Episode Summary
I joined the team at the NeuroNoodle Neurofeedback Podcast for a wide conversation about how remote neurofeedback actually runs, why consumer EEG gear succeeds or fails, what infraslow training does and does not do, and the strange edges of EEG where consciousness, gamma, and slow cortical potentials meet. You can watch the original conversation on their channel. What follows is drawn from my side of that discussion.
What does running a remote neurofeedback practice actually take?
I run a hundred or more remote clients at any one time, across India, Australia, Hong Kong, the East Coast of the US and Canada, Sweden, the UK, Dubai, and the broader Middle East. People assume the hard part of a remote program is the neurofeedback. It is not.
The neurofeedback itself is a practical process once you know the brain. Arousal-model and regulatory-model phenotype training off a QEEG is straightforward when you have done it with enough people. The hard part is the human infrastructure. You need staff seven days a week, twelve hours a day, because clients live in every time zone. You manage language barriers. I speak several of my clients' languages, but not all, and working through a daughter or an aunt to ask someone about sleep, attention, mood, or stress is its own skill.
Neurofeedback survives a language barrier better than therapy would. I am asking about regulatory shifts in day-to-day experience: fluctuating resources like sleep, mood, focus, and stress. Those are easier to track across a translation gap than the subtleties a talk-based session demands. If you want the mechanics of how this works for clients, I cover it in Remote Neurofeedback: How It Works and What to Expect.
You do not need a live internet connection to train. We provide live support, which helps, but training itself runs on intermittent connectivity. I use tiny battery-powered QEEG amplifiers now. Geography is rarely the limit. If you are in a country where I have an office, we rent you gear. If you are not, you travel to one or you buy gear.
Why is EEG so much harder in Europe?
The 50 Hz line noise across Europe is a real problem. The buildings are wired differently, and ambient 50 Hz floods many of them. My working theory on why nobody seems to care: 50 Hz does not stop your heart the way 60 Hz can, so it never got engineered out the same way. Many older European buildings were built without grounded circuits throughout, and an ungrounded circuit near your recording site pulls noise straight in.
A notch filter handles this to a point. A notch from 45 to 55 Hz that drops at 120 dB will clean up a reasonable amount of contamination. If you are sitting too close to a transformer or a heavy source, nothing saves you. The filter is good up to roughly 100 dB; push past that and the noise wins. Move away from the source.
Can you build neurofeedback gear on a Raspberry Pi?
People keep asking whether shrinking hardware and cheap open AI models mean you can build your own system. The first question I ask back is which problem you are trying to solve. Machine learning can do many things; you have to name the one you want.
Some of these projects are genuinely doable. HRV biofeedback is trivial for processing. You can pull a heartbeat off someone's face with camera-based red-channel filtering, no contact needed, add a breath tracer, and build a working HRV kit on a small board. The constraint is the amplifier and the money, not the math.
For EEG, you need a good amplifier with high-pass, low-pass, and notch filtering and clean engineering. You can now buy an instrumentation amp on a single chip with a high sampling rate and a multiplexer, which is what we struggled to build on two boards decades ago. The economics are the wall. Manufacturing costs money, and this is a tiny field.
One technical caution if you go this route: if you get an amp that multiplexes through a single chip, you cannot remontage afterward without introducing error. For flexible montage work, you want a chip per channel.
Where I trust AI in clinical work, and where I do not
I do not yet trust these systems to do EEG analysis. There is no ground truth in EEG, and the literature is full of plausible-sounding output that does not hold up. When a colleague ran a generative model on far-field phase reversals, it produced text referencing "true sources," which do not exist in EEG. I would have graded it a C-minus.
Here is how I actually use it. I record 30 to 45 minute videos walking a client through their QEEG and continuous performance test, every time, systematically. I built a tool that ingests those videos, breaks down the likely phenotypes and goals, and pulls from my list of starting-point protocols that might match. It formats a summary off the analysis I already did. I put the model in the teaching role, after I have done the clinical work, and I can read the output to confirm it is valid. I will not hand it raw waveforms and trust it to flag transients, because that saves me nothing; I still have to look at every trace myself.
For the diagnostic side, the field is moving. A group recently published an algorithm that detects spikes human readers miss most of the time, and a related approach that finds transient patches of tissue suppressing spikes, brief frozen rafts of data that function like a spike without looking like one. Detection like that is useful. Interpretation is still mine.
What do brain maps show before and after training?
The most satisfying part of this work is the pre and post comparison. The electrophysiology in some cases is dramatic, and the maximum change shows up locally, right where we trained. The data plus the person's own story is where the meaning lands.
I reviewed a 75-year-old recently, four months apart. He developed a clean 10 Hz alpha rhythm at the back of the head, beautiful topography. He is sleeping deeply, dreaming, and remembering his dreams. He could not recall ever dreaming or remembering a dream before. He is a well-known sports figure who set up a foundation to fund neurofeedback for traumatic brain injury in football, including for people who cannot pay. If you want the background on what that posterior alpha rhythm means, I wrote Decoding Alpha Waves: Your Brain's Idle and Its Brakes. The broader picture of how QEEG drives these decisions is in the QEEG Brain Mapping guide.
TBI runs deep in sports most people do not flag. Umpires take unprotected balls and bats to the face for years and often carry head-injury histories worse than the players.
What does gamma have to do with consciousness?
People ask what I see in psychics, healers, and people in ecstatic states. For the first twenty years I did this work, I would have said there was little reliable correlation between what someone claimed to be doing and what their EEG showed. In the past five years, a handful of people have shown me the same phenomena again and again. A pet psychic voluntarily shifted her brain the same two ways with two different techniques, came back two months later, and did it again.
Two separate things get conflated here. Gamma 2, the 80 to 100 Hz band, is associated with long-term meditation traits, the acquired changes documented in Cliff Saron's work. That is trait-level change, not the same as an ecstatic state, which tends to be absorptive and dissociative. Gamma is part of consciousness itself rather than specific to the ecstatic state. If you are conscious during an experience, gamma will be present.
The cleaner statement is this: consciousness is the relationship between your slow cortical potentials and your gamma. When those two are time-locked or phase-locked, you have consciousness. When they decouple, you do not; you are deep enough for surgery, or in a state that cannot be described.
What I see in some of these performers is the opposite of what you would expect. Bursts of focal delta bloom out, as if they are shutting off parts of the brain, at the moment they appear to dissociate. Big delta is a suppression pattern. Years ago I watched someone put out a thousand microvolts of ten-second waves, which is subdelta, infraslow content, alongside massive gamma. That pattern is real, not a sweat artifact. Perspiration produces large slow sways, but each sweat gland fires independently, so there is no sync and no pattern. This was patterned.
Can two brains synchronize?
Yes, and the mundane version is genuinely useful. In couples work you can hook two people to one feedback system and downtrain shared arousal. You watch one person slow down or speed up to match the other. There is nothing spooky about it. It is good EEG-based couples regulation.
The stranger version we published in the Society for Scientific Exploration's journal. A healer and a healee at a distance, not touching. The healer produced standing waves that phase-synchronized the other person's EEG. Creating a standing potential takes large harmonics, and he hit the Schumann resonances, which are plural: 7.83 Hz and a long stack of harmonics, with the fourth harmonic up in gamma. The healee was not volitionally trying to sync. Phase synchrony appeared in under a second once the healer produced the frequency and its harmonics.
The limit is clear. This identifies a connection, not healing. If I call someone and they pick up, that is a connection, not a conversation. Something still has to be delivered for anything to happen. What we can do now is identify that connection blind, which is a step up from not seeing it at all.
Why does montage choice change what you see?
A few technical points came up that matter for anyone reading EEG. "Bipolar" as an electrode-placement term is a common misnomer. Your amplifier is a push-pull differential amplifier with two inputs and a ground. You can place it sequentially across two active sites and read the difference, or pick a single spot you assume is neutral. No spot is ever truly neutral. Ear references are particularly bad; they pick up temporal-lobe activity and smear it as though it comes from everywhere.
