ADHD is neurofeedback's flagship indication: the three standard protocols reach the highest evidence level, with effects that persist after training ends — a notable contrast with medication.
AAPB5Efficacious and specificNeuroLogic5Efficacious and specificsame vs AAPB
Level 5 maintained. Randomised superiority over semi-active, waitlist and treatment-as-usual controls in independent multicentre trials, with remission rates of about 32-47 % and effects sustained at 6-12 months (Arns 2020; Van Doren 2019); no significant difference from bona fide active treatments (working-memory training — Hasslinger 2022, 'no indication that NF was superior to WMT'; self-management training — Korfmacher 2022, similar improvement; EMG biofeedback); protocol-specific neurophysiological change. Sham-controlled trials (ICAN, Wang 2026) show comparable improvement in both arms: under the Parsons (2026) framework this is not read as absence of a specific effect, since the sham condition is itself a partially active control; ICAN's own 25-month follow-up describes a package effect comparable to the MTA reference treatments. Acknowledged limit: NEWROFEED (2022) did not demonstrate non-inferiority of neurofeedback to methylphenidate — but that was at-home training with no clinician present, automatic thresholds and short (< 30 min) sessions, i.e. neurofeedback far removed from clinical practice; the result therefore carries little weight for neurofeedback delivered by a qualified practitioner.
AAPB5Efficacious and specificNeuroLogic4Efficaciouslower vs AAPB
Level 4. In adolescents the evidence rests on one large pragmatic trial (Hasslinger 2022, ages 9-17, n = 202: SCP and z-score superior to treatment as usual on some measures, not superior to working-memory training) and on pooled data from five RCTs spanning ages 6-60 (Fan 2022, mean age 23: inattention SMD −0.48, 95 % CI −0.90 to −0.06, vs waitlist / treatment as usual; no hyperactivity effect). Defined population and protocols, results found across settings, but thinner replication than in children and no demonstrated superiority over a bona fide treatment — hence Level 4 rather than 5.
AAPB5Efficacious and specificNeuroLogic4Efficaciouslower vs AAPB
Level 4. Several independent randomised trials in adults (Schönenberg 2017, Barth 2021, Mayer 2016) show substantial symptom improvement with no significant difference from bona fide active treatments (meta-cognitive group therapy; EMG biofeedback — Barth 2021, n = 67, three arms); superiority over waitlist / treatment as usual for inattention (Fan 2022). Defined population and protocols, result found in independent settings: Level 4 criteria met. Level 5 is not assigned because superiority over a reference treatment has not been shown and the adult base remains small (Ostinelli 2025: 10 neurostimulation/neurofeedback trials, 194 participants, effects inconsistent across raters; Courrèges 2025: no technique reached significant pooled efficacy).
Level 1. For ADHD, LORETA neurofeedback rests on case series and clinical reports (Koberda 2014) with no controlled study or adequately powered sample. The surface z-score arm of Hasslinger 2022 (LZS) is not LORETA and is counted in the main section.
Level 2. One single-blind randomised trial against sham feedback (Ölçüoğlu 2024, n = 100, 60 sessions) shows gains on the WISC-R (verbal, performance and full-scale IQ) but did not measure ADHD symptoms; no independent replication.
Level 2. Two randomised trials in adolescents: Alegria 2017 (n = 31, right inferior frontal target vs active control region: symptom improvement in both arms at post-intervention and 11 months, no between-group difference; only the rIFG arm showed target-region transfer and a trend to better sustained attention) and Lam 2022 (n = 88, double-blind vs sham: no between-group clinical benefit; no overall group × time effect on inhibition performance, exploratory left-IFC activation change only — Lukito 2024). Feasibility and target engagement shown; clinical efficacy not established.
Level 2. Marx 2015 (pilot, n = 27, three arms of 9: NIRS-NF vs EEG-NF vs EMG-BF): significant within-group symptom reduction only in the NIRS arm, no significant between-group difference; Wu 2022 (NIRS-NF vs atomoxetine, parallel groups); more recent fNIRS and HEG pilots. Encouraging but not replicated on clinical outcomes by an independent group.
Level 2. Tinello 2022 systematic review: attentional gains in ADHD subgroups of small trials; Barbini 2026 pilot (n = 17, waitlist). No adequately powered controlled trial.
By age
Child (under 12)
In children, neurofeedback is one of the best-studied non-drug approaches to ADHD. Children who receive it improve substantially and the gains tend to last; it holds up as well as other structured approaches such as working-memory or self-management training. On its own it is less effective than medication. Trials comparing it with a look-alike 'sham' training found similar improvement in both groups — a result that is hard to interpret, because the sham condition is itself an active training rather than an inert placebo (see the recommended reading). Children who also have an anxiety disorder respond less well to the theta/beta protocol.
AAPB Level 5 confirmed by NeuroLogic for TBR/SMR/SCP. Durable superiority over waitlist and TAU; no significant difference from active comparators (WMT, SMT, EMG-BF); inferiority to methylphenidate as monotherapy (NEWROFEED); ICAN: no difference from the pre-recorded-EEG control at 13 and 25 months, but d = 1.6 pre-post and a package effect comparable to MTA arms. Moderators: comorbid anxiety (poorer response), ODD (better response at 13 months), computational cognitive markers.
