No Data
TED2016

Moran Cerf: This scientist can hack your dreams

Filmed:

What if we could peek inside our brains and see our dreams -- or even shape them? Studying memory-specific brain cells, neuroscientist (and ex-hacker) Moran Cerf found that our sleeping brains retain some of the content we encounter when we're awake and that our dreams can influence our waking actions. Where could this lead us? "Neuroscientists are now giving us a new tool to control our dreams," Cerf says, "a new canvas that flickers to life when we fall asleep."

So we spend a 12th of our life dreaming,
00:12
and most of it is forgotten.
00:18
What if we could peek inside our brain
and see our dreams,
00:21
maybe even shape them?
00:24
I am a neuroscientist, and I study
how thinking works inside the brain.
00:26
But this story starts five years ago,
00:32
when my work was about to be
published broadly,
00:34
and one slip of a tongue
made me take a big turn.
00:39
It was October 28th, and I woke up
because my phone was buzzing.
00:43
I look at the phone and I see
that I have hundreds of missed calls.
00:47
So I pick up the phone,
00:51
and on the other line
is a producer from the BBC,
00:53
and in a thick British accent,
he asks me about my work
00:55
and specifically about my team's ability
to record people's dreams
00:59
for the first time in history.
01:03
Now, just to be clear, my work had nothing
to do with recording people's dreams.
01:05
(Laughter)
01:10
Apparently, he saw a short movie
01:11
that I made to accompany the publication
that's about to come out,
01:14
and I'm explaining the results there,
01:18
but I want to end with something
that's like a TED-like uplifting message,
01:21
so I ask one of my colleagues
to explain what the future may hold.
01:25
And he says, "In the future,
01:29
we'll be able to decode
people's intentions, people's memories,
01:31
people's emotions --
and maybe even their dreams."
01:34
And the camera fades out.
01:37
So I explained to the producer
that we didn't actually do that.
01:40
(Laughter)
01:43
And he asks me,
01:44
"But is it possible
to record people's dreams?"
01:45
And I say, "In theory, it's possible."
01:48
"Thank you, bye."
01:49
(Laughter)
01:51
Five minutes later, the headline on BBC:
01:52
"Scientist claims
dream recording is possible."
01:55
(Laughter)
01:58
Within minutes, the entire world
picks up on this headline,
02:01
and everyone speaks about my team's
ability to record people's dreams.
02:05
(Laughter)
02:08
It doesn't matter how much I explained
to everyone that this was impossible.
02:10
The story now has a life of its own.
02:13
(Laughter)
02:16
So I call my dad, who is a journalist,
02:17
and I ask him,
"Dad, how can I kill a story?"
02:20
And he says, "No problem,
just turn off the phone.
02:23
No one cares about science."
02:25
So I turn off the phone --
02:27
(Laughter)
02:28
(Applause)
02:30
and people start emailing me their dreams.
02:31
(Laughter)
02:33
A famous chef asks me if I can
open his brain and extract a recipe
02:35
he's been dreaming of for a while.
02:40
(Laughter)
02:41
People ask me to look
into their spouse's dreams.
02:42
Others, who finally have a proof
that the CIA is spying on them,
02:45
want me to go and testify in Congress.
02:48
Warner Brothers calls me.
02:51
They want me to be the face
of the upcoming DVD release
02:53
of the movie "Inception,"
02:56
to show the science behind it.
02:58
(Laughter)
03:00
And a big computer company
03:02
wanted to option
the "dream-recording machine,"
03:03
because my device now has a name,
03:08
it has specifications,
03:10
it had Wikipedia pages
03:12
and fans all over the world.
03:14
(Laughter)
03:16
To make things worse --
03:18
(Laughter)
03:20
my students had the idea
to go out to a Halloween party
03:21
to forget about things,
03:24
so I decided, in state
of a depreciative sense of humor,
03:26
to dress up like Sigmund Freud.
03:29
(Laughter)
03:32
So now there is not only
just a story about me
03:33
but also a nice visual to go with it,
03:35
and I am deemed by the media now
the modern-day Freud.
03:38
This story went on for a few weeks,
03:43
until, fortunately, Prince William
proposed to his girlfriend,
03:45
which was much more important,
03:49
and I got to go back to my work.
03:50
(Applause)
03:52
Now, if you were interested
in studying dreams,
03:56
I would recommend starting first
04:00
by just looking at people's thoughts
when they are awake,
04:02
and this is what I do.
04:05
So I am indeed a neuroscientist,
04:07
but I study the brain
in a very non-traditional way,
04:09
partially inspired by my background.
04:12
Before I became a neuroscientist,
I was a computer hacker.