You need more than one montage to know what you are looking at, the way you need more than one angle to judge a face. The Laplacian montage is the one I keep coming back to, even though newer practitioners treat it as old-fashioned. It is somewhat insensitive to global distortions, so medication and fatigue effects wash out and you gain spatial precision. It does not eliminate a medication effect; it reduces it. You will see an ADHD theta signature, or a benzodiazepine intoxication pattern, more accurately localized in the Laplacian than in linked ears.
On training across two sites: a C4-PZ sequential placement tends to speed up alpha at PZ. Faster alpha than you are used to feels like higher function, not quite like training beta in the same spot. The protocol picks up content from both locations, so you are effectively shaping two sites rather than synchronizing them. If you want to understand the focus-and-calm band most people train, I cover it in SMR Neurofeedback: Train Sleep, Focus, and Self-Control.
A note for anyone building feedback: high-order FIR filters give you very accurate frequency estimates but heavy latency, which smears the time relationship between the wave leaving the head and the beep coming back. Push past third-order filtering and learning falls apart, because the feedback is no longer tied to the brain in time. Wavelets are the fix; fast and accurate, and you build a whole bank of them.
When is infraslow training the right tool?
Infraslow is not my modality, so I will stay in my lane. Infraslow content is critical for brain function. Europe historically filtered out the high frequencies and looked only at slow cortical potentials; in the US we filtered out the slow cortical potentials and looked only at the oscillatory EEG. You do not understand brain function until you put both together, because consciousness lives in the relationship between them.
In practice, infraslow seems to work well for a couple of cases. The profoundly over-aroused nervous system, including profound autism, looks like a good target, and trauma-informed therapists are reporting success. My honest view is that it is a forced-quieting tool rather than a finely controllable one, and the field is still working out where to train and what frequency tuning to use.
Most infraslow trainers work symptom-based and do not look at a QEEG, though that is changing. Whichever modality you pick, the better question is what your brain actually needs, and the only way to know is to look. An evaluation should point to what needs to come up, what needs to come down, and which relationships between regions need to change. That guides the training. Nobody has run the full horse race comparing every technique head to head, and the combinatorics make it nearly impossible to fund. Most studies also force one protocol on a whole group, which means they are studying a technique, not the personalized clinical practice that neurofeedback actually is.
Are consumer EEG devices worth buying?
It depends on the task. If you want to float a ball up and down on a screen, most consumer devices are adequate. For an ERP, the amplifier barely matters, because the ongoing EEG is noise relative to the event-related signal. The real question is reliability and validity for your specific purpose. The professional-grade signal-quality picture is in my research overview on neurofeedback legitimacy.
Realistically, you cannot enter the EEG space in a usable way for under a thousand or two. You can buy an OpenBCI Ganglion board for around a hundred dollars, but you have to already be an EEG expert to get anything out of it. The bigger danger is people who spend thousands on gear they do not understand.
Two things you can spot with the naked eye on a wearable. Look at the electrode. A solid shiny metal disc has contact jitter and is a weaker design. A polymer-coated sensor means they thought about contact jitter; it is not truly dry, it is a damp sensor holding a little moisture as a partial salt bridge, which is a step up. And get an amp with a chip per channel rather than a single multiplexed chip if you ever want to remontage cleanly.
This is a niche field. There are maybe 15,000 providers and perhaps 30,000 active installs of hardware and software worldwide. There is no development pressure and no financial pressure, and the field is smaller now than two years ago as its founders age out. That is why product maturity stays low and why service, not hardware sales, is where the durable value sits.
Where to go next
If you want to see this work in action, I run weekly live streams on YouTube at Dr Hill, demoing neurofeedback and walking through how to biohack the brain, and you can find the clinic at peakbrainla. For the foundational reads tied to today's conversation, start with the QEEG Brain Mapping guide, then Biohacking with EEG Phenotypes to understand how we match training to your specific frequency patterns.
The throughline of the whole conversation is simple to state and hard to execute: look before you train, use more than one perspective on the data, and personalize the protocol to the brain in front of you. The gear keeps shrinking and the models keep improving, but the clinical judgment is still the part that does the work.
Full Transcript
Welcome to Neuronoodles neuro feedback and neurossychology podcast featuring tech legend Jake Gungkelman. He's the man who has read well over a half a million brain scans and Dr. Maurice Swingle, author of I Minds. Our goal is to provide information and promote options for better mental health. Hey everyone, if you're enjoying the show, want to help us keep it going, please consider supporting us on Patreon. Your support helps improve the content and brings on great guests and grow the show. Join us on Patreon at patreon.com/neuronoodle. But hey, if supporting financially isn't in the cards right now, no worries. You can still help us out for free by liking, subscribing, and sharing this episode. It really makes a difference and helps more people find the show. If we've earned it, hit that subscribe button and turn on notifications so you never miss an episode. The Neuron Noodle Podcast is supported by listeners and businesses just like you. And our uh uh we had a student presentation would always start the morning and uh there were a few of them but they only had a very brief time period to present and um they they had to do it on time and the student was running out of time and I would give them a 5minute mark from the back of the room and I'd walk up to the side and then as they time was running out, I'd get closer and closer to them, and as their time was out, I would take the mic off of them and switch the next person on. Well, as I got closer, he he put another slide on the overlay, so he wasn't going to stop. And I I stepped right next to him, and I literally chased him around the podium, and uh the audience thought it was funny as can be. But uh uh when it came to the award ceremony later on, we we chased him around the podium again. It turned out this is Ephemus Angelicus. He's an now a main editor for a major neuroscience publication in Europe. Uh he was a Greek student at Lubar's lab at the time. So we we always had fun. But after that the students would look at a stopwatch and time their presentation you know to be done on time. Uh >> is that operant or classical conditioning? And I thought this was last week. >> Uh well it's u is conditioning them. But you know we did it to Barry Sturman too. You know Barry the the EG for the new millennium in 2000. We had the the aquarium uh for our party in Monterey, but the during the meeting uh he he had a full hour. We gave him an hour cuz you know he's a major presenter. Time came to 10 minutes. I gave him a 10-minute sign from the back of the room. He said, he shook his head no. And uh so I I was doing my normal thing. I walk up next to the side and as I was getting closer to uh to him, it was obvious he wasn't going to stop. He he did another acetate overlay. He was just barely past the cat story. He wasn't really into anything new. And uh I looked over to the side and uh then uh uh Nolan White and uh Big Joe u Horvat were on the front row and I just motioned to him. They both came up, one on either side of him, picked Barry up by the arms. His feet were in the air kicking and he said, "Let me get my slides." I grabbed his stuff and the next person was already on. So, he finished on time as scheduled. Uh, but we gave him a special presentation in another room to finish his talk uh because he wasn't anywhere close to being done. Uh so anybody who wanted to catch the last part of it could get another shot at at uh finishing it. But um that was the most popular video we had was was him being carried off stage. So >> I understand. I understand. It's funny though with timing. It was what was it two years ago now? I think it was two years ago. Uh there were around six of us in the same uh conundrum. We all thought that we had uh been awarded uh an an an hour each. Uh we were our presentations and then we got there and we literally had 15 minutes each. You know, it was one of the scientific programs. Um and it was obviously massive communication because all six of us different presentations had prepared for an hour that oh that was that was something that was something just trying to go on the spot. >> Well, I've always said I would walk out of a talk I wasn't enjoying >> even if it was mine. And I was on a panel discussion. There were three of us and we talked at front 15 minutes each and then 15 minutes for questions. You know, it makes good sense. >> And the first person went like 20 minutes easily. And uh the second person was running long and uh I realized he was going to go over as well. >> Uh I grabbed my material and I said, "Obviously, you need more time. You can have mine." I got up and