Sign in to Peak Brain Path
Sign in to access your courses, books, and progress tracker. New here? Signing in creates your account automatically.
Want to explore courses first?
Browse courses and booksSign in to access your courses, books, and progress tracker. New here? Signing in creates your account automatically.
Want to explore courses first?
Browse courses and booksModule 14
Chapter 14 · 1 h · 8 quiz items · pass at 80%
BCIA Domain VII expects fluency in the workhorse protocols. This module gives the rationale, sites, bands, and thresholds for theta/beta, alpha/beta, and SMR, the three protocols a practitioner runs most often, and names the beginner errors that undermine them. The quiz proves the learner can configure each protocol and match it to the right presentation.
These three protocols are the foundation of clinical neurofeedback. They are the oldest, the best-evidenced, and the ones a BCN candidate is most likely to be tested on and a new practitioner most likely to run first. Theta/beta training built the field's reputation in ADHD. SMR training is where the field began, in Sterman's cats, and it remains the safest first-line choice for a wide range of presentations. Alpha/beta sits between them, a smaller and more situational tool. Master these three and you can responsibly treat most of the clients who walk through a general neurofeedback practice. The later chapters on alpha-theta, z-score, and slow cortical potential work build on the operant logic you learn here.
This chapter takes each protocol in turn: where it came from, what it is doing to the brain, what the evidence supports, and then the practical parameters, the electrode sites, the frequency bands, the thresholds, and the session structure, that you will set at the chair. It closes with the beginner errors that account for most failed early sessions, because knowing the parameters matters less than knowing how to set them so a client can actually learn.
The theta/beta ratio is the single most-studied number in clinical neurofeedback, and the protocol built on it is the reason the field has the foothold it has in attention disorders.
The story runs through Joe Lubar, who in the 1970s extended Barry Sterman's sensorimotor work into the question of attention and arrived at the observation that has anchored ADHD neurofeedback ever since: children with attention problems show excess slow activity, specifically theta, over the central and frontal cortex, relative to the faster beta activity that accompanies engaged processing (Lubar & Shouse, 1976). The ratio of theta power to beta power became a marker of cortical arousal. A high ratio means too much slow activity and not enough fast: a cortex that is, in functional terms, underactivated, drifting, disengaged. A low ratio means an alert, engaged cortex.
The marker held up. In a frequently cited dataset, an elevated theta/beta ratio appeared in roughly 78 percent of subjects with ADHD and was absent in about 97 percent of neurotypical controls (Monastra et al., 1999), which made it one of the more reliable single EEG features associated with the disorder. The functional story is coherent: the elevated ratio reflects reduced metabolic support for the cortex, less blood flow and slower processing, which presents behaviorally as the inattention, drift, and need for stimulation that define the inattentive picture.
Two honest cautions belong here, because the field oversold this number for years. First, the theta/beta ratio is not a diagnosis. It is a marker that is elevated in a majority, not all, of people with attention problems, and a meaningful minority of people with clear ADHD do not show it. The FDA clearance of a theta/beta-based device as an aid to ADHD assessment did not make the ratio a standalone test, and you should not present it to clients as one. Second, the average theta/beta ratio in the ADHD population has declined across decades of studies, and part of the elevation in any given client can be an artifact of a slow individual alpha peak bleeding into the theta band rather than true theta excess (Lansbergen et al., 2011). Check the individual alpha frequency before you treat an apparent theta excess as the real thing. A client whose alpha peak sits at 8 to 9 Hz can show an inflated theta/beta ratio without being underaroused in the way the protocol assumes.
The protocol that follows from the marker is straightforward in concept: drive the ratio down by rewarding beta and inhibiting theta.
Reward band. Reward low beta, commonly in the 15 to 18 Hz range. This is the band associated with focused, engaged processing without straying into the high beta (above 20 Hz) that accompanies anxiety and tension. The aim is to teach the cortex to produce more of the activity that accompanies attention.
Inhibit band. Inhibit theta, typically 4 to 8 Hz, the slow activity whose excess defines the underaroused picture. Many protocols add a high-beta inhibit (roughly 22 to 30 Hz) as well, so that you are not inadvertently rewarding the client into an anxious, over-activated state while chasing beta. Rewarding low beta while inhibiting both the theta below and the high beta above keeps the trained state in the productive middle.
