Neurofeedback therapy is a type of brain training that measures your brain’s electrical activity in real time and rewards you when your brain shifts toward desired patterns. Sensors placed on your scalp record brainwaves using an EEG (electroencephalogram), and software translates that activity into something you can see or hear, like a video game, animation, or music. When your brain produces the targeted pattern, you get a reward signal (a screen brightening, a character advancing, a tone playing). When it drifts away, the reward stops. Over repeated sessions, your brain learns to reproduce those patterns more reliably.
How the Feedback Loop Works
The core learning mechanism is operant conditioning, the same principle behind training any behavior through rewards. At first, your brain hits the target pattern essentially by chance. These random fluctuations occasionally cross a reward threshold, and when they do, the feedback signal arrives. Your brain links that distinct neural state to the reward, creating what researchers call an internal set-point. A reward-related chemical signal, primarily dopamine, then supports the synaptic changes that help encode that state.
With each subsequent session, your brain gets better at comparing its current activity to the target and adjusting. Early sessions feel largely random. Over dozens of repetitions, the process becomes more efficient and the desired brainwave pattern easier to reproduce. This is why neurofeedback typically requires 20 to 40 sessions, sometimes more, to produce lasting changes.
What a Session Looks Like
A typical session lasts 30 to 60 minutes. A technician or clinician places one or more electrodes on specific spots on your scalp using conductive paste, then attaches reference electrodes (often on the earlobes). You sit in a chair facing a screen. The practitioner selects a training protocol based on your goals and baseline brain mapping, then starts the software. Your job is simply to watch the screen or listen. You don’t consciously “try” to change your brainwaves; the feedback does the teaching below conscious awareness, though some protocols involve deliberate mental strategies.
Sessions are painless. The electrodes only read electrical signals; they don’t send anything into your brain. Most people describe the experience as relaxing, though it can also feel mentally tiring, similar to sustained concentration.
Common Training Protocols
Not all neurofeedback is the same. The protocol your practitioner chooses determines which brainwave frequencies get rewarded and where on the scalp the sensors go. The most researched protocols include theta/beta ratio training, which aims to reduce slow-wave activity and boost faster, attention-related waves; sensorimotor rhythm (SMR) training, which targets a specific frequency band associated with calm focus and motor stillness; and slow cortical potential (SCP) training, which teaches you to shift the overall electrical charge of your cortex up or down on demand.
These traditional protocols use one or two electrode sites. More advanced approaches use 19 or more channels simultaneously. One such method, called LORETA neurofeedback, uses a full cap of electrodes to estimate the three-dimensional source of electrical signals deeper in the brain, rather than just reading activity at the surface. The idea is to train the actual neural networks generating the patterns, not just the signal that reaches the scalp. These multi-channel approaches aim for the kind of spatial precision that brain-imaging-based neurofeedback can achieve, though they remain less extensively studied than standard protocols.
Conditions It’s Used For
Neurofeedback has the strongest research base for ADHD and epilepsy, particularly using the standard protocols mentioned above. Growing evidence also supports its use for autism, sleep disorders, head injuries, and learning difficulties, though the strength of that evidence varies.
ADHD
The ADHD picture is nuanced. A 2024 meta-analysis pooling 20 studies with over 1,200 participants found that when outcomes were rated by evaluators who didn’t know whether the child received real or sham neurofeedback, the overall improvement was not statistically significant. However, when the analysis was restricted to studies using established standard protocols, there was a small but meaningful effect. This suggests that the type of neurofeedback matters considerably, and that not all commercially available approaches have equal support.
PTSD
For post-traumatic stress disorder, the results are more striking. A systematic review found that neurofeedback produced a median symptom reduction of 42.5% in PTSD groups, compared with changes in control groups that ranged from a 37% decrease to a 12% increase. About 79% of participants in neurofeedback groups achieved clinically meaningful improvement (defined as more than a 30% reduction in symptoms), compared to 24% in control groups. That remission rate is comparable to the 62% seen in meta-analyses of established PTSD therapies. Neurofeedback also significantly reduced anxiety symptoms in the PTSD studies that measured them.
Side Effects
Neurofeedback is generally well tolerated, but it can cause transient side effects. A double-blind, sham-controlled study found that most participants reported experiencing some side effects during training. These divide into non-specific effects that come from the training process itself (fatigue, headache, difficulty sleeping) and protocol-specific effects tied to which brainwave frequencies are being trained. The SMR protocol appears to be the most sensitive to producing side effects. In most cases, these effects are mild and resolve between sessions. Serious adverse events are rare in the published literature, but poorly targeted training, especially without proper brain mapping, carries a higher risk of unwanted shifts in mood, sleep, or arousal.
Clinical-Grade vs. Consumer Devices
The equipment gap between a clinical office and an at-home headband is significant. Clinical EEG systems typically use 19 to 64 electrodes with conductive gel, high sampling rates, and sensitive amplifiers designed to detect the tiny voltage fluctuations (measured in millionths of a volt) that the brain produces. Consumer devices from companies like Muse or Emotiv use far fewer electrodes, dry sensors, and lower sampling rates. This means they capture a less complete picture of brain activity and are more susceptible to noise from movement, poor electrode contact, and environmental interference.
That doesn’t make consumer devices useless. They can be suitable for basic relaxation training or meditation support. But for clinical neurofeedback targeting specific conditions, the precision of electrode placement, the quality of signal acquisition, and the expertise of the practitioner interpreting the data all matter. A consumer headset doing generalized “calm” training is a fundamentally different intervention than a clinician running a theta/beta protocol on specific scalp locations based on a quantitative EEG assessment.
Cost and Insurance
In-clinic neurofeedback sessions typically cost $50 to $200 each in the U.S., with rural and suburban clinics on the lower end ($50 to $100) and major metro areas like New York or Los Angeles running $200 to $250. A full course of 30 to 40 sessions can therefore range from roughly $1,500 to $8,000 or more.
Insurance coverage is inconsistent. Many plans still classify neurofeedback as investigational or alternative, which means the sessions themselves are not reimbursed under standard behavioral health benefits. Some insurers will cover related components, such as the initial psychiatric evaluation, the diagnosis, or the general biofeedback billing codes, without covering the full course of neurofeedback training. If coverage matters to you, ask your provider’s office whether they bill under biofeedback CPT codes and check with your insurer before starting. Some clinics offer sliding-scale fees or package discounts for committing to a block of sessions.