I left. I walked out of my own talk. I didn't do a talk. I got I got up and left. So I wasn't having fun. >> Dr. Hill, >> I'm a devout nurse. So >> Dr. Hill, what's the update, man? You all right? You still got the same place? >> Still got the same place. We're still here in California. Yeah. Um, >> no, we're glad you're okay. All All of our LA peeps, all of our California peeps. >> Yes, exactly. >> It rained over the weekend. We It rained just enough to put like to put some water on the ground without making ma massive flows of chemical sludge everywhere. So, that's nice. >> Good, good, good, good. Yeah. >> Uh, you know, it's a Tuesday. I'm hanging out with some neurogeogs. Neuronoodling. It's what we do. Hey, >> hold on. Did I just fly around the world? Is it Tuesday or Wednesday? I'm >> Oh, Wednesday. Sorry. Sorry. Wednesday. >> I thought it was Wednesday. >> I have clients in India. I have clients in Australia. I have clients in Hong Kong, >> on the east coast, US, on the east coast of Canada, and I never know when it is in my own time zone because I'm always thinking about I'm talking to somebody, it's that time there, and I lose track of what day it is completely. >> I hear you. I think if you know if I were ever you know knock on wood in some kind of an accident and to orientate me they they ask me what day it is I'll fail you know I'll totally my weekend now is Monday Tuesday because I work Saturday Sunday it's like I don't know what day it is >> what kind of people are coming to the uh office or in your case you know 20% Dr. ill. >> Yeah. Well, the new year, right, we have this gym uh new year, new brain kind of focus where everyone decides that they're going to do some stuff. Um I haven't seen as much of it this year, the kind of New Year's uh resolution uh push, but there is some of that. And there's a certain amount of people hitting back to college and back to uh secondary school that are getting their first like wakeup call in new classes, new teachers, new uh progress reports. And so we're starting to get some of the kid uh wave that comes uh a few weeks after the school year starts. Um but I I find I get a a New Year's wave in February. Sometimes people like put off, >> you know, the transformation, put off looking at their brain, put off their dentist appointments, put off their oil changes, and then they're like, "Ah, it's already February. Oh crap, I better get on some stuff." >> So >> So that's next week. >> What? Yeah, next week. Exactly. I'm hoping for a big a big rush. Come on in. We'll on the re on the remote stuff. Anything different? >> Not really. We're expanding that. We now have a remote services out of Florida. Um and we're also on boarding a partner in Nova Scotia. And so we sort of have the entire East Coast locked up uh from >> New York City, Nova Scotia, New York City all the way down to Florida now. Um but just because of uh we have so many of our clients run remote. I tend to run about a hundred or more remote clients at any one point right now and it's just a lot of customer support. I mean the hardest part of a remote program is not the neuro feedback. >> Um we're doing you know arousal model regulatory model phenotype training off of QEGs. That is a somewhat straightforward process. If you know the brain and if you've done a thousand you know done nerf with a thousand people or something you kind of get the hang of doing that style neuro feedback. It's a practical approach. Hard part of remote neuro feedback is having to have staff on seven days a week, 12 hours a day, have staff have clients in different time zones. Uh language barriers, uh most of my clients, I speak their languages, but you know, there's occasionally a few that I don't speak their languages. That makes it hard to do neuro feedback. You know, how how do you ask somebody about how they're feeling or nuance things or explain things about how a brain works or a resource works when there's a language barrier? So >> yeah, >> in languages that that are in in a couple languages I speak uh I have to do some work in, but I've mostly lucked out in my other big center outside of the UK and the US is Sweden and Swedes all speak English for the most part and I speak Swedish. So like I there's no there's no there's no issue there generally with like communication. But then I have a few clients in like Hong Kong who only speak Cantonese and I'm working through their like daughter or their you know aunt or something. Uh and we had in the in the US I have a few clients in LA and I don't speak Spanish. I should I should learn Spanish but don't speak Spanish. I have a few clients where I work through their their daughter or their you know their friend uh because of the language barrier being significant. But you know it's it it's still easier to do neuro feedback than it would be to do therapy in that context because I'm asking about regulatory shifts in people's experience and stress, sleep, attention, mood, whatever. These are fluctuating resources dayto-day or experiences dayto-day. So you know the language barrier is not too hard to cross. >> Yeah. radar. Yeah. So good. >> More more meatballs. >> Yeah, Kotf. That's meat meatballs. >> Thank you. >> Yeah. >> Is that your go-to? >> No, I'm not a huge fan of Swedish meatballs, actually. I have some of the same time zone difficulties with international clients and some of the groups I I I do have people from all over some from Europe and from Australia in the same group. So it's you know it's odd times and odd days. Um >> but uh the good thing is um when you're reviewing EEGs live with a group occasionally you get a pre and a post >> and be able to actually show you know this is what we saw. this is what you trained. Here's the results. And the the the electrphysiologic results are fabulous in in some of these cases and quite dramatic and actually show locally the areas that were being trained are the areas that had maximum change. But the the person changing is I think the uh the story that goes along with all the data ends up giving it the kind of the emotion. And uh >> yeah, I do 10 of those sometimes a day. Today's what? Wednesday. I've lost track already. Um Tuesday is my big day. >> Wednesday. Yes. >> Yeah. Tuesdays I do 10 or 12. >> You determined that 10 minutes ago. So I know. and and Wednesday I usually do about nine or 10 direct consults going over often change maps with experiences saying here's a feature that looks changed often it means this what is your experience and validating that and getting those stories and we were talking about I think AI a couple weeks ago I now have I think pushing 10,000 videos of me discussing people's brains with them over live data for 30 to 40 minutes validating hypotheses testing ideas talking about experience teaching them they're nerd their neuroscience. Um, so that's going to be a really good training uh data set at some point. If people want to let me use their data, I'll have to ask them. But uh, you know, >> I've learned a lot. That's how I got good at doing it was by doing that exact thing that that like correlative um, >> yeah, >> review. >> Today I looked at a 75year-old that I saw about four months ago for the pre and the post. He he has a perfect 10 hertz alpha at the back of the head. Beautiful uh uh the topographies, everything. He's sleeping like a baby. He's dreaming and he's remembering his dreams. >> And what's striking about that, I mean, you tell people you're dreaming and you can remember your dreams. Some people say, "Well, so what? I I I dreamed too, you know, but he can't remember ever in his past >> dreaming or remembering his dreams >> and uh he's a very famous sports figure and uh he's set up a foundation >> uh because of the the dramatic impact on his life. He set up a foundation to fund neuro feedback for traumatic brain injury for >> uh other people in football, but they can pay, but also people who don't have money that don't have to pay because the foundation will cover it. Anyway, >> a couple folks in uh >> the MLB who are in the umpire association about trying to find ways to get nerf feedback supported for the umpires because they all have major head injury experiences from being umpires. like they all get balls to the face, unprotected, un you know, unflinching, no no response balls to the face or bats to the face or, you know, and so a lot of them have TBI kind of history worse than than the baseball players. >> Yeah. >> Speaking of baseball doctor, here's a softball for you. How far away can you help people? I've talked with people in Switzerland and Qoutar or Karatar. >> Either orQatar. I don't know which is I don't know which is the worst American pronunciation of >> it. Um I've got uh clients in the Middle East who have opened a a clinic and um were the first ones to be allowed by their government which apparently gives some like a state licensing board sort of a thing only for the country. uh he's the first person who's not a neurologist to be able to get a EG credential and uh he's the first and only uh neuro feedback uh uh person. He's a psychologist. >> Uh but um he's done really quite well. He gets people from all over the M East coming to him. Uh and then all of his EGs come through me. So >> great. >> Um and he's he's been doing it for a number of years now. So >> EG is pretty big in the in the Middle East. Um I have a lot of clients in Dubai. >> Um currently and I work with both uh a lot of the Dubai community of clients I work with go back and forth between London and Dubai. Uh and we end up we getting brain maps when they're in London. We also I know at least two different neuro feedback and QG centers in Dubai that send me data. >> Yeah. Um, and gosh, in Switzerland, I mean, I I work with clients in Switzerland, but if you're in Switzerland and you're anywhere near uh Chur, you should see Andy Mueller, Andreas