I run a weekly livestream where I do neurofeedback on myself and answer whatever questions come in. This article pulls together one of those sessions. I cleaned up the live setup talk and kept the teaching: how I build a compound protocol, why two-channel training works the way it does, what aging does to your brainwaves, and where neurofeedback earns its keep with seizures, Parkinson's, and stress. Audience questions are paraphrased and anonymized.
What is a compound neurofeedback protocol?
Most people picture neurofeedback as one thing at a time. One electrode, one site, one band you push up or pull down. A compound protocol runs several short training segments in sequence inside a single session.
For this session I set up a C3-A1 segment followed by a C4-A2 segment. C3 sits over the left sensorimotor strip, C4 over the right. On the left I rewarded beta in the 14.6 to 17.6 Hz range while inhibiting theta at 4 to 7 Hz and inhibiting high beta at 24 to 36 Hz. On the right I trained 12 Hz, a little slower than my usual reward, for a short six-minute block to feel how it lands.
The software adjusts thresholds automatically every 30 seconds. For each band it parks the comparison just below the things I want to grow and just above the things I want to shrink. When my brainwaves hold in range, the game runs and beeps. When they drift, the feedback stops. You can read more about the band I lean on most in SMR Neurofeedback: Train Sleep, Focus, and Self-Control.
Why is neurofeedback involuntary?
This is the part people get backwards. I am not steering my brainwaves on purpose. I am sitting in front of a screen, hearing beeps, watching a highway unspool. The feedback is audio and visual, and my brain notices it below the level of my conscious attention. That makes neurofeedback a form of operant conditioning where the conditioned system is your cortex, rather than your intentions.
People rarely notice anything for the first three or four sessions. A handful of sessions in, an after-effect starts to show up and lingers. As you repeat the protocol, the effect grows stronger and lasts longer. That after-effect is the steering wheel. I watch how a client's sleep, stress, and attention shift between sessions and tune the next protocol from what I see. If you want the bigger picture on what the technique can and cannot do, see Is Neurofeedback Legitimate? A Research Overview.
How does two-channel training differ from running two protocols?
When you sequence segments, you train each piece of tissue on its own, confirm you are getting good effects, then combine them only if the combination makes sense.
C3-A1 and C4-A2 combine well because they are homotopic, opposite sides of the head, with strong natural communication between them. The left sensorimotor strip likes to make beta. The right likes to make SMR. When you put them into a dual protocol, the left leans into beta and the right leans into SMR, and a contingency forms between them. Both sites get more training information, and they have to move together. You start building a relationship of support and stability between the two locations.
This is also why you cannot dual everything. PZ, back midline, has a healthy mode of making lots of alpha and failure modes in both excess theta and excess beta. Run a beta-manipulation dual at PZ and it makes beta too easily, which tends to produce anxiety. The real question in two-channel design is what the brain already does and whether it wants to do it. You lean into existing modes and existing communication, then build a relational, contingent effect on top.
Where did neurofeedback come from?
The technique most of us use, beta-manipulation training, traces to the mid-1960s. Joe Kamiya in Northern California was doing voluntary alpha training, a biofeedback-flavored mindfulness approach, in the early-to-mid 1960s. It was interesting but lacked the punch of what came next. Barry Sturman's work with cats around 1965 to 1966 uncovered SMR, the sensorimotor rhythm with its strong inhibitory tone. SMR training raises the seizure threshold, supports executive function, and helps sleep. The alpha-theta work for addiction and trauma came later, mostly in the late 1970s and 1980s.
What does aging do to your brainwaves?
A viewer asked whether SMR slows with age the way alpha does. Beta in general does slow and shrink with age. You get reduced amplitude from cell body loss and reduced frequency from myelin loss. We build myelin lifelong, but at some point the new myelin can no longer overcome oxidative stress on the existing myelin, and the integrity of cell bodies drops. Beta dips and slows.
SMR does not change with age the way alpha does. Alpha is a frequency that matures. A six- or seven-year-old has a much slower alpha peak than a twelve- or fourteen-year-old. You can read someone's native alpha speed with eyes closed at PZ-A1. For a long time, the field assumed you could extrapolate from alpha to set everyone's reward frequency, the individualized optimal reward frequency. That logic does not track for beta. By early life you have enough connectivity and myelin to produce the beta frequencies you will use for most of your life. For more on this, see Decoding Alpha Waves: Your Brain's Idle and Its Brakes and The Critical Aging Window: Why Your Brain Starts Aging at 44, Not 70.
The encouraging part: even at 70 you are still making 600 to 700 new neurons a day. The brain stays plastic far later into life than we believed even ten or twenty years ago. Biohacking Plasticity: Unlock Your Brain's Adaptive Potential covers the mechanisms.
What does "stabilization" mean in neurofeedback?
It depends on the practitioner. In an arousal-model or infraslow framework, stabilization usually means helping a brain learn to calm itself on demand, often through interhemispheric training at sites like T3-T4 or T5-T6. It can also mean stabilizing against seizures by raising the seizure threshold. Newer infraslow practitioners use the term for deeper energy-management questions, how much sympathetic load you can carry before you ramp up too much. Both traditional and infraslow approaches reach for those interhemispheric protocols when stabilization is the goal.
Is the arousal model separate from mainstream neurofeedback?
