SMR vs. Theta/Beta Neurofeedback for ADHD

The Cat That Accidentally Changed ADHD brain patterns Treatment

In 1967, a neuroscientist named Barry Sterman at UCLA was doing something unremarkable: recording brainwaves from cats. He'd implanted electrodes over their sensorimotor cortex, the strip of brain tissue that controls voluntary movement, and he was watching the EEG squiggles as the cats went about their day.

Then he noticed something odd. When the cats sat perfectly still but remained wide awake, waiting for a food reward, their brainwaves produced a distinctive rhythm right around 12 to 15 Hz. Not the slow waves of drowsiness. Not the fast, jagged activity of movement. Something in between. A hum of calm, focused stillness.

Sterman called it the sensorimotor rhythm, or SMR. And he found he could train cats to produce more of it using simple rewards. When the SMR rhythm appeared, the cat got a food pellet. Within weeks, the cats could slip into this calm-alert state on demand.

That would have been a minor footnote in the history of neuroscience. Except for what happened next.

NASA contacted Sterman. They were testing the effects of rocket fuel (monomethylhydrazine) on the nervous system, and they needed cats for toxicity experiments. Sterman provided some from his lab. When exposed to the fuel, most of the cats developed seizures. But the cats that had undergone SMR training resisted seizures far longer than the others. Some didn't seize at all.

Sterman had accidentally discovered that training one specific brainwave rhythm could change the brain's fundamental stability. That discovery eventually led to SMR neurofeedback for epilepsy, then for ADHD, and it sparked a 50-year research program that's still running.

But SMR training wasn't the only neurofeedback protocol that emerged for ADHD. On the other side of the field, researchers were taking a completely different approach, one focused not on the sensorimotor cortex but on the frontal lobe, and not on a single rhythm but on the ratio between two frequency bands.

Two protocols. Two brain targets. Two theories of what's actually wrong in ADHD. And two very different clinical experiences for the people who try them.

What's Actually Going On in the ADHD Brain

Before you can understand why these two protocols exist, you need to understand what they're both trying to fix.

Here's the short version: the ADHD brain has a signature. It shows up on EEG with remarkable consistency, and it looks like this: too much theta, not enough beta, especially over the frontal cortex.

theta brainwaves (4-8 Hz) are the slow, rolling rhythms your brain produces when you're drowsy, zoning out, or in that half-asleep state right before you drift off. beta brainwaves (13-21 Hz) are the faster, crisper rhythms associated with active thinking, problem-solving, and sustained attention.

In a typical brain, when you sit down to work on something that requires focus, frontal theta drops and frontal beta rises. Your brain shifts gears. It transitions from idle to engaged.

In many ADHD brains, this shift doesn't happen cleanly. Theta stays elevated. Beta stays suppressed. The frontal cortex, the region responsible for executive function, planning, impulse control, and sustaining attention on boring things, is essentially running in low gear when it should be in high gear.

Researchers captured this pattern as a simple metric: the theta/beta ratio (TBR). Take the power in the theta band, divide it by the power in the beta band, and you get a number. In 2013, the FDA actually cleared a TBR-based EEG test (the NEBA system) as an aid in diagnosing ADHD, the first time a brainwave measure had been granted that status.

Now, here's where it gets interesting. The elevated theta/beta ratio tells you what is different in the ADHD brain, but it doesn't tell you where to intervene. And the answer to that "where" question is exactly what split neurofeedback into two camps.

Protocol One: Theta/Beta Ratio Training

The logic of theta/beta training is beautifully straightforward. If the ADHD brain has too much theta and not enough beta at frontal sites, then train the brain to produce less theta and more beta at frontal sites.

Here's how it works in practice.

Electrodes are placed at Fz or Cz, midline positions sitting right over the frontal and central cortex. These locations were chosen because they sit above the prefrontal and anterior cingulate cortex, the regions most directly implicated in ADHD's executive function deficits.

The software monitors two things simultaneously: the amplitude of theta (4-8 Hz) and the amplitude of beta (13-21 Hz). The patient watches a screen. When theta drops and beta rises, something rewarding happens. A video plays smoothly. A spaceship accelerates. A tone sounds. When the ratio drifts back toward the ADHD pattern (too much theta, too little beta), the reward stops.