Placement. The classic sites are along the midline and frontal cortex: Cz, Fz, and F3, referenced to linked ears or a single ear. Cz, at the vertex, is the traditional anchor and the default. Fz and F3 move the training more frontally, toward the executive and attentional cortex, and F3 in particular is chosen when you want a left-frontal emphasis. The 10-20 nomenclature here is worth keeping precise: odd numbers are left-hemisphere, even are right, and z denotes midline, so F3 is left-frontal, F4 right-frontal, and Cz the central vertex. For a first uncomplicated inattentive case without a QEEG, Cz or Fz is a defensible starting site.
This is a single-channel protocol in its standard form, which is part of why it is a good first protocol: one active site, clean to keep artifact-free, easy for the client to understand.
What the evidence does and does not support. Be honest with clients and with yourself about where theta/beta training stands. Meta-analytic work on neurofeedback for ADHD has found benefits, with the effects strongest for inattention and for the standard frequency-band protocols (theta/beta and SMR) when judged by raters who know the child, and more modest when judged by blinded raters or active-control comparisons (Arns et al., 2009; Cortese et al., 2016). That gap, larger effects from unblinded raters than from blinded ones, is the central honest caveat of the ADHD literature, and a candidate who can state it will sound like a clinician rather than a salesperson. The fair summary is that theta/beta training has a real evidence base, that it is one of the better-studied behavioral interventions for ADHD, and that the magnitude of benefit is debated and depends on how outcomes are measured. Present it as an evidence-supported option with a documented mechanism, not as a cure.
Alpha/beta training is the smaller, more situational sibling. The bands involved sit close together, which is exactly why precision matters.
The logic is about arousal regulation through the alpha and low-beta range. In the configuration aimed at calming an overaroused client, you reward the SMR-adjacent low band, around 12 to 15 Hz, while inhibiting the slower alpha and theta beneath it (8 to 10 Hz and below) and the high beta above. This nudges an over-activated cortex toward calm alertness rather than either drowsiness or tension. The configuration can be reversed for an underaroused presentation, rewarding the faster activity and inhibiting the slow, but in practice, when you are reaching for an arousal-down effect, SMR proper is the cleaner and better-evidenced choice, and when you want to restore rest capacity in a ruminative client, posterior alpha uptraining is more direct.
For that reason alpha/beta occupies a narrow band of clinical situations: the client whose arousal sits in between, where you want fine control over the 8 to 15 Hz region without committing to a full SMR or full posterior-alpha protocol. Know it for the exam and keep it in the toolkit, but recognize that for most overaroused presentations SMR is the first reach, and for most ruminative presentations posterior alpha is the first reach. The evidence base for alpha/beta as a distinct protocol is thinner than for either SMR or theta/beta, and the bands are close enough together that sloppy filter settings can blur the contingency you are trying to teach.
SMR, the sensorimotor rhythm, is where clinical neurofeedback started and the protocol I reach for most often when a presentation is mixed or I am unsure. Its history is the field's origin story and its mechanism is the clearest of the three.
Barry Sterman discovered the sensorimotor rhythm in cats: a 12 to 15 Hz rhythm over the sensorimotor cortex that appeared when the animal was still but alert, motionless and waiting rather than drowsy. He trained cats to produce it through operant reward, and then, by serendipity, those same cats turned out to be resistant to seizures induced by a convulsant chemical they were later exposed to in an unrelated study (Sterman & Friar, 1972). The cats that had learned to produce more SMR had raised their seizure threshold. That accidental finding launched the application of SMR training to human epilepsy and, from there, to the broader regulatory uses the rhythm serves today.
The motor-strip rationale. SMR is an idling rhythm of the sensorimotor cortex, analogous to how posterior alpha is an idling rhythm of visual cortex. It appears when the motor system is quiet but the cortex is alert: physically still, attentionally engaged, low muscle tension. The rhythm reflects strong thalamocortical inhibition, the thalamus effectively gating sensory and motor traffic so the cortex is not flooded. When SMR is strong, the brain filters irrelevant sensory input well, the startle response is damped, and the motor system is calm. This is why training it produces calm alertness rather than either sedation or activation, and why it sits at the bridge between the underaroused and overaroused extremes.
Where the rhythm lives and why C3/C4 matters. SMR is generated along the sensorimotor strip, the band of cortex running roughly ear to ear over the top of the head. The standard training sites are C3 and C4, sitting over the left and right sensorimotor cortex respectively, sometimes Cz at the vertex. C3 (left) and C4 (right) are chosen because they sit directly over the generator. This is the same strip whose function Chapter 5 covered as the seat of the motor and somatosensory homunculus.
Parameters. Reward 12 to 15 Hz at C3, C4, or Cz. Inhibit the theta below (4 to 8 Hz) and the high beta above (20 to 30 Hz), so you are rewarding the narrow sensorimotor band while keeping the client from drifting into drowsiness or tension. The narrowness of the reward band is a feature: SMR is a specific 12 to 15 Hz rhythm, not broadband beta, and the protocol works by training that specific activity.