04:15
I used to break into banks
and government institutes
04:18
to test their security.
04:21
And I wanted to use the same
techniques that hackers use
04:23
to look inside black boxes
04:28
when I wanted to study the brain,
04:30
looking from the inside out.
04:32
Now, neuroscientists study the brain
in one of two typical methods.
04:35
Some of them look at the brain
from the outside
04:39
using imaging techniques like EEG or fMRI.
04:41
And the problem there is that the signal
is very kind of blurry, coarse.
04:45
So others look at the brain
from the inside,
04:48
where they stick electrodes
inside the brain
04:52
and listen to brain cells
speaking their own language.
04:54
This is very precise,
04:57
but this obviously can be done
only with animals.
04:59
Now, if you were to peek inside the brain
05:02
and listen to it speak,
05:06
what you would see
is that it has this electrochemical signal
05:07
that you can translate to sound,
05:10
and this sound is
the common currency of the brain.
05:12
It sounds something like this.
05:17
(Clicking)
05:18
So I wanted to use this in humans,
05:22
but who would let you do that?
05:27
Patients who undergo brain surgery.
05:31
So I partner with neurosurgeons
across the globe
05:35
who employ this unique procedure
05:38
where they open the skull of patients,
05:41
they stick electrodes in the brain
to find the source of the problem,
05:44
and finding the source
can take days or sometimes weeks,
05:47
so this gives us a unique opportunity
05:50
to eavesdrop on the brains of patients
while they are awake and behaving
05:53
and they have their skull open
with electrodes inside.
05:58
So now that we do that, we want to find
what triggers those cells active,
06:03
what makes them tick.
06:07
So what we do is we run
studies like this one.
06:09
This is Linda, one of our patients.
06:13
She is sitting here
and watching those clips.
06:15
(Video) ... can't even begin to imagine.
06:17
(Singing)
06:19
Morgan Cerf: What you hear is the sound
of one cell in her brain.
06:21
Try to see if you can figure out
what it cares about.
06:27
(Clicking)
06:31
(Video) ... I have a dream that one day --
06:32
(Clicking increases)
06:34
The Simpsons.
06:39
(Laughter)
06:40
This is one cell out of a little network
06:41
that's bound hand and fist
with the concept of The Simpsons
06:44
in Linda's brain.
06:48
Now, if you know what The Simpsons are,
06:49
then you all have
similar cells in your brain
06:51
that came to life right now
when you saw this image.
06:53
Now, those cells are amazing.
06:57
They're very precise.
06:59
They don't fire for any cartoon character
or any yellow creature.
07:01
(Laughter)
07:04
They only fire when you think
of The Simpsons.
07:05
They fire when you see
any picture of The Simpsons.
07:10
Even if you hear the sound
of The Simpsons,
07:13
or just read the text, "The Simpsons,"
these cells will fire.
07:15
In fact, they fire
even if you close your eyes
07:18
and just imagine The Simpsons
in your own head.
07:21
We can remove the world,
and the show still goes on.
07:24
And we know that because
we came back to Linda
07:27
and asked her now,
without saying anything,
07:30
to just close her eyes
and recall from her memory
07:32
the things she has seen before,
07:34
and what you see is that
as she remembers The Simpsons,
07:36
the same cell fires.
07:39
In fact, it fires seconds
before she speaks.
07:42
(Clicking)
07:45
(Video) The Simpsons.
07:47
(Laughter)
07:49
So we get to see her thoughts
coming to her.
07:51
Now, this is remarkable,
because when I speak to you right now,
07:54
I feel that everything that I have to say
comes out in real time.
07:57
But now you can tell
that we're actually a little bit behind.
08:00
My brain plans the next word.
08:03
That's the slowest I could ever speak.
08:08
(Laughter)
08:10
So once we find such cells in her brain,
Bart Simpson, mom, dad,
08:12
the Eiffel Tower, we can essentially
know what she's thinking,
08:16
and then we can do all sorts of fancy
stuff that is based on those cells.
08:19
For example, we can have her
think of her mom
08:23
and just move a spaceship
on a computer screen
08:26
with her thoughts alone.
08:28
Or, if I can see her thoughts arising,
08:30
I can actually do things
before she's conscious of it.
08:34
So I can ask her to press a button,
08:37
but beat her to it and turn
the lights on before she gets there.
08:39
(Buzz)
08:43
(Laughter)
08:45
And it doesn't matter
how smart she is or how fast she is.
08:50
I'm always going to be there earlier
because I'm inside her head.
08:53
(Laughter)
08:57
(Laughter)
09:00
(Applause)
09:03
And finally,
09:06
if we can actually show your thoughts,
09:08
we can just project them on a screen
in front of your eyes.