Adolescent (12-17)
In adolescents the data are thinner. One large Swedish trial shows gains above usual care on some measures but not above working-memory training; pooled adolescent/adult analyses show benefit mainly on inattention. Most conclusions are still extrapolated from child trials.
NeuroLogic Level 4 (AAPB 5 pooled). Hasslinger 2022 (ages 9-17, SCP and LZS in high-frequency format, 5 sessions/week); Fan 2022; ESCAlate 2026 (16-45): no added benefit of NF over counselling or MPH.
Adult (18+)
For adults, neurofeedback has been tested in several good trials. People improve substantially, as much as with other structured approaches such as muscle-relaxation biofeedback or group therapy — which makes it a supported option. Studies remain fewer than in children, and neurofeedback has not been shown to be better than those alternatives or than medication.
NeuroLogic Level 4 (AAPB 5 pooled). Schönenberg 2017 (triple-blind, NF = sham = group therapy, all improved), Barth 2021 (NF = EMG-BF), Mayer 2016 (SCP = EMG-BF), Fan 2022 (inattention vs waitlist); Ostinelli 2025 and Courrèges 2025 inconclusive.
In short
Clinical reading
Level 5 for TBR, SMR and SCP, with medium to large effect sizes, 32-47% remission rates and sustained effects at follow-up. Personalised multimodal neurofeedback shows remission rates comparable to medication monotherapy in controlled settings.
Protocols
TBR or SMR at central sites (Cz, C3/C4), frontocentral SCP; 30-40 sessions with active transfer, often qEEG-informed.
Limits
Several recent trials depart from clinical neurofeedback (at-home training in NEWROFEED, 2-channel headsets and short courses in the mobile trials), though ICAN was clinician-delivered over 38 sessions; sham-controlled trials show no between-group difference (not decisive under the framework used here); EEG learning does not predict clinical improvement (ICAN); little adolescent- or adult-specific data.
Study base
More than 40 RCTs and around fifteen meta-analyses (2016-2026); 2022-2026 base: 75 publications indexed in the archive.
Brendan's perspective
Level 5 in children, same as the AAPB, and I would defend it. Level 4 in adolescents and adults is not a demotion of the method; it is the honest state of two less developed literatures. The sham-controlled trials move me less than their design suggests they should: a sham keeps a substantial share of contingent reinforcement, so a null between arms is not an absence of specific effect (Parsons 2026). ICAN and NEWROFEED get quoted in the same breath and should not be — ICAN was clinician-delivered, 38 sessions with trainers coaching in the room, while NEWROFEED trained children at home with no clinician present, and that second one is not the intervention I deliver. What I deliver is a protocol read off the qEEG rather than picked from a menu: the research says which targets carry the evidence, the map says which of them this child actually needs, and personalising inside what the literature supports is the whole of it. I do not train slow cortical potentials. Nor do I treat comorbidity as background — anxiety, sleep and mood change which protocol I select and how I pace it, and a large share of my protocol decisions are made right there. Thresholds set by hand, twice a week, 30 to 40 sessions, transfer tasks from the first block. What I tell parents: in children the place of neurofeedback is settled; in adults, Schönenberg 2017 and Barth 2021 show real improvement with no demonstrated superiority over a bona fide active treatment, and I say so before the first session.
Gevensleben et al. (2010) Neurofeedback training in children with ADHD: 6-month follow-up of a randomised controlled trial doi:10.1007/s00787-010-0109-5
Gevensleben et al. (2009) Is neurofeedback an efficacious treatment for ADHD? A randomised controlled clinical trial doi:10.1111/j.1469-7610.2008.02033.x
Krepel et al. (2020) A multicenter effectiveness trial of QEEG-informed neurofeedback in ADHD: Replication and treatment prediction doi:10.1016/j.nicl.2020.102399
Arns et al. (2020) Neurofeedback and attention-deficit/hyperactivity-disorder (ADHD) in children: rating the evidence and proposed guidelines doi:10.1007/s10484-020-09455-2
Van Doren et al. (2019) Sustained effects of neurofeedback in ADHD: a systematic review and meta-analysis doi:10.1007/s00787-018-1121-4
Westwood et al. (2025) Neurofeedback for Attention-Deficit/Hyperactivity Disorder: A Systematic Review and Meta-Analysis doi:10.1001/jamapsychiatry.2024.3702
Neurofeedback Collaborative Group (2023) Neurofeedback for Attention-Deficit/Hyperactivity Disorder: 25-Month Follow-up of Double-Blind Randomized Controlled Trial doi:10.1016/j.jaac.2022.07.862
Purper-Ouakil et al. (2022) Personalized at-home neurofeedback compared to long-acting methylphenidate in children with ADHD: NEWROFEED, a European randomized noninferiority trial doi:10.1111/jcpp.13462
Hasslinger et al. (2022) Slow Cortical Potential Versus Live Z-score Neurofeedback in Children and Adolescents with ADHD: A Multi-arm Pragmatic Randomized Controlled Trial with Active and Passive Comparators doi:10.1007/s10802-021-00858-1
