Matt Angle:

Welcome, everybody to Neurotech Pub. Today, we have from Prime Movers Lab, Amy Kruse, who is a General Partner there, and she's actually joining me to help introduce the episode. Amy, welcome.

‍

Amy Kruse:

Thanks, Matt. Happy to be here.

‍

Matt Angle:

In this episode, we're talking about sleep. This is your wheelhouse.

‍

Amy Kruse:

Yeah. This has been an area that I have been interested in since my time at DARPA. As a Program Manager, I ran the Preventing Sleep Deprivation program there. To say I've been obsessed with sleep for a decade or more is accurate.

‍

Matt Angle:

Our guests today are Luis de Lecea, Professor at Stanford. I believe that you know Luis, right Amy?

‍

Amy Kruse:

Yeah. Actually, I met Luis, originally back in the DARPA days. I recently reconnected with him.

‍

Matt Angle:

Oh, perfect, and Ram Gurumoorthy, who is the CTO at StimScience. StimScience is actually developing a product that is supposed to aid with sleep. It's a wearable that passes electricity into the brain and activates sleep centers, modulate sleep. Quite interesting. He talks a little bit about it in the episode. In the time since we've recorded this podcast, I noticed that Ram was actually featured in Fast Company talking about their sleep device.

‍

Amy Kruse:

Excellent. Yeah, it's great to see coverage of sleep and neurotech in the popular press.

‍

Matt Angle:

We talk a little bit about the science of sleep, but also, I guess it's worth saying that sleep has started to become an interest in the startup world, too.

You know, things like The Oura Ring, other wearables tracking sleep. StimScience is now offering a product that can help you sleep by stimulating your brain. Have you had a chance to look at this at all, Amy?

‍

Amy Kruse:

I have, actually. I got really interested. Of course, have been interested in sleep for a long time. I'm actually quite a poor sleeper myself. Have been working on sleep hacking for quite some time. From the VC perspective, got really interested in this last year, as you say.

Just started to see a bunch of companies, techniques, approaches come to the fore. Have definitely been looking at it. Sleep is heating up.

‍

Matt Angle:

Did you see anything that you thought was particularly interesting?

‍

Amy Kruse:

Yeah. I think there's an interesting dichotomy that's happening. I think there are some things that are consumer-facing, like some of the bands that you referenced. Whether they're something you wear on your wrist, something you wear on your finger or something you wear on your head, something that goes in the mattress. Those are all consumer-facing devices.

Then I think there's a renewed interest in clinical sleep and insomnia, shift work disorder, some of the things that are there. I've really taken a little bit more of an interest on the clinical side, initially because we typically don't do direct-to-consumer investments. That's not been a focus area for us.

Also, I really, really like the applicability of doing something in the clinical space and then potentially branching out from there.

‍

Matt Angle:

What was your favorite part of the conversation that we had?

‍

Amy Kruse:

Sleep is fascinating, right? People are really intrigued by sleep because essentially, we're unconscious. We don't have any control over it. I do think that the notion that, sleep has different stages. These stages are associated with different potential outcomes in the brain, and that you can actually modulate the stages of sleep with electricity, or optogenetics, or ultrasound, or whatever.

The notion of actually changing sleep. Not just recording it, but actually perturbing it or perturbing those signals in some way, I think is a lot of fun. I think we had a lot of fun with that.

‍

Matt Angle:

Amy, one thing that we've talked about in the past is just how much momentum the neurotech field is getting right now. It must be really interesting for you, starting off in DARPA, then being here in venture during this period of intensity, honestly, in investment and advancements.

I'm curious, from your perspective, what do you see happening right now?

‍

Amy Kruse:

Yeah. The neurotech industry, it's finally becoming the industry that I hoped it would, back when we started making all these investments. Whether it's what's happening in brain-computer interfaces, the work that we're even seeing with peripheral nerve stimulation and vagal nerve stimulation, and frankly, what we're talking about today, in sleep.

I actually think that sleep applications are probably going to be one of the breakout uses for neurotechnology, not five years from now, not 10 years from now, but imminently. Really excited to have folks closely track that space as well as all the other things that we're working on, but excited to see where this lands.

‍

Matt Angle:

Actually, after we recorded, our producer, Ali actually sent me this paper from Sydney Cash who's been on our show before, and works at Mass General Hospital and does a lot of interoperative recordings.

They actually have looked at sleep data in patients with tetraplegia who have a BrainGate implant. Yeah, it's actually quite interesting. There's a convergence of BCI and sleep, where you have these patients with Utah arrays that have been doing motor tasks. Then they go to sleep, and then you see, just like in other forms of motor learning, there's a replay during slow-wave sleep.

‍

Amy Kruse:

Yeah. Yeah.

‍

Matt Angle:

We'll link that paper in the show notes.

‍

Amy Kruse:

That'd be great. Yeah. I mean, the evidence around consolidation of memories during sleep, and sleep as that replay, rehearsal, and then, again, stimulating, or something to enhance that there were DARPA efforts in that as well, after I left. The RAM Replay program highlighted some of that work as well.

Yeah, sleep is doing a lot for the brain, which is I think is one of the reasons why we should all prioritize it.

‍

Matt Angle:

Well, okay, leading into this, we're actually going to have one of our interns, Zoe recorded a explainer on sleep, since we just jumped straight into the meat of it and started talking about course level 300, a brief talk about the different phases of sleep.

Amy, if people want to learn more about Prime Movers Lab, what's the best way to follow you or follow them on the socials?

‍

Amy Kruse:

On the socials. Absolutely. First of all, primemoverslab.com is our website. There you can find a list of our portfolio companies, a little bit about our philosophy of investing, some of our thought papers that are quite detailed in some of the areas that we've dug into.

Then on Medium, we are prolific bloggers. Really recommend you follow us on Medium. I myself have a couple of blogs on, obviously neurotechnology and sleep and neurostimulation that could be of interest to this audience that, would be great to share. So, you can find us on LinkedIn and Twitter.

‍

Matt Angle:

If anyone's interested in learning more about Neurotech Pub, our handle is @neurotechpub. You can also subscribe to Apple, Spotify, or wherever you get your podcasts. You can check out paradromics.com/podcasts for video, show notes and references.

Thanks, everybody for joining us. Before we kick off this episode, one of our interns, Zoe has actually prepared a short introduction to sleep.

