Neurofeedback for beginners: a clear, practical guide
Understand what EEG neurofeedback measures, what a first visit may involve, how to read the evidence, and how to choose and review care.
AI-generated illustrative scenes and narration. EEG traces are synthetic teaching examples.
Read the video transcript
A signal you can learn from
Right now, your brain is producing electrical patterns as you watch, listen, and think. Imagine being able to see part of that activity, and use it as feedback. That is the idea behind neurofeedback. Let's explore what it measures, how training works, and what to expect.
The feedback loop
In E. E. G. neurofeedback, sensors on your scalp record electrical activity. Software measures selected features of the signal and translates them into a changing picture, a sound, or a game. When the chosen target is met, the feedback changes. With practice, the aim is to help you learn to influence that activity. Learning and results vary from person to person.
What EEG measures
E. E. G. stands for electroencephalography. It records tiny voltage differences at the scalp, reflecting the combined electrical activity of many brain cells. The sensors record signals; they do not send electrical stimulation into your brain. An amplifier makes these small signals easier to measure, and software displays them as traces. Eye movements and muscle tension can also affect the recording.
Understanding frequency
One way to describe these rhythms is frequency: how many cycles occur each second. Four cycles in one second means four Hertz. Slower rhythms have fewer cycles; faster rhythms have more. Researchers group frequencies into bands, although their boundaries vary between sources and systems. Several bands can be present at the same time.
Delta, theta, and alpha
Delta, about half a Hertz to four Hertz, is prominent during deep sleep.
Theta, about four to eight Hertz, appears during drowsiness and some stages of sleep. It also participates in memory and other mental processes.
Alpha, about eight to twelve Hertz, often becomes stronger when you close your eyes while awake. It is usually most prominent toward the back of the head. Think of these descriptions as common associations, rather than a way to read someone's thoughts.
Low beta, beta, high beta, and gamma
For this explanation, we divide beta into three ranges.
Low beta, about twelve to fifteen Hertz, includes the sensorimotor rhythm, or S. M. R., when recorded over sensorimotor areas. It is associated with being awake and physically still.
Mid beta, about fifteen to eighteen Hertz, is often discussed in relation to active thinking and attention.
High beta, about eighteen to thirty Hertz, can accompany alertness and arousal. Its meaning depends on context, and muscle tension can contribute to the signal.
Gamma, about thirty Hertz and above, is studied in perception, attention, and how information is brought together. Scalp gamma is especially difficult to separate from muscle activity.
These rhythms overlap. No band is simply good or bad, and faster does not automatically mean better.
An individual training plan
A provider chooses what to measure and how to set the feedback, based on the assessment and your goals. Some use quantitative E. E. G., or Q. E. E. G., which applies numerical analysis to the recording. A colorful map is one way to display those measurements. It needs interpretation alongside other information about you.
What a visit can look like
You sit comfortably while sensors are placed on your scalp. Depending on the approach, you may watch a screen, listen to sounds, or take an active role in a task. Sessions often last twenty to forty minutes, and training generally involves repeated sessions. The schedule and feedback should be explained before you begin.
Evidence and expectations
People explore neurofeedback for attention, sleep, stress, and other goals. Research is ongoing, and evidence varies by condition and training method. Ask about expected benefits, limitations, costs, and possible unwanted effects. Agree on how you will track changes in daily life and when to review the plan. A qualified provider should help you make an informed decision.
Continue exploring
Neurofeedback turns measured brain activity into feedback for practice. To explore the research, learn what to expect, and find questions to bring to a qualified provider, visit neurofeedback dot com.
What neurofeedback is
Neurofeedback uses information about brain activity to provide feedback while a person practices. It is a form of biofeedback, a process that makes a normally hard-to-observe body signal easier to notice and respond to. With EEG neurofeedback, the measured signal is electrical activity recorded at the scalp.
A useful analogy is practicing a skill with a live display, like a musician using a tuner or an athlete reviewing a movement cue. The display can guide practice, but the display alone does not prove that the practice will improve a particular health condition. That requires separate evidence.
You do not need a neuroscience background to ask good questions. Start with three: What are we measuring? What are we trying to change? How will we know whether it helps me?
Measure, provide feedback, practice, review.
The loop describes a process. It does not by itself establish a clinical benefit.
- MeasureScalp sensors record small electrical voltage changes.
