Translator: Tanya Cushman
Reviewer: Queenie Lee You know, for a sleep scientist, I actually don't sleep very well. Any little chink of light in the room,
and I'm awake all night. And my eye-mask is just
as important to me as my laptop. But really, I take sleep very seriously, and I'm hoping this talk will be
a sort of wake-up call to all of you to make some of you
feel the same about it. Now, what I'm not going to do is preach to you about how you
should get more sleep. We all know that. We all know we live
in a sleep-deprived society. Instead, I'm going to talk about something
which I think is much more interesting. This is how we can manipulate
the sleep that we do get in order to get the most out of it, in order to improve our quality of life. And I call this "the new science
of sleep engineering." But let's start from the beginning. As humans, we spend
roughly a third of our lives asleep, eight hours a day. That's more time than we spend
doing anything else; that's a huge amount of time. And just the pure fact
of that time investment suggests that sleep must be doing
something incredibly important. But what is this? Well, it turns out that sleep
is all about the brain. Contrary to popular opinion, the brain doesn't just switch off
when we go to sleep; instead, it goes through a series of highly specific,
different types of activities. We can measure these by putting
electrodes all over the scalp, like this. This, by the way, is
a sleep scientist's idea of a selfie. (Laughter) With these electrodes, we can measure
the electrical activity of the brain. And during wake, it looks something
like this - just a wiggly line, with time going from left to right. And what that tells us
is the brain is active - good - and that the activity
is not particularly synchronized, so things aren't summing up
in any particular way. But as we fall asleep,
the pattern changes a little bit. It slows down a little bit, and the amplitude of those brainwaves
gets a little bit higher, showing a bit more synchrony. And we also start to see occasional bursts
of high-frequency activity that we call "sleep spindles." These spindles don't occur
across the whole brain; they just occur in localized
areas at any one time. I'm going to come back to them
several times during the talk, so try to remember what those look like. Now, as we go deeper into sleep,
the activity slows down still more, and we start to see these
high-amplitude, slow oscillations that we call "slow waves." And this shows a high degree of synchrony
in the firing across the cortex. So many neurons are all firing together,
then pausing, then firing together. It's very different than the kind
of activity that we see during wake. And if we go still deeper, we go into a sleep stage
that I'm sure you've all heard about - rapid eye movement sleep. This is famous for the way
the eyes dart around under closed lids, and it actually looks very similar
to the brain activity that we see during wake, probably because of the dreaming
that's happening, not much cortical synchrony there. So why do we do this? Why do our brains
spend a third of our life going through these highly precise
different types of activity in a cycle from one stage to another? Well, there are two main answers to this. One of them relates to sleep's role
in maintaining a healthy brain, and the other to its role
in learning and memory. And I'm going to start
by talking about the healthy brain. Sleep plays a sort of housekeeping role: it cleans our brains,
it helps us to remove toxins. And some of the most
interesting studies of this have shown that the spaces
between brain cells expand during that slow-wave sleep
that I showed you by as much as about 60%. This allows cerebrospinal fluid,
the fluid in the brain, to flush through and efficiently clear
away toxins that build up during wake. One of these toxins
that's particularly interesting is something you might have
heard about - beta amyloid. This is a protein that can build up not only during wake
but actually across a lifetime. And buildups of beta amyloid are linked
to a formation of plaques in the brain that are predictive
of cognitive impairment, particularly problems with memory. If it gets bad, it's also linked
to dementia and Alzheimer's Disease. Beta amyloid is also linked
to cell death in the brain and a gradual degeneration
of some parts of the cortex that can happen with aging, again, in dementia
and Alzheimer's Disease. So it's obvious it's important for us
to flush this out of the brain if we can. Now, interestingly, as we age,
our sleep patterns also change. Our sleep becomes more fragmented, and those high-amplitude,
slow oscillations that I told you about gradually stretch and flatten out, and after the age of 65 or so, it's quite common not to get
any more slow-wave sleep at all. Problem, right? Furthermore, this gradual decline
in slow-wave across the lifecycle has been shown to predict the extent to which the cortex
actually atrophies and shrinks, so [for] some of the prefrontal
regions of the cortex that shrinkage is predicted
by the decrease in slow-wave sleep. So wouldn't it be great if there was a way that we could maintain
those slow-waves as we got older, and not have that decline? And this is where we come
to sleep engineering. Very recent research has suggested
a way that we can do this. If we play sounds to people - just click, simple click sounds - while they're in slow-wave sleep, and if we place those sounds near the peaks of those
high-amplitude, slow oscillations, it turns out that can enhance them. Let me show you what this looks like. (Clicking) (Clicking) So the clicks occurring just
near the peaks boost the amplitude, and they've also been shown
to improve memory the next day. That works very well
in healthy young people. And I've got several quite
sleep-deprived graduate students who are working hard
on taking this to the older population. Right now, our results are very promising, so we're hoping that in not too many years we might be able to offer
a sort of preventative treatment that could help people to maintain
their slow-wave sleep as they get older, and possibly might slow down
some of this decline - cortical and cognitive - that happens. So, let me move on now
to talk about learning and memory. [Why Sleep?] In his famous book
"One Hundred Years of Solitude," Gabriel Garcia Marquez wrote about a plague of insomnia
that swept across a land. People just couldn't sleep. At first, they didn't mind that at all. But eventually, negative symptoms
started to manifest. And these symptoms were the fact
that they lost their memories, they couldn't learn new things, and they started to forget
what objects were. They had to cover things
with notes, saying things like "This is a cow, it gives milk. Pull here." (Laughter) Given that this book
was published in 1967, when we knew almost nothing
about sleep's role in memory, it's really quite remarkable
that Garcia Marquez had insight into this. But subsequent research has shown
that he was absolutely right. Sleep is very important
for forming new memories and also integrating those memories
