Neurofeedback: Training Your Brain for Mental Health and Focus
How real-time EEG brainwave feedback trains the nervous system to regulate itself, with evidence for anxiety, PTSD, depression, and cognitive performance
Neurofeedback is a type of biofeedback that uses real-time brainwave monitoring to teach your nervous system to self-regulate. Sensors placed on the scalp read your EEG (electroencephalogram) — the electrical activity of your brain — and a computer translates that activity into audio or visual feedback, usually a game or a sound, that changes based on your brain state. When your brain produces the target patterns, you get a reward signal; when it drifts away, the signal dims. Over many sessions, the brain learns to sustain more balanced activity on its own. Research shows clinically meaningful benefits for PTSD, depression, and anxiety, while evidence for ADHD remains mixed. [2][3][5]
How Neurofeedback Works
The brain is always producing rhythmic electrical activity at different frequencies. Each frequency band is associated with different mental states:
- Delta (1–4 Hz): Deep sleep, unconscious processing
- Theta (4–8 Hz): Drowsiness, mind-wandering, emotional processing
- Alpha (8–12 Hz): Relaxed alertness, calm focus
- Beta (12–30 Hz): Active thinking, concentration
- Gamma (30+ Hz): High cognitive processing, perception
In many mental health conditions, certain frequency patterns become dysregulated. People with anxiety often show excess high-frequency beta or insufficient alpha. People with depression commonly show asymmetry in alpha power between the left and right frontal lobes — a pattern linked to withdrawal and low mood. PTSD is associated with disrupted alpha-theta balance and hyperarousal signatures.
Neurofeedback targets these specific imbalances. In a typical session, you sit comfortably in a chair while watching a video or listening to music. The audio or visual feed is subtly controlled by your brainwaves — when your brain produces the desired pattern, the experience is rewarding and uninterrupted; when it doesn't, the experience briefly dims or pauses. This is operant conditioning applied directly to neural oscillations.
Sessions typically run 30–45 minutes and are delivered in courses of 20–40 sessions. Results from any single session are usually subtle; the cumulative effect over weeks of training is where clinically meaningful changes emerge.
Anxiety
Anxiety is associated with excess high-frequency arousal and suppressed alpha activity, particularly in the frontal and temporal regions. Alpha-theta training — designed to increase slower, calmer frequencies — has been the most studied approach for anxiety.
A controlled study involving 34 participants (17 with diagnosed anxiety disorders, 17 healthy controls) found that a structured three-stage neurofeedback protocol significantly reduced anxiety symptoms and increased gamma power in ways suggesting improved mindfulness and emotional regulation. EEG markers of anxiety shifted measurably over the course of training. [5]
Clinical neurofeedback practitioners report consistent anxiety reductions across a range of anxiety diagnoses, though the research base is still catching up to clinical observation. Unlike medication, neurofeedback has no systemic side effects and does not require ongoing prescription; unlike therapy, it does not require verbal processing of difficult material — which can make it valuable for people who struggle to engage with talk-based interventions.
PTSD
PTSD may be the condition where neurofeedback has accumulated the strongest evidence. Trauma reorganizes the nervous system toward hypervigilance, startle reactivity, and intrusive re-experiencing — patterns that map onto measurable EEG signatures including elevated beta and disrupted alpha-theta transitions.
A 2023 systematic review and meta-analysis of ten clinical trials found that neurofeedback produced moderate beneficial effects on PTSD symptom severity across different trauma populations. Secondary outcomes including depression and anxiety also improved. The review included seven randomized controlled trials, representing a more rigorous evidence base than many complementary interventions have at this stage of research. [2]
Alpha-theta training is the most commonly used protocol for PTSD. In this approach, the client enters a deeply relaxed, near-drowsy state and the feedback guides the brain through the alpha-theta crossover — a point associated with hypnagogic imagery and emotional processing. Some researchers hypothesize that this state facilitates the re-processing of traumatic memories at a neurological level, similar in some respects to what EMDR achieves through a different mechanism.
Depression
Depression research has focused on asymmetric frontal alpha activity. The left prefrontal cortex is associated with approach motivation and positive affect; the right with withdrawal and negative affect. In depression, right frontal alpha tends to be lower (meaning the right hemisphere is more active), creating an asymmetry that tracks with depressive states.
Alpha asymmetry training — increasing left frontal alpha relative to right — attempts to directly correct this imbalance. A 2023 review of 12 studies found that patients with depression showed significant cognitive, clinical, and neural improvements following EEG neurofeedback training, with minimal side effects. The review noted that multiple sessions (typically 20 or more) were needed for sustained effects. [3]
Neurofeedback is not a replacement for evidence-based depression treatments including antidepressants and psychotherapy, but it may serve as a useful adjunctive treatment, particularly for people with treatment-resistant depression or those who cannot tolerate medication side effects.