09:11
And this is what we did five years ago.
09:15
So naturally, when people saw
thoughts on a screen,
09:17
they started imagining the possibility
of decoding dreams in the same way.
09:20
Which we did not do.
09:24
(Laughter)
09:27
Until two years ago.
09:29
(Laughter)
09:30
When I got a call from the BBC again,
09:31
and they asked me about dream recording,
09:33
I told them, "You know,
we've been through it before."
09:35
And they said, "No, we actually
want you to comment
09:38
on the work of someone that just came out
09:40
that shows that this can be done."
09:43
So it turns out, three years
after I explained to everyone
09:45
why this was only theoretically possible,
09:49
a colleague of mine,
Professor Yukiyasu Kamitani from Japan,
09:51
was inspired by it and did it.
09:55
And that's when I decided
I want to do it myself.
09:57
So I decided I'm going to dive
into dream research in two paths.
10:00
The researcher in me said I'm going
to try to see if we can access them
10:05
and show them,
10:08
and the hacker in me said
I want to see if we can influence them
10:09
and change people's behaviors.
10:12
Looking into dreams
is really interesting, obviously,
10:15
but it's also very helpful.
10:18
And the reason is that we don't really
know what we're dreaming precisely
10:20
if we just rely on what you tell us,
10:25
because the language that you use
is unreliable for two reasons.
10:27
One, they fade away when we wake up.
10:30
And this is by design:
10:32
this is our brain's way of making sure
we don't confuse them with real memories,
10:33
so you don't think
you actually scored a touchdown
10:37
and had dinner with Aaron Sorkin.
10:40
And if you actually remember something,
10:41
we only describe them in the vocabulary
that we have when we're awake.
10:45
So blind people usually don't describe
their dreams visually,
10:49
because sight isn't in their vocabulary.
10:53
Or, if anyone here is in their 70s,
10:56
it's very likely
that when you were younger,
10:59
you described your dreams
in black and white.
11:01
Only when people started seeing
movies in color
11:03
did they think that dreams
should be in color too.
11:06
So looking into dreams
is really interesting.
11:09
It's allowing us to actually
get access to something that you forget.
11:11
Imagine giving Einstein, Shakespeare
or Picasso access to the dreams
11:14
they forget in the morning.
11:18
We were very excited about it,
11:19
and we said, let's try
to do it. But how?
11:20
And then we had the idea:
11:22
maybe the same neurons
that fire when you're awake
11:24
survive even when you're sleeping.
11:26
After all, it's the same brain.
It's us, awake or asleep.
11:28
So my colleague Florian Mormann
had the idea to go to the patient
11:31
and tell them a story that involved
the things that we find during the day.
11:34
So Bart Simpson visits Paris,
11:37
he goes to a Beatles concert
with Al Gore, and so on.
11:39
Then he asks the patient
to memorize the story and go to sleep,
11:41
and tells them, when you wake up,
I'll ask you to tell it again.
11:44
And what he shows
is that when they're sleeping,
11:47
the brain replays the same story
11:50
using the same neurons
11:54
in the same order, sequence and timing,
11:55
as if you see the story in your mind,
and this is remarkable.
11:57
Now we don't know
if the story that she sees
12:00
is the same as the one that we decode.
12:03
Maybe in her mind, Al Gore is holding
a balloon and Bart Simpson is blue.
12:05
But something about the content
is preserved, and we can now decode it.
12:09
And this is the first time
someone could show a dream
12:12
in such precise accuracy in humans.
12:15
So we were very excited.
12:18
But we said, this is still people
with open brains, electrodes inside.
12:20
Can we do it with everyone?
12:25
And here, our colleagues from Japan
show us that the answer is yes.
12:27
We can zoom out from the brain
12:30
and actually bring people
and try to look at their dreams.
12:32
Now, when you zoom out,
the resolution decreases,
12:35
so instead of seeing
Bart Simpson and Al Gore,
12:37
you will see maybe that it's a person
12:39
or an object or some character,
12:41
but something about
the content is preserved.
12:43
And maybe we can look at it.
12:45
So we brought people to the scanner
12:47
and had them fall asleep,
12:49
and fortunately, from my experience,
12:50
people fall asleep in my studies
all the time.
12:52
And while they are sleeping,
12:55
the computer tries to guess
what they're dreaming
12:56
using those high-level concepts.
12:59
Is it a person? Is it a character?
13:01
And once the computer
is getting good in making a guess
13:03
on what you're dreaming,
13:07
we wake you up
13:09
and we ask you to tell us the story.