The arousal model is mainstream neurofeedback. It is baked into Loreta-based work, NeurOptimal, SMR training, infraslow, and slow cortical potential training. The arousal model and the phenotype model overlap heavily. One reads valence, stress, and attention, plus the after-effects of training, to guide protocol choices. Traditional neurofeedback also uses QEEG brain maps to set starting places, and those starting places are interpreted through the same arousal logic: low alpha or slowed alpha in a region reads as over-arousal in that tissue, high beta reads as over-arousal too.
Done well, you combine the rigor of QEEG with the validating feedback of symptom change. You design protocols for the physiology, then use shifts in sleep, stress, and anxiety to tell you how close your frequencies and speeds are. If you go 100% into the arousal model and drop QEEG entirely, you leave something on the table. To understand what a brain map actually shows, see QEEG Brain Mapping: What It Is, What It Shows, and What to Expect and Biohacking with EEG Phenotypes: Predict Your Brain Function.
Is it dangerous to train delta and theta?
Training delta carelessly is genuinely risky. You can release seizures in someone prone to them, and you can produce fog and anxiety. You almost never need to train anything below about 3 Hz unless you are deliberately working below 1 Hz in infraslow. Work on the betas, alphas, and thetas, and the delta usually takes care of itself.
Theta training carries its own specific contraindication: rewarding theta when theta is already elevated, or when beta is already high in the same location, is disinhibitory and tends to make things worse. There are exceptions. In alpha-theta training, eyes closed on the edge of sleep, you train delta down to keep yourself awake and theta up to ride the hypnagogic state.
Location matters as much as band. Tonic or chronic front-midline theta at FZ is never good, so you would not train theta up there. PZ with eyes closed makes theta without trouble. The principle holds across the board: figure out what the tissue is meant to do, then help it do more of that or regulate it better.
Can neurofeedback help seizures, Parkinson's, and CP?
These are the questions where I separate regulatory targets from structural ones.
Seizures. Sturman's review of the seizure literature found an average reduction around 50% across studies, with about 5% of people reaching complete control. In my own practice I rarely see results as weak as 50%. People with chronic or constant seizures often get strong reduction. SMR training changes how the brain resists the seizure state, which is why this is one of the most powerful targets we have. People dealing with seizures rarely care about the published averages; they have heard it works, and it usually does.
Parkinson's. My confidence here drops to roughly 50% for significant change, because Parkinson's is several different things and there is often real tissue loss. The strongest effects show up in younger people, in their 40s through 60s, where I have seen progression slow dramatically or stop. In elders, the most reliable win is improving how their dopaminergic medication lands and lasts. Tremor medication that stops working after three hours wrecks sleep; people wake, redose, and never sleep well again. Help the medication hold through the night and you get a follow-on improvement in brain health. We can influence the dopaminergic system, but we cannot rebuild tissue that is already gone, which is why these are moderate effects rather than remediation.
Cerebral palsy. Like any focal brain injury, it depends on what is injured. I have seen good reduction in seizures and gains in executive function, sensory regulation, and anxiety, especially in kids where CP is one of several diagnoses. In people with severe motor impairment, the motor changes were minimal. Language, visual fusion, and movement are not meant to keep changing late in life, so focal damage to that tissue is hard to move. Attention, anxiety, sensory filtering, sleep, stress, cravings, and seizures are regulatory phenomena that keep tuning, which is why neurofeedback does so well there.
What is the best biofeedback tool for stress and adrenaline surges?
For adrenaline and cortisol surges, think about the vagus and the cingulates. Fast alpha at PZ can produce a blood-sugar-like surge; you can get panic phenomena at PZ and hyperfocus at FZ, and these drive the heart and gut through the vagus.
The tool I reach for is heart rate variability biofeedback. HRV is a skill, practiced rather than switched on. Practice with a device like a HeartMath emWave for 15 minutes, four or five times a week, away from any stressful situation. Learn the state shift, the drop into parasympathetic dominance, then reach for it when you are not sitting in front of the machine. With practice you can ride sympathetic activation back down rather than adrenalizing into fight or flight. Biohacking Fight or Flight: Mastering Your Stress Response goes deeper on this.
Temperature and GSR biofeedback are useful, temperature for headaches, both for general stress, but they are weaker than HRV. With handwarming or GSR you train a distal consequence rather than the control signal itself, so it is slow and indirect. HRV gives you a fast, obvious, repeatable state shift you can learn to recreate on purpose.
A note on equipment
People ask constantly which system is best. I can do good neurofeedback with almost any system if I do the right thing with it. Some systems restrict what you can control, which limits you. The bigger constraints are the cost of clinical-grade equipment and, more so, knowing how to use it. If you understand how the brain works and how neurofeedback works, the tools in front of you matter less than the thinking behind them. One practical tip: avoid gold-plated electrodes, which fail fast. Use solid silver or, for DC-coupled and infraslow work, silver-silver-chloride pellets.
Most of the clients at my practice work from home with remote support seven days a week, which lets us catch sessions and tune protocols as life happens. If you are curious how that works, see Remote Neurofeedback: How It Works and What to Expect. And if you want a clinic, ask them to teach you to read your own brain maps and explain why they choose each protocol. The more you understand your own physiology and report what you notice day to day, the better your results will be.
If you are dealing with seizures, ADHD, or chronic stress and want to know what your own brain is doing, get a QEEG and have someone who understands the arousal model look at it. That is the starting place for everything else.