Biofeedback combined with cognitive-behavioural therapy reduces tinnitus-related distress. On the neurofeedback side, alpha/delta amplitude training and fMRI neurofeedback each now have randomised trials behind them — encouraging, but not yet replicated.
Updated :
What the research shows
EEG neurofeedback — standard amplitude training
Alpha/delta ratio, alpha enhancement
Ages
Adult (18+)
Techniques
Alpha training, Other neurofeedback
AAPB2Possibly efficaciousNeuroLogic3Probably efficacioushigher vs AAPB
Level 3, raised from the AAPB rating, which is a pooled neurofeedback level resting on quasi-experimental alpha/delta work (Dohrmann 2007, Crocetti 2011) and one 12-participant RCT (Saki 2016, named Khoramzadeh 2016 in the AAPB text). Since then Jensen 2023 has run a three-arm RCT (n = 94, 10 sessions, 3-month follow-up): no time × group interaction on the Tinnitus Handicap Inventory, but a significant interaction on the Tinnitus Magnitude Index, with both neurofeedback arms (alpha/delta and beta/theta) reducing tinnitus intensity relative to a non-neurofeedback minimal-treatment condition — alpha/delta training was sufficient but not necessary. Ma 2022 (n = 56, 15 sessions, participants blinded, untreated control): THI, THQ and visual analogue scores improved only in the trained group, with an occipital alpha increase. Multiple controlled studies from independent groups therefore support benefit, but the randomised evidence misses the distress primary outcome and no trial has shown superiority over a bona fide treatment: Level 4 is not reached.
AAPB2Possibly efficaciousNeuroLogic3Probably efficacioushigher vs AAPB
Level 3. The AAPB Level 2 is a pooled rating covering both EEG and real-time fMRI neurofeedback; on the fMRI side it rested on a six-participant proof-of-concept study (Haller 2010). Gninenko 2024 has since published a prospective randomised trial (n = 43) comparing 15 weekly real-time fMRI sessions downregulating the auditory cortex with 10 weekly group cognitive-behavioural therapy sessions, the current standard treatment: greater reduction in Tinnitus Handicap Inventory scores in the fMRI arm at 6 months (mean change −28.21 vs −12.09 points, P = .005) and at 12 months (−30 vs −4, P = .01), with sleep, trait anxiety, depression and general functioning improving in that arm only. The mechanistic analysis of the same cohort confirms auditory-cortex downregulation and reduced connectivity with parietal operculum 3 (Gninenko 2026). This is superiority over a bona fide treatment, but from a single trial at one centre with small arms: replication is required before Level 4.
Multifocal tomographic neurofeedback (LORETA and variants)
Ages
Adult (18+)
Techniques
LORETA
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2, a new row: AAPB did not review source-localised neurofeedback for tinnitus. Elmer 2026 trained two matched cohorts of older people with chronic subjective tinnitus to raise the alpha/delta ratio in the bilateral auditory cortex and the theta/beta ratio in the dorsal anterior cingulate cortex and anterior insula, comparing one against two weekly sessions for the same total number of sessions, with resting EEG to 6 months. Tinnitus distress and subjective general health improved in the low-intensity group, but subjective tinnitus perception did not change and no effect was found in the trained regions. One controlled study with well-identified outcome measures, no non-neurofeedback comparator and a null result on the trained neural target: possibly efficacious, no more. Brendan's published position is that source-space methods multiply parameters and reduce interpretive clarity; nothing here separates them from amplitude training.
Biofeedback
EMG, thermal and respiratory biofeedback, embedded in cognitive-behavioural or relaxation treatment
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained. The rating rests on Weise 2008, which randomised 130 patients with chronic tinnitus to twelve one-hour sessions combining EMG down-training of the most reactive cervico-facial sites and autonomic feedback with cognitive-behavioural work, or to a waitlist: annoyance, subjective loudness and perceived controllability improved, with coping cognitions and depressive symptoms improved at 6 months. The companion analysis (Heinecke 2009) shows frontalis, trapezius, masseter and sternocleidomastoid activity falling after the intervention, with no skin-conductance change and no relation between physiological and psychological change. Buizza 2025 adds a second independent setting — 431 outpatients given brief multichannel biofeedback with breathing and relaxation exercises or usual ENT care, with lower Tinnitus Handicap Inventory total and subscale scores at 3 months — but allocation was neither randomised nor described, so it does not support a raise to Level 4.