‍

Zoe Lalji:

This episode of Neurotech Pub deals with an essential topic to all of our lives, sleep. Sleep can be broken down into four stages, three stages of non-rapid eye movement or non-REM sleep, and one stage of rapid eye movement or REM sleep. These four stages together form one sleep cycle, which lasts approximately 90 minutes.

Human beings ideally need about four to six of these cycles per night, adding up to about seven to eight hours of sleep. Now let's break down each of these stages and how they individually contribute to our overall health and well-being.

The first three stages of sleep are considered non-REM sleep, and are often referred to as N1, N2, and N3 respectively. Non-REM sleep is characterized by the relaxation and restoration of the mind and body.

The first stage of non-REM sleep, N1 is often referred to as light sleep because it occurs as the body is transitioning from being awake into a state of sleep. This stage typically lasts from five to 10 minutes during which the eyes begin to close, the muscles begin to relax and melatonin begins to release.

The second stage of non-REM sleep, or N2 is the longest stage in the sleep cycle, typically ranging from 10 to 25 minutes at the beginning of the night, and increasing with each subsequent sleep cycle, making up nearly half of your total sleep time. This stage is characterized by a further slowing down of the body, lowering of the body temperature and heart rate, and a decreased awareness of your environment. This stage brings you deeper into sleep.

The final stage of non-REM sleep is N3. N3 is the deepest and most restorative stage of sleep. It is longest towards the beginning of the night, lasting 20 to 40 minutes, and becomes shorter as the night goes on. During this stage, the brain is relatively inactive and there is increased blood flow to the body, aiding in muscle growth and repair, as well as strengthening of the immune system. This is where the body does most of its healing. N3 also plays a large role in helping us meet our fitness goals.

Being woken up during this stage can often result in disorientation and grogginess since it is the furthest state from wakefulness. It is also the stage where sleepwalking is most seen. Following non-REM sleep, the body transitions into REM sleep, during which the mind lights up with activity. REM sleep is characterized by rapid eye movement, near paralysis of the muscles referred to as atonia, increase brain activity, and most of all, vivid dreaming.

This stage is essential to cognitive function, creativity, emotional well-being, and formation and consolidation of memory. It is also most prominent towards the end of the night.

Now, why is this important? Our sleep architecture, or the amount of time we spend in each stage of sleep dictates our everyday function. Not enough REM sleep, for example, can cause declines in cognitive ability and creativity, increased fatigue, and higher rates of anxiety and depression. Less time spent in N3 may cause harm to the immune system and an increased risk for dementia and chronic diseases, like cancer.

Later in the podcast, Matt gives us insight about the consequences of not getting enough sleep, on exam performance. Let's say you lose two hours of sleep, studying the night before an exam, or in Matt's case, playing video games. You may not think it was that big of a deal, but in reality, it is towards the end of our sleep when more time is spent in REM sleep, the stage where the most cognitive benefits are found.

As a result, even though you only lost two hours of sleep, you have lost much more than that in cognitive function, which is especially detrimental on the day of an exam. This is just one of the situations where understanding sleep can help place us in the driver's seat of our own lives.

Now let's jump in and hear from our guests, Ram Gurumoorthy, Amy Kruse and Luis de Lecea.

‍

Amy Kruse:

Okay. Ooh, what do you have? Something cork finished. Very nice. I have the Southern Tier Pumking.

‍

Matt Angle:

Excellent.

‍

Amy Kruse:

My husband's going to be like, "Where'd the Pumking go?" I'll be like, "I had a podcast pub recording."

‍

Matt Angle:

Maybe we should just start by doing intros. Maybe, Amy, can you start? Can you kick things off? Tell us a little bit about yourself. Name, affiliation.

‍

Amy Kruse:

Amy Kruse. I'm a General Partner at Prime Movers Lab and my background is neuroscience by training.

‍

Matt Angle:

Ram, do you want to go next?

‍

Ram Gurumoorthy:

Yeah. I am Ram Gurumoorthy. I'm the Founder and CTO of StimScience. We are a company that focuses on using stimulation to improve sleep. I got my PhD from Berkeley, long back. I'm like a river that's meandered around.

I'm a technologist. Anything, any technology, doesn't matter. I keep meandering around. My latest passion for the last 15 years have been neuroscience.

‍

Luis de Lecea:

Yeah. My name is Luis de Lecea. I'm a Professor of Psychiatry at Stanford. I've been studying sleep circuits for the last 30 years.

‍

Matt Angle:

Okay. Well, maybe we can get started with our sleep stories. Amy, do you want to kick it off? Do you have a story about sleep, about oversleeping or undersleeping?

‍

Amy Kruse:

Yeah. The first thing that came to mind was, in a previous role, I was a Chief Technology Officer of a global defense company and we had a big office in Australia. The Australians offered to take me out for the night.

I had learned a few things about Australians, but not everything yet. We were in Melbourne. We went to some dance club. The dance club didn't even open until 11:00 PM. My flight, by the way, was the next morning, back home. 14-hour flight back to L.A. or whatever.

We're dancing, 80s night, whatever, whatever, here we go. I look at my watch and it's 2:30 in the morning. My flight's at ... I don't know? Let's say my flight's at 8:30. I'm like, "I got to go or I'm going to miss my flight."

I race back to my hotel room. I pack, I sleep for ... I don't know, let's just say I sleep for three hours or something like that. I set every alarm I have. I wake up, I make it to the airport. I'm fine. I get on the ... Good. Everything's good.

One year later, my colleague does, almost exactly the same trip. I told him. I said, "Whatever you do, don't miss your flight." Well, he got taken out to a much more interesting place than I did, with a bunch of rugby players. Let's just say he missed his flight by oversleeping.

I slept, barely made the flight. He overslept, missed the flight. Same town, same hard drink, and Australians. Just be careful when you're in Australia with Australians.

‍

Matt Angle:

That's good.

‍

Ram Gurumoorthy:

I mean, I started way back with GE Corporate R&D. The interesting thing is, I used to find, any major initiatives, be it a Jack Welch initiative, or a firefight at a business for large contracts, they will reach out to Corporate R&D.

It almost used to be, every three to four months there'll be one call like that, and it never is during the day. Right? It's always, it's around ... By the time I'll be winding up to go do my kung fu class around 7:00, and that's when the call will come in.