Mueller, Dr. Mueller, um, who worked with uh, Yuri Cropov and helping develop some of the HBI med stuff. I believe that's misremembering. and it has a much larger set of analysis they've worked on that's >> beautiful in distinguishing ADHD states >> but I work with clients all over the world radar so you know if you're in a country that I have an office and we rent you gear and if you're not in a country I have an office and you travel to one or you buy gear >> it's it's hard to get far enough away to not have access you know I mean >> yeah and you don't need live internet to actually train although we do live support so that helps but you just need you know some internet here and there. And I use battery powered little QEG amplifiers now that are tiny. So, it doesn't really, you know, the the hardest part I was I was griping about this on on the Nerf feedback Discord channel yesterday. The hardest part is uh the 50 Hz noise in Europe. I hate I really hate EEG in Europe because of the 50 Hz noise in all of the buildings in Sweden and in the UK and other parts of Europe. It's just so bad that the the buildings are built differently. And I mean 50 Hz I think the I think the reason is 50 Hz won't kill you and 60 Hz will. You grab a live wire in the US of DC 60 Hz and it won't stop your heart. But 50 Hz won't do that. Doesn't interfere with physiology the same way for some reason. So there I don't I think that's why there's just ambient 50 Hz noise flooding Europe because no one cares because there isn't really >> a lot of the European buildings weren't built with grounded circuits all over and like in the US if you have an ungrounded circuit and you you get a 60 Hz in that area >> because of the the circuit not having a ground. So, but the you know um 50 and 60 Hz have notch filters that handle them fairly well, but uh a notch filter is only good up to, you know, about 100 dB and sometimes you get more than that of the noise, which means you've got too close to a source is all. >> What about any impact of um even with a notch filter? Uh, is there a there must be a frequency roll off for where it is sensitive and where it's not? Like if you're getting 50 Hz bleeding across all frequencies, which happens to high amplitude, are you able to notch filter it out and affect frequencies that are lower? >> In fact, a notch filter from 45 to 55 that drops at 120 dB will take care of it if it's a reasonable amount. If it's too much, nothing can handle it. if it's, you know, you're too close to a a transformer. >> This is something kind of interesting, too. I'll hold it up. I'll cover the name. And well, here we go. >> The name's been changed. Protect the >> Well, no, it's just this is currently on my desk. They're doing renovations upstairs. >> Oh, and you're catching the >> But it's really interesting in terms of where we get the interference and where we completely don't. Okay. So, I'm with the landlord. Okay. Like literally, what are they doing on the seventh floor that's causing this? And we're on the second floor, right? >> Yeah. And it actually it it sounds like running water down here. I know it's not, but it's it's it's fascinating in terms of where this happens. And we've talked about this before in terms of how the interference is almost always stronger on F4. And again, look at that. It's stronger on F4 compared to what it is on F3. >> Compared to F3. >> Yeah. >> Perhaps impedance difference. Who knows? >> It's universal. Jay, it's a mystery that be before I Well, I'm never retiring, but you know how it is. I I do trust that uh you know, some way sometime we'll figure out why whenever >> if you can figure out how to retire, you you let me know because I'm really not good at it at all. >> No, I I mean I'm not planning to. I plan to do less of what I'm doing and integrate more fun things. Um, well, this is fun, too. That's the problem. This is too much fun. >> Since we have so many shy people, then you're they're relying on me to ask the question. So, I'm going to throw one out there. Dr. Hill, with uh I'm holding a little device that's hard for you to see. Little Raspberry Pi here. with Raspberry Pies out there. Of course, I mispronounced it. And this this new thing with Deep Six, a thing that's challenging. >> Yeah. Deepc. >> Yeah. >> Do you see with things getting smaller, the the Raspberry Pi? Long story short, it's about a hundred buck computer that has most of most of the inards with it. And with deep six which is a better going to be a better AI chatpt can neuro feedback build their own AI run their own um their own chat sheet for lack of a better >> let's back up. >> Yeah. What problem are we trying to solve with that AI? Yes. There's lots of things you can do with machine learning or AI but which particular problem are you trying to frame? I I think uh the first one >> a simple chatbot where you you're going through symptoms intake and and and all that to get you along in a in a process. >> There's a couple people already on the discord uh near fit discord who've built custom GPTs loaded in with 20 30 40 libraries. Lisa Tatarin's in the same thing with her data sets and her way of processing and mat lab and they're and they're getting decent analysis out of their data sets. I'm I'm I think it's possible. I'm not sure that I would um I'm not sure I trust the ability of the systems to do analysis yet. So, my solution, what I've I'm pretty close to finishing. I I built a tool to you know, I do these these videos, these 45 sometimes 30 to 45 minute videos describing, you know, orienting someone to what a what a QEG is, what a CPT is, walking through it every single time, being very, very systematic. And what I've done is set up a tool so I can take those videos, ingest them, uh, break down the likely phenotypes, the likely goals, and then pull from my list of starting place protocols that might match the particular things we found and create a formatted sort of summary, but off of the analysis that I've done and describe to them, not doing the analysis. So, I'm putting the AI in the teaching role essentially. After I've taught, it goes and summarizes and creates another pedagogical, if you will, kind of instrument. I find it's really useful for that kind of thing, bunging, reformatting, reframing, expanding stuff. And I trust it to do that. I can also read that and make sure it's valid. I don't trust it to throw in waveforms and have it tell me, hey, there's some transients on this site without it won't save me any time. I'll still have to spend all of the time looking at the EEG. John Hopkins just today posted news that they've got an algorithm to spot spikes. 70% of the time the doctors miss >> and they can spot them in the EEG. And I'm just saying, well, that's because half the doctors were in the bottom half of their class >> and the ones that are neurologists got six months of training. So the 70% missed >> are today because the data had the it was in the data but they couldn't see it >> because of the lack of training >> or or because it was a nictal events. There was a maybe this maybe it's a different group but I saw a paper yesterday or today talking about how there was an algorithmic approach to spike detection be but it wasn't looking for spikes. It was looking for transient batches of tissue that were suppressing spikes that weren't visible as a spike, but as like a little raft of data that was suddenly getting frozen briefly and then letting go again. And that was essentially it's equivalent to a spike apparently. >> Is that this group or different paper? >> That it it's the same group. >> Interesting. means >> you you need to look at the percentages um that you know the again 70% higher detection rate than the doctors looking at it with their eyes. >> Well, I mean I'm going to put a little politics in here though. It kind of irks me that we put the money into, you know, training a machine, but we don't put the money into training a person, you know. And I think, you know, where we are technologically, >> what Jay is saying is we can't put them into into both, right? because we still need I don't know I'm I'm all for the humanistic, you know, persontoperson approach balanced with with the AI. I just Yeah. Again, it bothers me that we just jump over the um not putting time into training people. >> Yeah. >> I refer to it as PI, plagiarized intelligence. >> Yeah. Yeah. >> It's it's not really artificial intelligence. They just plagiarize a bunch of data and put it out as though it was intelligent. >> Wait, what happened to coopetition? Jay, >> no, this is different. I'm with I'm with Jay. I mean, I talked about art. >> Lisa Tatran just uh heard my talk on farfield phase reversals and she had her algorithm uh spit out as they spit out long treatises about stuff one paragraph at a time uh all it could come up with about farfield phase reversals. and she asked me what I thought of it. And going through it, it had texts that talked about true sources. In EG, there's no ground truth. >> What are you talking about? A true source in Eg, you know. >> Yeah. Scientific analysis is full of hallucinations. >> I would have given the the the paper that it spit out a C minus in as a student basically. They they it should it should have done better. Uh, but it wasn't bad. It was a C minus. I mean, it wasn't, you know, >> talking about the bottom half of the class again, you know, if if you're working on my brain. I don't want anybody with a C minus. You know, that means you've absorbed >> all all of the, you know, all of those complaints are are looking at a snapshot in time. And this stuff is not that. >> This stuff is accelerating exponentially. And Pete pointed out that, you know, the Raspberry Pi, uh, Jeff Gearling demoed installing DeepSeek on the Raspberry Pi on YouTube recently. It's kind of slow and he's using a reduced, uh, model to get