This piece comes from my Monday night livestream, where I run a live neurofeedback demo and then teach a topic in depth. This week I covered what actually changes your results when you train your brain: medication timing, session frequency, supplements, sleep, and the protocol decisions that separate good outcomes from disappointing ones. I run Peak Brain Institute, and most of what follows is observation from about 14 years of coaching and reading brain maps, not published trial data. I will flag which is which as I go.
What gets in the way of a clean QEEG?
Before you train, you map. A QEEG (quantitative EEG, or brain map) puts 19 electrodes across the scalp plus ear references, and you sit with eyes closed for about 10 minutes and eyes open for about 10 minutes. If there is any seizure history or neurological concern, record longer to hit the 20-minute total a neurologist wants to see. The point is to compare your resting brain activity to an age-matched population, find the features that stick out, figure out which frequency bands they live in, and decide what is worth working on. You can read more in my QEEG brain mapping guide.
What you put in your system distorts that comparison. The database you are matched against did not have caffeine on board, and caffeine moves the EEG hard. Stop all caffeine by 4pm the day before your map: coffee, green tea, chocolate, orange soda, tiramisu, all of it. By the next morning it is mostly cleared in EEG terms.
Psychostimulants need more lead time. Adderall, Ritalin, and similar drugs take about 48 hours to wash out, and getting them out of your system reveals the patterns you actually want to train. Regular cannabis smokers need about 24 hours; regular edible users need 48, because the effect lasts longer. You can also do a deliberate with-and-without comparison, mapping both states to see in real time what the stimulant does to your resting EEG and your executive function on an attention test.
Most other medications do not need washout. SSRIs, mood stabilizers, and other blood-level drugs shift the EEG without hiding the patterns. Map in the first half of the day, before your meds and supplements, well-rested and uncaffeinated, get the timing right on the acute stuff, and you will know what you are working with. There is no reason to pull someone off long-acting medication for a brain map.
Do you have to change your medications during training?
For the training itself, no. When I run a session I am training the brain relative to itself, moment to moment, not against a database. So whatever you need in your system can stay. Theta and beta still respond the way they respond, second by second.
There is a real effect to watch for. After a few weeks of regular training, usually in the three to five week range, tolerance to some substances can drop. The same Adderall dose can feel like twice the dose. The same cannabis can hit two or three times harder than you expect. In my experience that is a signal the brain has become flexible and responsive. It is also an opening to ride the dose down with a prescriber, and to renegotiate a relationship with these substances.
For SSRIs and mood stabilizers, there is no potentiation like that. What I hear about instead is the floor coming up to meet you. Resources strengthen. Once mood is more resilient, that is the time to talk with a doctor about tapering, because antidepressants do not appear to interfere with the learning.
A few things do blunt the learning. High-dose benzodiazepines (Xanax) and high-dose opiates (Vicodin) both slow neurofeedback down. Progress gets harder and session count matters more. If someone is overusing those, the lower-dose conversation with their physician is worth having, because neurofeedback works on the same anxiety and pain those drugs target. Even active drinking seems to slow training only modestly; people can often train straight through it.
How often should you train, and when does it become permanent?
Train three times a week. Keep at least one rest day between any two sessions, so the sleep and next-day consolidation cycles can build. Three times a week sits in the sweet spot.
The shared data I have seen at conferences and from colleagues (not published, treat it as clinical observation) suggests three sessions a week is roughly twice as impactful as two, and four is only slightly better than three. My working explanation rests on something that is well established: a single neurofeedback session produces a measurable plasticity burst. Stimulate the hand area of motor cortex with a TMS coil and you get a motor evoked potential, the hand twitching at a certain energy level. After a session, the energy needed to fire that tissue can drop. The cells are primed to fire. Train every other day and you keep that plasticity elevated broadly, not just briefly after each session. Single-session neurofeedback has been shown to alter cortical excitability measured this way (Ros et al., 2010).
This is also why other interventions start working better. I get calls from physical therapists asking what I did to their client, because her balance is suddenly different and she walked in without a cane. Meditation, weightlifting, language learning, and trauma therapies like EMDR all start producing more change. Raise plasticity in one area and everything that depends on plasticity moves. If you want to understand the mechanism more, see biohacking plasticity.