They'll be like, "We really are stuck in this, so can you please take a look at this?" Then we'll rally up couple of people and then it becomes an all-nighter. Right? It used to be ... It never happened during the day. That's something ... That's why I used to think, "A firefight has to be only in the nights. Okay."

‍

Amy Kruse:

That's funny. Luis, you're Spanish. There has to be some, you stayed up too late story, right?

‍

Luis de Lecea:

Well, that's the default.

‍

Matt Angle:

Exactly. We have to normalize too late to Luis.

‍

Amy Kruse:

Yeah. Yeah. Baseline is, he stayed up too late.

‍

Luis de Lecea:

Yeah, it's sort of a layman anecdote, actually. I remember, now that you mentioned, I was invited to teach this course in Spain a couple of years ago. Of course, now I'm used to the Western American schedule.

I was worried that I was going to land in this small town at 9:00 PM and I was going to have to order room service. Of course, it so happened that there was a festival in that town when I landed, and the dinner started at 11:00 PM. That was not a problem.

Then, of course there was another party that started at midnight, and then we ended up going to bed at 4:00 AM or so. That was a typical...really, that's a typical Spanish schedule. Not much of an anecdote. It's just, that's how it works there.

‍

Amy Kruse:

Right. Then if you showed up for class the next morning and you were a little bit late, it was not an--

Luis de Lecea:

Oh, yeah. No big deal. Everyone was late. Yeah. Exactly.

‍

Amy Kruse:

Okay. Not the German style, right Matt?

Luis de Lecea:

No. Yeah. In that course, there was, I don't remember her name, someone from Harvard. She was flying the same flight with me, and she couldn't believe her eyes. "What is going on here? They're starting dinner at 11:00, and when there were fireworks at midnight. What are they thinking?" Anyway, that's just the cultural differences.

‍

Amy Kruse:

How about you, Matt?

‍

Matt Angle:

Well, I remember...this is going to--of course, everyone, probably already knows at this point, I'm a nerd. But this will just cement that narrative. I'm remembering, I really liked my biochemistry class in college. Gordon Rule was the teacher.

I remember, I was not super diligent with homework and things, but I really knew this stuff, so I could always make it up on the big tests. I was, kind of cocky, I guess.

One night, I was playing a video game called Knights of the Old Republic, or it was Knights of the Old Republic II. It was some Star Wars role-playing game where you run around with the lightsaber and fight people.

I got to about midnight and I was like, "You know what? If I go to sleep now, I'm just going to be playing this game all week. If I push it, if I just beat the game tonight, I'd just get it over with. I'm done. Then, that seems like, actually the responsible thing to do, because then I can channel all of my energy, the rest of the week into my schoolwork."

I said, "All right, I just got to do it. It's the only responsible thing to do." I stayed up till morning and I crashed hard. Then my alarm at 7:30, to go to my biochemistry midterm woke me up.

‍

Amy Kruse:

Oh, no.

‍

Matt Angle:

I went there. I fell asleep. I fell asleep during the test.

‍

Amy Kruse:

In the test?

‍

Matt Angle:

I lost an hour. I lost an hour from that test.

‍

Amy Kruse:

Brilliant.

‍

Matt Angle:

I think I still got a B or something like that though. It wasn't a terrible performance, but it was pretty unnerving to fall asleep, and wake up and have lost an hour from your test-taking time.

‍

Amy Kruse:

Here you are.

‍

Matt Angle:

Yeah. I still made it through. Just to kick things off, big question, why do we sleep? It would seem, from an evolutionary standpoint that, being defenseless and semi-unconscious for long periods of time at night would be undesirable. Yet, we see that it's selected for, pretty strongly across a large number of different animals. It must be doing something.

‍

Luis de Lecea:

Yeah, that's, of course a million-dollar question, but the prevalent thought in the field is that there has to be, really a fundamental property that is served by sleep. There's recent, I think, very attractive hypothesis in addition to the lymphatic clearance, which is quite attractive, but it has its question marks.

Another hypothesis is that DNA repair is really at the core of sleep. Why? Well, DNA repair, of course is essential to keep the cellular homeostasis. Why would you need to sleep? Why would you need to stop, or reduce, or change the overall neural activity to allow DNA repair?

There's a metaphor that I think is quite appropriate for this. You know, you fix the potholes, usually during the night when there's no traffic. Otherwise, you actually make a big mess. That would be equivalent to what is going on with DNA repair. You need, essentially an offline mode in order for DNA repair to be efficient, and not mess up with the normal processing of information that occurs.

‍

Matt Angle:

That's actually pretty fascinating to me because it wouldn't have been obvious to me on its face that neural activity would be taxing on DNA, and lot of transcription associated with neural activity.

I could imagine, where the sodium potassium ATPase is more active. I could imagine, calcium signaling or local dendritic protein synthesis. All of those things being related to, either activity or plasticity, but I wouldn't have thought—I wouldn't have immediately thought that it would be taxing on the DNA machinery. Is it known, why?

‍

Luis de Lecea:

No. That's, of course, again, that's still a hypothesis. There are others, a lot of science that needs to improve in that sense. What's clear is that neural activity does cause DNA damage over time.

At the circuit level, it's also interesting because the GABAergic cells which are "more active" during sleep, they're actually being repaired during the day. There seems to be, really a inverse relationship between DNA activity and the DNA damage, and that fits the hypothesis that DNA repair is essential for...it's one of the essential features of sleep.

There needs to be more than that, of course. What we see, all the mechanisms that we see that, happen during sleep, now in evolution, probably are very different from what the origin, why sleep evolves in other species. I think it's, essentially an unanswered question.

‍

Matt Angle:

Is there a complexity threshold at which animals sleep or don't sleep, or is there some ... If we trained a machine learning class, if I were to sort the sleepers and the non-sleepers, what would its classification criteria be?

‍

Luis de Lecea:

Yeah, that's a great question. I don't think there's a clear answer to that either. Jerry Seigal has been trying to extract features of species that sleep more versus less, and so forth. Of course, the ecosystem, the ecological niche is one of the determinants of how much a species sleep and what kind of rhythms they have.

