Hi again and we’ll continue our next panel on methodology issues in um, in studies and practices. So our panelists Chris Berka and Dan Levendowski.

And Dan president and cofounder of Advanced Brain Monitoring, both Chris also cofounder of Advanced Brain Monitoring. This company is, um, just medical device company based in Carlsbad, California.

We um so um Advanced Brain Monitoring specializes in the acquisition and analysis of the brain electrical signals in the development of biomarkers that access cognitive functions and emotions, sleep, neurodegenerative disorders, and depression. And Chris also cofounder of Advanced Brain Monitoring, and Chris has over thirty years of experience managing clinical research and developing and commercializing new technologies.

So um Chris and Dan will talk today about the issues in creating these technologies and in doing research with these technologies. So, Chris and Dan, please, oops sorry.

Uh, Maria, can you allow me to share the screen, please?

Um, you cannot, right? If you try, can you do it? Because I think you should be able.

Yeah, I can now.

Okay – silence while he gets the slides up

Okay, are you seeing the slides now? Are you seeing that slide?

Yes, yes

Okay. So, um, what Chris and I are gonna do is a brief slideshow, and the idea is we looked at what Ken and what Ben were talking about at least from my side with sleep. So that’s why we put me first, and Chris will follow talking about her waking state EEG. And, you know, I was thinking uh we heh with what we call the sleep profile, and Maria was our very first user of this that she took back in 2012 up in a monastery in Tibet trying to acquire the very first sleep data.

And the good news is technology has moved forward, and power consumption and everything has changed dramatically. So this is gonna be a little bit technical about sleep, but what I'm hoping to do is show some of what we might be able to measure to possibly answer some of these questions that were raised in the previous panel.

So, the sleep profiler, um, is a device that’s worn on the forehead, so we’re not talking about high density sleep. What we measure is from three frontal polar leads. So we’re gonna look at AF7 to FPZ and AF8 to FPZ, and that’s gonna give us our EEG and ocular activity. And then from FP7 to FP8, that is our EEG signal that we do our staging on.

We can, we have an acoustic microphone, so we can measure snoring. Uh, we can measure submental EEG, pulse rate, head movement, head position. And how we present the signals with the left and the right EOG, EEG, we also show the power spectral characteristics. And as I walk through this you’ll begin to see how that provides some interesting information by zooming out and looking at changes over time.

So to kind of walk you through what that means is that the four power spectra characteristics that we find most useful in looking at changes in sleep stages or changes in patterns is to look at the relative power between alpha, sigma, beta, and EMG. If we added delta or theta, it would quash the power of these smaller amplitude signals, and we wouldn't really be able to see what's going on.

So, here’s an example of somebody in REM sleep transitioning to a brief REM period, they awaken, you can see the submental EMG and so on, and they go back to sleep. So everything we do is kinda color coded by our stages, and I’m trying to kinda set the framework for some of these other photos or some of the other slides I'm gonna show you. The device is very simple to use. It was designed to be used on patients with insomnia. But we also have developed the capability of adding cardiorespiratory.

So, because for some of our experiments, sleep disorder breathing could be a confounder, uh, we have the capability of being able to do a home polysomnography as well as gathering these signals. So here's some examples of some of the signals. I know, Ken, when I watched your presentation, your video, you were talking about spindles and slow wave activities. So we do the automated detection of spindles. And what you’re seeing here are the bursts of alpha and sigma power.

And this is what we use, and then we mark the duration of each one. And we also do compute the capability, if the peak like in this example where the sigma power comes before the alpha, we consider that a faster spindle. And so far it hasn’t been meaningful in any of our research. I’ve seen some other work that’s talked about that so there may be some benefit in that. But the point is it can be automatically tallied, the duration across the night, as well as by thirty second epics.

Um, we also tally slow wave sleep, and we identify by color coding each one of those periods that meet our criteria for slow wave activity. So if you’re going through, you can count how many periods with slow wave activity, and we also tally those as well for every thirty second period, or you can count the total number of slow waves during non-REM sleep over the course of the night. So, again, trying to provide some thoughts or some guidance where we could do repeated measure analysis, um, with some of these different conditions and seeing what’s changing or what isn’t changing.