The patient doesn't consciously think "reduce my theta." That's not how operant conditioning works. Instead, their brain stumbles into the right state, receives a reward, and gradually learns to reproduce that state more reliably. Session after session, the frontal cortex gets better at making the shift from idle to engaged.

A typical theta/beta protocol runs 30 to 40 sessions, each lasting about 30 minutes, delivered two or three times per week. That's 10 to 20 weeks of training.

The Evidence

Theta/beta training is the most-studied neurofeedback protocol for ADHD, period. The evidence base includes dozens of randomized controlled trials, several large meta-analyses, and follow-up studies extending out to two years.

The landmark trial is the one by Arns and colleagues, published in European Child and Adolescent Psychiatry in 2014. This multicenter randomized controlled trial compared theta/beta neurofeedback to a sham (placebo) neurofeedback condition in children with ADHD. The real neurofeedback group showed statistically significant improvements in inattention that persisted at the 6-month follow-up. The sham group did not.

A 2019 meta-analysis in the Journal of Child Psychology and Psychiatry, covering 13 RCTs, confirmed that neurofeedback (primarily theta/beta protocols) produced significant improvements in both inattention and hyperactivity/impulsivity. The effect sizes were moderate, comparable to non-stimulant ADHD medications.

The American Academy of Pediatrics has rated neurofeedback, with theta/beta training as the primary protocol studied, as a Level 1 "Best Support" intervention for ADHD since 2012.

Protocol Two: SMR Training

SMR training comes at ADHD from a completely different angle. Instead of targeting the frontal brainwave imbalance directly, it trains a rhythm over the sensorimotor cortex that produces calm, physically still, mentally alert attention.

The electrode goes at C3 or C4, positions sitting directly over the left or right sensorimotor cortex (the same strip of tissue Sterman was recording from in his cats). The software rewards increases in the SMR band: 12 to 15 Hz.

This frequency range sits right at the border between alpha (8-12 Hz, associated with relaxed wakefulness) and low beta (13-21 Hz, associated with active thinking). SMR is neither relaxation nor activation. It's something more specific: the brain's signal for "my body is still, but my mind is ready."

Think about what that means for someone with ADHD. One of the hallmark symptoms, especially in the hyperactive-impulsive presentation, is the inability to sit still. Fidgeting. Restlessness. A body that won't stop moving even when the task demands it. SMR training directly targets the neural circuitry that controls this. It teaches the sensorimotor cortex to maintain a calm, inhibited state, and that physical stillness appears to create a platform on which sustained attention can rest.

The Evidence

SMR training's evidence base is smaller than theta/beta's but remarkably consistent. A key study by Arns and colleagues in 2012, published in Brain Topography, found that SMR neurofeedback produced specific improvements in inattention and impulsivity in children with ADHD, with effects that persisted at follow-up.

A 2020 study in Human Brain Mapping showed that SMR neurofeedback improved sustained attention and working memory not just in ADHD patients but in healthy adults, suggesting that the mechanism works on attention circuits broadly, not just on ADHD pathology.

Here's the finding that really caught researchers' attention: in head-to-head comparisons, SMR training often works better for the hyperactive-impulsive subtype of ADHD, while theta/beta training often works better for the predominantly inattentive subtype. This makes intuitive sense. If your primary struggle is physical restlessness, train the motor inhibition circuit. If your primary struggle is mental fogginess, train the frontal activation circuit.

The Head-to-Head Comparison

Now that the trunk is built, let's put these two protocols side by side. Because the differences are more than academic. They determine where the electrodes go on your head, what your brain practices during each session, and which symptoms improve fastest.

The Surprising Finding That Changed Both Camps

Here's the part that neither side of the debate expected.

In 2014, Martijn Arns and his team at the Brainclinics Foundation in the Netherlands published a study that should have ended the argument, or at least reframed it entirely. They conducted a retrospective analysis of over 1,000 EEG recordings from ADHD patients and matched them against treatment outcomes from both SMR and theta/beta protocols.

What they found was this: the theta/beta ratio isn't elevated in all ADHD patients. In fact, only about 30-40% of people diagnosed with ADHD show the classic elevated TBR pattern. The rest have perfectly normal theta/beta ratios, or even suppressed ones.

Let that sink in. The brainwave pattern that the entire theta/beta protocol was designed to fix isn't present in the majority of ADHD cases.

This didn't mean theta/beta training doesn't work. It does. But it suggested something more nuanced: patients with elevated TBR responded dramatically well to theta/beta training, while patients with normal TBR responded better to other protocols, including SMR.

