Anxiety is where the evidence varies most by precise diagnosis. For panic disorder, respiratory biofeedback (learning to correct hyperventilation with a capnometer) is one of the best-demonstrated approaches in the whole field. For generalised and everyday anxiety, biofeedback and amplitude neurofeedback (alpha or SMR training) both have a solid base, with a large recent trial in which biofeedback outperformed an antidepressant. For phobias the evidence remains preliminary.
Updated :
What the research shows
EEG neurofeedback — standard amplitude training
(alpha, alpha/theta, SMR, frontal alpha asymmetry) — GAD and anxiety as a life problem
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 (AAPB level assigned jointly to biofeedback and neurofeedback.) Randomised superiority over waitlist or no treatment, replicated by independent groups over four decades (Plotkin & Rice 1981; Rice, Blanchard & Purcell 1993; Mennella 2017, frontal alpha asymmetry NF vs an active-control training condition, n = 32, reduced negative affect and anxiety); no significant difference from a bona fide active comparator (Fink 2023, n = 56: alpha/theta NF vs mindfulness group therapy in cancer patients, both improved); Tolin 2020 quantitative review (large pooled effects, heterogeneous quality); protocol-specific EEG change (alpha, asymmetry). Panic disorder has no neurofeedback literature of its own: its Level 5 belongs to respiratory biofeedback (see the biofeedback box).
(alpha, alpha/theta, SMR, frontal alpha asymmetry) — GAD and anxiety as a life problem
Ages
Adolescent (12-17)
Techniques
Alpha training, Sensorimotor rhythm (SMR)
AAPB4EfficaciousNeuroLogic3Probably efficaciouslower vs AAPB
Level 3 (AAPB level assigned jointly to biofeedback and neurofeedback.) Small controlled trials in secondary-school and university students (alpha and SMR protocols, test and performance anxiety) with within-subject or waitlist comparisons; no adequately powered randomised trial in a clinical adolescent sample. Protocols and training logic are those validated in adults.
(alpha, alpha/theta, SMR, frontal alpha asymmetry) — GAD and anxiety as a life problem
Ages
Child (under 12)
Techniques
Alpha training, Sensorimotor rhythm (SMR)
AAPB4EfficaciousNeuroLogic3Probably efficaciouslower vs AAPB
Level 3 (AAPB level assigned jointly to biofeedback and neurofeedback.) Case series and small controlled studies in school or clinical settings; no 2022-2026 publication on neurofeedback for childhood anxiety passed screening. Level assigned by continuity with the adult and adolescent evidence, since the physiology of alpha/SMR training does not differ fundamentally in children.
AAPB2Possibly efficaciousNeuroLogic2Possibly efficacioussame vs AAPB
Level 2 (AAPB level assigned jointly to biofeedback and neurofeedback.) No EEG neurofeedback trial; the only controlled study is decoded fMRI neurofeedback (see fMRI section).
AAPB1Not empirically supportedNeuroLogic1Not empirically supportedsame vs AAPB
Level 1 (AAPB level assigned jointly to biofeedback and neurofeedback.) Recent consumer-headset 'neurofeedback-assisted meditation' trials (Jahani 2026, Fernandez-Crespo 2025, Lee 2024) are null or marginal and are not clinical neurofeedback.
Other neurofeedback methods
LORETA and variants (sLORETA, z-score LORETA)
Ages
Adult (18+)
Techniques
LORETA, qEEG-guided (incl. z-score)
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Wu 2024: swLORETA z-score neurofeedback vs treatment as usual, n = 48 adults with major depression and anxiety symptoms, 10 sessions; anxiety (BAI) reduced vs control. Single study, non-randomised assignment.
Infra-low frequency (ILF) / infra-slow (ISF)
Ages
Adult (18+)
Techniques
Infra-low frequency (ILF)
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Perez 2025 (ISAD): double-blind sham-controlled randomised pilot, n = 60 women with internalising disorders, 6 sessions — large improvement in all arms, no active-vs-sham difference on HADS. One adequately powered controlled study; equal-to-sham is not disqualifying, but there is no treatment-as-usual or active-comparator evidence.
fMRI neurofeedback (specific phobias)
Ages
Adult (18+)
Techniques
fMRI neurofeedback
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. Cushing 2024: decoded fMRI neurofeedback (multivoxel neuroreinforcement), double-blind, 23 adults with specific phobias; amygdala reactivity reduced for the target but not the control phobia. Neural rather than clinical outcome; small sample.
Biofeedback
Respiratory (capnometry) and HRV biofeedback — panic disorder
AAPB5Efficacious and specificNeuroLogic5Efficacious and specificsame vs AAPB
Level 5 maintained. Capnometry-assisted respiratory training (CART): Meuret's randomised trials (2008, n = 37, vs delayed treatment; 2010, n = 41, vs cognitive training) — a bona fide comparator — with demonstrated mechanism specificity (mediation by pCO2), replicated by Kim 2012 (two opposing breathing therapies, both effective). HRV biofeedback: Herhaus 2022, randomised against HRV-sham (n = 52) — panic symptoms reduced and HRV increased; Meuret 2023 (secondary subgroup analysis of a 120-patient asthma trial, 21.7 % with high anxiety): CART superior to slow-breathing feedback on anxiety sensitivity and negative affect, d ≥ 0.54 at 6 months.
