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 20
Chapter 20 · 1.5 h · 8 quiz items · pass at 80%
BCIA Domain VIII expects the practitioner to manage a full course, not just a single session. This module gives the hold-versus-advance rules, the typical session counts by presentation, the standardized outcome measures, and the non-response decision that triggers referral. The quiz proves the learner can monitor outcomes objectively and recognize when neurofeedback is not the answer.
The previous chapter ended with the client standing up from the chair after a single session. This one is about the forty sessions on either side of it. A neurofeedback course is not a sequence of identical visits. It is an arc with a shape: a beginning where the protocol is a first approximation, a middle where it stabilizes and you start to taper and layer, and an end where you decide whether the work is done. Running one session competently is a technical skill. Running a course is a clinical one, and most of it lives in two judgments you make over and over: hold or advance, and is this working.
Those two judgments depend on reading the client at the right time scale and on having something to read other than your own impression. A protocol that produced no change in the chair can be reshaping sleep two nights later. A client whose brain map has normalized can show no behavioral gain on a continuous performance test. The session-by-session display tells you almost nothing about either. What tells you is a structured outcome picture, sampled at intervals, paired physiology with paired behavior, and read against the client's lived report. This chapter builds that picture: the typical arc by presentation, the hold-versus-advance decision, threshold tapering and multi-site progression, re-assessment timing, the standardized measures you will use, how to recognize the client who is not responding, and the documentation the course generates for insurance and supervision.
Clients and referral sources ask the same question at intake: how many sessions. The honest answer is a range that depends on the presentation, the client, and what counts as done. The ranges below are the field's working consensus, not guarantees, and a candidate who states them with the caveat attached sounds like a clinician rather than a salesperson.
ADHD and attention work: roughly 40 sessions. Attention training with theta/beta or SMR builds slowly, and the gains in these protocols tend to continue through the course rather than plateau early. Forty sessions, two to three times a week, over three to five months, is the standard arc. A course cut to twenty sessions for an attention case is often stopped at the point where the trait change is just consolidating. Set the expectation to forty at the outset so neither you nor the client mistakes the normal slow build for failure.
Anxiety and stress: 20 to 30 sessions. SMR and arousal-regulation work tends to land faster than attention work. Sleep onset and stress reactivity, the two foundations that flex earliest, often show movement in the first ten to fifteen sessions, and a 20-to-30-session course is frequently enough to consolidate it. Some anxiety cases need more, particularly when the anxiety sits on top of something else.
PTSD with alpha-theta: 30 to 40 sessions. Alpha-theta is slow by design, the depth of the work and the consolidation it depends on take time, and the trauma courses in the literature run long (Chapter 15). This number assumes alpha-theta as the deep phase, sitting on top of a stabilization phase of arousal-regulation work that may add sessions of its own. A trauma course is rarely short.
Peak performance: 20 to 40 sessions. Performance work varies most because the target varies most. A specific, narrow goal (a musician's performance anxiety, an athlete's pre-competition arousal) can resolve in twenty sessions. A broader optimization goal runs longer. The range is wide because the contract is wide.
Two structural points apply across all of these. First, the early sessions are calibration. The first several sessions of any course are individual-response calibration, the period where you find the thresholds, the site, and the bands that fit this brain, and the trait change has not begun. A client at session three is in a different place than a client at session fifteen, and reading session three as if it were the verdict on the protocol is the error of premature discouragement. Second, the count is sessions, not weeks. A client training three times a week and a client training once a week reach session forty at different points on the calendar, and the calendar matters less than the count for the trait change but more for the relationship and the cost.
The central running decision in a course is whether to keep the current protocol or change it. New practitioners get this wrong in both directions: they over-adjust, changing the protocol every two or three sessions so that no direction ever accumulates the evidence that would tell them whether it works, or they under-adjust, running the same protocol for twenty sessions without ever checking whether it is producing what it should. The discipline sits between the two, and it rests on the smallest-useful-adjustment principle: change one thing at a time, and only when the case picture calls for it.
A protocol unchanged for several sessions has accumulated evidence. A protocol just changed is in its early-evidence phase and cannot yet be evaluated. So the first question before any change is how long the current direction has been running. A protocol that has run four or five clean sessions and is producing the expected movement is a protocol to hold. A protocol that has run the same four or five sessions and produced nothing readable is a candidate for adjustment, but adjustment of one variable, not a wholesale change.