Arns called this the "EEG phenotype" approach. Instead of giving every ADHD patient the same protocol, match the protocol to the individual's actual brainwave signature. If the EEG shows frontal theta excess, use theta/beta training. If it shows poor sensorimotor regulation, use SMR training. If it shows something else entirely (and there are at least three or four other common ADHD EEG patterns), use a protocol matched to that pattern.

This is the "I had no idea" moment in ADHD neurofeedback: the same diagnosis can have completely different brain signatures, and the right protocol depends on which signature you have. Treating all ADHD brains the same way is like prescribing the same glasses to everyone who says "I can't see well." Some people are nearsighted. Some are farsighted. Some have astigmatism. The symptom is similar. The underlying optics are different.

The Crown captures brainwave data at 256Hz across 8 channels. All processing happens on-device. Build with JavaScript or Python SDKs.

Explore the Crown

What Actually Happens During a Session

The clinical experience of these two protocols feels remarkably different, even though the setup looks almost identical from the outside. Understanding that difference matters if you're deciding which one to try.

An SMR Session

You sit in a chair. An electrode is placed at C3 or C4, over the sensorimotor cortex, with a reference electrode on one earlobe. The screen in front of you shows some kind of visual feedback. Maybe a bar graph. Maybe a slowly moving animation. Maybe a video that plays when you're in the zone and pauses when you're not.

Your job is to sit still. Not tense, rigid stillness. Relaxed, calm stillness. The kind of stillness a cat has when it's watching a bird through a window. Alert but not twitchy.

When your brain produces SMR in the 12-15 Hz band, you get a reward. A tone. A point. The video plays. Over 30 minutes, you cycle through periods of success and periods of drift. At first, the successes feel random. You don't know what you did differently. But after several sessions, something shifts. You start to recognize the internal state. It's a feeling of settled readiness. Your body is quiet. Your mind is clear. You're not trying to focus on anything specific. You're just... present.

Many people describe SMR training as meditative but sharper than meditation. There's an active quality to it that pure relaxation lacks.

A Theta/Beta Session

Same chair. But now the electrode is at Fz or Cz, right on the forehead or at the top of the head. And the feedback is tracking something different: the ratio between slow theta waves and fast beta waves.

This one feels more like waking up. As your theta drops and your beta rises, the feedback tells you you're getting closer to a state of focused engagement. Many patients report that it feels like "the fog lifting." Like the mental cloudiness they've lived with for years is temporarily clearing, and for the first time they can see what sustained, easy attention feels like.

The challenge is different too. In SMR training, the enemy is restlessness. In theta/beta training, the enemy is zoning out. You'll notice yourself drifting into that familiar ADHD haze, where your eyes are open but nothing's getting in, and the feedback will drop. That drop is information. It tells your brain "you just did that thing again," and over time, your brain learns to catch the drift earlier and pull itself back.

The Electrode Placement Question

One practical detail that matters more than most people realize: where you put the electrode determines which protocol you can run. And not all consumer EEG devices have electrodes in the right positions for both.

Theta/beta training requires frontal or central midline coverage. Clinical protocols typically use Fz (frontal midline) or Cz (central midline). These positions sit directly above the prefrontal cortex and anterior cingulate, the executive function regions where the theta/beta imbalance is measured.

SMR training requires coverage over the sensorimotor cortex, typically at C3 (left hemisphere) or C4 (right hemisphere). These positions sit directly above the motor strip, where the sensorimotor rhythm originates.

Most clinical neurofeedback setups use one or two active electrodes that can be placed at any standard 10-20 system position. But consumer EEG devices have fixed electrode positions, so the positions that come with the device determine which protocols are possible.

The Neurosity Crown has electrodes at eight positions: CP3, C3, F5, PO3, PO4, F6, C4, and CP4. That's significant because C3 and C4 are the exact positions used for SMR training, and F5 and F6 provide frontal coverage near the positions used for theta/beta training. The Crown can run both protocols without any hardware modifications, which is unusual for a consumer device.

Who Should Try Which Protocol

The research points toward a practical decision framework. It's not perfect, and it should not replace clinical assessment, but it's a useful starting point.