Cognitive Performance in Healthy Adults
Neurofeedback is used not only in clinical contexts but also for cognitive enhancement in healthy people. The most studied applications are working memory, sustained attention, and executive function — the set of mental control processes that govern planning, impulse control, and cognitive flexibility.
A 2021 systematic review assessed EEG-neurofeedback studies targeting executive functions in healthy adults. While the number of high-quality studies was still limited, the available evidence suggested that alpha and theta protocols can improve working memory and certain executive function measures in people without clinical diagnoses. The effects were modest but consistent across studies. [4]
This is an area of active development — consumer-grade EEG headsets and neurofeedback apps have brought home training within reach, though clinical-grade equipment and professional guidance remain the standard for therapeutic applications.
What to Expect
A typical neurofeedback course involves an initial qEEG (quantitative EEG) brain map — a full 19-channel recording that identifies which frequency imbalances are present and informs the training protocol. This is followed by 20–40 sessions of training, usually delivered 2–3 times per week.
The process is non-invasive and generally comfortable. No electricity is delivered to the brain — the sensors only read brain activity. Side effects are rare; some people report temporary fatigue or mild headaches after sessions, which typically resolve with rest.
Progress is gradual. Most people report noticing changes in sleep quality, anxiety levels, or mental clarity within the first 10–15 sessions, with the most significant improvements appearing after 20 or more sessions. Effects can persist long after training ends, particularly for conditions where neuroplasticity is harnessed to establish new regulatory patterns.
Cost is a significant barrier — professional neurofeedback typically runs $100–$200 per session, and sessions are rarely covered by insurance. Consumer-grade home systems exist but the evidence base for their efficacy lags behind clinical-grade protocols.
See our biofeedback page and meditation and breathwork page for related approaches to nervous system regulation.
Evidence Review
ADHD: The 2025 JAMA Psychiatry Meta-Analysis
Westwood et al. (PMID 39661381), published in JAMA Psychiatry in 2025, is the most comprehensive and methodologically rigorous meta-analysis of neurofeedback for ADHD to date. The review analyzed 38 randomized controlled trials comprising 2,472 participants aged 5 to 40 years.
The headline finding was that probably-blinded assessments of ADHD total symptoms showed no statistically significant improvement with neurofeedback. When analyses were restricted to RCTs using established, standard protocols, a small but significant improvement emerged. Processing speed was the one neuropsychological measure that showed a small, significant improvement across the full dataset.
The conclusion — that neurofeedback does not appear to meaningfully benefit individuals with ADHD at the group level, based on rigorously blinded outcomes — is an important corrective to earlier more optimistic reviews that relied on less blinded assessments (where parent and teacher ratings, which are susceptible to expectation effects, showed larger improvements). [1]
This finding does not mean neurofeedback has no role in ADHD care, but it suggests that any beneficial effects are modest and that many previously reported gains reflected non-specific (expectancy, attention, and structured engagement) rather than specific neurofeedback effects. ADHD currently represents the weakest evidence base among the conditions for which neurofeedback is commonly used.
PTSD: Systematic Review and Meta-Analysis, Askovic et al. 2023
Askovic et al. (PMID 37732560) conducted a systematic review and meta-analysis of ten clinical trials examining neurofeedback for PTSD, with seven of those trials being randomized controlled designs. The review covered diverse trauma populations and multiple neurofeedback protocols.
The meta-analysis found moderate beneficial effects on PTSD symptom severity (pooled effect sizes in the moderate range). Secondary outcomes — depression and anxiety — also improved significantly. Neurophysiological outcomes were heterogeneous across studies, reflecting differences in EEG targets and measurement approaches, making definitive conclusions about mechanism difficult.
The quality of PTSD neurofeedback research is notably better than in some other areas: the presence of multiple RCTs, effect sizes that survive blinding, and improvements in secondary outcomes consistent with the primary neurophysiological rationale lend the PTSD evidence base more credibility than the ADHD literature. Limitations include small sample sizes in individual trials and variation in protocol (alpha-theta vs. beta suppression vs. real-time fMRI neurofeedback). [2]
Depression: Review of EEG-Based Neurofeedback, Patil et al. 2023
Patil et al. (PMID 36682174) reviewed 12 studies of EEG neurofeedback for depression, focusing specifically on frontal alpha asymmetry training and related protocols targeting depression-associated EEG signatures.