A very good example is the one of these bats that sleep 22 hours a day and they only are awake when their food is available. That is, of course an extreme, and probably an extreme adaptation as well. Some people argue, well, the default mode is sleep, and we only need to be awake for a certain amount of time. I think that's an extreme because you really need to feed and you need to reproduce. Indeed, if you can fulfill those functions in a couple of hours a day, the rest, one could sleep, right?

‍

Matt Angle:

I had some friends in college that lived like that.

‍

Luis de Lecea:

Looks like we all know a few examples.

‍

Amy Kruse:

Yeah. Yeah. Well Matt, on that thread ... Thanks for bringing that up, Luis. Back in the day when I was at DARPA, as a Program Manager we had a program called, Preventing Sleep Deprivation. It actually started before I got there, and I inherited it. It started out as a basic science research program.

One of the things that was in the program was actually looking at migratory birds, and fur seals, and dolphins. There is some really interesting variation across the animal kingdom, on how animals sleep, when they sleep, why they sleep.

The fur seals, for example had totally different behavior when they were on land versus when they were in the water. They actually shifted their behavior and their sleep cycles. Dolphins, you've heard can put one half of their brain to sleep while the other keeps swimming. That's obviously an evolutionary adaptation.

One of the reasons is because that little clock I was talking about is migrated off the ventricle of the brain, which is where the cerebral spinal fluid and all of the communications signaling is. The brain, the hemispheres as well as the lack of deeper, connected structure like human brains have, allows the dolphin to be able to put half of its brain to sleep while the other half still functions.

I think there are some really interesting ... That was the question. Are there really interesting lessons to be learned from the animal kingdom about sleep, sleep regulation, and those types of things?

I think it's only appropriate that we bring in travel to Mars and Elon Musk, and how convenient would it be if we had the ability to put humans in, if not stasis or hibernation, some sort of suspended sleep state. Yeah, just wanted to throw that out there, you know?

‍

Matt Angle:

Elon should really consider making a serious play in the neuro space. I bet he'd be pretty good at it.

‍

Amy Kruse:

Yeah. No, he might be interested. I think neuroscience is slightly cooler than space, but I'm open to both.

‍

Ram Gurumoorthy:

I'll take a shot at one perspective. The interesting thing is, see, I, lot of times, analogize our brains to a computer processor and drive, right? Hard drive. It's one, together. It's a processor on a drive, together. Of course, you have all the peripherals. The sensory systems attached back into it, like our USB drives and the microphones and the speakers.

The computer, if you leave it on, running for a long time, it starts hiccuping. It starts having issues. The same thing. Our brains and our body, if we don't have the reset periodically, where we do the defragmenting, where we do the cleaning up, the garbage collection, pretty soon we will see that our wake time even, we will not be efficient.

From an evolutionary perspective, I think, probably learned it the hard way saying, "I may be defenseless, but the rest of the time when I'm awake, I need to be able to defend myself." If we don't get any sleep ... The normal guideline people talk about is, if you sleep less than six hours, your morbidity rate or mortality rate actually goes up. That's the general rule of thumb.

Of course, by now I should be dead because I don't sleep more than four hours. Having said that, it looks like it is the body's way of cleaning up and resetting and getting ourselves set for the next 18 hours.

‍

Amy Kruse:

Yeah, I totally agree with Ram on the cleanup hypothesis. There is a lot of people who think that sleep is related to, what they call the glymphatic system. Actually, the waves within sleep itself are resonsible for cleaning out the brain.

I have asked this question myself to many sleep researchers, and one answer that often comes to the surface from very famous sleep researcher, Giulio Tononi, Giulio would say, "It's the price we pay for plasticity."

If you think about the functions of sleep, it's not only to clean up the brain, which is obviously a huge function that would be conserved, but it's also the time in which our brains engage in memory consolidation, rewiring, and other activities, probably as evidenced by dreaming and other functions like that.

I think there are probably a couple of reasons why we sleep, but they're both very important.

‍

Matt Angle:

Do you have a sense about how far down the filogenetic or how far up the filogenetic tree that goes? I don't think worms sleep. Is there some level of complexity at which ... Just like you don't have to defrag your 1980s digital watch, but you have to defrag your laptop.

‍

Amy Kruse:

There is, Matt. You know, I actually wrote a blog post on this, earlier in the year on sleep, and I did the research. I was like, "Wait, before I say, everything sleeps," it was pretty far down. It wasn't into ... I think, even fruit flies have–

‍

Matt Angle:

Oh, interesting.

‍

Amy Kruse:

... little mini sleeps. It's pretty conserved. I think that probably gives us a big clue to its importance.

‍

Matt Angle:

A lot of the time I hear about sleep, it's about environmental cues, or let's say broad-acting biological signals. Melatonin is an example. If we had free rein of the brain, if we had the ability to flick the lights on and off in whatever neurons we wanted, is it clear, what sub-population of neurons or what structures you'd want to be able to just reach out and touch in order to let someone go to sleep immediately, or wake up immediately?

What if I was staring at a screen all night, and I ate too late, and I've be giving myself all the wrong signals, but nonetheless, I know I have to get up early tomorrow morning. Do we know where some of the shortcuts are? How mapped out are the pathways?

‍

Luis de Lecea:

Well, that's a little bit of what we've been studying for the last 20 years. Yes, there are critical nodes in the brain that determine when and how the brain wakes up. There's a hierarchy to those structures.

Hypothalamus connecting to the catecholamines, the catecholamines to the cortex, and the cortex feeding back to other structures, sensing how much sleep has happened. That's, of course the overall scheme, but there are many, many nuances and many sub-circuits and many iterations that make sleep and wakefulness happen in a reliable and efficient way.

I'm not sure if the question was, if I were to design a new brain? I'm not sure what the question–

‍

Matt Angle:

Yeah. Let's imagine that Karl Deisseroth gave you a very specific virus to transect very specific neural subtype, and a very small fiber optic that you could run within a light's distance of those neurons.

Would it be clear what neurons that you'd want to be able to promote or suppress activity?

‍

Luis de Lecea:

Well, that experiment was done by us, exactly as you said.

‍

Matt Angle:

Well, that's good.

‍

Luis de Lecea:

How long ago? In 2007.

‍

Matt Angle:

Okay.

‍

Luis de Lecea:

14 years ago. Yeah. That was the first opto experiment ever done in video. We targeted our favorite neurons, the hypocretin neurons in the lateral hypothalamus, which are the modulators for sleep. They integrate all of this information that is required to prevent an unwanted sleep event. They're modulated by stress signals.