In short
Clinical reading
NeuroLogic Level 3 for biofeedback (AAPB 3: RCT of biofeedback plus CBT against a waitlist, plus a large non-randomised study against usual care). Level 3 for alpha/delta amplitude training and for fMRI neurofeedback, both raised from the pooled AAPB 2: the first on a three-arm RCT positive on intensity but not on distress, the second on an RCT against group CBT with superiority at 6 and 12 months. Level 2 for tomographic neurofeedback.
Protocols
Cervico-facial surface EMG, thermal and respiratory biofeedback within a structured CBT or relaxation frame; in EEG neurofeedback, alpha/delta or alpha-enhancement protocols at fronto-central or temporal sites, 10 to 15 sessions in the published trials; in fMRI, auditory-cortex downregulation over 15 weekly sessions.
Limits
Biofeedback is never tested apart from CBT or relaxation, and a waitlist control cannot establish specificity. In EEG neurofeedback the best-controlled randomised trial is null on distress and finds the same benefit from a beta/theta protocol: specificity of the alpha/delta target is not established. The fMRI result comes from a single centre. Tomographic neurofeedback produced no neural change in the trained regions. All data are adult: there is no paediatric literature.
Study base
Two historical biofeedback RCTs (same sample) and three randomised or controlled neurofeedback trials since 2022; 2022-2026 base: 6 publications indexed in the archive.
Brendan's perspective
Two raises here, both from the AAPB's pooled Level 2, and they are not the same kind of raise. Alpha/delta amplitude training goes to 3 on Jensen 2023 and Ma 2022 — but read Jensen properly: the distress outcome did not move, and a beta/theta protocol did as well as alpha/delta. Alpha/delta was sufficient, not necessary, which means we have an effect without a demonstrated target. The fMRI row goes to 3 because Gninenko 2024 did something rare: superiority over group CBT, a bona fide treatment, held at twelve months, with target engagement verified in the same cohort. Good trial. It is also a single centre, and fMRI neurofeedback remains science rather than a clinic option. The tomographic row stays at 2: Elmer 2026 moved distress and did not move the trained regions. Tinnitus is not a condition I see often; if I did, I would train what the qEEG shows and keep the CBT frame around it, because that is where the biofeedback evidence actually sits.
Weise et al. (2008) Biofeedback-based behavioral treatment for chronic tinnitus: Results of a randomized controlled trial doi:10.1037/a0013811
Heinecke et al. (2009) Psychophysiological effects of biofeedback treatment in tinnitus sufferers doi:10.1348/014466508X386207
Gninenko et al. (2024) Functional MRI Neurofeedback Outperforms Cognitive Behavioral Therapy for Reducing Tinnitus Distress: A Prospective Randomized Clinical Trial doi:10.1148/radiol.231143
Jensen et al. (2023) Does it matter what is trained? A randomized controlled trial evaluating the specificity of alpha/delta ratio neurofeedback in reducing tinnitus symptoms doi:10.1093/braincomms/fcad185
Ma et al. (2022) Self-Directed Neurofeedback Treatment for Subjective Tinnitus Patients Evaluated by Multimodal Functional Imaging doi:10.1155/2022/5114721
Buizza et al. (2025) A Brief Biofeedback Training, Integrated with Breathing and Relaxation Exercises, in Treating Tinnitus Disorders within Routine Medical Care doi:10.1007/s10484-025-09694-1
Elmer et al. (2026) EEG-Based Multifocal Tomographic Neurofeedback in Older Individuals With Chronic Tinnitus Does Not Lead to Persistent Electrophysiological Changes doi:10.1002/brb3.71293