- FeedbackSelected signal features affect a visual or sound cue.
- PracticeYou engage with the cue while the provider monitors the session.
- ReviewYou compare the experience and daily-life goals at agreed checkpoints.
A first appointment, step by step
A thoughtful first conversation begins with your goals, history, current care, and questions. The provider may discuss sleep, attention, mood, daily functioning, medications, diagnoses, and other factors relevant to the service. Neurofeedback should not replace needed medical or mental health assessment simply because it uses technology.
Sensors are placed on the scalp using a cap, paste, gel, or individual electrodes. In conventional EEG neurofeedback, these sensors record; they do not stimulate the brain. The provider checks contact and signal quality and watches for artifacts such as eye movement, muscle tension, or motion.
During practice, selected features of the recording affect a visual or sound cue. A video might brighten, music might continue, or a game element might respond. The clinician or trained provider supervises the signal, explains the task, notices comfort and engagement, and decides when the plan needs review. The experience and rules vary by approach.
Before leaving, you should understand what was done, what you are being asked to notice between visits, what changes to report, and when progress will be reviewed. Bring a medication list, relevant diagnoses, current providers, your main daily-life goals, and questions about costs and alternatives.
- 1
Connect the service to your history, current care, and concrete goals.
- 2
Prepare the scalp sensors and check recording quality and comfort.
- 3
Explain what signal feature controls the visual or sound feedback.
- 4
Practice while the provider monitors the recording, artifacts, and your experience.
- 5
Review what happened, what to notice between visits, and when the plan will be reassessed.
Terms that make the conversation easier
EEG is the recording method. Brainwaves are patterns described within that recording. QEEG is a set of numerical analyses applied to EEG data. A protocol is the planned set of measurement and feedback rules. These terms describe different parts of the process and should not be treated as interchangeable.
QEEG results require context. A color map can look precise while still depending on recording conditions, artifact removal, the reference database, the measures selected, and the question being asked. A difference from a database average is not automatically a disorder, a treatment target, or proof that a symptom comes from that location.
Channel count describes how many voltage differences are recorded, not the quality of a provider or the likely outcome. Ask why a certain setup fits the purpose. Cost, visual complexity, or the number of sensors should not substitute for a clear rationale.
Measurement
EEG sensors record voltage changes at the scalp. Good signal quality and artifact handling matter.
Feedback
Software uses selected features of that recording to change a cue during practice.
Stimulation
Electrical or magnetic stimulation delivers energy. It is not the same process as conventional EEG recording and feedback.
Brainwaves
A familiar name for rhythmic patterns observed in EEG. Labels such as alpha, beta, theta, and delta describe frequency ranges, not separate substances or simple measures of how well a brain works.
Channel
One EEG signal recorded or calculated by comparing electrical activity between electrode inputs. A conventional single-channel neurofeedback setup commonly uses three electrodes: active, reference, and ground. The active and reference provide the signal comparison; the ground supports the recording circuit. Multichannel systems may share reference and ground electrodes, so channel count is not the same as sensor count. More channels do not automatically mean better care.
EEG
Electroencephalography, a method for recording small electrical voltage changes at the scalp. EEG reflects synchronized activity from populations of neurons along with noise and artifacts that must be managed.
Feedback
Information returned during practice. It may be a changing sound, animation, game, video, or other cue linked to selected features of the measured signal.
Protocol
The planned rules for a session, including what is recorded, which features affect feedback, where sensors are placed, thresholds or targets, session structure, and how the plan is reviewed.
QEEG
Quantitative EEG, which applies numerical analysis to EEG data and may compare selected measures with a reference database. The result is not a stand-alone diagnosis and depends on recording quality, artifact handling, methods, context, and qualified interpretation.
How to think about research and evidence
The most useful evidence question is not simply “Does neurofeedback work?” Ask: for which problem, in which population, using which protocol, compared with what, and measured by whom? A study of one protocol in children with one diagnosis does not establish results for a different protocol, age group, or goal.
Study design changes what a result can show. A waitlist comparison can estimate whether offering a program is better than waiting, but it cannot separate the specific effect of feedback from attention, expectations, time with a provider, and repeated practice as well as a convincing active or sham comparison. Blinded outcome ratings reduce the chance that expectations influence symptom scores.