Consider theta/beta training if:

• Your primary symptom is inattention rather than hyperactivity
• You experience "brain fog" or mental sluggishness
• You have difficulty starting and sustaining focused work
• A QEEG assessment shows elevated frontal theta relative to beta
• You have the predominantly inattentive ADHD presentation (formerly called ADD)

Consider SMR training if:

• Your primary symptoms include hyperactivity and impulsivity
• You struggle with physical restlessness, fidgeting, or feeling "wired"
• You have difficulty sitting still even when you want to
• You have the hyperactive-impulsive or combined ADHD presentation
• You also deal with sleep onset difficulties (SMR training has shown benefits for sleep)

Consider combining both if:

• You have the combined ADHD presentation with both inattentive and hyperactive symptoms
• You've tried one protocol and gotten partial but not complete improvement
• You want to train both frontal executive function and sensorimotor regulation

The real answer, though, is that the best protocol is the one matched to your individual EEG signature. And the only way to know your signature is to measure it.

Running Your Own Protocol With Real Data

This is where the conversation shifts from clinical history to something you can actually do.

For most of neurofeedback's history, accessing either protocol meant visiting a clinician's office two or three times a week at $100 to $200 per session. A full course of 30 to 40 sessions could cost $3,000 to $8,000. That price barrier kept neurofeedback in the domain of the wealthy or the desperate, even as the evidence kept getting stronger.

Consumer EEG has changed this equation. But not all consumer EEG is created equal.

For meaningful neurofeedback, you need three things: sufficient spatial resolution (enough channels in the right positions), sufficient temporal resolution (a fast enough sampling rate to distinguish theta from beta from SMR), and a way to process the signal and deliver feedback in real time.

The Neurosity Crown checks all three boxes. Its 8 channels at 256Hz provide the resolution to separate the SMR band (12-15 Hz) from theta (4-8 Hz) and beta (13-21 Hz) with high precision. The C3 and C4 electrode positions enable SMR protocols. The F5 and F6 positions enable frontal monitoring for theta/beta work. And the open JavaScript and Python SDKs mean you can build a feedback application that responds to exactly the frequency bands and electrode positions your protocol requires.

You could, for example, write a script that reads the power spectral density from C3, extracts the 12-15 Hz band, and triggers an audio tone when SMR power exceeds a threshold. That's an SMR neurofeedback session. Or you could compute the theta/beta ratio from F5 and F6, set a target ratio, and provide visual feedback when the ratio drops below the threshold. That's a theta/beta session.

The Crown's on-device N3 chipset handles the signal processing at the hardware level, which means the latency between your brain producing a pattern and the feedback reaching your senses is measured in milliseconds. That's critical, because neurofeedback relies on the feedback being fast enough for your brain to associate the reward with what it just did. Delays of more than about 250 milliseconds start to degrade the learning signal.

The Future Is Personalized, Not Standardized

The SMR-versus-theta/beta debate has been productive, but it's also been a distraction. It assumed there was one right answer for all ADHD brains, and it sent researchers searching for a winner in a contest that didn't need to have one.

The real lesson from 50 years of ADHD neurofeedback research is that the brain is not a monolith. Two people with identical ADHD diagnoses can have radically different EEG signatures, and the protocol that transforms one person's attention might do nothing for the other. The future isn't picking the "best" protocol. It's measuring the individual brain and matching the training to what that specific brain needs.

This is exactly the kind of problem that consumer EEG devices with open data access are built to solve. When you can see your own brainwave data, in real time, from multiple electrode positions, you're not guessing which protocol might work. You're looking at the evidence. You can measure your own theta/beta ratio. You can observe your own SMR production. You can even run a few sessions of each protocol and compare which one produces more reliable changes in your EEG.

Barry Sterman didn't know, when he was rewarding cats for sitting still in 1967, that he was launching a half-century scientific debate. He was just watching the data and noticing a pattern. The cats that learned to control their own brain rhythms became fundamentally more resilient.

That's not a metaphor. That's the mechanism.

Your brain, like those cats' brains, can learn to regulate its own rhythms. It can learn to suppress theta and boost beta over the frontal cortex. It can learn to produce SMR over the sensorimotor strip. It can probably learn to do both, if you give it the right feedback and enough practice.

The only question that matters is the one that Sterman answered 60 years ago: can you show the brain what it's doing, and will it learn from what it sees?

The answer, across every protocol, every study, and every patient who sat in that chair and watched their own brainwaves dance across a screen, is yes. It will. Every time.