Obviously, you don't want to go to sleep if you have a predator behind you. They're modulated by metabolism, by circadian rhythms, blood pressure. A whole bunch of homeostatic signals, again, that prevent falling asleep when you can't fall asleep.

In terms of the other side of the equation, how would I induce sleep? That is actually a more elaborate answer because we thought, a couple of decades ago that, there was one or two or maybe three sleep centers that would induce sleep all over the brain when activated.

We're learning now that, that's not the case. The sleep centers are distributed. If you engage one, partially then sleep doesn't happen. You need to engage at least several of them in order to have a full-blown sleep.

Again, knowledge of this is evolving on a yearly basis. We are learning so much more. We know so much more now, than we knew only a few years ago.

‍

Matt Angle:

Are there certain pathologies, like narcolepsy that can give us a hint about sleep mechanisms, or do you think that those are weird phenomena that don't mimic natural sleep?

‍

Luis de Lecea:

No, absolutely. I mean, you're giving me the same example. The hypocretin neurons in the lateral hypothalamus are deficient in narcolepsy. That has been, really the Rosetta Stone in the sleep field for the last 20 years.

The fact that these neurons degenerate in narcolepsy, has allowed us to dissect the neurocircuitry and the language of sleep-to-wake transitions and how the brain decides when to sleep and when to wake up.

‍

Matt Angle:

I see. You say it's not as simple as just suppressing the activity of those neurons. That it's actually a larger network?

‍

Luis de Lecea:

Yeah, that's right. There are, as in any complex engineered system, you have alert signals, like a switch that will turn on or off, everything. Under normal operation, you have a division of labor and redundance to make the system robust. That's, essentially what the brain is about.

‍

Matt Angle:

Amy, if there's any group that has really, really looked at sleep, I'm sure it's the Department of Defense. I'm curious, over the decades of work, to try to keep soldiers awake for weeks at a time, or let them sleep just before a big fighter pilot run or something like that.

Even if not mechanistically understood, are there important existence proofs or empirical findings from that work?

‍

Amy Kruse:

Oh, well, sure. You know, I think, probably one of the things that came out of some of the early work that had been done in the military populations in terms of sleep and sleep deprivation is just how bad sleep deprivation is. Right?

Just, we used to quote this statistic that, essentially being sleep deprived for 24 hours was the equivalent of being intoxicated, in terms of your reaction time and performance and everything else. That's been a really critical interest for the military. Yeah.

‍

Matt Angle:

I'm curious, do you think that's because of the effect of what you described as sleep pressure, or do you think that, that's because the system starts falling apart, already for lack of sleep?

Is it for lack of sleep, for a lack of restorative function, or is it simply because the body is just ramping up the signal, to go to hell, to sleep?

‍

Amy Kruse:

Yeah. That's a really interesting question. We did some work, originally on that same, Preventing Sleep Deprivation program where we tried to get rested baseline, because one of the things, if you're going to do an experiment, you want to get rested baseline in individuals. We couldn't find a marine trainer that had rested baseline.

I don't necessarily think that it's just like the body or system falling apart because there are, certainly people who are able to perform at a pretty high level in the instances of sleep deprivation.

I think it is really the struggle between attention systems, and push and pull of the populations of neurons battling it out in the brain. It's just not a coordinated system. Right? The more sleep deprived you get, I think the less coordinated and functioning that system gets.

I will tell you, by the way that, when we did those experiments, we did experiments in military populations, there were a very few, but a few people who, you could sleep deprive them for 24 hours, and it was nothing. They would just shake it off.

Then there were people who melted down and started seeing visual illusions and other stuff. I do think there is some variability in individuals. When people say, "Oh, I don't need that much sleep," mostly that's just them telling themselves that. It really doesn't relate to the brain.

I would say, the other piece is that, one of the reasons we were so interested in short ... I had a little effort I was playing around with. We were calling it power nap, right? Is there ability to give someone a really short restorative bout of sleep, prior to performance or prior to a mission or whatever, because the problem with keeping people awake a really long time is that you, generally have to give them a stimulant that could be as simple as caffeine, or that could be something else?

Then they need to come back down, right? They end up in this terrible energy/brain cycle of being up and down. Then the thought shifted towards, rather than these extended bouts of wakefulness, maybe you could do something that would give them short bouts of sleepfulness that, would be restorative enough to keep them going.

I do think the military is still interested, and playing around with some of those concepts. That's been an interest for a long time.

‍

Ram Gurumoorthy:

Just to add to couple of the things that Amy was mentioning, and Luis was also talking about. If you think about our systems, lot of the systems that are the subcortical, like the hypothalamus and all these systems are giving you signals and distractions to tell you, "Hey, why you don't want to go to sleep? Something else going on that, you don't want to go to sleep." From a sleep perspective, that's distractions, right?

They are giving you a signal, not to go to sleep, but there are ways. For example, when you typically think about, there are ways to just say, "Can I override that system? Can I override that system?" It's like the Bose noise cancellation system.

What does that say? "As you are hearing something, there is a lot of other noise that's coming in." For purely listening just to music, to that purpose and objective, the rest of the sounds that are coming in are noise. Same way, if you want to go to sleep, of course you may not survive if you went to sleep at the wrong time. If you want to go to sleep, the rest of the systems that are telling you, "Hey, don't go to sleep right now," could be suppressed by canceling them at this level.

When you are trying to ... It's like the noise cancellation headphones. It just says, "Hey, there are other sounds you need to know when you're listening to music. If you're driving, you don't want to completely cancel out everything around you. That's going to be important. If you want to, you should be able to cancel that out." Right? "You can negate those."

It's a reference signal that's being given to the sleep system. It says, "Can I look at some of those references and say, I'll ignore them? I'll choose to ignore it by canceling it right now." That's another interesting perspective people have started thinking about now.

That is something that, is it, evolutionary going to be good if we just do it without considering all the possibilities? Like Luis said, there are the reasons why these have developed, to make sure that we don't fall asleep at the wrong times and we don't fall asleep when we need to be awake. If we are, right now in a nice, safe environment, we should be able to override them.