Promising and mixed can both be accurate. Small studies can justify more research without settling a clinical question. Reviews can find no meaningful average benefit while identifying a narrower subgroup or protocol worth testing. The balanced response is to preserve the main result, the uncertainty, and the boundaries of the evidence.
ADHD, 2025: A meta-analysis of 38 randomized trials and 2,472 participants found no significant improvement in the main probably blinded symptom analysis. A subgroup using established standard protocols showed a small improvement. The subgroup is a research signal, not evidence that all neurofeedback works for all ADHD.
Open the PubMed recordPTSD, 2016: A randomized trial of 52 adults reported symptom improvement with neurofeedback compared with a waitlist. That is encouraging, but a waitlist cannot isolate feedback-specific effects as well as a credible active or sham comparison. A 2019 correction replaced one figure.
Open the PubMed recordInsomnia, 2017: In a double-blind, placebo-controlled study of 25 people, sensorimotor-rhythm neurofeedback did not improve EEG or objective sleep beyond nonspecific effects. This illustrates why expectation and the training experience need a convincing comparison.
Open the PubMed recordThese examples are selected to teach evidence reading. They are not an exhaustive or continuously updated review of every condition or protocol.
Choosing a provider and a plan
Start with the professional service being offered. Clinical treatment may require a license or other credential appropriate to the jurisdiction and condition. A certification such as BCIA documents training in a modality, but BCIA states that certification is not a substitute for an appropriate professional license or credential. Ask who is responsible for assessment, diagnosis, treatment decisions, and supervision.
Relevant experience matters more than a long list of technologies. Ask how often the provider works with people who share your goals, how they select and change protocols, how they coordinate with existing care, and what evidence supports the proposed approach. A provider should be able to explain both what is known and what is uncertain without guaranteeing a result.
Request the full practical picture: assessment fees, session fees, likely range of visits, equipment or subscription costs, cancellation policies, insurance documentation, and follow-up. Compare this with alternatives, including watchful waiting, established treatments, accommodations, coaching, or other supports appropriate to your situation.
Questions worth bringing to a provider
- What license, credential, certification, and supervision apply to the service you are offering?
- What exactly will you measure, and what feature will control the feedback?
- Why does this protocol fit my goal, and what evidence supports this specific use?
- How will my health history, medication, diagnoses, and existing care be considered?
- What unwanted changes should I report, to whom, and how quickly?
- What are the total expected costs, review points, alternatives, and stopping rules?
Certification documents training in a modality. It does not replace the professional license or credential required for clinical diagnosis or treatment.
Measuring progress and knowing when to reconsider
Agree on a few individualized goals before training. “Improve my brain” is too broad to evaluate. Examples of clearer goals might be following a conversation with fewer lapses, falling asleep within a defined range, completing a routine with fewer prompts, or reducing a validated symptom score. The goal should matter in daily life and fit the reason for care.
Record a baseline and choose review points in advance. Use the same measures when possible, while remembering that sleep, medication, school or work demands, other treatments, and ordinary variation can affect them. EEG or QEEG changes do not automatically mean a person feels or functions better. Daily-life benefit and unwanted changes both belong in the review.
Ask what would count as enough improvement to continue, what would lead to a protocol change, and when the plan would stop. If training is not helping, is burdensome, causes concerning changes, or delays more appropriate care, revisit the rationale and alternatives. Safety reporting in neurofeedback trials is still inconsistent, so a plan for reporting and responding to changes is important even when serious problems are not expected.
Set a baseline
Choose a few meaningful daily-life goals and record the starting point.
Notice the whole picture
Track function, symptoms, burden, and unwanted changes, not only EEG measures.
Use agreed checkpoints
Decide whether to continue, adjust, pause, stop, or consider alternatives.
Frequently asked questions
Is neurofeedback the same as electrical or magnetic stimulation?
No. Conventional EEG neurofeedback records scalp electrical activity and uses information from that recording to change feedback. It does not send an electrical current or magnetic pulse into the brain. Stimulation approaches such as tDCS or TMS deliver energy and should be evaluated under their own evidence, safety, and regulatory frameworks. Some products combine technologies, so ask exactly what the equipment records and what, if anything, it delivers.
Does a QEEG tell a provider which treatment will work?
Not by itself. A QEEG may help describe selected EEG features, but interpretation is affected by recording quality, artifacts, alertness, medication, reference methods, and clinical context. A map or database comparison is not a diagnosis or a guarantee that a particular protocol will help. Ask how the result fits with your history, symptoms, examination, and other assessment information.