EMG, thermal, EDA and HRV biofeedback — GAD and anxiety as a life problem (adult)
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 maintained. AAPB base (EMG/thermal/EDA trials vs waitlist and relaxation) strengthened by Kotelnikova 2026: 188 inpatients with anxiety disorders, three arms, assessor-blind HAM-A — biofeedback superior to escitalopram alone (mean difference −6.30, 95 % CI −9.12 to −3.48) with no significant difference from combined therapy (retrospective registration, per-protocol primary analysis). HRV biofeedback: Goessl 2017 (24 studies, n = 484, between-group g = 0.83 on stress and anxiety) and Lehrer 2020 meta-analyses, multiple RCTs; the 2022-2026 trials are mixed (Saito 2024, active-controlled, 10 sessions: vagal HRV improved but no STAI difference in anxious students; several single-session or app-based studies).
Anxiety in children and adolescents
Ages
Child (under 12) · Adolescent (12-17)
Techniques
HRV — resonance-frequency breathing, EMG biofeedback, Other biofeedback
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 maintained, with a caveat: the AAPB level is pooled across ages and the paediatric literature of its own is thin. Thabrew 2026 (CBT + biofeedback game vs placebo game, n = 60, ages 8-18): improvement in both arms, no difference; VR-biofeedback meta-analysis (Kothgassner 2022: 7 studies, n = 191, self-reported anxiety g = −0.28; no advantage over 2D biofeedback). Little clinician-delivered biofeedback in children.
AAPB1Not empirically supportedNeuroLogic1Not empirically supportedsame vs AAPB
Level 1. Premkumar 2024: heart-rate + frontal-alpha-asymmetry biofeedback added to self-guided VR exposure for social anxiety (n = 72 randomised, 25 completers per arm; corrigendum issued) — both arms improved, biofeedback steadied FAA and self-rated arousal but gave no additional symptom benefit.
By age
Child (under 12)
In children, anxiety has been little studied with these methods: a few small studies and one recent trial in which a biofeedback game did no better than a placebo game. The levels shown rest mostly on the adult evidence.
Adolescent (12-17)
In adolescents, alpha or SMR training has been tested mainly for test and performance anxiety in small, encouraging studies; respiratory and HRV biofeedback rests on the adult evidence.
In short
Clinical reading
AAPB rating with joint biofeedback/neurofeedback levels: panic 5, GAD and anxiety as a life problem 4, phobias 2, other 1. NeuroLogic separates the modalities: the panic Level 5 is carried by respiratory (Meuret, Kim) and HRV (Herhaus 2022) biofeedback; amplitude neurofeedback holds Level 4 in adults for GAD and anxiety as a life problem (Rice 1993, Mennella 2017, Fink 2023, Tolin 2020), 3 in adolescents and children. Kotelnikova 2026 (n = 188, assessor-blind HAM-A) is the strongest trial in the AAPB base.
Protocols
Respiratory biofeedback and capnometry (correcting hypocapnia), resonance-frequency HRV, frontal EMG, EDA and thermal; in neurofeedback, alpha training (eyes closed, auditory feedback), alpha/theta, SMR, frontal alpha asymmetry, 10-20 sessions.
Limits
No 2022-2026 randomised trial of clinical neurofeedback in a diagnosed anxiety sample; the new consumer-headset and app studies do not inform clinical practice; little paediatric data; categories other than panic and GAD remain little studied.
Study base
Mature biofeedback literature (four decades, several HRV meta-analyses); older and smaller neurofeedback base. 2022-2026 base: 50 publications indexed in the archive (14 neurofeedback, 36 biofeedback/HRV).
Brendan's perspective
The adult level 4 for amplitude training is the same as the AAPB's, and it is earned rather than generous: four decades of randomised superiority over waitlist, Mennella 2017 against an active training control, Fink 2023 finding no difference from mindfulness group therapy. I drop children and adolescents to 3 because the trials there are students with test anxiety, not clinical samples. In my practice anxiety is mostly alpha work done properly — eyes closed or dim, auditory feedback, individual alpha frequency, the protocol taken from the qEEG rather than from a menu — with SMR where arousal is the driver, and HRV training first if the breathing is disorganised — HRV is the autonomic work I actually run. Panic is the exception: that level 5 belongs to capnometry, not to the EEG, and I say so. What I would like to see tested there is something I have started doing and cannot yet defend with a trial — exposure work delivered simultaneously with HRV biofeedback and neurofeedback, the regulation training running during the exposure rather than as a preparation for it. I present that to clients as experimental, because it is. It is also the direction in this indication that I think holds the most real promise, and it needs proper research. What nobody can promise yet: there is no assessor-blind trial of clinical neurofeedback in a diagnosed anxiety sample, and consumer headsets do not fill that gap.
Kim et al. (2012) Opposing breathing therapies for panic disorder: A randomized controlled trial doi:10.4088/JCP.11m07068
Tolin et al. (2020) Biofeedback and neurofeedback for anxiety disorders: a quantitative and qualitative systematic review doi:10.1007/978-981-32-9705-0_16
Kotelnikova et al. (2026) Mindfulness Development via Biofeedback for Anxiety in Hospitalized Patients: A Randomized Controlled Trial doi:10.3390/brainsci16070748
Herhaus et al. (2022) Effect of a Biofeedback Intervention on Heart Rate Variability in Individuals With Panic Disorder: A Randomized Controlled Trial doi:10.1097/PSY.0000000000001031
Mennella, Patron & Palomba (2017) Frontal alpha asymmetry neurofeedback for the reduction of negative affect and anxiety doi:10.1016/j.brat.2017.02.002