Um, as far as REM, so, here’s an example, we compute REM density at fifteen second increments. So what I'm showing you are four different examples of REM periods. Here, they started out with, we’ll call it moderate density, and then it went to a tonic period where there isn't much ocular activity going on here. And so you see the difference in values. Here, there’s a lot of phasic activity, and then it becomes more moderate. So, again, these data are available, and what might be interesting, at least as I was listening to the presentations, are, you know, what sort of phasic density is occurring during the dreamless sleep.

Is it, is it phasic activity as you would expect during lucid dreaming or is it more of a tonic nature? And I was thinking, well, you might be able to design an experiment where they purposely were trying to do the dreamless sleep on one night and not the next night and see how the REM patterns change from night to night.

It’s pretty easy to wear and put on. So we have three sensors that snap in, you wash the forehead, you pull the covers off, you put it on. Um, assuming they remove makeup and get, um you know, clean skin, we’re gonna get impedances that will be down below five kilohms, we’ll get a good quality signal. We measure the impedances. We provide voice messages so that we know if the person has put it on properly.

Um, and we can provide messages say put it on again. Um and potentially some of these message we, for example for our sleep disorder breathing if the oximeter comes off at night we wake the patient up with the voice message to put the device back on. So it does have a speaker. It does have some capabilities that way. We also have the capability of doing both submental EMG and ECG.

So one of our research collaborators down in Australia was recently looking at parasympathetic and sympathetic activation during different REM stages, and we would have the potential of acquiring that and potentially looking at that as well during REM and non-REM sleep. Um, in the morning they simply turn it off and have to replace the sensors.

The beauty of this with now the low energy bluetooth is we can do three nights or over thirty hours of recording without recharging. Back when Maria first went to Tibet with this, she had to charge it every night and there was no electricity, um, and it was challenging trying to charge it off her computer. So, you know, now this provides the capability of doing three full nights or even for some of the work we’re doing they’re wearing it, uh at UCSF, they put it on a patient who claimed um to have hypersomnia. They put it on, they wore it for twenty-four straight hours and were able to look at the sleep and the wake patterns and were able to confirm that the person was really sleeping fifteen, sixteen hours a night.

Uh, what I want to do now is just briefly show you some of the examples, um, of some of the signal patterns. So this is a one hour recording, and you can see how the power spectral characteristics change as they transition from non-REM to REM. and this provides a way of being able to zoom out in ways that not necessarily have been done when we’ve previously done sleep experiments. So this is, when we do the, um, the technical review of the autostaging, we typically start out zooming way out like this to make sure that we understand what the REM patterns are and what the non-REM patterns are. And you can see the little dots here, the marks here, that’s the spindle activity.

And here’s what it looks like on four hours where you can see much more dramatically the alpha, the sigma, the beta, and the EMG power transitioning in and out of REM. What’s kind of interesting in the frontalis muscle, there isn’t much difference in EMG power during REM and non-REM. So unlike the chin EMG, so you can see in this particular case that black line is pretty steady across all the sleep stages other than when they go away.

And in a paper we published a couple years ago, we were able to demonstrate the between subject variability of the frontalis EMG is much much lower across all stages than what you see from the submentalis. So it provides a good way for us to characterize when somebody’s awake and when they’re not awake or when they’re transitioning between stages.

Now what I want to do is briefly show you a couple of examples. This is, I, when I look at the cases, and I’ve looked at thousands of records, and I see differences across people, I’m always fascinated trying to understand why these two individuals who have fairly similar non-REM going in, or or, non-REM going into the REM period, where you can see here the beta isn’t that high, the sigma’s very, very low. This is kind of a classic example of the relationships between the power.