On permanence: I suggest remapping every 20 to 30 sessions. You can usually see change in the data after 15 to 18 sessions, sometimes a full standard deviation of z-score shift, which is enough for teachers and family to notice on their own. For regulatory goals (attention, anxiety, sleep, sensory and social processing, speed of processing), durable change tends to kick in around 40 sessions. That is roughly three months at three times a week, often a couple of standard deviations of change in both executive function testing and resting QEEG. These resources tend to lock in because you use them every day; the brain keeps practicing them. My ADHD neurofeedback guide goes deeper on those goals, and the maintenance question is worth reading once you hit a stable floor.
Severe and unusual presentations behave differently. Seizure disorders, language difficulties in autism, significant traumatic brain injury, large developmental issues: these move slowly and less predictably, and may need ongoing or repeated rounds. A child with non-verbal autism, a seizure disorder, bedwetting, ADHD, and sensory issues might see the sensory, executive, and bedwetting goals settle by 40 sessions while the seizure work is nowhere near done.
Why does sequencing protocols matter so much?
The single most common reason people come to me after a disappointing course of neurofeedback elsewhere: their protocols never changed. One protocol run 20 times, or a closed "magic box" system the provider could not adjust, pushing the brain further in one direction. That should not happen.
Good neurofeedback is iterative. I run two to four sessions of a small set, test, see what moved, and shift to avoid overtraining in any one direction. I rarely run the exact same thing more than two to eight times in a row, and the longer stretches are usually for very young clients or acute needs where something is clearly working. Most people do better changing every three or four protocols. This is exactly why I ask about your sleep, stress, mood, and what you noticed. If the protocol was working and then session four brought worse sleep or a flatter effect, you are starting to overtrain, and it is time to switch.
In the demo I ran a C4 minus A1 SMR protocol: an SMR reward at 12 to 15 Hz, a slow inhibit at 4 to 7 Hz (theta), and a high inhibit at 22 to 34 Hz, training above the right precentral gyrus. SMR (sensorimotor rhythm) sits at the center of how I do this work. I tend to start or finish with it, and often blend it into protocols that are otherwise not purely SMR. SMR adds gravity to a session: it gives the brain control over whatever process it just went through, so a strongly activating beta protocol does not leave the system overreacting. It also encourages plasticity and learning, which keeps things changing, and SMR training has been associated with improved sleep, which in turn consolidates the day's learning (Hoedlmoser et al., 2008). One of my mentees told me clients feel my protocols more than others, and a big part of that is careful QEEG tailoring plus a generous layer of SMR. More on that band in SMR neurofeedback.
Sequencing creates genuinely different effects. A protocol I call a "sweep" runs bilateral hemisphere protocols, physically moving the wires through adjacent locations and training each with its own frequencies, producing heavy co-activation of tissue. Running an F3/F4 dual into a C3/C4 dual feels nothing like running those four segments alone. If you have protocols you like, experiment with combining them and changing the order.
What about the deep, trauma-release protocols?
The Sebern Fisher style attachment-trauma protocols deserve a separate warning, because they break the usual rule. Most protocols feel gently positive afterward: a little calm, a little focused, a little relaxed, or nothing at first, with a subtle positive effect emerging over the next few hours. The attachment-trauma protocols feel crunchy, irritated, and disconcerting right after, because they reach deep posterior and subcortical tissue.
I developed a version of these, a "double down Alpha" protocol, and handed it to three clinicians I was supervising. Every one of their clients said some version of "I did not like that, do not do that again," and then came back within a day or two asking for more, because the later reorganization felt good. The theory comes from a small fMRI study Fisher discusses: a group of people with dissociative PTSD and developmental trauma, training above the posterior cingulate to reach the periaqueductal gray and deeper tissue, with reported changes in posterior cingulate-amygdala coupling after a single session. These are demanding protocols. If you are going to try them, go slowly, start with three to six minutes, and expect to feel worse before you feel better.
Can you stack neurofeedback with psychedelics, ketamine, or TMS?
This came up repeatedly in the Q&A, so here is my position.
Psychedelics and microdosing: unnecessary alongside neurofeedback, and risky. People microdose to raise plasticity, and a few sessions a week already has your BDNF high and your brain primed. Within a few weeks, neurofeedback potentiates the psychedelic, so your microdose stops being a microdose. I have had professors and programmers ignore that warning and end up hallucinating mid-lecture because the same dose hit far harder than expected. If you want more plasticity, add meditation or 10 to 20 minutes of yoga. That plus neurofeedback puts you at about as much plasticity as your brain can use.