What you read to make the call is the convergence of three streams. The training data tells you whether the brain is engaging the contingency: is the reward band shifting in the direction the protocol is shaped toward, are the within-session traces showing the brain finding the rewarded state. The outcome data, sampled in the two-to-twenty-four-hour window after sessions, tells you whether anything is propagating: sleep, stress reactivity, and attention are the three foundations that flex earliest, and a change in any one of them across the first ten to fifteen sessions is the early signal that the training is engaging the system at all. The client's report is the third stream, weighted as the chapter on time courses weights it: a single within-session report is weak evidence, a four-week trend in the daily reports is moderate evidence of emerging trait change, and a structured re-assessment is the strong evidence.
The "stuck" client is the one where these streams disagree or flatten. A common and readable case: the training data shows the brain engaging the contingency cleanly, but none of the three foundations is flexing after ten to fifteen sessions. That pattern points you to look for a confound before you change the protocol. A client with untreated sleep apnea, a substance affecting the EEG, or a major life stressor consuming the gains as fast as the training produces them will look like a non-responder when the protocol is fine and the context is the problem. Another readable case: clean sessions producing little effect, which suggests the protocol is well-run but not well-fitted, and a site or band change is warranted. The judgment is whether the absence of change is telling you to change the protocol, change the context, or wait, and it is the judgment that separates a course that is being managed from one that is merely being run.
A threshold set well early in a course is set wrong by the middle of it, because the client has improved and the target has not moved with them. Threshold tapering is the discipline of keeping the challenge live as the brain gets better, and it follows the same logic as adding weight to a lifter's bar as their strength grows.
Early in a course, the threshold is set so success comes readily, a reward rate well over half the time, commonly around sixty percent, because an unrewarded client cannot detect what produced the reward and no learning happens (Chapter 14). As the client improves, that same threshold starts to deliver rewards too easily; the client is meeting it ninety percent of the time, the contingency has gone slack, and the training is no longer teaching. The taper tightens the threshold to bring the reward rate back into the teaching range, so the client is again working for a meaningful fraction of the rewards.
Two ways to run the taper. Manual titration means you watch the reward rate across sessions and tighten the threshold by hand when it drifts high, re-anchoring it to a fresh baseline of the client's now-improved activity. Auto-thresholding means the software continuously adjusts the threshold to hold a target reward rate, doing the taper for you. Auto-thresholding is reasonable and common, but it has a specific failure mode you have to watch: because it holds the reward rate constant by definition, the client's apparent performance stays flat across the course even when their underlying activity is improving, and a coach reading only the reward rate will see no progress where there is plenty. The fix is to read the underlying amplitude trajectory, not just the reward percentage, and a manual review of where the auto-thresholds have drifted across sessions sometimes reveals what the protocol is actually doing. The other auto-threshold failure is the reverse: the thresholds drift to a place where the contingency is no longer effective at all, rewarding noise or rewarding nothing, and only a manual look catches it.
Most courses start single-site, one active electrode, clean to keep artifact-free, easy for the client to understand. As the first site stabilizes, the question of whether to add a second arises. The principle is sequencing: add a second site when the first has stabilized, not before, and read the new site against the case picture rather than adding it because more sites feels like more treatment.
A worked example of the logic. An SMR-led course at C4 stabilizes sleep and settles reactivity over the first fifteen to twenty sessions. The sleep onset has come down, the stress reports have softened, the foundation is holding. The attention goal, though, is still in progress, and the case picture points toward adding left-side activation to bring up the engaged, alert processing that C4 alone does not target. At that point a second site, C3 or a frontal site, gets layered in, run on its own days or alternated, while the C4 work continues as maintenance. The sequence matters because the first site built the regulatory floor the second site's more activating work can land on. Layering the activation in before the floor was stable would have been the wrong order.
The opposite of disciplined progression is the over-modality trap: layering many protocols at multiple sites so that the variables co-act and the case picture becomes unreadable. When a course is running three sites and two adjuncts and the client changes, you cannot say what produced the change, and when the client stops changing you cannot say which variable to adjust. Simpler stacks are more interpretable. Add the second site when the first is stable and the case picture names a specific job for it; resist adding the third and fourth unless each has a named job too.
A course is built on an assessment, and an assessment is a snapshot. Over the weeks of training the brain changes, which means the picture that defined your protocol goes stale, and re-assessment is how you keep the targets honest and document the change. The standard rhythm has three structured points.
Intake, before training begins. The full baseline: the physiological track and the behavioral track together, plus the intake inventories with the client's history, goals, and current complaints. This is the reference point for everything that follows.
Mid-course, commonly session 10 to 20. The first measure of trait-level change. The same physiology track, the same behavioral track, the same inventories re-administered. The window sits at the inflection between the early-program calibration and the late-program consolidation, which is why it is the standard re-map point. For a QEEG-guided or z-score course, the mid-course re-map also tells you whether the trained metrics are converging on a full map rather than only on the live training display, and whether new deviations have emerged or old ones resolved that should change the targets (Chapter 16).