That's something I just wanted to add to that conversation. From a rhythms perspective, the brain rhythms perspective, when people start getting drowsy, there are these, I would say, the end of a lower frequency, that's the high end of the lower frequency spectrum, which is the alpha rhythms that start heightening.

That's getting you into the drowsy state, right? Non-focused, unfocused, drowsy state. Then when you get into the light sleep, then slightly lower rhythms, which are the theta band activity. The alpha band being eight to 12, eight to 13. Different people see it differently. Then there is the four to eight.

As you get into the slightly lower frequency range, you are actually falling asleep, right? You get into the light sleep mode. Then the theta rhythm starts enhancing and increasing, and it slowly starts shifting into the slower wave rhythms, which is the deep sleep rhythms, which is, it starts with the delta, which typically people think of, in the range of 1 to 4 hertz, and then even lower.

As you get to more recuperative sleep, you even get lower to half a hertz, you start seeing these systems. This is, first from a perspective of, as you are looking at the sleep and how you are getting into sleep, you see these rhythms shifting. It's like a see-saw.

If you look at the power of activity going on in these various frequency ranges in our body, if you've had to plot it like the equalizer in a sound system, that see-saw keeps going up towards the left. That is, the lower frequency start playing a bigger role, and the higher frequencies start dying down as you get to more and more deep sleep.

Then, of course when you get into the REM rhythms, that is not a pure chatter. That, actually are specific bands of activity in a specific shaped activity. That's why it is not ... Lot of people, if you don't consider the morphology of that REM activity in the brain, lot of people find it hard to separate it from wake.

Really, when you look at the morphology of the signal, not just as, "Hey, is there enough power in the slightly higher beta frequencies, et cetera?" then you will not be able to separate out wake from REM.

Actually, the interesting thing is, wake is the one that's the chatter, broad spectrum chatter while REM, actually is a very specific shaped activity in a band, right? From a frequency perspective, you actually can see these variations go through. Of course, why do we need those frequencies? That's another interesting thing, right?

The reason being, I think some of them are related to suppression of external sensory activity, right? It's like the noise cancellation headsets. Some of the slightly higher frequency range, which is in the mu to alpha range, are things that are getting your system geared to say, "I'll start helping you cancel out any arousals or other sensory events that happening around you. I'll start canceling them out."

Then you start going into the lower frequencies. Some of them are related to actual memory consolidation. Some of them are related to, actually the finer reparative activities that are going on.

‍

Matt Angle:

This is, maybe for Luis. Do we have an understanding of where those signals are coming from and what the underlying circuits are? Is it naive of me to think that there are just two states of sleep, or light sleep, slow-wave sleep and REM sleep? Is it exist on a continuum? Is it relatively, homogeneously distributed through the brain, or do you find that there are particular areas controlling particular modes?

Can you give me some insight into the mechanisms there, or is it known?

‍

Luis de Lecea:

Yeah. It's starting to be known with high detail. The big picture is that those oscillations are being driven by thalamocortical loops. Those neurons in the thalamus project to broad areas of the brain. Sleep states are not monolithic. We learned a few years ago that you can have modules of cortical areas undergoing sleep and other areas undergoing other brain states. There's some sort of dissociation of brain states at a given time.

The prevailing concept is that those sleep modules travel from front to back in chaotic patterns driven, again by thalamic activity. Also, driven by the activity of neuromodulators, like catecholamine, subcortical structures, and also driven by interneurons, the intrinsic network in the cortex.

The result of this combination is what we see in cortical waves, that are traveling waves. It's just a beautiful pattern to see, now that we have the tools with calcium and voltage sensors that are so powerful and–

‍

Matt Angle:

Do you see the same spacial dynamics, the same kind of wave propagation in, both slow-wave sleep and REM sleep, or do you find–

‍

Luis de Lecea:

Yeah, it's also a great question. So far, we've only been able to follow those waves with some precision in slow-wave. REM sleep has a very different pattern. We see that as a disruption of the slow-wave pattern. Obviously, with more time and technological advantages, we'll be able to figure that out in REM sleep as well.

‍

Matt Angle:

Are those patterns of activity, are those modes distinctly different from anesthesia?

‍

Luis de Lecea:

Oh, yeah, yeah, yeah. Very different. Anesthesia is, as you know, it's chemically induced and it's so much more uniform than natural sleep. It's just very, very different.

‍

Matt Angle:

Now, for a lot of engineering applications, a cortical target would be preferable than a deep target because it's more easy to access. It's easier to target by non-invasive stimulation. If you have to cut open the brain, while certainly the less deep you have to go, the better.

I'm curious, do you see opportunities to modulate shallow targets? Do you think that you could control sleep with a purely cortical interface?

‍

Luis de Lecea:

You know, people have tried that for a long time without success because that traveling wave is quite elusive. It's really a moving target. If you try to modulate activity in one small cortical area, and the other will come up, it's just very difficult to harness.

I favor manipulating deep structures. Even as difficult as that is, I think it's actually a better strategy.

‍

Matt Angle:

How close are we to being able to engineer a solution to give us sleep after a hard night?

‍

Luis de Lecea:

I hope, not too far. As Amy knows, we have different strategies with ultrasound and electrical stimulation and other non-invasive technologies that allow us to target deep structures, and with the precision that was really unthinkable only a few years ago.

Our knowledge of the circuitry is so much better now that, I think we're really at a time where that intervention is possible.

‍

Matt Angle:

Okay. So for each of you, I'm going to ask you for a number. It's like, how many months or how many years until...forget cost. I have a lot of money. I pay myself a lot of money in Paradromics. A huge amount of money.

‍

Amy Kruse:

Well, wait a minute.

‍

Matt Angle:

Cost is no issue. I need it. I want something that doesn't involve opening the skin. I can put it over my head. At the end of a hard day, I can fall asleep in five minutes' time, regardless of how I've been mistreating my body. How long do you think I'll have to wait until that's a consumer product?

‍

Amy Kruse:

A consumer product?

‍

Matt Angle:

Okay, fine. It could be available by prescription too, but it's something that a person could get. It's something a person could get. Not just existing in one Stanford lab, but something that anyone with a difficulty could get?

‍

Luis de Lecea:

Amy, you want to go first?

‍

Amy Kruse:

I mean, I am optimistic in this space. I certainly see that happening within five years' time. If we were using your unlimited budget, Matt that, apparently you've got, you could probably shorten that to three years.