What it is good for. SMR's clinical reach is wide because calm alertness is useful across many presentations. The well-supported uses include the regulatory and arousal-stabilizing work that opens many treatment courses: somatic anxiety with physical tension, hypervigilance and exaggerated startle, sensory sensitivity, and motor restlessness including tics and restless legs. It has a strong place in sleep, particularly difficulty staying asleep. The link is mechanistic, not coincidental: the 12 to 15 Hz sensorimotor rhythm shares the frequency and the thalamocortical machinery of the sleep spindles that punctuate Stage 2 sleep, so training daytime SMR strengthens the same spindle-generating circuitry that stabilizes sleep at night. Clients who train SMR often report falling back asleep more easily after a night waking, which fits the spindle account, and SMR training has improved sleep quality in chronic insomnia in controlled work (Cortoos et al., 2010). And it retains its original application in seizure work, where the Sterman lineage and subsequent reviews support it as an adjunct (Sterman & Egner, 2006). Hammond's review of SMR for anxiety and related presentations summarizes much of the clinical literature (Hammond, 2005).
No neurofeedback protocol is hazardous in the way a drug can be, but each has presentations where it is the wrong reach, and SMR is no exception despite its safety.
SMR cautions. Do not lead with SMR in acute panic. Stabilize the immediate crisis first with breathing or HRV work, then bring SMR in for the longer arc. Be careful in depression with low arousal: SMR's calming, motor-quieting effect can worsen lethargy in a client who is already underaroused and slowed, where a beta-up approach would serve better. The common side effects are mild and worth warning clients about: early sessions can produce a relaxation "hangover," a spacey feeling for an hour or two afterward, and clients report vivid dreams as sleep architecture shifts. These settle within the first weeks, and effects from SMR training emerge after 10 to 15 sessions rather than immediately.
Theta/beta cautions. The main hazard is treating an apparent theta excess that is in fact a slow-alpha artifact, which sends you inhibiting activity the client needs. The other is rewarding low beta in a client who is actually anxious rather than underaroused, which can increase arousal in the wrong direction. This is why the high-beta inhibit matters and why sorting the presentation correctly in Chapter 13 comes first.
General contraindications. The protocols in this chapter are surface arousal-regulation training and are broadly safe, but the whole-person contraindications still apply: active psychosis and acute mania are not situations for elective neurofeedback, and a seizure-disorder client should be trained in coordination with their treating physician rather than as a solo neurofeedback case. Chapter 11 covered the intake-stage contraindication screen that gates all of this.
Thresholds are where the most sessions are won or lost, and they are what beginners get most wrong. A threshold is the level the trained activity must reach to trigger a reward. Set it well and the client succeeds often enough to learn; set it badly and you either reward noise or starve the client of the feedback that drives learning.
The standard deviation method. A common approach takes a brief baseline of the reward band, computes its mean and standard deviation, and sets the reward threshold relative to that baseline, for instance roughly one standard deviation in the direction you want to train. This anchors the threshold to the client's own activity rather than to an arbitrary microvolt value, which is necessary because amplitudes vary widely between people and sites.
The percentile method. Equivalently, set the threshold so the client meets it some target percentage of the time at baseline, commonly around 60 percent. A reward rate near 60 percent means the client succeeds more often than not, which keeps the operant contingency learnable, while still having to work for a meaningful fraction of the rewards. This is the principle that matters more than the exact number: the client must succeed often enough that the brain can detect what produced the reward.
Start conservatively and titrate. Begin with thresholds set so success comes readily, then tighten as the client improves. A threshold that gives a 60 to 70 percent reward rate early in a course is teaching; a threshold set so tight that the client succeeds 20 percent of the time is not teaching, it is frustrating, and an unrewarded client does not learn the contingency. Across a training course you taper the threshold to keep the challenge live as the client gets better, the same way you add weight as a lifter's strength grows. Auto-thresholding, where the software continuously adjusts to hold a target reward rate, automates this and is reasonable, but understand what it is doing so you are not surprised when a client's apparent performance stays flat while their underlying activity is in fact improving.
A clinical session has a shape, and the shape matters as much as the parameters.
Length. Surface arousal-regulation sessions run 20 to 40 minutes of actual training. For theta/beta and SMR, 30 minutes is the standard. Twenty minutes can suffice for younger children or fatigue-prone clients, and 40 minutes is toward the upper end before session fatigue starts working against you. More is not better. A tired client late in an overlong session is consolidating fatigue, not the trained state.