it to run, but it runs. >> And this stuff's going to keep changing. So, you know, whenever we start talking about the the problems with AI, you know, we might as well be talking about, oh my god, cell phones are so disruptive. Ah, movies. The talkies. Now, Andrew, we're we're going to get into a little bit of battle here. My issue is that it's going too fast and we're not acknowledging um uh the errors and it's being put out before it's ready and it can cause harm. I mean, I'm not against it. I'm >> Well, it's not it's not done. That's the thing. This is not a piece of technology like other technology you've ever experienced. It it moves so fast. >> Yeah. >> That it's going to be >> It's also falsely accelerated. That's my issue. Yes. it's moving fast, but if you have an unfinished product that's moving fast, um I think it can compound the potential for help as well as compound the potential for harm. So, I you know, caution here. You know, we we we can get a bit over excited. >> Well, caution's gone because the the models now open the models are open source. Y >> DeepSeek R1 is performing almost as well as the pro model, the $200 a month model from from OpenAI >> and it's free and open source. So a beefy computer and a bit of time and you can be running something as good as the best AI essentially. >> That has implications that are quite broad for changing >> industries, education, technology. >> I mean I think we're heading to a place in a couple years that software is no longer going to be a an industry. Just ask your your your piece of tech to become something briefly. >> Turn to Nerf feedback software right now please for me. >> It does that. Hey planning. Hey D. Hey recipes and it reconfigures itself. So that's probably where we're going. >> Yeah. >> So hey BJ couple folks showing up. >> Sorry we're ranting about AI. >> Well PJ's late. They're relying on my questions. Everybody's so shy. So here we go. spiritual stuff from BJ. >> Working on enlightenment, huh? Yeah. So, our job is to get BJ to develop his uh how would you describe what we're doing, VJ? >> Ghostbusters >> something. So, I've talked to BJ a little bit about this question. He was asking me about what I see when we do psychics or spiritualists or people in ecstatic states because I' I've this point I've done a bunch. You know, I did kind of humorously Tyler Henry, Dr. Drew and I did a play-by-play on Tyler Henry's brain when he was doing his eat his Hollywood show. Um, and I've done a few other psychics and I had a couple other recently a pet psychic was able to voluntarily change her brain in the same two ways with two different techniques she does for her psychic practice and I was like, "Wow, that's really interesting. She changed it and two months later came back then did another set of recordings and did the exact same thing." So, I've seen these kind of phenomena in these people with atypical, if you will, spiritual experiences. And I guess that's what maybe VJ is wondering is what have what what have other folks seen because you guys >> quite a bit of quite a bit of the spiritual stuff is going to need something beyond your standard medical grade amp. Uh DC to 70 hertz doesn't really catch the gamma. That is the gamma that they look for if they really are researching gamma. gamma 2 between 80 and 100 hertz which is outside the DC to 70 range you need to go up to 150 hertz and get gamma 2 and ripple >> you're talking about mostly trait level stuff not state stuff right this is cliff sarin's work around gamma this is traits this is acquired traits through long-term meditation that's not necessarily the same thing as ecstatic experiences which are absorptive and dissociative so I I would argue that gamma is probably not part of the ecstatic state as much as it is It's part of consciousness. >> Yeah. But the ecstatic state is something specific >> as such. It'll be part of if you're conscious during something, it'll be there. It's it's a major piece of anything that's got consciousness in it. The the um if you're conscious, you have a active relationship between your slow cortical potentials and your gamma. And if they're related, you've got consciousness. If they're not if they're not related, you're not conscious. You're you're deep enough to have surgery sometimes or you're in a state that's not effable. >> For folks that are wondering, related means time locked or or phase locked basically. >> Yeah. One of my questions too as well, I mean, gosh, we were, you know, kind of curious and doing some of the stuff well 20 25 years ago, people would say, "Hey, record my brain because I can do this." Now, one of the things we did see is the recordings were different, but there wasn't necessarily any correlation between the the differences in the brain waves and what they thought they were doing. So, yes, >> I would say the same thing for the first 20 years that I did this, but in the past five years, there's been a handful of people and I've been seeing the same phenomena show up again and again. And Tyler Henry was the first one and I'm like, "Wow." I mean, Drew and I talked about it on air. were like, "You look like he's dissociating when he's just sticulating, moving his hands, and sweating." And his and he has these massive swells of delta showing up >> uh at the same time, like he was checking out. >> And I thought that's interesting. And I I've seen that since that same phenomena where there's either focal >> um well, it's it's usually focal delta, big bursts of focal delta blob right out as if they're shutting off parts of their brain. And and I would normally expect somebody to be in a dissociative state >> turning on. >> You know, big delta isn't necessarily off. Uh uh years ago, uh John Duca uh took Salvia Deonorum at a IS&R meeting. There it was a group of people that met in a room. Uh they brought me in to watch the EGB recorded to see if it was really EG. And uh he he was putting out a thousand microvolts of 10second long waves. Uh and it was >> uh well 10-second long waves is kind of subdelta that's infrasow. >> And uh uh the amount of gamma and the amount of infrasow content was massive. uh so massive that uh um you you would think of it as some sort of an artifact if it hadn't been observed really. So um you know you can get gigantic perspiration artifact. You can get all sorts of other gigantic slow waves but this was not >> uh perspiration has random phase initiated at a point in time by the sympathetic trigger but after that each sweat gland is independent as to when it absorbs and re remits. And so you get large slow sways but they're not there's no sync to them. There there's no pattern to them. And this was all patterned. And uh it it was uh the altered state section of ISNR ass and u >> I haven't heard of that. I must have missed that meeting and I I was it was re >> three years after we did the event we actually presented the data as a workshop at the and uh uh sometime if you want to watch it I've got John Duca's uh PowerPoint which is a rather interesting uh journey in and of itself. My guess is that video is on the archive the ISNR. So for those of you who are not ISNR members, if you sign up, you can I think >> I seriously doubt that it's in ISNR's possession. >> Oh, okay. Okay. >> I don't think it made it out of that that video didn't make it out of the workshop to get into their archives. >> Interesting. >> All right, we're keeping on Ghostbusters here. Have you guys ever seen anything spooky like in an EG of two people's EEG synchronizing? >> Yep. >> I want to take this out of spook. Um you know there's some really interesting things like we we do know for example that you can start to uh obviously you can regulate and match respiration you can start to regulate heartbeat um and to a certain extent in some of the couples work we were doing um and I know uh Castle has started to take this up but we were doing this years and years ago where you you know you would hook two people up um with with one feedback system even uh though the the the two signals were were different and it was just in terms of frequently downtraining the arousal in a couple going through great difficulty and you I I mean it's not like overt synchronization but you definitely see one person slow down or one person speed up to to match. There's nothing spooky about that. It's just really good um EEGbased couples counseling right >> now. and want to take it into spook. >> We we did a publication uh published in society for scientific explorations journal um and uh we were looking at a healer and a heal couple >> and the healer heal were at a distance. They weren't like touching uh but uh the the healer was was putting out standing waves that phase synchronized the healey to his EEG. >> Mhm. >> Now uh creating a standing potential requires uh gigantic harmonics and he hit the Schuman resonances >> which is plural. It's not one frequency. It's 7.83 83 and lots and lots of harmonics. The fourth harmonic is in gamma. Uh but it um the the healey uh is is not uh volitionally trying to sink. Uh th this is if you're a swimmer and you were in a wave pool wave table and you're a swimmer or your little cork in a tiny wave table uh and somebody was creating a standing potential in the pool it would look like there were waves but they weren't moving. And if you were a cork you would float to the bottom of a wavelet. So you you phase in train with standing potentials and uh the the healey uh was phase synchronized within less than a second as soon as the healer was putting out this specific frequency and all of its harmonics. And um the analysis was done by Luke Hris in Minneapolis. he doesn't really know much about EEG and uh he said well now we got it collected you can analyze it I said no you're going to analyze it and uh he said well I don't know anything about it I said well just turn