#neurofeedbackpodcast #eegpodcast Dr Andrew Hill Founder of Peak Brain Institute joins Jay Gunkelman the man who has read over 500,000 EEG's and Pete Jansons on the NeuroNoodle Neurofeedback Podcast to catch up on what Dr Hill has been up to since he last was on the show. Key Moments 0:00 1:37 Show Start 3:02 How Long has Jay Gunkelman been doing EEG? 5:00 Jay Gunkelman read 100-200 EEG's a day 7:00 Eran Zaidel UCLA 7:34 Blind Men and Elephants Analogy early Days of Neurofeedback 8:37 Dr Larry Hirshberg 10:13 LANT Lateralized Attention Network Task 11:10 Jack Johnstone 13:30 EEG Phenotypes Clinical and Normal 14:48 We all have weird brains 15:43 Its not a question of whats wrong with you but what is different about you 16:00 Compensatory mechanisms 16:33 Extreme Athletes 17:14 Dr Hill last appearance 17:20 What has Dr Hill been up to since we saw him last on the NeuroNoodle Podcast? 18:16 Peak Brain Institute 18:40 How to control multiple Neurofeedback offices 21:50 Showing Brainmap helps clear up the mystery of what is going on mentally with a person 22:56 Pre Post EEG Canabis Use and Brain Fog 24:37 Artificial Intelligence coms into play with Neurofeedback 25:08 Constraints of Artificial Intelligence 27:00 Open Source EEG Databases 30:40 Adding Dimension of time to EEG 32:29 How can we combine databases? 33:28 Need to look at brain like you do you lipid panels at a physical 38:44 Dr Hill new podcast 39:01 Link 40:11 What do people want to know about their mental health? 41:55 Sleep Stress 44:18 referring out Clinical Cases 45:40 Training someone with epilepsy 46:34 ISNR Article show On Neuroregulation link 49:08 Medication Management 50:00 Attention all Stoners and Weed Smokers 50:34 Consciousness becomes clear you get more sensitive to meds 51:06 SMR Training ERP Potential 54:30 Rat Play Fun ? Cocaine 54:15 Most Addictive stuff 56:56 There is a sea change occurring to the credibility of Neurofeedback https://peakbraininstitute.com/about-pbi/ Dr. Hill is one of the top peak performance coaches in the country. He holds a Ph.D. in Cognitive Neuroscience from UCLA’s Department of Psychology and continues to do research on attention and cognition. Research methodology includes EEG, QEEG, and ERP. He has been practicing neurofeedback since 2003. In addition to founding Peak Brain Institute , Dr. Hill is the host of the Head First Podcast with Dr. Hill and lectures at UCLA, teaching courses in psychology, neuroscience, and gerontology. @peakbraininstitute7638
Episode Summary
I sat down with Jay Gunkelman and Pete Jansons on the NeuroNoodle Neurofeedback Podcast to catch up on what I have been building at Peak Brain Institute, where the science of EEG is heading, and what I see in brain maps every week. Jay has read well over half a million EEGs across his career, so this was a conversation between people who have spent decades staring at the same signals. You can watch the original conversation.
Here is what I want to walk you through: what a brain map actually shows you, why everyone has a weird brain, how neurofeedback changes addiction and epilepsy, and why I think you should track your brain the way you track your cholesterol.
What Does Peak Brain Institute Actually Do?
I am not a therapist, and I do not run a therapy practice. My colleagues who taught me are mostly clinicians with a roster of 20 or 30 patients they see for a blend of therapy and neurofeedback, usually within one population they specialize in: autism, eating disorders, anxiety, trauma. That model works. It is also a different job than mine.
At Peak Brain we operate like personal trainers and coaches for the brain. We give you back your own agency and teach you how your brain works. We see roughly ten times as many clients as the average solo neurofeedback practitioner because I teach people to read their own QEEG brain maps. We do not hand you a long diagnostic report full of labels that might be true about you. We sit down with the data, the continuous performance test, the brain map, and we work through it together.
We have four offices in the US now: New York City, St. Louis, Los Angeles, and Orange County. We opened in London. About 80 percent of our clients work from home with equipment, with live coaches available to jump on and help. That remote model turned what used to be a bug of neurofeedback, the problem of tracking how a client's brain changes day to day, into a feature. We teach you to run your own sessions and track your own progress.
Why Does Everyone Have a Weird Brain?
Jay has spent more than 50 years looking at EEG mostly through the lens of clinical pathology. When I look at endophenotypes, the patterns that show up in the EEG across people, I find the same patterns in someone who walks in with no complaints as in someone with severe dysregulation.
The phenotypes exist in normal people and clinical people alike. The same pattern is just more severely expressed in the clinical case. Statistically, you reduce the proportionality of the Mahalanobis distance in the multivariate data scatter. Less divergence within a pattern reads as normal. More divergence within that same pattern reads as clinical. Same phenotype, different intensity.
