial. A null result in a feasibility-powered trial against a neurofeedback placebo is not a reason to go below 2 (Parsons 2026), and is not a reason to go above it either. No LORETA row is created: the trial is source-localised but trains effective connectivity.
Biofeedback
Surface EMG biofeedback (trapezius relaxation, forehead and forearm sites)
Ages
Adult (18+)
Techniques
EMG biofeedback
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained (AAPB level assigned jointly to biofeedback and neurofeedback.) AAPB bases its level on "two successful RCTs using NFB and one using SEMG feedback" (AAPB ch. 20): that single trial is Babu 2007, randomised and double-blind, 15 patients receiving surface EMG biofeedback at forearm, forehead and upper shoulder sites versus 15 receiving sham feedback over six daily sessions — both groups improved, but the true-biofeedback group had significantly fewer tender points and significantly lower VAS scores, with the FIQ and six-minute walk test also differentiating the groups. Baumueller 2017: 18 per group, 14 sessions of visual trapezius SEMG feedback versus treatment as usual — the FIQ improved slightly in both arms with no significant between-group difference, trapezius tender-point pain thresholds improved in the treatment group only with assessors blinded, and overall pain experience did not improve. Glombiewski 2013: seven studies, 321 participants, effects mainly on short-term pain intensity, with a subgroup analysis attributing the positive outcomes mainly to SEMG rather than EEG biofeedback. One sham-controlled positive over six days and one 14-session trial null on its primary outcome do not make Level 4.
Resonance-frequency HRV biofeedback
Ages
Adult (18+)
Techniques
HRV — resonance-frequency breathing
AAPB3Probably efficaciousNeuroLogic2Possibly efficaciouslower vs AAPB
Level 2 (AAPB level assigned jointly to biofeedback and neurofeedback.) This row is a disaggregation of the biofeedback box: the AAPB's Level 3 rests explicitly on "two successful RCTs using NFB and one using SEMG feedback" (AAPB ch. 20) — HRV biofeedback is not among them. Its only fibromyalgia-specific study in the AAPB base is Hassett 2007, an uncontrolled pilot in 12 patients: 10 weekly resonance-frequency sessions plus 10 weeks of daily home practice, with significant reductions in pain intensity, improved FIQ and reduced depression at post-intervention and three-month follow-up, limited by the absence of a control group. The 2022-2026 window makes this weaker rather than stronger: Carta 2024, a phase II randomised trial in 64 patients, 10 HRV biofeedback sessions added to standard pharmacotherapy versus standard pharmacotherapy alone, found no between-group differences over time on any primary outcome (FIQ impact, sleep regularity, sense of coherence, depression, pain) or on quality of life. This is the only level proposed below the AAPB's across these two indications, and the reason is specific: the joint rating never tested this modality, and its one adequately controlled test is null.
In short
Clinical reading
AAPB gives a single, joint biofeedback/neurofeedback level for the whole indication: "Level 3 – Probably Efficacious… because of two successful RCTs using NFB and one using SEMG feedback". NeuroLogic separates the modalities. Amplitude neurofeedback rises to 4: Kayiran 2010 (versus escitalopram, rater-blind, with the theta/SMR ratio reduced) and Wu 2021 (n = 80, 20 sessions of SMR and alpha training versus phone support, assessors blinded), two independent groups, supported by Ribeiro 2023's fibromyalgia subset. Surface EMG stays at 3 (Babu 2007 against sham, Baumueller 2017 null on its primary outcome). HRV biofeedback comes down to 2: the joint level does not rest on it and Carta 2024 is null. Two new rows under other methods: amygdala-derived EEG fingerprint at 3 (Or-Borichev 2025) and effective-connectivity neurofeedback at 2 (Anderson 2025). Adult indication only.
Protocols
SEMG over painful regions, resonance-frequency HRV, SMR or qEEG-guided neurofeedback; typically 20 sessions.