But this individual has a lot more beta power, and it’s intermixed with a lot of sigma power. Have no idea what that is other than we’ve seen in some cases it might be a trait effect, but I'm kinda curious on if there’s any interventions, or what might be done that could change this. 00:13:04 --> 00:13:45 Here’s another example where, again, these patterns look quite similar, but here there’s a lot more alpha power in the REM as compared to this case, and they start having a little bit more EMG activity during non-REM sleep. And here’s the case, what I did was I moved back a little bit in time. Here's that elevated EMG during the non-REM sleep, and here’s what happened before, and this is a new pattern that we found, and we just submitted a paper to the Journal of Sleep for this pattern that we call non-REM sleep with Hypertonia.

And we only spotted it because we were looking at these signals at such a zoomed out level. And what you can see is this elevated EMG and beta power that occurs during non-REM sleep, and then it’ll drop down. And, uh, the research that we found is that this non-REM sleep with Hypertonia we found this very, very prevalent in patients with Synucleinopathies, but not in patients with MCI, AD, or normal controls.

Um, the same pattern of being able to zoom out and look at these characteristics, um, also helped us identify a pattern that we call Atypical Stage N3. So what happens is in these cases, and we found this initially in patients in the intensive care unit that’s it’s been called polymorphic delta. Some have associated it with delirium. But when we took that algorithm and applied it to our neurodegenerative disease patients, it was higher in the patients with dementia than those that had the prodromal risk or normal controls.

So, what I’m trying to do is just explain, I think that there are some tools that we have that may be able to identify some of these biomarkers or the conditions that we’re talking about the changes of this lucid dreaming that we might be pick up some patterns just zooming out a little bit.

The other thing that I think is kind of exciting is we now have the capability of seeing this in near-real time. So we’re transmitting from the device bluetooth to an android tablet. What we do is we hold the signal steady and then every thirty seconds it refreshes. It takes about fifteen seconds for the algorithms to do the staging, but it also would provide time for a technician to say, “yes they're absolutely in REM.” Um, it doesn’t do the absolute timing of trying to catch the upside of a slow wave. Um, but if you’re trying to do an experiment that you need to know are they in slow wave sleep, are they in REM sleep, this could provide the capability for that.

These are just some of the research partners that we’re currently working with with this technology. Um and so the other thing is if you use this for research, we’ve done, we’ve published a number of studies where we’ve compared the accuracy of the autoscoring and/or with technical review the simultaneous PSG. So it has the capability if we do some work together and want to publish it, that we don’t necessarily have to rely on polysomnography. We could go into the home or into the monastery and collect some data.

Um and we’ve also done work, and like I showed you, with the neurodegenerative disease, uh the group in Australia that was looking at sleep stage related changes in heart rate variability, it’s been used to look at medication affects, and also sleep, memory, and depression.

The last thing, I’m pretty naive at this, so I’m kinda going in this fingers crossed that maybe um I might have some questions that might hold water, I’m not sure. I know when Maria and I were first talking about what we might be able to contribute, what I was hoping to do was say, “look if we have some tools that might help us answer these questions, what sort of questions might be answered?”

And a couple things after watching the other videos is I’m curious if dream yoga will change slow wave sleep or spindle activity or duration of phasic REM. And even is there a cognitive benefit of increasing phasic REM duration even if it’s not lucid dreaming - something that we could measure with our system. And then could lucid dreaming frequency be increased if in the morning somebody had a profile of how their dream patterns looked and how much phasic REM activity they had.

And then I was, uh, with what Ben was talking about, I was wondering if the medication that you were um providing to, to uh, for the lucid dreaming, is that something where the practice of lucid dreaming, could you flip it around and in someway could some of these changes in REM sleep or through meditation affect a neurodegenerative disease risk profile. You know if you have a condition where you don’t have a lot of spindle activity or you don’t have a lot of slow wave activity and you have a family profile of neurodegenerative disease, are there any things that we might be able to do to help the quality of somebody’s sleep through practice, um, that this could be reinforced.

Chris, you’re up.

Alright, thanks. That was great, and this was such an exciting panel. It brings together two worlds for me. Twenty years I’ve been working at Advanced Brain Monitoring, developing, you know, EEG systems, and I think we’ve worked with Maria for over a decade now. Um, but for longer than that I’ve been a very active meditation practitioner, and I’ve actually, you know, participated in some lucid dreaming groups, where we attempted to do group dreams.