Back in 2005, with Jay, Jack Johnstone, and Joy Lunt, I published a clinical characterization of these patterns. We had been treating from the EEG, not the DSM, for years. If two people showed the same EEG pattern, they got the same treatment approach, regardless of what diagnostic label they carried. We suspected these were endophenotypic, with two of the 11 patterns having known genetic correlates at the time. The rest have since been filled in.
The practical consequence: a hot anterior cingulate can show up as perseverative, intrusive, and obsessive, or it can show up as a highly organized CEO who loves their own mind. Heavy back-midline activity can be a trauma response, or it can be an extremely skilled lifeguard who simply is not threatened by their own threat-detection system. The electrical pattern is the same. The question is what is different about you, and whether you want to work on it.
A deviation in the data can mean three different things: an abnormality, a compensatory mechanism, or a unique outlier state that functions as a special skill. Gerard Fitzmaurice in South Africa works with extreme athletes, the people who run across mountain ranges for six days or row the Atlantic solo. A lot of them carry singular patterns. Without those patterns they could not stay on task through that.
What Can You Learn From Showing Someone Their Brain Map?
When you see something on an X-ray, like a broken shoulder, you get annoyed and frustrated. You do not feel ashamed or overwhelmed by it. You say, that thing is broken, it is annoying, let me deal with it.
A brain map can do the same thing for the experiences that usually carry shame. I have seen more pre- and post-cannabis maps than I can count, along with Adderall, caffeine, nootropic stacks, post-COVID brain fog, and reactivated post-concussion patterns from a ski trip. When you can see the shift in the numbers and connect it to how you feel, your relationship with the suffering changes. The frustration stays. The shame leaves. Now you have a target.
QEEGs are hard to interpret, and people are weird, so I tell clients the same thing every time: something only matters if it is valid for you. If a finding maps onto something you actually experience, then you have a handle on it. You can change it.
The experiences people most want to understand cluster around the same few things. Sleep, stress, and attention are the foundational ones. Someone in their mid-fifties with word-finding trouble assumes it is aging. Often their alpha is running slow, their alpha speeds are scattered, their delta is intruding because they are not sleeping well, and their speed of processing has dropped. When you distill it down to how their specific brain works, the experience coheres. People go from "I don't want to know" to "wait, I can change it? Then I want to know."
How Will AI Change EEG and Brain Mapping?
Jay is a big proponent of AI. My read is that we have just discovered fire. We know it can burn things. We do not yet know how to cook a gourmet meal with it.
The constraint with large language models is that they are good only when you constrain the questions narrowly and train them on a narrow data set. To get them to do the skilled judgment we do takes work, because you have to capture an enormous number of clinical judgments and provide the data sets to train against. I have a charting system holding six-plus years of decision-making across more than 5,000 clients: every protocol, every survey, every attention test, every brain map. The path forward combines old-school machine learning and model fitting with the inferential reasoning of the newer language systems. You cannot pour domain expertise into these models directly. That is the central problem.
To characterize normal variability, you do not need half a million records. You need somewhere between 10,000 and 100,000, depending on what you are measuring. Jay made the point I agree with completely: you cannot pool everyone into one group. You randomize into two, so you can validate a finding from the first group independently on the second. Without that step you can end up in a box canyon and not know it. Replicability is one of the hardest problems in neuroscience, and we have to build it in from the start.
The clinical clusters are harder. Stanford handed a private spinoff 37,000 EEGs that were already cleaned and digitized, but they were nearly all epilepsy and encephalopathy cases, because that is what EEG gets used for clinically. That data cannot characterize normal variability. We need to start treating brains as normal variation, not pathology.
Can Adding Time Cut the Data You Need?
I think we can take the conceptual model of QEEG and add a dimension of time. Imagine wellness trackers for the brain: people logging sleep, stress, seizures, migraines, mood, drinking, exercise, with periodic brain maps against their own normal variability. Capturing that intra-individual variability with context changes the math. A power analysis would likely show you can cut the needed sample from around 200,000 down to roughly 20,000, because you are picking up the seizure that happens every few weeks, the recurrent migraine, the contextual data around each snapshot.
That same logic lets us stop demanding pristine EEG. Right now we control out caffeine, cannabis, fatigue, and stress to get clean data. I want people to walk in, sit down, do a map, and walk out, even if they had coffee that morning and smoked the night before. With enough contextual characterization, we can look right through the noise and still make judgments. That is coming.
Why Should You Track Your Brain Like Your Lipid Panel?
You manage your blood pressure and your cholesterol with a baseline and a year-over-year comparison. Above your neck deserves the same attention, and the irony is you cannot feel most of it directly.
If you have a data point distorted by your suffering, your medication, or your sleep, understanding why it is distorted gives you agency. That is the whole pitch. Your brain is part of your physiology that you manage, not a pathology you wait to be diagnosed with. We run a membership model for brain mapping for exactly this reason: people pay once and come back through the year to use mapping and attention testing as a tool, watching themselves change.
