This article comes from one of my weekly NFB & Chill live streams, where I set up a real neurofeedback session on camera and field questions from the chat. I run these every Monday to demystify neurofeedback and help you become your own expert on whatever brain topic you care about. I've removed names from the audience questions and kept the teaching.
What does a left-right-center neurofeedback protocol train?
The protocol I ran on myself is what my notes call LR2C, short for left, right, two, center. Three locations along the sensorimotor strip: C3 on the left, C4 on the right, and CZ at the vertex. One beta reward and two variants of sensorimotor rhythm (SMR) training.
Here is what each site does.
C3 minus A1 (left sensorimotor strip). A left-side beta reward. The left pre/postcentral gyrus tracks with task initiation and activation. Training beta here tends to support engagement and forward motion.
C4 minus A2 (right sensorimotor strip). SMR training for executive function. SMR sits in the 12-15 Hz range but only on the sensorimotor strip. The same frequency elsewhere is regular beta processing or fast alpha. On the strip, SMR behaves like a calming, regulatory rhythm despite its beta-like speed.
CZ minus A1 (vertex). A mix of executive function and sleep. SMR at the midline strengthens the thalamocortical circuits that generate sleep spindles, the 12-14 Hz bursts that maintain sleep stability. That overlap is why SMR training often improves both daytime focus and nighttime sleep. You can read more on the mechanism in SMR Neurofeedback: Train Sleep, Focus, and Self-Control.
Across all three I inhibit theta (4-7 Hz) and reward low beta. Theta gets in the way of voluntary, fine-grained control over the tissue. Low beta supports concentration, sleep, and relaxation at the same time.
How do thresholds and rewards actually work?
The basic loop of neurofeedback is simple to state. The software watches a brain wave in semi-real time. When that parameter moves briefly in the direction you want, usually on its own, you applaud it with a beep and a visual change. The brain does not care what the applause sounds like. A high beep or a low beep produces the same learning. What matters is that the reward arrives sometimes and is withheld other times.
To drive learning instead of just measuring, the threshold sits some range away from where the brain currently is, and it updates on a schedule. The settings I use are auto-goal values of 25, 65, and 15, recalculated every 30 seconds against the person's moment-to-moment average. During the session you can watch the percentage above threshold shift. When I made more beta, the percentage dropped from 74 to 66 and the threshold climbed from about 7.6 to 8.5 microvolts. The game got harder in real time. That is the training happening in front of you.
If you set the frequencies wrong on the fly, you can fix them without stopping. In EEGer you highlight the band and use the control key with the arrow keys to nudge one end of the range or the whole range. I had to do exactly that mid-session when a setting failed to save.
What does fine-tuning the SMR reward frequency look like?
Finding the right SMR frequency is one of the most useful skills in this work. One viewer can feel a quarter-hertz difference, where 12-15 Hz feels too activating but 11.75-14.75 feels slightly too slowing. Roughly one person in ten can feel that fine a distinction. If you are one of them, split the difference. EEGer can be set to eighth-of-a-hertz steps, so you can land between the two.
Here is the read-out you use to dial it in.
Immediate effect. How you feel right after the session and for about two hours. A real window, though a small one.
24-hour window. Subtle shifts in sleep, calmness, focus, or tiredness rolling through over the next day. Usually small, and it wears off unless you repeat the protocol and let it accumulate.
Sleep, as a tuning signal. Undershoot someone's SMR and they get calm but struggle to fall asleep. Overshoot and they also struggle with sleep onset. Reward too low at CZ and they wake repeatedly through the night, which means you downregulated too far. If 11 Hz reward produces all-night waking, they probably needed 12 Hz.
You can also check the QEEG. Look at the raw peak alpha frequency at PZ with eyes closed and linked ears. How far below 10 Hz the alpha sits tells you whether their alpha speed runs slow or fast. Adjust the alpha reward toward that, then adjust the neighboring beta frequencies by a smaller amount. Fatigue can slow alpha too, so do not chase the number all the way down without reason.
Once you index the optimal SMR on the strip, regional differences fall into place. The right hemisphere runs slower than the left. The back runs slower than the front. Move left or move forward and the optimal speed increases. I run C4 minus A2 at 11.75, then step down to 11.5 at CZ. That is a classic step-down pattern. This is the old arousal model, the original laterality-and-arousal framework, and it still drives roughly half the field through the EEG phenotype approach.
When should you use the C4 minus A2 reference, and when should you avoid it?
I use both C4 minus A1 and C4 minus A2 depending on the person. A2 references tend to be more activating and more right-hemisphere dominant. A1 can feel softer.
The caution is negative affect. If someone shows a frontal asymmetry, the right-dominant pattern Richard Davidson described, do not push C4 minus A2. You risk irritating the right front by broadly raising beta there. C4 minus A2 also often runs at a slightly slower hertz within the same person than C4 minus A1, possibly because you are tapping some faster-running left hemisphere. Once you know a person's reward frequency at C4 minus A1, you can use the A2 reference with more confidence.
Where do you train alpha-theta, and who is it for?