Are more sensors or channels always better?
No. More channels can capture more spatial information and may be useful for some assessment or training questions. They can also require more setup, processing, and expertise. A smaller montage may be appropriate for a specific protocol. The useful question is whether the recording and protocol fit the goal and whether their limits are explained.
How many sessions will I need?
There is no responsible universal number. Plans vary by goal, approach, individual response, setting, and the evidence for the intended use. Ask for the proposed schedule, total expected cost, early review points, and what would lead to continuing, changing, or stopping the plan.
Does neurofeedback work?
That question needs four details: for which problem, in which people, using which protocol, and compared with what? Research is mixed across conditions and protocols. A positive result for one defined approach should not be applied automatically to every device, condition, or person. A null result also may not answer every narrower question.
What should I report during training?
Report any new, worsening, or concerning change, whether or not it seems related. Examples to discuss include headache, unusual fatigue, irritability, sleep changes, dizziness, discomfort from sensors, or a meaningful change in the symptoms being tracked. Ask before starting whom to contact, what needs prompt attention, and how the protocol will be adjusted or paused if needed.
Quick glossary
- Artifact
- Activity in a recording that does not represent the brain signal of interest, such as eye movement, jaw tension, motion, poor sensor contact, or electrical interference.
- Brainwaves
- A familiar name for rhythmic patterns observed in EEG. Labels such as alpha, beta, theta, and delta describe frequency ranges, not separate substances or simple measures of how well a brain works.
- Channel
- One EEG signal recorded or calculated by comparing electrical activity between electrode inputs. A conventional single-channel neurofeedback setup commonly uses three electrodes: active, reference, and ground. The active and reference provide the signal comparison; the ground supports the recording circuit. Multichannel systems may share reference and ground electrodes, so channel count is not the same as sensor count. More channels do not automatically mean better care.
- EEG
- Electroencephalography, a method for recording small electrical voltage changes at the scalp. EEG reflects synchronized activity from populations of neurons along with noise and artifacts that must be managed.
- Feedback
- Information returned during practice. It may be a changing sound, animation, game, video, or other cue linked to selected features of the measured signal.
- Protocol
- The planned rules for a session, including what is recorded, which features affect feedback, where sensors are placed, thresholds or targets, session structure, and how the plan is reviewed.
- QEEG
- Quantitative EEG, which applies numerical analysis to EEG data and may compare selected measures with a reference database. The result is not a stand-alone diagnosis and depends on recording quality, artifact handling, methods, context, and qualified interpretation.
Clear questions are a strong first step.
- Neurofeedback is a family of approaches, not one uniform treatment.
- Measurement, feedback, and stimulation are different processes.
- Evidence should match the condition, population, protocol, comparator, and outcome.
- Credentials, realistic expectations, costs, alternatives, and review points all matter.
- Progress means meaningful daily-life change, not only a change on a display or EEG measure.
Sources behind this guide
Use these original records and professional guidance as starting points, not as a complete review.
- BCIA: choosing a biofeedback professional Consumer guidance on credentials, scope of practice, and questions to ask.
- Westwood et al., JAMA Psychiatry (2025) Systematic review and meta-analysis of neurofeedback for ADHD.
- van der Kolk et al., PLOS ONE (2016) Randomized trial of neurofeedback for chronic PTSD using a waitlist comparison.
- Schabus et al., Brain (2017) Double-blind, placebo-controlled insomnia study illustrating the importance of a credible comparison.
- AAN practice advisory on QEEG for ADHD (2013) Guidance on the limits of using a theta/beta power ratio to confirm an ADHD diagnosis.
- International QEEG Certification Board technical guideline (2025) Technical guidance on acquisition, artifact management, visual inspection, and reporting.
- Neurofeedback safety reporting in adult ADHD trials (2026) A systematic review noting that safety reporting remains limited across the included studies.
- Correction to the 2016 PTSD trial (2019) The indexed correction replaces Figure 2 in the original publication.
- Beniczky and Schomer: EEG biophysics and recording technology (2021) First published January 21, 2021. Educational review explaining differential inputs, references, and EEG montages.
Neurofeedback.com helps readers understand options and ask better questions. It does not promise that one method, device, or provider is right for everyone.