on the bspectrum the display and scan through 4second epox scan through the EEG and tell me if you see it well what am I looking for well if it's there you'll see it you know and the next day he shows me uh one uh pattern which is basically just the the healer's background alpha rhythm at about 11. It's it's just the bissectrum is a simple one line up at seven at at 11 one across and they intersect at 11. So, it's a simple pattern. And then he showed me one that looked like a checkered tablecloth of crossspectral coupling. And he said, "Is this it?" And I said, "I told you you couldn't miss it if it happened. You know, it's harmonics." And the bisspectrum shows the harmonics. You need to know what you're looking at. You just see, oh, lots of harmonics. And at that point, the the the harmonics were present and the phase shifting happened. We were invited to analyze a healer uh event where a person was going to be healed and the healer and there were apparently 10,000 people across the world that at a certain time we're going to have the image of lightning striking the sky and they and Schuman resonances are created by lightning by the way. Uh but the the they had a five-minute baseline and two five-minute periods and I would I simply looked which one of the five-minute periods is full of harmonics and that was the one where they were doing the healing uh blinded raider and they were all excited about the data. I said, "Well, for God's sakes, don't show the data to anybody. This is a 50/50 chance. You know, uh you know the next time you're going to do something like this, you know, don't do this after the fact showing it to me. ask me ahead of time. I would have said, "Do a five-minute baseline." And then nine periods of time, two of which were randomly picked, not ahead of time, but randomly picked as as you're going. And if you pick two out of the nine and still have it accurate, then you've got something. They did that. It was still obvious. So healer healy relationships are uh easily identifiable as uh a big uh bunch of human resonance happening uh and and that's the coupling between them. Now it doesn't mean that's the healing. It's the connection. >> Yeah. >> You know if I call Andrew on the phone and and he picks up the phone that's a connection. There's no conversation yet. you know that there's there's just a connection. There has to be something delivered for healing to happen. Uh and it it may be something really quite obvious but uh we've we've seenly seen the ability to identify the connection and that that's a step up from not being able to identify it blindly. So gamma awareness consciousness seems to be implicated for controlling these disassociative or remaining upright during them seems accurate because Hindus talk about awareness during dreamless sleep. BJ says or states like astral projection. >> Well, uh the you being consciously aware and controlling your dreams is you know the lucid dreaming uh which was a spin-off of Carl Pram uh lecture at Har at at Stanford and um you know a lot of stuff spun off of Carl. >> I think it's a spun off of a 12pack. Radar Ashwood says since since bipolar inhibits train coherence and why do we bipolar inhibits with the expectation that they will actually inhibit? Doesn't coherence increase power? >> Well, if you're inhibiting the difference between two locations, >> it if you're talking about bipolar, you're talking about an amplifier. For God's sakes, let's get this technical stuff straight here. Bipolar talking about an electrode placement is a misnomer. It's a common misnomer. It's so common everybody thinks it's okay. But technically your amp is a pushpull differential amplifier which is bipolar. It has two inputs and a ground. You can use it sequentially with two known active sites and you're looking at the difference between the sites or you can have one spot that you assume is a neutral reference point. It never will be. you know, you pick a spot to be a neutral reference. >> Yeah, the ears are active, unfortunately. Radar, so even an A1 or A2 is a is an active is >> ear references are awful. They're terrible references. They pick up stuff off the temporal lobe and manifest it as though it's coming from everywhere. >> Uh you need to switch more than one montage just to know what you're looking at. you know, the the the standards for interpreting eg more than one montage. >> You know, you can look at straight at somebody you don't know. Turn sideways. Do they have a big nose? They have a flat face. I mean, until you see 3D, you don't really know what you're looking at. So, in eg, you need more than one perspective. And the montage is your perspective. Um, if you pay attention to Nunes, who's not trying to sell you anything, he will tell you that Lelash and montages are the best way to see the the coherence that goes against the saying of other big names in the field who are trying to sell you machines. So, you know, um think twice about who you're believing, the the salesperson or the person who uh writes books about EEG. Nunes is awfully sharp. He talked about Lelash and Montages being the better way to look at it in 1997. uh and I I talked about that in 1998 at a meeting false frontal alpha using linked ears and um I didn't know about his paper. I mean there's so many journals out there you don't get them all you know but it's >> I still use lelassian and everyone you know who's an old school it's newer eg people like oh using lelassian that's so old I'm like no no no no no here's like four reasons you want to use lelassian because it has benefits it has a couple of drawbacks but it has some significant benefits the biggest one is it's a little bit insensitive to gross distortions that are global so medication fatigue wash that out and they get a little spatial precision >> and I hear I hear people say that Lelash and Montage gets rid of medication effect all the time. >> No, not not doesn't get rid of it, but it does reduce it. >> I want to see one EEG where medication effect is eliminated by lelassian or even >> not eliminated, but it it is reduced like like you'll see someone's theta in in an ADHD state when they're medicated. You'll see it in the plastian more than you will link ears. I I can pop up a benzoazipene intoxication case and switch montages around. You're going to see it no matter what montage you're in. If you want to see it accurately as to where it's coming from, you should use a lelassian technique which shows local function better than any of the others. Uh but it you know the the the but the term bipolar is commonly referred to as >> two active sites and it does look at the difference between the sites but um it it's quite often uh u uh capable of training uh differential uh uh the sequential placement can uh end up training SMR stuff at C4 and fast alpha stuff at PZ. Z and you're not going to synchronize C4 and PZ. So you're really training two sites to create any kind of SMR content that you >> Hey, I have a question about that that particular protocol. So I on Monday nights I do this live stream uh neuro feedback and chill where I set up myself and do a session and describe it and talk about the technique behind it and then I go into some other biohacking topic or answer questions. And a couple nights ago I did C4 minus PZ. And I do that protocol for me at 11 1/2 to 14 1/2 hertz. And I I generally notice I'm fairly sensitive to training. I notice it. I feel it when I when I do neuro feedback. And in this case I felt a little bit of you know the classic stuff, a little bit of focus, a little bit of activation, a little bit of calmness. But as I was talking because I'm answering lots of questions, um I noticed that my voice got tight and I didn't think too much of it in the half an hour, but it kept getting tight and I developed a little bit of vocal strider um in response to that protocol and then it wore off. But I was a little bit surprised that I had like a like a like literally a throat tightening kind of response to training C4 minus PZ. So I assume I did get some beta being brought up at PZ through the C4 minus PZ montage. Does that sound plausible? >> It depends upon what you were suppressing and what you were training as to what kind of an effect you got. >> 4 to 7 11 and a half >> C4PZ generally has a tendency to speed up alpha at PZ. C4PZ >> and fast alpha faster than you're used to having feels a little higher function. It's not quite like training beta in the same location, but it's sort of it it's a >> very noticeable like like voice effect though. Like I it was very I started to sound like RFK Jr. for a second. >> It might not be the protocol that you needed, you know, >> clearly. But I've done that before a lot actually. So I was a little surprised by that particular effect. Yeah, >> since no since nobody's asking questions, then it's everybody. They're it's the chat room's fault for for letting me come in. >> There's a question. Radar's got one. >> There we go. Dang it. Oh, by the way, I watched the Jakeman documentary. I was amazed by how much high praise Jay got. Popular guy. Well earned, I might say. >> Well, you know, you can't believe everything you see. You know, >> it would have to be popular to get a bobblehead bit out of you, right? >> Lots of bobbleheads. >> It um it it was um humbling to have the documentary done. Um and the first time I saw it, I didn't know, well, the editors can make you look like the hero or the fool, and I've been the fool a lot, so I kind of halfway expected that was going to be a fair amount of it. But uh they they they put a pretty good polish on on it looked like. So uh a lot less of the fool apparently. >> Good. >> Well, quick quick technology question. This can go to Jay and Dr. Hill. Jay, you were making and selling amplifiers when they were the size of an armwire. Um mid-century armoire, no less. So, so, so