Matt Walker is a sleep scientist at Berkeley, author of Why We Sleep, and the host of a new podcast called The Matt Walker Podcast. Please support this podcast by checking out our sponsors:
- Stamps.com: https://stamps.com and use code LEX to get free postage & scale
- Squarespace: https://lexfridman.com/squarespace and use code LEX to get 10% off
- Athletic Greens: https://athleticgreens.com/lex and use code LEX to get 1 month of fish oil
- BetterHelp: https://betterhelp.com/lex to get 10% off
- Onnit: https://lexfridman.com/onnit to get up to 10% off

EPISODE LINKS:
Matt's Podcast: https://themattwalkerpodcast.buzzsprout.com/
Matt's Twitter: https://twitter.com/sleepdiplomat
Matt's Instagram: https://www.instagram.com/drmattwalker
Matt's Website: https://www.sleepdiplomat.com/

PODCAST INFO:
Podcast website: https://lexfridman.com/podcast
Apple Podcasts: https://apple.co/2lwqZIr
Spotify: https://spoti.fi/2nEwCF8
RSS: https://lexfridman.com/feed/podcast/
Full episodes playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4
Clips playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOeciFP3CBCIEElOJeitOr41

OUTLINE:
0:00 - Introduction
2:05 - Putin moment: Lex takes Matt's sunglasses
2:26 - Fascination with sleep
6:35 - Why do we sleep?
15:06 - Computer vision for driver assistance
24:28 - Consciousness is fundamental
32:34 - Lex on human to robot connection
35:01 - Scent of a Woman is better than "John Wick"
46:42 - Distinction between coffee and caffeine
1:12:26 - The science of 'sleeping on it'
1:26:19 - Lex on his sleeping schedule
1:51:23 - Chronotypes
1:58:52 - How to overcome insomnia
2:16:15 - Diet and sleep
2:25:12 - Where do dreams come from?
2:38:50 - How sleep affects emotions
2:45:43 - Meaning of life

SOCIAL:
- Twitter: https://twitter.com/lexfridman
- LinkedIn: https://www.linkedin.com/in/lexfridman
- Facebook: https://www.facebook.com/lexfridman
- Instagram: https://www.instagram.com/lexfridman
- Medium: https://medium.com/@lexfridman
- Reddit: https://reddit.com/r/lexfridman
- Support on Patreon: https://www.patreon.com/lexfridman

The following is a conversation with Matt Walker, sleep scientist, professor of neuroscience and psychology at Berkeley, author of Why We Sleep, and the host of a new podcast called the Matt Walker Podcast.

It's 10 minute episodes a couple of times a month, covering sleep and other health and science topics.

It's up there with the greats like the Huberman Lab podcast with Andrew Huberman, and I think David Sinclair's putting out an audio series soon too.

I'm really excited by the future of science in the podcasting world.

To support this podcast, please check out our sponsors, Stamps.com, Squarespace, Athletic Greens, BetterHelp, and Onnit.

As a side note, let me say that to me, a healthy life is 1 in which you fall in love with the world around you, with ideas, with people, with small goals and big goals, no matter how difficult, with dreams you hold on to and chase for years.

Second to that is the understanding and the utilization of the best available science on diet, exercise, supplements, sleep, and other lifestyle choices.

To me, science in the realm of health is a guide for we should try, not the absolute truth of how to live life.

The goal is to learn to listen to your body and figure out what works best for you.

All that said, a good night's sleep can be a great tool in making life awesome and productive, and Matt is a great advocate of the how and the why of sleep.

We agree on some things and disagree on others, but he's a great human being, a great scientist, and is recently a friend with whom I enjoy having these wide-ranging conversations.

This is the Lex Friedman podcast, and here is my conversation with Matt Walker.

You should try these shades on and see what you look like.

It's the Same thing as Putin took the Super Bowl ring and it's now his ring.

Yeah, 1 wonders if he was offered it, but they are yours.

When did you first fall in love with the dream of understanding sleep?

Like where did the fascination with sleep begin?

So back in the United Kingdom, you can sort of start doing medicine at age 18 and it's a 5 year program.

And I was at the Queen's Medical Center in the UK.

And I remember just being fascinated by states of consciousness and particularly anesthesia.

Within seconds, I can take a perfectly conscious human being and I can remove all existence of the mentality and their awareness within seconds.

So I started to get really interested in conscious states.

I even started to read a lot about hypnosis.

And all of these things, hypnosis, even sleep and dreams at the time, they were very esoteric.

It was sort of charlatan science at that stage.

And I think almost all of my colleagues and I are accidental sleep researchers.

No 1, as I recall, in the classroom when you're sort of 5 years old and the teacher says, what would you like to be when you grow up?

No one's putting their hand up and saying, I would love to be a sleep researcher.