Obviously, there's a little bit of clinical work and confirmatory science, I think that, needs to be done, but I think we've got the piece parts in place to bring those to market. Obviously, one needs investors to help along the way.

Again, Matt, with your unlimited checkbook, we can just press ahead. I would say, definitely within five years. Optimistic within three.

‍

Matt Angle:

Ram, what do you think?

‍

Ram Gurumoorthy:

Before I talk about that, I'm going to step back a little bit and I'm going to say, I think just going to the deep structures is not the only solution for a different reason because if you think about it, like we talked about the swing example, if I go back to the swing example.

A kid is swinging in the park. Another kid comes to a swing next to him, starts swinging. Initially, they start going asynchronously, but they start wanting to look at each other, smile at each other, chat with each other as they are swinging. The system actually adapts. The brain waves and the functional systems in the brain are indeed adaptive.

I agree, like Luis pointed out, lot of the effort in trying to come at it from a cortical perspective has been, to try to go, exactly match it and shift it. That's wrong because the brain is adaptive.

If you try to go, and intervene and match it, you are going to shift. It's an elusive target. The moment you try to precisely locate it, you have added momentum to it and you don't know what its momentum is. That same idea. You are going to shift the brain activity and the brain states.

If you were to say, "Hey, if I just keep doing," in a bit, you actually are going to get tuned to that frequency. Right? I don't need to just say, "I know exactly where you are. I'll keep chasing you." No. You do it at the right waves, you can actually get the system to adapt to it. You don't chase it. Right?

Having said that. Having said that, so that's something I definitely, slightly differ in the view from what's going on right now in the industry. Having said that, I would say that these kinds of, going at that deeper structures and–

‍

Matt Angle:

No, no, no. I'm asking you how soon you think it could happen. I'm not asking you how it would happen. I'm saying, how soon do you think it'll be before I can put on a helmet and go to bed in five minutes?

‍

Ram Gurumoorthy:

Our hypothesis, or what we have, we have some good evidence in StimScience to show it's just stimulation at the cortical level. We are able to get people to have their onset latency, which is with, both insomniacs and non-insomniacs, right?

We think it's probably in the order of, couple of years. It's not much longer. We already are able to show that we can, with stimulation, electrical stimulation, we are able to get people to consistently reduce their onset latency and also increase the quality of sleep.

Meaning, can you actually lessen the toss and turn? Can you reduce those transitions? Like Luis pointed out, these are things, while you sleep, it's not a beautiful, nice rhythm of you going to N1 to N2 to N3 to REM and come back to N1. It's not that, right? You keep going, jumping back and forth.

We are able to even reduce those transitions, right? The choppy transits. The choppiness of your sleep. We think that it's, possibly the next couple of years.

‍

Matt Angle:

What is it like to wear a transcranial stimulation hat? I've never actually tried that. I'm a big neurotech advocate, but I've never actually put electricity into my brain in that way. Can someone describe the subjective experience of doing that?

‍

Ram Gurumoorthy:

Right. It's just like, let's say you put a band on. You need to have a good level of conductivity and contact because the skin and the face is something which is important. Whether you measure or you are stimulating, actuating, you need to make sure that your impedance at the skin level is good.

Let's say you've taken care of that. We have seen, there are different ways of doing that. Once you have taken care of it, I would say, when you stimulate it electrically, you could do it in different places. Like Luis pointed out, it starts from the frontal and goes backwards.

If you focus on the frontal stimulation, then I would say there are a couple of things. You will feel a little bit of a small tingle or warmth. Have you done those muscle stimulators, the TENS machines, right?

‍

Matt Angle:

Yeah. They're–

‍

Ram Gurumoorthy:

Yeah. Exactly, exactly. I was going to say, those are just like bang, bang, bang, bang, bang, bang. They just try to force twitch your muscles. Right? Whereas, the kinds of stimulation we are talking about is very minimal orders of magnitude below that, in terms of the currents that you're talking about. It's less than a milliamp. That's the kind of, orders of things.

Those kinds of stimulation, when you put it on, it feels a little bit ... Some different sensations, people usually feel are, number one, small tingling, pulsing that you will feel that, you get habituated to very quickly, within minute or two of you starting to stimulate. You do it for, let's say 15 minutes. In the first couple of minutes, you start habituating to that.

Then the other thing people feel is a little bit of warmth, because there's something going on and it's also closing, right? You have an open forehead and suddenly something is put on it. You feel a little bit of warmth.

The interesting thing. If you did it in the frontal as opposed to other parts of the brain, because you are so close to the dipole, the optical dipole, some people also see a little bit of phosphenes, the peripheral visual flashing. You do feel a little bit of that, but that's the extent of what people describe their sensations, because we have run couple of hundred people through these protocols. The sensations, we record them stating it, also. These are the kinds of sensations people typically describe.

‍

Matt Angle:

At that level of current, what is some ... I'm sure there's some animal studies. How much does that change the firing rates of cortical neurons? If you have one neuron that had a firing probability of .2 during a one-second interval, is the probability of it firing .22 or is it .8, just to give me a mental model?

Is this like a butterfly landing on my daughter's shoulder or is this a NFL linebacker trying to make her go upside down?

‍

Ram Gurumoorthy:

Right. Hopefully, it is somewhere in the middle, because the interesting thing is, a lot of people, even with not even these kind of frequencies of stimulation, people talk about the immediate excitation or suppression of the neurons under the area you're stimulating.

That, I think is just, it's basically, it's the mechanism to get to the end. The end, really is to get you to swing properly. When I'm pushing you initially, if I push you a little hard, you are actually, instead of getting the kid to swing back and forth normally, you are actually jerking that kid, right?

That, you have to fine tune at a personal level because each of us, it's little different, right? What level is actually a little jerky to us, it's a little fine tune needed. It's not a jerk that really gets you sleeping. It's the gentle oscillations that you need that, will make you sleep.

That's why I really think the suppression or excitation of the neuronal firing is an underlying phenomena, but that gets aggregated into the actual sleep waves. That's really what we are going after.

‍

Matt Angle:

Luis, when you're looking at, from a cortical activity, just for the audience, to try to keep a mental picture of this, how stable or precarious is the sleep-to-wake transition? Is it two states separated by a strong energy barrier or is it a relatively subtle balance?