you know a little bit about the engineering and Dr. Hill, you know a little bit about the pie. As I'm fiddling with this thing, it reminds me back in the old days of the old electronics 101 set where you can take something and plug it together. The the AI camera that you can plug into this thing. So, you can get uh facial recognition, you can see object recognition. So you have the camera, you know, looking at the face. >> Then you can do HRV bio feedback with it because you can do red level filtering. >> That that's where where >> you don't need any contact. You can do camerabased red filtering and get heartbeat off of someone's face and therefore you can do HRV with a breath tracer and uh a visual processing. You could actually build an HRV kit within within a Raspberry Pi really easily. Yeah. >> HRV doesn't require any processing either. It's super trivial for processing. Yeah. >> So, you have that and I'm going back to Jay. All an amplifier is is taking something really small and making it really big, right? >> Ideally, sort of a simplistic way. That's how you would explain it. But >> kind of a simplistic guy. Yeah. >> An amplifier. An amplifier is going to have high highpass, low pass, notch filters. I mean, there's all sorts of other things that you end up having in a good uh uh amplifier. Box car filter. How many people even have heard of a box car filter, you know? Um but the >> Yeah. >> But but it's amplifying, but it's it's it's filtering out all the crap, right? That's the big thing that you want out of it. >> It's supposed to. Last one. Supposed to. It sets up a window of frequencies for you to see, but it keeps other frequencies out of your view. >> And there's times that you want to see those frequencies. The reason that people want to go all the way down to DC is that there are slow cortical potentials that turn on and off networks. And if you filter out from one hertz on down, you don't see any of that. >> Uh, you know, you you can blind yourself to real content. Uh the reason to go up above 70 we mentioned today if you're interested in consciousness for God's sakes go up to 150 at least in order to get gamma. >> Yeah. Cliff Sarah's author project was showing 200 300 400. He pulled out some data about a thousand and showed us. >> Yeah. >> So like there's there's stripes of gamma just like I don't know how high it goes but it's it's definitely up to at least a thousand hertz. >> We got 50 kHz out of people before so you know uh but not all of it is >> basically radio transmitter at that point. >> Some some of it some of it is probably chemical uh neurotransmitters structurally uh uh morph. So they they they place like charged particles towards each other and they try to avoid each other. So they have an oscillation because of the structure and uh you dopamine you you can end up having very high frequencies generated by uh by dopamine it just the structure of it creating it >> possibly. >> Um anyway we've we've got um usually those kinds of frequencies are not of interest to the medical community. So, they want DC to 70 and that's what they've that's the window they're used to looking through. Um, but again, if you're interested in consciousness, you should look outside of the traditional medical window at some higher frequencies. And I suggest going to 150 at least because gamma and ripples can be seen. Gamma 2 and ripples can be seen. If you've got an amp that can go up to 800 like the old Nexus, um you can actually with Nyquis principal see up to like maybe close to 400 hertz. But you know Nyquis principle you have to have something sampling twice as fast as a signal you're interested in seeing is only for detection of the signal. It's not for actually being able to draw the waveform of the signal. You'll get a high and a low and you can see that there's something there. But what's the shape of the wave? You don't really know. So the Nyquis principle is okay for detection. But you have to have a five times oversampling if you want to start to draw some kind of a curve with some dots, you know, >> which means you can't do neuro feedback on a waveform like that because the five times over sampling smears the relationship between the wave leaving the head and the time and the processing and then the time feedback. So you for those of you who are geeky on signals if you sample if you do more than third or third third order sampling or filters you end up breaking you smear time relationships of signals leaving the head and the information coming back and learning will go away. So you can't like set up your bioexplore and put fifth and sixth and eighth order filters on it'll work. It'll seem like it's working. The beeps have very little relationship. >> Ain't no feedback. >> Yeah. The beeps no longer related to the brain all that much in terms of the time. the order of the FIR filter being so high gets you very very very accurate but it gives you a tremendous latency >> and uh I are quick but they're not accurate and the fix for that is wavelets which are fast and accurate. You need a whole bank of them but you know that's easy to construct. So >> the reason I was bringing it up, things are getting so small and I wonder what the the the constraint is on the amplifier side of things to keep that small because when I'm looking at my uneducated eye is I am I am coding with something albeit this one is from the Chinese and it does work a little bit better. I'm sure they're gonna give everything out, but uh the a computer, my laptop is now smaller than the size of a a cigarette pack. If all this stuff is coming down, what is the constraint of getting an amplifier together and somebody manufacturing this themselves and and using AI to to work to measure the signals? >> The problem is funding. >> You know, it t it takes money to manufacture something. We we made a pocket-siz brainwave analyzer in 1976. Nobody could believe you had a a full amp that slipped into your pocket, >> but we didn't sell one of them. Which is why I went to work in the EG lab that get me all the experience I had because the company I had, how do you make a small fortune in neuro feedback? You start with a large one. It I I spent more money trying to manufacture stuff and into this field than uh >> See, you made it too strong. when it was the size of an armwire and had a color TV in it and a record player. >> I'm I'm just going way back nostalgia lane, you know, literally hanging out in my father's lab in I don't know what year when it was it was all wallto-wall and I would literally assist with the plugging in and out and soldering, you know, when I was like that that tall. But no, I think it's you know this is just advancement and you know we'll get as small as as long as we can operationalize the the components you can go as small as you can go. So technology begets technology in terms of how nano we get. I you know >> the one somebody asked me for the circuit diagram of what we had made so that they could make one and I said well we were using op amps to build an instrumentation amp. You can get an instrumentation amp on a chip now. >> And so uh what we made with great difficulty with two boards squeezed into a into something that slip in your pocket. uh uh you can buy it on a chip and that chip if it's a good chip with a very high sampling rate and a multiplexer you can have one chip of an amp chip that is multiplex similarly to what Brain Master does with their amp. You you have one ship and you multiplex through it. So lots of channels sample and hold and and that you reconstruct them. So, you know, that that good instrumentation amp and a multiplexer and you you'd have, you know, lots and lots and lots of channels. >> VJ just p me a $20 bill. >> I I would suggest if you got if you get an amp that does multiplexing, you can't do remonaging without error being reduced. So, uh the uh get an amp that's got a chip for every channel. So, >> to ensure prompt service, uh VJ asked the last question, we'll split. Can we talk about infraslow? When is training when is training more beneficial than other waves? Are there downstream effects on infraslow from training other waves? >> Um what's what kind of training is more beneficial? Probably the one that your brain needs. And the the only way you know which one your brain needs is to actually look. uh you you need to have an an evaluation uh that can point to what needs to be trained up or down or the relationship between areas that needs to be changed and that's going to guide your training your brain and a good analysis will guide you as to what to train and uh as as far as >> do infraslow trainers even bother with a >> they don't look at they're they're symptom based they don't look at EG QEG But there's that's changing slowly. >> That's why I'm quiet on this. It's it's not it's not my modality. I mean, >> yeah, mine either. I'm familiar with it, but don't use it. And >> same same. >> My perspective, EJ, is there's a couple of use cases where it seems to work really well. Watching what Sue Omer did, I think the profound autism over aroused nervous system is a good target. And I'm seeing a lot of trauma-informed therapists using it with good success. But I don't think it's as I mean I'm biased as you know but maybe all of us here are but I don't think it's as nuanced a tool or as me as as as operatable a tool as as as controllable >> forced quieting >> it's forced quieting >> I think they're still learning learning what they're doing >> you know that where to train what frequency tuning to get and all of that is still being hammered out uh there there's some schools of thought that are more advanced than others. Uh but uh as a as a signal the infrasow content is is critical for brain function. Uh the the thing is Europe filtered out all the high frequencies and only looked at the slow cortical potentials. But in the