Insomnia is understood as chronic hyperarousal, which biofeedback targets directly. More than twenty adult trials show sleep improvements, but not yet at the level established for CBT-I; recent studies mostly concern HRV biofeedback in patients whose insomnia accompanies another condition.
Updated :
What the research shows
EEG neurofeedback — standard amplitude training
(central SMR with theta and high-beta inhibits, frontal beta-down, alpha) — adult
Ages
Adult (18+)
Techniques
Sensorimotor rhythm (SMR), Beta training, Alpha training
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained. (AAPB level assigned jointly to biofeedback and neurofeedback.) Randomised superiority over an internal control reported by four independent groups: Cortoos 2010 (SMR vs EMG biofeedback, n = 17: only the neurofeedback arm increased polysomnographic total sleep time and improved all sleep-log measures), Basiri 2017 (n = 40: SMR superior to CBT-I and to no treatment on PSQI and ISI), Jeon & Choi 2017 (frontal beta-down vs waitlist, n = 14: PSQI, ISI, pre-sleep arousal and device-derived sleep improved) and Schabus 2013 (SMR vs pseudofeedback, n = 24: slow-wave sleep up, awakenings and PSQI down). Lu 2025 adds a single-blind sham-controlled trial (n = 32, central alpha, remitted depression or anxiety with persistent insomnia): PSQI better than sham through 6 months, protocol-specific alpha change, few objective changes. Not raised to 4: the positive trials are small and mostly self-report; Schabus 2017 (double-blind crossover, n = 30) found subjective gains in both active and sham conditions and no objective improvement; Kwan 2022 (n = 17) found no significant difference from CBT-I; Recio-Rodriguez 2024 pooled five RCTs with controls (including CBT-I and other biofeedback) ahead on PSQI (MD 0.57, 95 % CI 0.13 to 1.01). Under the Parsons 2026 framework the sham-null limits demonstrated specificity without showing absence of effect; replication remains too thin and unblinded for Level 4.
Other neurofeedback methods
qEEG-guided and live z-score neurofeedback (surface)
Ages
Adult (18+)
Techniques
qEEG-guided (incl. z-score)
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Hammer 2011: eight adults with DSM-IV insomnia disorder randomised to an individualised qEEG-guided z-score protocol (n = 3, 15 sessions, four most abnormal sites, amplitude, asymmetry, coherence and phase lag) or modified SMR (n = 5, mean 13.4 sessions); both groups improved on ISI, PSQI and quality of life, total sleep time rose by an average of 61.88 minutes and post-qEEG showed reduced excess delta; AAPB reports within-group change only, with no untreated control. Pérez-Elvira 2019: a single case of 30 sessions of live z-score training, sleep-quality score 86 to 15, sustained at 6 months, with 90.63 % of z-scores in range at the end. This is the floor of Level 2 — one underpowered study without an inert control and one case. No z-score, LORETA or source-space insomnia trial appeared in the 2022-2026 window; surface z-score work is filed here, not as LORETA.
Biofeedback
EMG biofeedback (frontalis downtraining) — adult
Ages
Adult (18+)
Techniques
EMG biofeedback
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained. (AAPB level assigned jointly to biofeedback and neurofeedback.) Randomised superiority of frontalis EMG biofeedback over a no-treatment or waitlist control, replicated by independent groups: Freedman & Papsdorf 1976 (n = 18, sleep-onset latency reduced by 29.66 minutes, superior to control, p < .02, polysomnography-verified), Haynes 1977 (n = 24, superior to control at 1 week and 3 months), Sanavio 1990 (n = 40, all three treatments superior to waitlist, p < .01, gains extended at 3-year follow-up) and, quasi-experimentally, Nicassio 1982 (n = 40, superior to no treatment on sleep-onset latency, p < .01). Not raised to 4: EMG biofeedback never separated from pseudo-feedback (Hughes & Hughes 1978, n = 36, all four arms improved equally; VanderPlate & Eno 1983, n = 36, no between-group difference) nor from relaxation or CBT in any trial, and in Hauri 1981 (n = 48) the untreated control also improved on polysomnographic sleep latency. Nothing on EMG biofeedback for insomnia has been published since 1990.
HRV biofeedback (resonance-frequency breathing) — adult, including insomnia comorbid with a medical or psychiatric condition
Ages
Adult (18+)
Techniques
HRV — resonance-frequency breathing
AAPB0Not ratedNeuroLogic3Probably efficacious
Level 3. The AAPB's only HRV evidence is a single case (McLay & Spira 2009: PSQI 10 to 3 at one week), folded into the joint rating. The 2022-2026 window adds randomised, treatment-as-usual-controlled positives from two independent groups: Hasuo 2023 (open-label RCT, n = 50, incurable cancer with sleep disturbance, bedtime home HRV biofeedback with resonant breathing: actigraphic sleep efficiency, sleep duration and low-frequency HRV improved over 10-14 days, 96 % completion), Hasuo 2025 (exploratory RCT, n = 28, cancer patients with insomnia disorder: sleep efficiency 82.0 % to 87.8 %, p < 0.001; hypnotic use 88.2 % to 51.5 %, p < 0.001) and Yen 2023 (RCT, n = 61 methamphetamine users: HRV biofeedback plus TAU better than TAU on sleep quality and depressive symptoms, sustained at follow-up). Held at 3 rather than 4: all are unblinded, every positive is in a comorbid population rather than primary insomnia, and the two trials against an active background were null — Lin 2026 (n = 44, insomnia disorder: HRV biofeedback added to CBT-I gave no significant additional benefit on ISI or pre-sleep arousal through 6 months, a unique gain only in subjective total sleep time) and Carta 2024 (n = 64 fibromyalgia: 10 sessions added to standard therapy, no between-group difference on any outcome).