13:10
And what we show is that the computer
is amazing in just learning over time
13:12
to guess your dream
13:17
and predict something about the gist
of the story that you saw in your mind.
13:18
So now we're pretty excited,
because we have a way to get there,
13:23
working both internally and externally.
13:26
So this gets us now to thinking
about the ability
13:30
to maybe also change things
when you are asleep.
13:33
To change things when you are asleep,
13:37
you have to know something
about the function of dreams.
13:39
And here neuroscientists are debating,
and the range of options is wide.
13:42
Some say the function of dreams
is to forget things, erase memories
13:45
or suppress thoughts,
13:48
whereas the others say that this is
our brain's way of simulating futures,
13:50
so you're kind of thinking about
dumping him and moving to Vancouver
13:54
or marrying him
and starting a family in Alabama.
13:57
You don't know the answer.
14:00
Your brain is going
to simulate both options
14:01
and give you something
to think about when you wake up.
14:03
But regardless of whichever
option you believe in,
14:06
what's common to all of them
14:09
is that during the night,
we think that this is our brain's way
14:10
of keeping ourselves up
and crafting answers to problems.
14:13
So we said that if we define
the right moment,
14:19
we can maybe change things in your dreams.
14:23
Now, this of course is quite spooky,
14:25
the idea that something
happens to me when I'm sleeping.
14:27
But we're going to show you
how we're using that.
14:30
So a colleague of mine
brought people to the lab
14:33
who are addicted smokers,
14:35
and she had them go to sleep.
14:36
And during the sleep, she sprayed
the smell of cigarettes into their nose.
14:39
And right after,
14:42
she sprayed the smell
of rotten eggs into their nose,
14:44
making their brain think
they don't like smoking.
14:46
And then she wakes them up.
14:50
They have no idea what happened,
14:51
but somehow their brain
creates this feeling.
14:53
So when they leave the lab,
for a few days after,
14:56
they feel less of an urge to smoke.
14:58
And then we said,
15:01
let's see if we can do the same thing
with people whose awake realities
15:02
haunt them when they are awake
15:06
and when they're sleeping,
15:08
like Nicholas.
15:10
He was discharged from the Army
after his service in Iraq
15:12
because he has post-traumatic stress.
15:15
His awake reality bothers him,
15:18
and he is unable to solve it.
15:20
So we got him into the lab
15:22
and had him to go sleep
for several nights.
15:23
And every night, we paired the sounds
that trigger the traumas
15:26
in his brain when he is awake,
15:32
and immediately after,
sprayed positive smells
15:33
to create some changes in his brain.
15:36
(Gunfire)
15:39
So we slowly, overnight,
15:41
used the fact that our brains
and our memories are not fixed.
15:43
They're malleable.
They change as we use them.
15:46
So we make his brain
remember something and override it,
15:49
remember, and override it in his dream.
15:51
And then when he wakes up,
he again has no idea what happened,
15:55
but suddenly he feels
less of this anxiety when he is up.
15:58
And this has been going on
for a few months now with Nicholas.
16:03
So we are pretty excited about that,
16:08
and now we're trying to see if there is
a way to help other people in the same way
16:10
by working with them
when they are awake in one way
16:14
and with the same brain
when they are asleep.
16:17
And we are thinking about
other ways to enhance experiences
16:20
using the opportunity
that your dream provides.
16:24
For example, what if we can teach you
things when you are asleep?
16:27
So you go to sleep
and you wake up knowing kung fu.
16:30
(Laughter)
16:34
Or two people sleep side by side,
and we can share dreams simultaneously.
16:35
Or just help you navigate your dream
to a specific location.
16:40
A lot of things in our life
are owed to dreams.
16:43
The periodic table. Hitchcock's movies.
16:46
What if we can help you focus your dreams
on something specific?
16:49
Or, I know there are a lot
of engineers here.
16:52
What if I can build a device
that helps guarantee
16:54
a positive sleep for everyone?
16:56
We are not there yet.
16:59
But if there is one thing
I learned from the last five years,
17:01
it is that saying that something
is impossible before you know it
17:04
is a mistake.
17:08
Because science
isn't just about collecting facts,
17:10
it's also about enabling promises,
hunches and even dreams.
17:13
Neuroscientists are now giving us
a new tool to control our dreams,
17:19
a new platform to create experiences on,
17:23
a new canvas that flickers to life
when we fall asleep.
17:26
So designers, perfumers,
musicians, filmmakers
17:29
can now use this
to create new experiences.
17:32
Because in the end,
17:36
we are now able to, for the first time,
17:38
help people change behavior
and understand ourselves better
17:41
by essentially teaching ourselves
to dream bigger.
17:46
Thank you.
17:51
(Applause)
17:53