Alpha-theta is a back-of-the-brain protocol. The classic site from the literature is PZ. Some providers move to OZ, some split the difference at PO, and others run dual-channel variants like P3 minus A1 paired with P4 minus A2. The differences matter. PZ gives more emotional access through the cingulate. OZ turns more visual, since O1, O2, and OZ sit over primary occipital cortex (V1). PO taps both.
Never run alpha-theta in the frontal lobe. Theta rewards anywhere frontal produce abreactions. This is why some tag-sync protocols cause trouble when they use frontal sites. Alpha-theta invokes the tissue that quiets when you close your eyes. Shutting off the frontal lobe is the opposite of what you want.
For who benefits: alpha-theta has real value for creativity, flow state access, and reducing alcohol craving by re-educating your ability to downshift without a drink. It is poorly suited to anyone with a lot of active trauma response or strong anxiety. It can make you intensely aware of your emotions for a day, and it can leave you feeling loopy or stoned, which is disconcerting for an already activated nervous system. Build resilience first across the three legs of regulation, sleep, stress, and attention, so you can handle being thrown off if it happens. Most trauma-informed practitioners today have moved away from alpha-theta toward infraslow or Sue Othmer-style training at occipital and parietal sites.
How do you read the FC minus PZ cingulate protocol?
FC minus PZ trains the anterior cingulate in front and the posterior cingulate in back. These often get stuck together, with too much beta or theta revving the tissue and not enough alpha to return things to neutral. Depending on the brain, you inhibit a fast beta or a theta or both, then reward whatever alpha range is dropping out in absolute or relative power.
The two ends of this circuit produce different suffering. FC, the anterior region, is more obsessive and perseverative when stuck: songs lodged in your head, nail-biting, OCD patterns. PZ, the posterior region, is more about fearfulness, threat sensitivity, rumination, and trauma response. When I see much stronger activation at PZ than FZ, I often start with PZ minus A1 rather than FC minus PZ, because the strongest drivers of intense anxiety tend to live posteriorly.
Why do you have to adjust protocols so often?
You iterate every session or two based on what the person reports, not once a month. Neurofeedback is a skill and an iterative practice, closer to personal training than to a prescription.
If you train the data alone without listening to the person's experience, you can move the EEG without moving anything the person cares about. The goal is large experiential change, and experiential change brings data change with it. About a third of the time you are not certain what will happen, so you listen carefully to how the person felt after the session and refine from there. Iterating on experience gets you both: a better subjective response over time and a changed resting EEG when you remap.
Pushing someone in the wrong direction five or ten times creates a real long-term change you do not want. Anyone doing self-training without access to real-time mentoring is taking on that risk. If your only check-in is monthly, you do not have the feedback loop the practice requires.
I almost never see an EEG move backward when training is done well. When something does move oddly, there is usually a reason. Ratio features can move non-linearly because the numerator and denominator do not change together, which actually reveals what is driving a deeper pattern. Exhaustion and fatigue throw off a QEEG. Montage and filtering choices change how the same underlying data looks. And some conditions have a time course: a moderate concussion can show a blooming inflammatory response a month or two after the injury rather than immediately, and post-COVID brain fog is often clearest in the EEG one to three months out, not right after recovery.
Do binaural beats actually change brainwaves?
The frequency-following response people attribute to binaural beats does not work in human brains for auditory stimuli. Play 100 Hz in one ear and 110 Hz in the other and you will not produce a 10 Hz alpha wave. The "alpha tapes" and "beta tapes" function as a kind of meditation where you anchor attention to the sound. Visual entrainment is different. A flickering light can produce a steady-state visual evoked potential, and there is some evidence of a frequency-following response to 40 Hz visual flicker in mammal brains. Audio binaural beats have repeatedly failed to show the effect when researchers go looking.
Will neurofeedback move to phones and apps?
A truly easy phone or tablet system is a way off. Real-time signal processing needs low-level hardware access that phones still limit, and Bluetooth training amps have consistently introduced timing problems or dropped the training effect. The hard part of neurofeedback was never setting up software. It is knowing what to do with the data.
That is what I am building first: a low-friction tracker where you log sleep, stress, mood, and attention daily, and the clinician sees those trends plotted against the protocols, meditation, and sleep habits. The clinical bottleneck is the client understanding the day-to-day through-line and the clinician understanding the client's day-to-day experience, then pointing both in the same direction.
How do you find PZ and the inion on your own head?
PZ usually sits in the middle of the flat spot at the back of the head. The precise method uses bone landmarks. Find the auricular notch and go up to the vertex for CZ. Find the occipital bump and go up about an inch for OZ. PZ is halfway between CZ and OZ. The inion is the notch just under the occipital bump, marking roughly the lower edge of the occipital lobe. Being slightly high or low on PZ is forgiving. Being left or right of it is not.
Where to start
If you want to see your own patterns before training anything, a QEEG brain map is the place to begin. At Peak Brain we run a $250 brain map that also covers repeat maps for the year, and I teach clients to read their own flat maps, attention testing, and raw EEG. Develop the agency to understand yourself rather than handing the whole process to someone else. The hands-on skill, iterating frequencies and sites against how you actually feel, is the part that makes neurofeedback work.