In short
Clinical reading
AAPB Level 3 assigned jointly to biofeedback and neurofeedback, adults only, across more than two dozen quasi-experimental and randomised studies. NeuroLogic holds 3 for amplitude neurofeedback (Cortoos 2010, Basiri 2017, Jeon & Choi 2017, Schabus 2013; Schabus 2017 null vs sham; Lu 2025 positive vs sham), 3 for frontalis EMG biofeedback (randomised base 1976-1990) and 3 for HRV biofeedback (Hasuo 2023, Hasuo 2025, Yen 2023 vs treatment as usual; Lin 2026 null as an add-on to CBT-I); qEEG-guided z-score at 2.
Protocols
In adults: SMR uptraining (12-15 Hz) at central sites (C3, Cz, C4) with theta and high-beta inhibits, or frontal beta-down; frontalis EMG downtraining; resonance-frequency HRV at bedtime; 10-25 clinic sessions alongside sleep hygiene. No paediatric protocol is documented.
Limits
Methodological quality is often low: few objective measures (polysomnography, actigraphy), small samples, and several studies showing no superiority over relaxation or sham feedback; subjective and objective outcomes diverge repeatedly; no child or adolescent data in any modality; the positive 2022-2026 trials are in comorbid populations, while trials in diagnosed insomnia disorder are null or show no difference from CBT-I; no agreed session count.
Study base
More than 24 quasi-experimental and randomised trials, adults only; randomised but dated EMG base (1976-1990), small and recent SMR base. 2022-2026 base: 10 publications indexed in the archive (5 neurofeedback, 5 HRV biofeedback).
Brendan's perspective
Level 3 is right, and I would not push it further. Four independent groups randomised neurofeedback against an internal control and won — Cortoos 2010, Basiri 2017, Jeon & Choi 2017, Schabus 2013 — and Lu 2025 adds a sham-controlled trial holding at six months. But Schabus 2017 is a careful double-blind study, and while a null against a partially active comparator is not a null against neurofeedback (Parsons 2026), it is no reason to move up either. Most primary outcomes are self-report; where polysomnography is present it is less flattering. In my practice insomnia is one of the questions classic amplitude training answers well: central SMR with theta and high-beta inhibits, or alpha done properly — eyes closed, auditory feedback, individual peak alpha — with resonance-frequency HRV at bedtime, because an unregulated autonomic system is noise in the learning loop. What I tell clients: CBT-I is still the benchmark, and you will feel better rested before your sleep architecture looks different. I would add hypnosis to that first line — it is a shorter intervention than neurofeedback and it does good work in insomnia. My own position is that neurofeedback is a solid second-line option here, and I do not recommend it to someone who has not already been through CBT-I, hypnosis, or both. There is also a wider point about sleep that this chapter does not capture. A great many clients arrive with subjective sleep complaints that quietly settle over a course of neurofeedback started for something else. Whether those complaints were comorbid, primary or causal is usually impossible to say afterwards. What is certain is that we end up working with sleep in something like nine clinical neurofeedback cases out of ten, whatever the referral said. And the comparison matters: for all the new compounds that have reached the market, the pharmacological gold standard here has long been benzodiazepines, which are a problem even over the short to medium term.
Basiri et al. (2017) Comparison of the effectiveness of cognitive behavioral therapy and neurofeedback: reducing insomnia symptoms doi:10.5539/gjhs.v9n7p35
Cortoos et al. (2010) An exploratory study on the effects of tele-neurofeedback and tele-biofeedback on objective and subjective sleep in patients with primary insomnia doi:10.1007/s10484-009-9116-z
Schabus et al. (2017) Better than sham? A double-blind placebo-controlled neurofeedback study in primary insomnia doi:10.1093/brain/awx011
Sanavio et al. (1990) Behaviour therapy for DIMS: Comparison of three treatment procedures with follow-up doi:10.1017/S0141347300009654
Freedman & Papsdorf (1976) Biofeedback and progressive relaxation treatment of sleep-onset insomnia doi:10.1007/BF01001167
Lu et al. (2025) Evaluation of alpha neurofeedback training to enhance sleep in remitted depression and anxiety sufferers with persistent insomnia doi:10.1016/j.psychres.2025.116401
Recio-Rodriguez et al. (2024) Neurofeedback to enhance sleep quality and insomnia: a systematic review and meta-analysis of randomized clinical trials doi:10.3389/fnins.2024.1450163
Hasuo et al. (2023) Short-term efficacy of home-based heart rate variability biofeedback on sleep disturbance in patients with incurable cancer: a randomised open-label study doi:10.1136/bmjspcare-2020-002324
Lin et al. (2026) Can Addressing Autonomic Hyperarousal with Heart Rate Variability Biofeedback Enhance CBT-I Outcomes in Insomnia Disorder? doi:10.1080/15402002.2025.2549554