Ketamine: impressive for many people, and I have watched it lift depression fast. Read the QEEG first, though. Anxiety is usually a beta phenomenon, the gas pedal cramped up, sometimes with low alpha. For a subset of people, the anxiety is driven by excess theta, meaning specific tissue is disinhibited. In those people, ketamine can erode neurofeedback gains. I have seen it blow up hard-won progress twice in clients I knew well. You can repair it, but be cautious adding strong interventions on top of training.
TMS works well in its own right, but combining it with neurofeedback pulls at cross purposes.
The pattern across all of these: adding more is not automatically better, and the wrong addition can undo your work.
What lifestyle factors make training stick?
Sleep first. Neurofeedback almost always improves sleep, especially SMR work. But if your sleep is badly disregulated (narcolepsy, shift-work circadian disruption, years of poor sleep), the changes still build in the maps over a couple of months while you feel almost nothing, which makes the training harder to guide. Stabilizing sleep helps you feel the progress and steer it. My sleep biohacking guide covers the details.
Two sleep moves I lean on: stop eating before bed, since fasting before sleep allows more growth hormone and deeper sleep, and protect the first part of your morning. Get up early seven days a week and spend 10 minutes on something gentle before you eat and before you sit down: a walk, low-key yoga, sun salutations. That morning anchor sets your circadian timing and supports everything else, including your training. I lay this out in biohacking your morning.
On caffeine timing, delay your first cup for an hour or so after waking. The cortisol and blood sugar that woke you up will burn off during gentle morning movement, and you do not want to add caffeine on top of a system already running on those signals. One exception: if your only hard workout window is first thing in the morning and fat loss is the goal, caffeine before intense resistance training is lipolytic and useful. With gentle morning movement it does not help; with hard training it does.
Meditation is the lifestyle add I recommend most, both as a brief morning practice and, for those who want faster change, as a more substantial sitting practice. It dovetails with training instead of competing with it. See biohacking meditation.
One more environmental note from the Q&A. Lead exposure genuinely impairs the brain and can work against training. Lead is chemically similar to calcium, slips into calcium receptors on neurons, and binds essentially irreversibly, blocking calcium signaling at that site. We can work around some of it, not much. If you have an old lead water line, test it and address it.
How does neurofeedback compare to medication for ADHD?
A common misreading floats around online: that neurofeedback only works alongside medication, implying it is weak on its own. That is not what the studies show.
There are trials comparing neurofeedback, stimulant medication, and both combined. In one large randomized trial, combined training plus medication and medication alone both produced stronger short-term effects on some measures than neurofeedback alone (Gevensleben et al., 2009). One line of research followed adolescents over the longer term, and the picture shifts: neurofeedback effects tend to persist or grow over follow-up while medication manages behavior in the moment without producing as much durable change (Arns et al., 2009). Medication managed behavior without always producing lasting change.
There is also a critical limitation in how this research is run. Trials almost always pick one protocol and apply it to a group. That is not how neurofeedback works in practice, where it is iterative and constantly adjusted. Running one fixed protocol is close to designing the study to underperform, and even those studies show real effects on reaction time, processing speed, and executive function.
The effect may extend well beyond ADHD. Gary Schummer ran a study with HIV-positive participants in the mid-1980s, before treatments existed, comparing cranial electrical stimulation alone, neurofeedback alone, and both. The groups that received neurofeedback, with or without stimulation, were reported to show a change in CD4+ T-cells over the study; the stimulation-only group did not. Neurofeedback tends to make other interventions work better, from weightlifting to physical therapy to talk therapy. Therapists send me clients who are stuck, we train the brain, and the therapy starts moving again.
The bottom line
Map clean: caffeine out by 4pm the day before, 48 hours off stimulants, the right washout for cannabis, no need to drop long-acting meds. Train three times a week with at least one rest day between sessions, and expect regulatory goals to settle in around 40 sessions. Keep the protocols moving and tailored to your QEEG, lean on SMR, and report what you feel so your provider can steer. Stabilize sleep, protect your morning, and add meditation rather than stacking strong drugs on top of an already plastic brain. If you want a starting point, get a brain map and look at your own data before you decide what to train.