US we filtered out slow cortical potentials and only looked at the oscilly EEG. You don't know brain function until you put those two together. you know that consciousness is the relationship between those two. >> But to VJ's question, do you need to worry about training both of them or can since you're training a system, can you get away, you know, or can you benefit from just doing perhaps one? >> And in fact, uh there there isn't any systematic um uh method to say which technique to pick over the other. At this point, there's no gigantic uh Kentucky horse race that's been run with all of the techniques uh with >> not even a little bit of a horse >> presentations. I mean, that the combination permutation to to assess that is just too complicated. >> And when when they're studying neuro feedback, they usually require you to do a specific protocol for the group. Well, >> which basically means you're no longer studying neuro feedback practice. You're studying a technique, but you're not studying the the the clinical practice of neuro feedback. And that's that's my none of this is about a technique per se. I mean, that's that's >> what I was what I would suggest is that it has to be personalized and they say, well, that's you know, how can you study that? Well, you set up ahead of time how you personalize it. you have a set of rules individualized >> it could be done but good luck getting the funding for that >> they could funding neuro feedback really in the early 70s and although they've tossed a few million at it lately with again one protocol fits all the beta um you know it's it's not how the field works so um >> it's you know it's not really a good test of Last question. Last question. Last question. And this goes out to all the Patreon people have been calling up for support. Are there any consumer devices you would recommend for EEG measurement or do they just not have the clinicalbased accuracy? Are we recommending products here? >> Well, it depends upon what you're going to use it for as to whether it's adequate for the task. you know, if if what you want to do is sit in front of a computer and float a little image of a ball up and down, uh most of them are adequate for that. >> Yeah. >> You know, um if you want to do an ERP, the EG amp doesn't have to be very good at all. The EG is noise to ERP, you know. So, >> it's reliability and val validity to your purpose, right? I mean, all of us can up jump up and and down on that. some of the interesting debates amongst practitioners is just what you know Jay was saying you know is it good enough for for what we're trying to do but I think all of us agree that you know without naming specific producers but there there's a lot that's just this flushed hogwash >> um and and and those ones are dangerous right >> I'll name them if they support me >> yeah you can't you can't get into the EEG space in a usable way for under a grand or two generally Um there are devices out there. >> People do it for a buck, you know. >> Yeah. But there are there are devices out there. >> The problem is people who do it for thousands and thousands and don't know how to use it and don't know what it's doing. >> Right. Right. You can get an open BCI ganglen board for 100 bucks or something. And you have to learn a lot about EEG. You have to already be an expert to use it. >> Exactly. >> Or you can spend a BCIA because people ask me all the time what I want to come back with. Hey, look, >> we here's this manufacturer. Here's they spent an hour with us. Here's what they said. Check it out. Make your own decisions. >> Yeah, >> we're not going to have that in our field because it's all nichy fractional. >> I know my product's better than your product, but we're >> No, no, not even that. There's just not enough people buying it. And so there's just no product maturity. There's, you know, there's 15,000 providers max, including the proumers. >> There's probably only 30,000 installs of software and hardware in in the whole world, you know, actively being used. >> It's just a niche. And there's just no pressure, no development pressure, no financial pressure. Yeah, >> you know, the field is smaller now than it was last year or two years ago because of everyone aging out and dying who started the field. >> That's what I discovered. If you're selling into a niche market, you can go broke. >> You sell one to everybody, your market's gone. >> Yeah. So, >> when you don't have another one to sell because everyone's got one, you you don't really have a business anymore. But you can provide service to a niche market and people require service again and again and again and again. So I I quit trying to manufacture EEG stuff and uh allowed people to use whatever EEG stuff was out there. There were some that I wouldn't accept because they were just, you know, the amp wasn't safe to have hooked up to a person and I didn't want to have any liability. So I wouldn't accept anything from some amps. But there are some standards for little wearable things that you can look at that kind of tell you whether the engineering is kind of better than others. If you see a solid metal disc for a for an electrode, probably not a good one. Uh um solid metal has contact jitter and if it's got a polymer coating, then it's not really a dry sensor. It's a damp sensor. there's some moisture held in that polymer and that's that's a partial salt bridge. So if you see one that's got a coating on it, that's a step up from something that's a solid metal. And those are easy to spot just with the naked eye. Look at the amp. Look at the electrode that the device has. If it looks shiny and metal, you probably don't have a better design. If you see one that's got a a polymer coating, they they've thought a couple of times about contact jitter and how to get around it. >> Anything beats a needle. >> Maybe we should go back to needles. >> Dr. Hill, you got >> it's what we started with, but it's not what we wanted to stay with. That's for sure. So, >> Dr. Hill, you got a good product. Where do we send people to to see your remote or take a look at your remote? Uh >> yeah, come check us out at peakbrainstitute.com uh which is peakbrainla and all the socials or come watch me on YouTube at dr hill dr h i l where I do weekly live streams descri demoing neuro feedback and talking about biohacking the brain >> and that's Monday so we'll put the link right here >> I recommend highly >> little little coopetition Dr. Swingle. How do people check out IMines? >> Well, yeah, that you can kind of get anywhere, you know, check out in your local store. Uh, check it out on on online. Yeah. Yeah. >> And the Swingle app. >> Uh, the app Swingle Sonic swing.com. Yeah. Good day. >> And don't even try. >> Jay's too busy. >> All right, everybody. Hollow. Thanks for stopping by. got your questions on TMS here. We're going to do a show in the morning and may and I'll and I'll bring these up. Take care all. And hey, Dr. Hill, glad you're still there, baby. And all of our friends in Cali. >> That's right. We're we're we're still here. We're hanging in. >> Hey everyone, if you're enjoying the show, want to help us keep it going, please consider supporting us on Patreon. Your support helps improve the content and brings on great guests and grow the show. Join us on Patreon at patreon.com/neroodle. But hey, if supporting financially isn't in the cards right now, no worries. You can still help us out for free by liking, subscribing, and sharing this episode. It really makes a difference and helps more people find the show. If we've earned it, hit that subscribe button and turn on notifications so you never miss an episode. The Neuronoodle podcast is supported by listeners and businesses just like you. Hey everyone, here's a quick message from our good old buddy Joshua Moore at Alternative Behavioral Therapy. Uh I'm preparing to do an introduction to QEG phenotypes. This is my first workshop on the topic and this is just the introductory overview of the project. This is going to be very very basic. So let's review each phenotype. We're just going to go through them one by one. If you look at the DSM 1, 2 or 3 and you look at the first chapter, it actually acknowledges that the DSM in mental health is just quote a descriptive adaptation unquote. We're going to customize the interventions and tailor them to the individual's brain activity. Not just applying the treatments or trainings to the individual's phenotype and maybe even their variance, but also customizing the protocols themselves to fit that person's individual unique frequency patterns. Maybe a dozen ways to have excessive energy in your nervous system. And then there's an infinite amount of ways to relate poorly to that excessive energy. which means that there is a lot of variations in what someone might call anxiety and that doesn't mean that we're talking about the same thing when we're talking about anxiety. So complicated stuff. But when we look at a brain map, we get some basis for which we can start talking about what is this energy. Then we have to go to the client. We have to start talking to the client about what is your association to this specific kind of energy that I can see. Now then we get to care about things that insurance doesn't uh restrict us on. We get to care about like the person in a broader way. 85% um improvement in mental health report um following pairing a medication with a specific phenotype. You know, jeez. Okay. Have you seen 85% in any mental health studies for any medication? I'm going to be going over a whole bunch of different types of tests, uh assessments, and things that can be used that have things called neuro cororal, which you saw some of those already. better success outcomes. That's what we really want. We need better predict predictive power. We need better individualization. That's what we want.