All 12 studies reported significant improvements in cognitive, clinical, and neural outcomes following neurofeedback training. Despite this consistent directional finding, the review identified several methodological concerns: small sample sizes, heterogeneous populations, variable session number and protocol, limited control conditions, and inconsistent outcome measurement. The authors concluded that EEG-neurofeedback is a promising intervention for depression with minimal side effects, but that larger, better-controlled trials are needed before it can be recommended as a primary treatment.
The findings align with the neurobiological rationale: frontal alpha asymmetry is a replicated EEG biomarker for depression, and training that directly targets this asymmetry produces improvements in mood that co-occur with measurable EEG changes — suggesting the mechanism is at least partially specific rather than entirely due to placebo or expectancy effects. [3]
Healthy Adults and Executive Function: Viviani and Vallesi 2021
Viviani and Vallesi (PMID 34117795) conducted a systematic review of EEG neurofeedback studies targeting executive functions — working memory, cognitive flexibility, inhibition, and task switching — in participants without clinical diagnoses.
The review identified a limited number of sufficiently rigorous studies to allow meta-analytic pooling, so it remained a qualitative synthesis. The available evidence suggested that theta-alpha protocols can improve working memory performance in healthy adults, with results that sometimes persist beyond the active training period. Effect sizes were generally modest, and the heterogeneity in both protocols and outcome measures makes firm conclusions premature.
The cognitive enhancement application represents an important frontier for neurofeedback research. If even modest, reliable cognitive improvements can be demonstrated in healthy participants through standardized protocols, neurofeedback could become relevant not only as a treatment but as a performance tool — analogous to physical training for the brain. [4]
Anxiety: Controlled Study, Chen et al. 2021
Chen et al. (PMID 34776855) designed a controlled neurofeedback experiment with 34 participants — 17 diagnosed with anxiety disorders and 17 healthy controls. The three-stage protocol included resting-state EEG recording, a mindfulness-based neurofeedback training phase, and post-training assessment.
The study found that gamma wave power increased during and following neurofeedback training in anxiety participants, and that this correlated with reduced anxiety scores and improved mindfulness capacity. The authors interpreted the gamma increase as reflecting greater cognitive processing efficiency and improved top-down emotional regulation — a neurological signature of anxiety relief.
The sample size is small and the study does not include a sham or waitlist control, which limits causal conclusions. It does provide mechanistic insight into what EEG changes accompany the clinical improvements seen in larger anxiety trials. [5]
Overall Evidence Assessment
Neurofeedback occupies an interesting position in the evidence hierarchy: the mechanistic rationale is sound (EEG signatures of psychiatric conditions are real, replicated findings, and operant conditioning of neural oscillations is neurobiologically plausible), the clinical evidence for PTSD and depression is moderately strong, and the evidence for anxiety is promising but underpowered. The ADHD evidence, while the most researched, does not survive rigorous blinding.
For people with treatment-resistant conditions or those seeking alternatives to medication, neurofeedback offers a non-pharmacological option with a favorable safety profile. The main practical barriers are cost, the time commitment (20+ sessions), and limited insurance coverage. For healthy individuals seeking cognitive enhancement, the current evidence supports cautious optimism rather than confident recommendation.
References
- Neurofeedback for Attention-Deficit/Hyperactivity Disorder: A Systematic Review and Meta-AnalysisWestwood SJ, Aggensteiner PM, Kaiser A, Nagy P, Donno F, Merkl D, Balia C, Goujon A, Bousquet E, Capodiferro AM, Derks L, Purper-Ouakil D, Carucci S, Holtmann M, Brandeis D, Cortese S, Sonuga-Barke EJS. JAMA Psychiatry, 2025. PubMed 39661381 →
- Neurofeedback for post-traumatic stress disorder: systematic review and meta-analysis of clinical and neurophysiological outcomesAskovic M, Soh N, Elhindi J, Harris AWF. European Journal of Psychotraumatology, 2023. PubMed 37732560 →
- Review of EEG-based neurofeedback as a therapeutic intervention to treat depressionPatil AU, Lin C, Lee SH, Huang HW, Wu SC, Madathil D, Huang CM. Psychiatry Research: Neuroimaging, 2023. PubMed 36682174 →
- EEG-neurofeedback and executive function enhancement in healthy adults: A systematic reviewViviani G, Vallesi A. Psychophysiology, 2021. PubMed 34117795 →
- Efficacy Evaluation of Neurofeedback-Based Anxiety ReliefChen C, Xiao X, Belkacem AN, Lu L, Wang X, Yi W, Li P, Wang C, Sha S, Zhao X, Ming D. Frontiers in Neuroscience, 2021. PubMed 34776855 →
Transparency
View edit historyEvery change to this page is tracked in version control. If you have conflicting research or think something is wrong, we want to hear about it.