Um, also, um I had one teacher who really encouraged teaching in the dream state, so you would go into the dream with the intention of working on your meditation practice in the dream. So this is a very interesting merging of my two domains and I’m really happy to be able to participate.

Okay, I’m gonna share a couple slides. – silence while getting slides

So, as I said, for twenty years now we’ve been working at Advanced Brain Monitoring to develop neurotechnology that we could take outside of the laboratory and basically use EEG and ECG and other non-invasive physiological sensors to assess human cognition in any environment. And, I didn’t get to see all of your presentation, Maria, but thank you so much for showing our systems, and you can see various generations of this system in Maria's slides.

Um, and we’ve tried to accommodate her in even the harshest of environments.

It was very hard for Chris, very hard, so I wouldn’t be able to collect any data without your equipment. It’s the only system that worked in this harsh environment.

Well thank you for trying, for giving it a go. So this is our twenty four channel system. It is extremely lightweight, the entire system weighs less than three ounces. We print the sensors on a flex strip, you put the silver ** sensors right onto a flexible cable, and then all the electronics are in that small box that can be worn on the back of the head.

Um, and amplification, aided deconversion, and then the digitized data is sent via bluetooth to a laptop or handheld. Uh our goals with this - number one goal: highly accurate, um, repeated, reliable, medical grade EEG signals. So that’s always been number one, and these are FDA cleared products, so we have to maintain a high quality standard.

Number two, um, we wanted the systems to be as unobtrusive as possible. Now, everybody knows that they have an EEG system on their head, but we have been told in many, many of our studies that people forget that they’re wearing a system if they’re engaged in a task, even in a group discussion where they’re looking at other people who have the systems on. So, um, as best as possible, we’ve attained that goal.

So, for the last six years, I spent, I have spent a great deal of time, um, wrapping up clinical trials. We have done thirty clinical trials all over the world. And I think we’ve proven beyond a shadow of a doubt that we can successfully, accurately reproduce EEG in just about any environment.

Um, we’ve been looking very specifically at EEG based signatures for Alzheimer's Disease, Mild Cognitive Impairment, and aging in general - what happens in the brain when you age. And we have a very definitive set of biomarkers - um this has been funded by NIH, and Biogen, and J&J and several other partners - a very definitive set of EEG biomarkers for cognitive decline. And those include resting state EEG measures, so where you’re just sitting passively. And I think a number of the presentations showed resting state.

But in addition to that um we have a platform that we deliver on a laptop of cognitive tasks - so a verbal memory task, an attention task, a visual memory task, and then an emotional face recognition task, and from that we can derive the event related potentials. And I noticed that there were quite a few presentations on event related potentials too, so I won't go into those approaches in any detail.

But what we’ve found is if we really want to characterize cognitive state and/or cognitive impairment, we like to have a combination of both the resting state and the event related potentials during very specific types of tasks. And this slide is kind of a classic slide now, everyone who’s studied Alzheimer’s or progression into Alzheimer’s will tell you exactl;y the same thing about the EEG.

We’ve tried to keep the cost, um, moderate, effective. We’ve tried to work with as many researchers as possible to expand the applications of the system. And, most importantly, we wanted it to be easy to apply, so that someone who has no experience with EEG or neurophysiology could become an expert in a relatively short period of time.

Uh, you see signs of slowing so you see enhanced theta and enhanced delta in early stages primarily over the temporal parietal regions, and then as the disease progresses you see more and more slow activity and less and less gamma, beta, and alpha activity. – silence while switching slides

In our, um, attempt to get sensitive biomarkers that moved much earlier in the progression of the disease, we found that the event related potentials during a recognition memory task were extremely sensitive even to very early, um, memory impairment. Uh, this was a sample that we got from the Harvard Memory Clinic, and, the, everyone came into the memory clinic because they have an issue or a complaint - either a family member complained or they have self-identified - so they all were suffering some perceived memory loss, but only half of them actually ended up with a diagnosis of mild cognitive impairment.