End of course, around session 30 to 40 or at the contracted end. The read on whether the course produced the change the client and you were working toward, what remains, and whether renewal or discharge is the right next step.
The exact session numbers are clinical judgment, not a fixed rule; a two-month course may run intake plus end with no mid-point, and a longer course may run two mid-course assessments. The principle is structured assessment at meaningful intervals with the same pair of tracks each time. And one operational caution carries weight here: a re-assessment is only comparable to baseline if you hold the conditions constant. Same reference, same montage, same eyes-open and eyes-closed protocol, same artifact standard, and as far as possible the same medication status and time of day. A re-map run under different conditions is measuring the conditions, not the training.
The session log and the client's report are necessary but not sufficient. A defensible course pairs a physiological measure with a behavioral or performance measure, run at the same intervals, because the two together produce a case picture neither could produce alone. This is the two-track principle, and it is the cognitive-neuroscience and neuropsychology approach applied to neurofeedback: a practitioner who runs only physiological assessment sometimes misses what the client is doing in the world, and one who runs only behavioral assessment sometimes misses what the brain is doing under the hood.
Continuous performance tests: TOVA, IVA-2, Conners CPT, QbTest. A CPT is the standard behavioral measure for attention and impulse-control work. The client responds to a sustained, monotonous target-detection task for ten to twenty minutes, and the test produces a small number of scores: response control or commission errors, attention or omission errors, reaction time, and response-time variability. The Test of Variables of Attention (TOVA) and the IVA-2 are the most common in neurofeedback practice; the Conners CPT and the QbTest are alternatives. You administer it at intake and re-administer it at the same intervals as the QEEG, and you read the change in the scores against the change in the physiology and the report. A CPT measures one slice of executive function, attention regulation and response control under demand; for cases where the presenting complaint is broader, some practices add a brief neuropsychological battery (working memory, processing speed) to sample more.
Behavioral rating scales: BASC, CBCL, Conners. For children and adolescents especially, the rating scales completed by parents and teachers carry information the CPT cannot, because they sample behavior across the real-world settings where the complaint actually lives. The Behavior Assessment System for Children (BASC), the Child Behavior Checklist (CBCL), and the Conners rating scales are the standard instruments. Re-administering the same scale to the same rater at intake, mid-course, and end produces a behavioral trait read across the settings that matter, and the gap between rater types is itself informative: the neurofeedback literature for ADHD shows larger effects from raters who know the child than from blinded raters or active-control comparisons (Arns et al., 2009), so a parent-rated improvement is real data and not the whole story.
Self-report: symptom inventories, the PTSD Checklist, sleep diaries. For the conditions where the client is the best reporter of their own state, structured self-report inventories carry the behavioral track. A broadband symptom inventory, the PTSD Checklist (PCL-5) for trauma work, the GAD-7 or the Perceived Stress Scale for anxiety and autonomic cases, and a sleep diary for sleep complaints. The sleep diary deserves a specific note: a client's own report of their sleep sometimes diverges from a tracker's measurement, and the divergence (a client who reports waking four times when the tracker shows two) is itself data about the client's perception, which is sometimes the thing the training changes first.
The physiological track in all of these is the QEEG for neurofeedback in general, the HRV assessment for autonomic-led work, and the device-specific signal (HEG level, sleep-tracker data) for the modality-specific cases. The pairs change with the modality; the principle does not. Pair physiology with performance, run the pair at intervals, and read both tracks against the client's lived experience as a third stream.
Not every client responds, and an honest practice names the non-responder rather than running the course indefinitely on the hope that the next ten sessions will turn it around. The judgment has to be made carefully, because the most common error is calling non-response too early, before the protocol has had the sessions it needs or before a confound has been ruled out.
The decision rests on a sequence. First, has the protocol had enough sessions to be evaluated. Effects in most surface protocols emerge after ten to fifteen sessions, not immediately, and HEG effects often need fifteen to twenty-five, so declaring non-response at session eight is premature for almost any case. Second, is the brain engaging the contingency. If the training data shows the reward band shifting and the within-session traces showing the brain finding the rewarded state, the protocol is being learned, and a flat outcome picture points away from the protocol and toward something else. Third, has a confound been ruled out. The non-responder who is not responding because of untreated sleep apnea, a substance affecting the data, a medication change mid-course, or a major life stressor is not a neurofeedback non-responder; they are a client whose context needs addressing first. Fourth, has the protocol direction been adjusted at least once. A site or band change, made deliberately and given its own four or five sessions, is part of the evaluation; a client who has not responded to the first protocol may respond to the second.