And so when I was doing my PhD, I was trying to identify different forms of dementia very early on in the course.

And I was using electrical brainwave recordings to do that.

It was a disaster, just no result after no result.

And I used to go home to the doctor's residence with this sort of a little igloo of journals that at the weekend I would sort of sit in and read.

And which I'm now thinking, do I really want to admit this?

Cause it sounds like I had no social life, which I didn't, I was social leper.

But, and I started to realize that some parts of the brain were sleep related areas.

And some dementias were eating away those sleep related areas, other dementias would leave them untouched.

And I thought, well, I'm doing this all wrong.

I'm measuring my patients while they're awake.

Instead I should be measuring them while they're asleep.

Started doing that, got some amazing results.

And then I wanted to ask the question, is that sleep disruption that my patients are experiencing as they go into dementia, maybe it's not a symptom of the dementia, I wonder if it's a cause of the dementia.

And at that point, which was, cough, cough, 20 years ago, no 1 could answer a very simple fundamental question, why do we sleep?

And I at the time didn't realize that some of the most brilliant minds in scientific history had tried to answer that question and failed.

And at that point, I just thought, well, I'm going to go and do a couple of years of sleep research and I'll figure out why we sleep.

And then I'll come back to my patients in this question of dementia.

And what I realized is that hard questions care very little about who asks them.

They will meter out their lessons of difficulty all the same.

And I was schooled in the difficulty of the question, why do we sleep?

But in truth, 20 years later, we've had to upend the question.

And by the way, the answer then was that we sleep to cure sleepiness.

Which is like saying, we eat to cure hunger.

That tells you nothing about the physiological benefits of food, same with sleep.

Now we've actually have to ask the question, is there any physiological system in the body or any major operation of the mind that isn't wonderfully enhanced when we get sleep or demonstrably impaired when we don't get enough?

And so far for the most part the answer seems to be no.

So why does the body and the mind crave sleep then?

How can we begin to answer that question then?

So I think 1 of the ways that I think about this or 1 of the answers that came to me is the following.

The reason that we implode so quickly and so thoroughly with insufficient sleep is because human beings seem to be 1 of the few species that will deliberately deprive themselves of sleep for no apparent good reason, biological.

And what that led me then to was the following.

Mother nature as a consequence, so No other species does what we do in that context.

There are a few species that do undergo sleep deprivation, but for very obvious, clear biological reasons.

1 is when they're in a condition of severe starvation.

The second is when they're caring for their newborn.

So for example, killer whales will often deprive themselves.

The female will go away from the pod, give birth and then bring the calf back.

And during that time, the mother will undergo sleep deprivation.

And then The third 1 is during migration when birds are flying trans oceanographic to 3000 miles.

But for the most part, it's never seen in the animal kingdom.

Which brings me back to the point, therefore, mother nature in the course of evolution has never had to face the challenge of this thing called sleep deprivation.

And therefore, she has never created a safety net in place to circumnavigate this common influence.

And there is a good example where we have, which is called the adipose cell, the fat cell.

Because during our evolutionary past, we had famine and we had feast.

And mother nature came up with a very clever recipe, which is how can I store caloric credit so that I can spend it when I go into debt?

And the fat cell was born, brilliant idea.

Where is that sort of banking chip for sleep?

And unfortunately, we don't seem to have 1 because she's never had to face that challenge.

So even if there's not some kind of physics

fundamental need for sleep, that physiologically or psychologically, the fact is Most organisms are built such that they need it.

And then mother nature never built an extra mechanism for sleep deprivation.

So it's interesting that why we sleep might not have a good answer.

But we need to sleep to be healthy is nevertheless true.

Yeah, and we have many answers right now.

In some ways, the question of why we sleep was the wrong question too.

It's what are the pluripotent many reasons we sleep?

We don't just sleep for 1 reason, because from an evolutionary perspective, it is the most idiotic thing that you could imagine.

When you're sleeping, you're not finding a mate, you're not reproducing, you're not caring for your young, you're not foraging for food, and worse still, you're vulnerable to predation.

So on any 1 of those grounds, but especially as a collective, sleep should have been strongly selected against in the course of evolution.

But in every species that we've studied carefully to date, sleep is present.

I think I've heard arguments from an evolutionary biology perspective that sleep is actually advantageous.

You know, maybe like some kind of predator-prey relationships.

But you're saying, and it actually makes way more sense what you're saying, is it should have been selected against.

Because, and you know, there was an energy conservation hypothesis for a while, which is that we need to essentially go into low battery mode, power down because it's unsustainable.

But in fact, that actually has been blasted out the water because sleep is an incredibly active process.

In fact, the difference between you just lying on the couch but remaining conscious versus you lying on the couch and falling asleep, it's only a savings of about 140, 150 calories.