with what we know already, also strengthening memories. Let me give you an example. I want you all to hold up your left hand, and we're numbering your fingers
from one - pinky, to four - index finger, and now I want you to press your fingers
on your thumb in this pattern: 4, 1, 3, 2, 4; 4, 1, 3, 2, 4; 4, 1, 3, 2, 4. Okay, I think you all got it. This was the task that was used in the experiment
I want to tell you about. People were asked to press
"4-1-3-2-4 sequences" like this as quickly as they could for about two hours. (Laughter) They were pretty bored, but they also stopped getting faster,
as the experimenters wanted, and when they got to that point,
they were given a test. 30 seconds. Press as many
4-1-3-2-4 sequences as you can. And this all happened around about 10 AM, and people did just fine -
they pressed about 21 sequences. And then, they were told to go away,
come back 12 hours later, and do it again. And their performance didn't change much. Now, here's the interesting bit: they went away a second time,
and they slept overnight, and they came back
and did this a third time. And now their performance
improved dramatically, about 20% improvement. Furthermore, the extent
to which they improved was predicted by the sleep spindles - remember those high-frequency
oscillations I told you about - that occurred over the bit of their brain, the motor cortex, that's particularly associated
with hand movement. So, this is the kind of data that suggested sleep is doing
something really important for memory. It's helping us to strengthen up memories. But it turns out it doesn't
just help us to strengthen memory. Sleep also helps us to integrate memories and to make connections between things that we might not otherwise
have realized were connected. And this is critical for solving
some kinds of problems. It's critical for creativity
and forms of innovation, and I bet everyone here
has experienced this: you've woken up in the morning with a solution to something that you
hadn't been able to solve the day before. So, in fact, our history is peppered
with examples of this. (Music: The Devil's
Trill Sonata - Giuseppe Tartini) So this sonata by Tartini is an example. This was inspired by a dream that he had of the Devil
playing violin on his bed. A more scientific example
comes from Friedrich Kekulé, who won the Nobel Prize for discovering the chemical
structure of benzene is cyclical rather than acyclical, more linear. And again, this was inspired by a dream in which he saw
a serpent biting its own tail. Interestingly, this type
of associative problem-solving is linked to REM sleep
rather than spindles, or slow-wave, sleep. So, what you might be wondering
is what's going on in sleep. How does sleep
allow us to do these things? And the answer seems
to be linked to the fact that memories are spontaneously
replayed during sleep. So the neural activity associated
with something that you've done replays spontaneously when you're asleep. Let me explain to you
an experiment which shows this. This is a maze which participants were asked
to navigate around in a video game. And while they were doing this,
their brain activity was measured, and unsurprisingly, the hippocampus -
amongst other areas - which is associated
with spatial memory and navigation, was active while they did this. Then, the same participants
stayed in the scanner, and they were asked to sleep. And during slow-wave sleep,
those high-amplitude, slow oscillations I told you about, the same structure was active again. And the extent to which it was active actually predicted how much better
they got at navigating around the town when they were tested again the next day. Furthermore, the sleep spindles
that I told you about are thought to be a marker
of this type of reactivation. So not only is this experiment
a nice example of how reactivation happens during sleep
and is linked to strengthening memories but also that finger-tapping experiment - this explains why the spindles
of the motor cortex were greater because, actually, that people were probably
reactivating those memories as well. So, what has this got to do
with sleep engineering? Well, what's exciting about it
is now that we know reactivation during sleep
is important for strengthening memories, we've also learned how to manipulate it. So we don't have to sit and wait
for reactivation to happen spontaneously, but instead, we can control it. Let me show you how this works. If I show you a cat (meow)
in this part of the screen and a dog (bark)
in this part of the screen, and then tonight, when you're asleep
and you enter slow-wave sleep, I'm watching, and I play (meow). Then the next day when I test you, you'll be much more likely to remember
where the cat was on the screen than the dog. And that's because that sound cue will have triggered reactivation
of the memory and strengthened it. So this doesn't end with just
simply strengthening memories. It's been shown that if we trigger
some kinds of problems while people are asleep, then they can be better
at solving those the next day as well. So reactivation
can also help with association and potentially with creativity
and innovation as well. But as we all know, we don't want
to remember everything always. So I'm sure some of you
have had a traumatic experience at some point in your life. You may have been mugged, you may have been in a car crash, maybe other things that happened
to you that you'd really - you might not mind
remembering the details, but you really don't want to be as upset
about it every time you remember it as you were right after it happened. We were interested in this and how reactivation
during sleep might help with it. It turns out that if we show people
upsetting pictures or videos and then play them the associated sounds
when they're in slow-wave sleep, (Chattering) (Gunshots) (Screaming) then ask them the next day
how upsetting that was, the things that have been replayed
to them during slow-wave sleep will be less upsetting. So it looks as though triggering
replay of memories can actually help to disassociate
the emotional response from them as well. We live in a time when we
are hyper-aware of our bodies. We're all very aware of how important
exercise is and how important diet is. We have gadgets to measure
every calorie that we take in and every calorie that we expend. We even have gadgets to measure our sleep. But I'd like to finish here by suggesting that we could be
taking this to the next level. Instead of just measuring our sleep, we could be using the information
that we now have about sleep to manipulate it in order to enhance things
like reducing our aging, improving aspects of our memory, enhancing our creativity, and also potentially controlling
aspects of our emotional responses. So, as a sleep scientist, my hope is that in not too many years, when your boss walks into your office and gives you a very difficult
problem to work on, you'll feel like the most
appropriate response to make to show her you're taking this seriously is to pull out a pillow (Laughter) and tell her you'll sleep on it. Thank you very much. (Applause)