How Does Neurofeedback Help With Epilepsy?
I am not a neurologist, so I cannot treat epilepsy. I can train someone who has epilepsy to operate their brain more optimally, working alongside the neurologist or epileptologist managing the case. I never work behind the scenes. The physician treats the seizures. I teach the brain to function better. When the seizures resolve, the doctor pulls the medication, because no physician wants you on anti-seizure meds when you are not having seizures.
We just published a case in Neuroregulation with Rusty Turner, Sue Wilson, and the treatment team. The patient, Isabella, had a little neurofeedback in the US as a child, then developed intractable temporal-lobe seizures as a preteen after her family moved to Barcelona. The US EEG had shown the seizure activity in the temporal lobe. Neurologists and a neurosurgeon in Barcelona recommended removing her right temporal lobe to control the intractable epilepsy. Her family chose neurofeedback and stopping medication instead. They were told that choice would kill their daughter. She has been seizure-free and medication-free for about seven years. She earned a Division I tennis scholarship, made four years of Dean's List, and graduated with honors from Baylor. She wanted her case published so she could give lectures about intractable epilepsy and life outcomes. She bloomed.
What Happens to Medication When You Train the Brain?
With antidepressants and anti-seizure meds, the pattern is consistent: neurofeedback brings the floor up to meet the person, and the medication question takes care of itself between the client and the doctor.
I am more pointed with my warnings around psychostimulants and cannabis. Somewhere between three and five weeks into training, tolerance to stimulants and to cannabis often gets abolished. The person suddenly gets three, four, five times the subjective effect from the same dose. That can cause real problems: the chronic cannabis user who cannot get off the couch, the kid on Adderall who is suddenly too irritable to sleep or eat. So I warn parents of kids on stimulants and I warn chronic cannabis users: a couple weeks in, you are likely to get potentiated effects as you become more sensitive. Make an appointment with your doctor and plan for it.
The mechanism is that as consciousness clears, sensitivity climbs. When you are clouded, you cannot spot the effect of a drug. As you clear, it becomes obvious. The literature on SMR training shows this directly. Look at a visual event-related potential and you can see how much cortical recruitment a visual stimulus gets. Run SMR and the size of that visual evoked potential jumps, because you have removed somatosensory interference. Lower the internal noise floor and signals you could not detect before become visible.
There is a plasticity effect too. Motor evoked potential work shows that a single session of SMR lowers the threshold needed to make the hand twitch when you deliver a small magnetic pulse to motor cortex. One session, threshold way down. SMR makes the system more sensitive, more plastic, more changeable.
Why Do People Lose the Taste for Their Drug?
It is true. People often lose cravings for alcohol, cannabis, and sugar after neurofeedback. We ran a research project on 30 addicted individuals across a range of substances and found two primary drive mechanisms toward addiction.
The first is over-arousal, with three EEG signatures: fast alpha, low-voltage fast, and beta spindles. If your drive was over-arousal, alpha-theta training was part of your protocol. When you remove the drive toward the substance and then get exposed to it, you often have a bad reaction, because there is no longer any reason for it. Basic psychology: remove the drive, the behavior tied to the drive goes with it.
The other third of the addicted group had anterior cingulate drive, an obsessive-compulsive pull rather than over-arousal. If you do not address that, you get symptom substitution. They get clean and sober and lock onto something else. Treat the cortico-striatal pattern and the substitution stops.
I think it is simpler than the mechanisms suggest. The most addictive behavior is driven by learning, and neurofeedback changes learning. The Rat Park work makes the point: when the park is enriched and fun, the rats live; when the park is empty, they do cocaine until they die. Give the brain an enriched, flexible learning environment and you can shape the direction it goes and break the stuck reinforcement loops. You are self-regulating, even when you are not aware of it. Even involuntary behavior is still yours.
Is Neurofeedback Becoming More Credible?
When I was interviewing for grad school, people rolled their eyes at the word biofeedback. Two or three years in, senior scientists across different departments were coming to me, the neurofeedback guy, offering test suites for my research. There has been a real change in legitimacy as the technology matured.
Internationally it has exploded. European neuroscience adopted neurofeedback as an experimental manipulation without the pejorative baggage it carried in the US. The Schwarz lab in Salzburg brought it into their experiments, and PhD dissertations across Europe now use it as a primary tool. There are Spanish, Italian, and pan-European applied-neuroscience societies. In Asia, Korean teams built a full database with a dry headset and are running AI scrubbing to predict mild cognitive impairment and dementia.
A little hangover remains in the US from the early hippie reputation. That was partly Jay and the first generation. We have come a long way since then. The field needs a few more conversations like this one to keep moving the word along.
If you want to go deeper into the science we covered, my podcast Head First with Dr. Hill is back with weekly episodes, and you can start by getting your own brain mapped to see where your patterns actually sit.