In other words, you just go out and club another baby seal or whatever it was and you wouldn't worry.

So it has to be much more to it than energy conservation, much more to it than sharing ecosystem, space, and time, much more to it than simply predator-prey relationships.

If sleep really did, and looking back, even very old evolutionary organisms like earthworms, millions of years old, they have periods where they're active and periods where they're passively asleep.

And so what that in some ways suggested to me was sleep evolved with life itself on this planet, and then it has fought its way through heroically every step along the evolutionary pathway, which then leads to the sort of famous sleep statement from a researcher that if sleep doesn't serve an absolutely vital function or functions, then it's the biggest mistake the evolutionary process has ever made.

And we've now realized Mother Nature didn't make a spectacular blunder with sleep.

You've mentioned the idea of conscious states.

Do you think of sleep as a fundamentally different conscious state than awake-ness?

So when you're into it, you're understanding of what the mind can do.

Do you think awake state, sleep state, or is there some kind of continuum?

There's a complicated state transition diagram.

Like how do you think about this whole space?

I think about it as a state space diagram.

And I think it's probably more of a continuum than we have believed it to be or suggested it to be.

So we used to think absent of anesthesia, that there were already 3 main states of consciousness.

There was being awake, being in non-rapid eye movement sleep or non-dream sleep, and then being in rapid eye movement sleep or dream sleep.

And those were the 3 states within which your brain could percolate and be conscious.

You know, conscious during non-REM sleep is maybe a stretch to say, but I still believe there is plenty of consciousness there.

And the reason is because we can have daydreams and we are in a very different wakeful state in those daydreams than we are when we are, as we are now together, present and extraceptively focused rather than intraceptively focused.

And then we also know that as you are sort of progressing into those different stages of sleep, during non-REM sleep, you can also still dream.

Depends on your definition of dreaming, but we seem to have some degree of dreaming in almost all stages of sleep.

We've also then found that when you are sleep deprived, there are even individual brain cells will fall asleep.

Despite the animal being, you know, behaviorally from best we can tell awake, individual brain cells and clusters of brain cells will go into a sleep-like state.

When we are sleep deprived, we have what are called micro-sleeps, where the eyelid will partially close and the brain essentially falls, lapses into a state of sleep, but behaviorally you seem to be awake and the danger here is road traffic accidents.

So these are the, what we call these sort of microsleep events at the wheel.

Now, if you're traveling at 65 miles an hour in a two-ton vehicle, it takes probably around 1 second to drift from 1 lane to the next, and it takes 2 seconds to go completely off the road.

So if you have 1 of these microsleeps at the wheel, you know, it could be the last microsleep that you ever have.

But I don't now see it as a set of, you know, very binary distinct, you know, step function states.

So I've for 5, 6 years at MIT, really focused on this human side of driving question.

And 1 of the big concerns is the micro-sleeps, drowsiness, these kinds of ideas.

And 1 of the open questions was, is it possible through computer vision to detect or any kind of sensors?

The nice thing about computer vision is you don't have to have direct contact to the person.

Is it possible to detect increases in drowsiness?

Is it possible to detect these kinds of micro-sleeps or actually just sleep in general?

These are all words that have so many meanings and so many debates, like attention is a whole nother 1.

Just because you're looking at something doesn't mean you're loading in the information.

Just because you're looking away doesn't mean your peripheral vision can't pick up the important information.

There's so many complicated vision science things there.

So I wonder if you could say something to, you know, they say the eyes are the windows to the soul.

sleepiness through computer vision, just looking at the video of the face?

And Andrew Huberman and I, your friend, have talked about this.

There's blinking and there's eye movement.

And those are the ones that can be picked up with computer vision.

Do you think those are signals that could be used to say something about where we are in this continuum?

And I think there are a number of other features too.

I think, you know, aperture of eye, so in other words, partial closures, full closures, duration of those closures, duration of those partial closures of the eyelid.

I think there may be some information in the pupil as well, because as we're transitioning between those states, there are changes in what's called the automatic nervous system or technically it's called the autonomic nervous system, part of which will control your pupillary size.

So I actually think that there is probably a wealth of information.

When you combine that probably with aspects of steering, angle steering, maneuver, and if you can sense the pressure on the pedals as well, my guess is that there is some combinatorial feature that creates a phenotype of you are starting to fall asleep.

And as the autonomous controls develop, it's time for them to kick in.

Some manufacturers, auto manufacturers sort of have something beta version, or maybe an alpha version of this already starting to come online, where they have a little camera in the wheel that I think tries to look at some features.

Almost everybody doing this, and it's very alpha.

So, you know, the thing that you currently have, some people have that in their car, there's a coffee cup or something that comes up that you might be sleepy.