But you can see the differences in what we would call the memory related, uh, component of the event related potential are pretty striking and very dramatic even in this group that is relatively homogeneous. Um, so these two, uh, the resting state as well as the memory related event related potentials are two that we use in all of our clinical trials.

We have been involved with many unfortunate failures of, um, cognitive enhancing drugs. Um, and, you know, Biogen was one of the earliest funders of our work, and we know that that drug is still not yet achieving the success that we had all hoped for.

So part of, um, our goal now is, you know, in retrospect six years ago I really believed that we would be going into the clinic with one or more drugs that were effective in reducing or mitigating cognitive decline, and we could be the biomarker associated with them. Well, we don't really have anything yet. Um, so the, in that vein, I mean, what we’ve decided to do is to start to, we’ve been investigating exercises in mitigation, we’ve been investigating nutrition as a mitigation.

We’ve been looking at various cognitive brain training exercises. Um, we’re embarking on a study of music. So I think that, um, in the absence of a true disease modifying pharmacological treatment, all of these approaches need to be employed early on. Um and I think that again bringing my other world and bringing your world into this, using meditation as part of an early intervention or a prevention strategy makes a lot of sense.

And we have objective biomarkers for measuring changes in cognition that we now have a library of about 20,000 EEGs that we’ve acquired. So, we’ve also identified EEG based markers for Parkinson’s disease and several, uh, EEG biomarkers for major depressive disorder. One of them is, um, patients with depression show an NI70 component response in the event related potentials to emotional faces, so they show a much larger response to sad faces than happy faces.

Um, and so, this gives us another set of EEG based biomarkers, and again, meditation as an intervention for depression is already being implemented, and now we have measurement tools to look at that in the EEG. so, you know, basically what we’ve been doing is promoting the notion of a brain health profile.

So just like you get a colonoscopy at a certain age, you get a mammogram at a certain age, maybe you do it earlier if you have family history. Um, let’s institute brain health screening early and often. You know it’s harmless, it’s non-invasive. And now we have these libraries and databases of both healthy controls as well as the people of cognitive decline and psychiatric disorders.

Now I think what would be really interesting is for us to be able to run these same protocols on a group of advanced meditators, because now we’re talking about perhaps hypernormal or above-normal or advanced cognition. Um, we have done work with expert athletes and expert musicians and some other experts who do have very unique brain profiles and may have enhanced attention or enhanced memory as a function of skill sets.

But again, I think, this is, it’s an interesting set of tools that we can use to actually quantify changes in the brain that are associated with changes in cognition. Um, and as i said, we’ve been working on a variety of interventions - robot interventions, prosthetic interventions. Um, we're really interested in enhancing brain function. We’ve worked with some virtual reality feedback systems, um, which gives us another way of inducing different types of mental states and mental experiences and measuring the effects of those.

Um, also combining, uh frequently, we will combine with eye tracking or heart rate or other measures of autonomic nervous system. Um, we did a study recently where we looked at the synchrony between cardiac activity, heart rate variability, and some brain neuro measures, EEG measures,um, and I think that’s another area that’s very interesting in terms of meditation because you will start to see some synchrony between the heart and the brain.

And I think Maria's probably delved into that a little bit, but we don't want to ignore the heart, it’s so important. And then we did a little bit of neurofeedback training, and, um you know, you can see using a smaller set of sensors, we can do some things with this. I don't think we need to do neurofeedback because there’s plenty of other people that are working on it that have, you know, achieved some beneficial results.

Um, the last thing I want to point out is another area we’ve been very active in is looking at group brain - so a group of people performing a task together, what does it look like when we look at the synchrony across their brains as they’re performing a certain task. And one of my particular interests, and again having experienced the power of group meditations versus you know meditating on your own, is uh certainly we could apply these same techniques and measure a group of people meditating and they could be in the same room or they could be connected virtually, but would we see a synchronization across the brains and a cardiovascular synchronate as well. So many opportunities to do interesting things and serve a variety of communities. So, thank you.