When all of those are accounted for, no change across a meaningfully long course with the contingency clearly engaged and confounds ruled out, the case is a genuine non-response, and the move is to say so and act on it. Acting on it means a referral decision: back to the referring clinician with what the course showed, or onward to a different modality or a different kind of treatment the case may need. A practice that runs a non-responder for sixty sessions because stopping feels like failure is not serving the client. The defensible position is honest: this protocol, run competently over an adequate course, did not produce the change we were working toward, here is what we tried, and here is what I recommend next.
A course generates a record, and the record serves three audiences beyond the client: the next practitioner who covers the case, the supervisor or mentor who reviews it, and, where it applies, the third-party payer. The structure that serves all three is the same clinical note structure the medical professions have used since the 1960s, the SOAP note, and it organizes the course better than any session-by-session ticket can.
A ticket says where a case sits in a workflow. It cannot tell the next practitioner what the client reported, what the last protocol was, what the supervisor thought at the last review, or what the renewal context is. A SOAP note carries the case. Its four sections map onto what a neurofeedback course actually produces. The Subjective carries the client's report in the client's words: the recent symptom-scale and sleep-diary trends, the messages and concerns the client has raised, the life context affecting training. The Objective carries the measured data: the session counts and protocols run, the survey and inventory scores summarized numerically, the most recent QEEG and CPT findings to the level current decisions need, the signal-quality flags, the adherence record. The Assessment carries your clinical judgment: where the case sits in the arc, what the recent training appears to be doing, whether the Subjective and Objective agree, and the risk flags, paradoxical responses, stagnation, a life event affecting the work. The Plan carries the next move: the next one to three protocols with sites and parameters, the communication and scheduling moves, the conditions and contingencies, the next re-assessment date.
For insurance, the Objective and Assessment sections carry the weight. A payer is looking for the medical necessity and the measured progress, and a record that pairs the standardized outcome measures (the TOVA composite moving from 87 to 96, the PCL-5 down fourteen points) with the protocol rationale and the session-by-session course is a defensible clinical record in a way that "trained the client, doing well" is not. z-Score and QEEG-guided courses add the obligation to record the specific quantitative targets and their movement, because "trained z-scores at Fz and Pz" is not a record and "Fz absolute theta from z = +2.4 to z = +1.1 over twelve sessions, PCL-5 down 14 points" is.
For supervision and BCIA mentoring, the record is what makes the review productive. A mentor or supervisor reviewing your cases reads the SOAP notes to see your clinical reasoning, not just your protocol choices, and the case-picture sentence in the Assessment, the one that says what this specific case is showing and what you are making of it, is the part that demonstrates you are reasoning rather than running a recipe. Writing it well is part of developing the judgment the mentoring is meant to build. The pre-review refresh of the note, brought to current state in the workday before a scheduled review, is the single most consequential piece of documentation you produce around a supervision meeting, because it lets the meeting be a conversation about the clinical question rather than a reconstruction of the case from scratch.
A course ends, and how it ends is part of the clinical work. The end-of-course assessment is the anchor: the final QEEG and CPT (or modality-appropriate pair) read against intake and mid-course, the symptom inventories re-administered, the trajectory laid out. From that picture the discharge conversation runs on evidence rather than impression, the client started at A, moved to B at mid-course, and is now at C, with the gains documented and what remains named.
The discharge decision has three common outcomes. The client has reached the goals and stops, with the option of maintenance sessions if symptoms drift back. The client has reached most of the goals and renews for a shorter course to consolidate or to address what remains. Or the client has not reached the goals, the case is a partial or non-response, and the discharge is a referral with a clear account of what the course produced. Maintenance sessions, when they apply, are a thinner cadence than the active course, a session every few weeks or a brief tune-up course when the client notices the gains softening, and they suit clients whose presentation is the kind that drifts back without periodic reinforcement. Some clients who have completed a course opt for an annual re-assessment, a year-over-year trait read with the same pair of tracks, even when they are not actively training, which gives both the client and you a longitudinal picture of whether the change held.
When you sit down at the end of a course to write the discharge note, the question that decides whether you have planned the discharge well is not whether the client liked the work. It is whether the next person to read the chart, the client themselves, the referring clinician, the payer, or you in a year when the client comes back, can see from the record what changed, how it was measured, and what the client should do now. The session-by-session display is gone the moment you close the software. The structured outcome picture, sampled at intervals and written into the chart, is what the course leaves behind.