The arc within a session. A workable structure runs: a brief baseline recording to set thresholds and see where the client is starting that day, then training delivered in blocks of several minutes with short rest intervals between them rather than one unbroken stretch, then a brief settling period at the end. The rest intervals matter because attention and the trained rhythm both fatigue, and a short break lets the client reset and re-engage. Watch the running display across the session for the trend you predicted in your rationale note.
Frequency and course length. Sessions are scheduled two to three times per week. The total course depends on the presentation: attention work with theta/beta runs toward 40 sessions, anxiety and stress work with SMR 20 to 30, with gains continuing through the course rather than plateauing early. Set the client's expectations to this arc at the outset, as Chapter 18 covers, so that neither of you mistakes the normal slow build for failure.
The numbers above are starting points, not fixed settings. The individual alpha frequency is the single most useful thing to check before tailoring: if a client's alpha peak sits low, the theta band and the alpha band are closer together than the textbook assumes, and both your theta/beta interpretation and your band edges may need adjusting. Reward and inhibit band widths can be narrowed or widened: a narrow reward band trains a more specific rhythm but is harder for the client to hit, while a wider band is easier but less precise. Software platforms differ in their defaults and their labels (BrainMaster, Thought Technology's BioGraph and BioTrace lines, and others each present these parameters with their own conventions), so confirm what a given system means by "SMR" or "low beta" rather than assuming the band edges match the textbook. The underlying physiology is the same across platforms. The parameter labels are not.
For combining with adjuncts, the best-supported pairing is HRV biofeedback layered onto an SMR or alpha/beta protocol for an anxious or stress-driven client, addressing autonomic regulation alongside the cortical training. The two work on complementary systems and are well tolerated together. These protocols also coexist routinely with psychotherapy and with medication, and the medication point bears repeating from Chapter 10: a stimulant shifts the EEG toward beta and away from theta, so a client's theta/beta ratio and their response to a theta/beta protocol both depend on their medication state, which you must document and account for.
Most failed early sessions trace to a short list of mistakes, and recognizing them is more useful than any parameter table.
Thresholds too tight. The most common error. The beginner sets a threshold so demanding that the client rarely succeeds, reasoning that a hard target means more effort. The opposite is true: an unrewarded client cannot detect what produced the reward, so no learning happens. Loosen the threshold until the client is succeeding well over half the time, then tighten gradually.
Feedback too fast or too dense. Reward delivered too quickly or too continuously gives the client no time to register the contingency. The feedback has to be readable as feedback, with enough structure that the client's brain can associate the rewarded state with the reward.
Sessions too long. A 60-minute attention-training session in a 9-year-old is training fatigue, not attention. Match the length to the client's capacity and stop while they are still able to engage.
Rewarding artifact. If the session is not artifact-gated, a client who tenses their jaw or moves can drive the reward band with muscle activity rather than real cortical activity, and you will be reinforcing EMG. Keep the signal clean, as Chapter 8 detailed, or you are training the wrong thing.
Treating an artifact as a finding. Setting up a theta-inhibit protocol on a theta excess that is in fact a slow individual alpha peak. Check the alpha frequency first.
When you sit down to run one of these protocols, the sequence is the same regardless of which: confirm the presentation and the site, take a brief baseline, set the threshold so the client will succeed around 60 percent of the time, deliver training in blocks across 20 to 40 minutes with rest intervals, watch the running trend against the change you predicted, and keep the signal artifact-clean throughout. Default to SMR at C3 or C4 when a presentation is mixed or you are unsure. Reach for theta/beta at Cz or Fz when the picture is underaroused and inattentive and you have ruled out a slow-alpha confound. Keep alpha/beta for the narrow in-between arousal cases.
For the BCN exam, fix the core numbers and their sources. Theta/beta: reward roughly 15 to 18 Hz, inhibit 4 to 8 Hz, at Cz, Fz, or F3, rooted in Lubar's ADHD work and the theta/beta ratio as an arousal marker. SMR: reward 12 to 15 Hz at C3 or C4, inhibit theta and high beta, rooted in Sterman's seizure work and serving regulation, sleep, and seizure-adjunct uses. Alpha/beta: the in-between arousal protocol, reward around 12 to 15 Hz with the configuration depending on the direction you want arousal to move. Know that thresholds are set for a roughly 60 percent reward rate so the client can learn, that sessions run 20 to 40 minutes two to three times a week, that a slow individual alpha frequency can masquerade as theta excess, and that the most common beginner error is a threshold set too tight for the client to succeed.