‍

I'm just trying to get a sense of this for when we talk about some of the technologies that would modulate these states, is it the case that we walk around on the edge of sleep, and we can easily be nudged one way or the other? Is it the case that we have really strong system dynamics that hold us in one state or the other, and we have to come in with a sledgehammer?

‍

Luis de Lecea:

It's more the second. Of course, in conditions like narcolepsy, it's more like the first. It's like, narcoleptics are walking on really thin wire because there's nothing to secure and stabilize one state or the other.

Under normal conditions, it's like, there's a safety system that solidify, that stabilize one state versus the other. That's why, usually we don't feel like we are going to fall asleep anytime in the middle of the day. Narcoleptics do feel that all the time. That's one of the experiences they report.

‍

Matt Angle:

If there were one ask, something that you would like to see happen in the field that, you think would accelerate sleep therapies, what do you think is the missing piece of the puzzle that would allow everyone to move faster?

For instance, would it be a better stimulation technology? Would it be a specific experiment to help identify targets? Would it be bringing the cost of goods down to make this more widely available? Would it be Amy putting more money into the field? What do you think is the accelerant that you'd all like to see?

Luis de Lecea:

All of the above. I think you hit all of the critical points.

‍

Matt Angle:

No, no, that's cheating. That's cheating. You have to pick. What's your one thing? This genie only grants one wish.

‍

Luis de Lecea:

Yeah. Technology, I think is the big issue, because in animals, we are able to switch back and forth relatively easily. We can make animals fall asleep within five seconds and we can wake them up at will even if they're sleep-deprived. We can do many, many things with invasive technologies and genetically precise technologies.

‍

Matt Angle:

Do those same genetically precise technologies exist scientifically in humans? Is it really just a regulatory barrier?

‍

Luis de Lecea:

Well, yeah. I mean, there are many technical barriers. If one could inject a virus into a human brain, and with a very precise technology, you would still have lots of issues. I mean, in monkeys, for instance, in non-human primates, optogenetics is not as successful as–

‍

Matt Angle:

I see. I see. It's not like purely a societal or a regulatory barrier. The real technical issue is moving from–

‍

Luis de Lecea:

Yeah.

‍

Matt Angle:

...system to–

‍

Luis de Lecea:

Not–

‍

Matt Angle:

... system.

‍

Luis de Lecea:

Not insurmountable, I think. Of course, if one had unlimited funding, yes, I think that those could be solved. Still, they're not easy to fix.

‍

Matt Angle:

Are you confident that the same mechanisms that put a mouse to sleep and wake it up are conserved across humans and primates?

‍

Luis de Lecea:

Yeah. The hypothalamus is extremely conserved in, it's a lizard's brain within our brain. That is really what drives sleep-wake transitions and controls stability. Yeah, absolutely.

‍

Matt Angle:

What about you, Amy? You're wearing a different hat today. What do you think? What–

‍

Amy Kruse:

Hopefully it's a sleep hat.

‍

Matt Angle:

Given Luis thinks that the underlying fundamental science is really strong, what do we need to make a practical therapy?

‍

Amy Kruse:

Yeah. I would probably err on the side of, maybe what I would call the systems integration technology side. I think a lot of the general tools and techniques are available, but they need to be miniaturized, or algorithmatized, or made into something that somebody could use in a deployed/mobile/home setting.

I agree with Luis that, a lot of the fundamental sciences, I think ... You know, nothing's ever perfect, but I think it's well-described. I think it's really about the design of closed loop systems that are wearable and usable in those mobile environments that, would be my wish for the sleep genie.

‍

Matt Angle:

Do you think that integration is a function of getting the right people at the table, or do you think it's cash-limited, or…

‍

Amy Kruse:

I mean, money always helps. You know, I have a famous saying, "There's things that money makes go faster and there are things that money doesn't." These things seem to be clearly in the category of, money could make them go faster.

I would say, getting the right people on the team, having the right capitalization to go after these problems. We've certainly seen individuals that were well-capitalized and highly capitalized, take on some really important systems engineering challenges and nail them. That would be my recommendation.

‍

Matt Angle:

Ram, what about you? What do you think is, because you have a different...you have a contrarian hypothesis about stimulating sleep. Perhaps you have a, also different take on what is needed.

‍

Ram Gurumoorthy:

I would say there is a lot of interesting knowledge, already about the functional, phenomenological model of sleep versus the cellular level and the cellular level models of sleep, I would say.

Also, the technologies, like Amy said, in terms of actual technology of their hardware, or what is needed, is out there, needs to be properly put together. I would say that one of the key things is, it's a data science problem. It's a data science problem of taking all this knowledge, taking those models and putting them together, because it's actually ... All said and done, some of these subsystems are actually inputs and references to our sleep. Some of these are actual actuators of sleep. Some of those subsystems are actual indicators or measurements or outputs of, really where we are in the sleep.

It's a data science problem of putting them together in the right way and getting us more efficacious in shaping sleep, because I keep coming back to that phrase, it's about shaping sleep.

Like Luis pointed out, to really say, "Somebody's awake. Get them into sleep," is a tough proposition, right? It needs that sledgehammer, kind of jerk. If we get this data science model of combining that to say, "Is there a way to shape the sleep?" That's what I come back to the swing analogy. Is there a way to actually shape you into the right oscillation that, you will dose into a good, high-quality sleep?

It's not against the other perspectives. It's just an additional perspective to say, there is a lot of focus that needs to come into the data science and modeling aspect of bringing these various knowledge and models together to say, "How can I shape sleep?"

The actual mechanisms of, what is that going to look like? Is it a helmet that's actually stimulating deep structures, deeper structures, or surface level structures, those, I think are out there. We just have to pull the right pieces, but it's actually, to pull this underlying actuation subsystem information, the reference subsystem information and the output system. That becomes a data science focus. I think that's my wish at this point.

‍

Matt Angle:

Amy, Luis, thank you so much for your time today. This has been really interesting for me.

‍

Luis de Lecea:

Thank you.

‍

Amy Kruse:

Thanks for inviting us.

‍

Ram Gurumoorthy:

It was fantastic meeting all of you.

‍

Amy Kruse:

Yeah. Thanks for having us, Matt. We really appreciate it.

‍

Matt Angle:

I love this. I get to selfishly choose topics that I want to learn about and then invite world experts.

‍