Limits
Adult indication only: every trial in the AAPB base concerns adult fibromyalgia and the only juvenile record in the window is an observational HRV study with no intervention — no child or adolescent band is created. The modalities share a principle but are not interchangeable; central sensitisation and the absence of an established mechanism complicate interpretation. No randomised amplitude neurofeedback trial has been published since 2022, and Wu 2021 did not retain its EEG data, so the link between brain change and clinical change remains untested. HRV biofeedback, long supported by a single uncontrolled pilot, has since been tested and the result is null. Trials combining biofeedback with virtual reality or cognitive-behavioural therapy do not isolate the biofeedback component.
Study base
Two amplitude neurofeedback RCTs with blinded assessment from independent groups (Kayiran 2010, Wu 2021), one sham-controlled surface EMG RCT (Babu 2007) and one against treatment as usual (Baumueller 2017), one recent double-blind dual-control limbic neurofeedback RCT with long follow-up (Or-Borichev 2025), and one null phase II HRV trial (Carta 2024). 2022-2026 base: 30 records returned by the search, 9 kept (4 neurofeedback, 3 biofeedback, 2 HRV), 3 with a PDF on file.
Brendan's perspective
Level 4 for amplitude training is a raise I would defend: Kayiran 2010 against escitalopram with a blinded rater, Wu 2021 against an attention control with blinded assessors, two independent groups, twenty sessions in the second. Note what is missing — Wu did not keep the EEG, so we have a clinical effect and no way to tie it to the trained band. That is honest bookkeeping, not a complaint. The row that will annoy people is HRV biofeedback coming down to 2. I use HRV work constantly, and I still think an unregulated autonomic system is noise in the EEG learning loop, but the AAPB level never rested on it and Carta 2024 tested it properly and found nothing. My preference is not evidence. In clinic: qEEG-guided SMR and alpha training, twenty sessions, surface EMG over the regions that actually guard, and HRV as an adjunct I no longer present as training aimed at the pain itself. A word about the condition rather than the methods. Part of what makes fibromyalgia research hard — biofeedback and neurofeedback research included — is that we are still not defining and diagnosing it well. In practice it is very often a diagnosis of exclusion: when it is not the other things, we settle on fibromyalgia. That lets in samples that are almost certainly heterogeneous — some cases psychogenic, some cases the disorder itself, and no clean way to separate them at enrolment. I want to be unambiguous, because this argument gets misused: I believe in the diagnosis. What I want is a better way to define and identify true fibromyalgia, because until we have one, every trial in the indication is measuring a mixture and reporting an average.
Kayıran et al. (2010) Neurofeedback intervention in fibromyalgia syndrome: A randomized, controlled, rater blind clinical trial doi:10.1007/s10484-010-9135-9
Wu et al. (2021) Effects of neurofeedback on fibromyalgia: A randomized controlled trial doi:10.1016/j.pmn.2021.01.004
Baumueller et al. (2017) Electromyogram biofeedback in patients with fibromyalgia: A randomized controlled trial doi:10.1159/000454692
Glombiewski, Bernardy & Häuser (2013) Efficacy of EMG- and EEG-biofeedback in fibromyalgia syndrome: A meta-analysis and a systematic review of randomized controlled trials doi:10.1155/2013/962741
Hassett et al. (2007) A pilot study of the efficacy of heart rate variability (HRV) biofeedback in patients with fibromyalgia doi:10.1007/s10484-006-9028-0
Or-Borichev et al. (2025) Targeted limbic self-neuromodulation for alleviating central sensitization symptoms in fibromyalgia doi:10.1186/s12916-025-04138-3
Anderson et al. (2025) A safety and feasibility randomized placebo controlled trial exploring electroencephalographic effective connectivity neurofeedback treatment for fibromyalgia doi:10.1038/s41598-024-83776-8
Carta et al. (2024) The efficacy of heart rate variability biofeedback training on sleep disorders and impact of fibromyalgia: Results of a phase II randomized controlled trial doi:10.1016/j.jpsychores.2024.111664
Ribeiro et al. (2023) Clinical applications of neurofeedback based on sensorimotor rhythm: a systematic review and meta-analysis doi:10.3389/fnins.2023.1195066
Torres et al. (2024) A systematic review of EEG neurofeedback in fibromyalgia to treat psychological variables, chronic pain and general health doi:10.1007/s00406-023-01612-y