The primary signal that they're comfortable using is a steering wheel reversals.

So basically using your interaction with the steering wheel and how much you're interacting with it as a sign of sleepiness.

So if you have to constantly correct the car, that's a sign of like you starting to drift into micro sleep.

I think that's a very, very crude signal.

There's a whole nother component to this, which is it seems like it's so driver and subject dependent.

How our behavior changes as we get sleepy and drowsy seems to be different in complicated, fascinating ways where you can't just use 1 signal.

There has to be a lot of different signals that you should then be able to combine.

The hope is there's the searches for like universal signals that are pretty damn good for like 90% of people.

But I don't think we need to take necessarily quite that approach.

I think what we could do in some clever fashion is using the individual.

So what you and I are perhaps suggesting here is that there is an array of features that we know provide information that is sensitive to whether or not you're falling asleep at the wheel.

Some of those, let's say that there are 10 of them, for me, 7 of them are the cardinal features.

For you, however, 6 of them, and they're not all the same sort of overlapping, are those for you.

I think what we need is algorithms that can firstly understand when you are well-slept.

So let's say that people have sleep trackers at night and then your car integrates that information.

And then you've got the data of the individual behavior unique to that individual snowflake like when they are well-slept.

This is the signature of well-rested driving.

Then you can look at deviations from that and pattern match it with the sleep history of that individual.

And then I don't need to find the sort of, you know, the 1 size fits all approach for 99% of the people.

I can create a very bespoke, tailor-like set of features, a Savile Row suit of sleepiness features.

You know, that would be my, if you want to ask me about moonshots and crazy ideas, that's where I'd go.

But to start with, I think your approach is a great 1.

Let's find something that covers 99% of the people, because the worrying thing about microsleeps, of course, unlike drugs or alcohol, which certainly is a terrible thing to be behind the wheel, with those, often you react too late, And that's the reason you get into an accident.

When you fall asleep behind the wheel, you don't react at all.

You know, at that point, there is a two-ton missile driving down the street and no one's

That's why those accidents can often be more dangerous.

Yeah, and the fascinating thing is, in the case of semi-autonomous vehicles, like Tesla Autopilot, this is where I've had disagreements with Mr.

Elon Musk, is, and the human factors community, which is this community that 1 of the big things they study is human supervision over automation.

So you have like pilots, you know, supervising an airplane that's mostly flying autonomously.

The question is when we're actually doing the driving, how do microsleeps or general, how does drowsiness progress and how does it affect our driving?

That question becomes more fascinating, more complicated when your task is not driving, but supervising the driving.

So your task is to take over when stuff goes wrong.

And that is complicated, but the basic conclusions from many decades is that humans are really crappy at supervising because they get drowsy and lose vigilance much, much faster.

The really surprising thing with Tesla autopilot, it was surprising to me, surprising to the human factors community, and in fact, they still argue with me about it, is it seems that humans in Teslas with autopilot and other similar systems are not becoming less vigilant, at least with the studies we've done.

So there's something about the urgency of driving.

I can't, I'm not sure why, but there's something about the risk, I think the fact that you might die is still keeping people awake.

The question is, as Tesla autopilot or similar systems get better and better and better, how does that affect increasing drowsiness?

And that's when you need to have, that's where the big disagreement was.

You need to have driver sensing, meaning driver facing camera that tracks some kind of information about the face that can tell you drowsiness.

So you can tell the car if you're drowsy so that the car can be like, you should be probably driving or pull to

Right, or I need to do some of the heavy lifting here.

Yeah, so there needs to be that dance of interaction of a human and machine.

But currently it's mostly steering wheel based.

So this idea that your hands should be on the steering wheel, that's a sign that you're paying attention is an outdated and a very crude metric.

I think there are far more sophisticated ways that we can solve that problem if we invest.

Can I ask you a big philosophical question before we get into fun details?

On the topic of conscious states, how fundamental do you think is consciousness to the human mind?

I ask this from almost like a robotics perspective.

So in your study of sleep, do you think the hard question of consciousness, that it feels like something to be us, Is that like a nice little feature, like a quirk of our mind, or is it somehow fundamental?

Because sleep feels like we take a step out of that consciousness a little bit.

So from all your study of sleep, Do you think consciousness is like deeply part of who we are or is it just a nice trick?

I think it's a deeply embedded feature that I can imagine has a whole panoply of biological benefits.

But to your point about sleep, what is interesting if you do a lot of dream research and we've done some, it's very, very rare at all, in fact, for you to end up becoming someone other than who you are in your dreams.

Now you can have third person perspective dreams where you can see yourself in the dream as if you're sort of, you know, you've risen above your physical being.

See all Lex Fridman transcripts on Youtube

Matt Walker: Sleep | Lex Fridman Podcast #210