Can we edit memories?

Memory is such an everyday thing
that we almost take it for granted.

We all remember what we had
for breakfast this morning,

or what we did last weekend.

It's only when memory starts to fail

that we appreciate just how amazing it is,

and how much we allow
our past experiences to define us.

But memory is not always a good thing.

As the American poet and clergyman
John Lancaster Spalding once said,

"As memory may be a paradise
from which we cannot be driven,

it may also be a hell
from which we cannot escape."

Many of us experience
chapters of our lives

that we would prefer
to never have happened.

It is estimated that
nearly 90 percent of us

will experience some sort of
traumatic event during our lifetimes.

Many of us will suffer acutely
following these events, and then recover,

maybe even become better people
because of those experiences.

But some events are so extreme

that many, up to half of those
who survive sexual violence, for example,

will go on to develop
post-traumatic stress disorder,

or PTSD.

PTSD is a debilitating
mental health condition

characterized by symptoms
such as intense fear and anxiety

and flashbacks of the traumatic event.

These symptoms have a huge impact
on a person's quality of life

and are often triggered
by particular situations

or cues in that person's environment.

The responses to those cues may have been
adaptive when they were first learned --

fear and diving for cover
in a war zone, for example --

but in PTSD,

they continue to control behavior
when it's no longer appropriate.

If a combat veteran returns home
and is diving for cover

when he or she hears a car backfiring,
or can't leave their own home

because of intense anxiety,
then the response to those cues,

those memories,

have become what we
would refer to as maladaptive.

In this way, we can think of PTSD
as being a disorder of maladaptive memory.

Now I should stop myself here,

because I'm talking about memory
as if it's a single thing.

It isn't.

There are many different types of memory,

and these depend upon different circuits
and regions within the brain.

As you can see, there are two
major distinctions in our types of memory.

There are those memories
that we're consciously aware of,

where we know we know

and that we can pass on in words.

This would include memories
for facts and events.

Because we can declare these memories,

we refer to these as declarative memories.

The other type of memory
is non-declarative.

These are memories where we often
don't have conscious access

to the content of those memories

and that we can't pass on in words.

The classic example
of a non-declarative memory

is the motor skill for riding a bike.

Now, this being Cambridge,
the odds are that you can ride a bike.

You know what you're doing on two wheels.

But if I asked you to write me
a list of instructions

that would teach me how to ride a bike,

as my four-year old son did
when we bought him a bike

for his last birthday,

you would really struggle to do that.

How should you sit on the bike
so you're balanced?

How fast do you need to pedal
so you're stable?

If a gust of wind comes at you,

which muscles should you tense
and by how much

so that you don't get blown off?

I'll be staggered if you can give
the answers to those questions.

But if you can ride a bike,
you do have the answers,

you're just not consciously aware of them.

Getting back to PTSD,

another type of non-declarative memory

is emotional memory.

Now this has a specific
meaning in psychology

and refers to our ability
to learn about cues in our environment

and their emotional
and motivational significance.

What do I mean by that?

Well, think of a cue
like a smell of baking bread,

or a more abstract cue
like a 20-pound note.

Because these cues have been pegged
with good things in the past,

we like them and we approach them.

Other cues, like the buzzing of a wasp,
elicit very negative emotions

and quite dramatic
avoidance behavior in some people.

Now, I hate wasps.

I can tell you that fact.

But what I can't give you
are the non-declarative emotional memories

for how I react
when there's a wasp nearby.

I can't give you the racing heart,

the sweaty palms,
that sense of rising panic.

I can describe them to you,
but I can't give them to you.

Now importantly,
from the perspective of PTSD,

stress has very different effects on
declarative and non-declarative memories

and the brain circuits
and regions supporting them.

Emotional memory is supported
by a small almond-shaped structure

called the amygdala and its connections.

Declarative memory, especially the what,
where and when of ?? memory,

is supported by a seahorse-shaped
region of the brain

called the hippocampus.

The extreme levels of stress
experienced during trauma

have very different effects
on these two structures.

As you can see, as you increase
a person's level of stress

from not stressful to slightly stressful,

the hippocampus,

acting to support the event memory,

increases in its activity
and works better to support

the storage of that declarative memory.

But as you increase
to moderately stressful,

intensely stressful,
and then extremely stressful,

as would be found in trauma,
the hippocampus effectively shuts down.

This means that under
the high levels of stress hormones

that are experienced during trauma,
we are not storing the details,

the specific details,
of what, where and when.

Now, while stress is doing that
to the hippocampus,

look at what it does to the amygdala,

that structure important
for the emotional, non-declarative memory.

Its activity gets stronger and stronger.

So what this leads us with in PTSD
is an overly strong emotional,

in this case fear, memory

that is not tied
to a specific time or place

because the hippocampus
is not storing what, where and when.

In this way, these cues
can control behavior

when it's no longer appropriate,
and that's how they become maladaptive.

So if we know that PTSD
is due to maladaptive memories,

can we use that knowledge
to improve treatment outcomes

for patients with PTSD?

A radical new approach being developed
to treat post-traumatic stress disorder

aims to destroy those maladaptive
emotional memories

that underly the disorder.

This approach has only
been considered a possibility

because of the profound changes

in our understanding
of memory in recent years.

Traditionally, it was thought
that making a memory

was like writing in a notebook in pen.

Once the ink had dried,
you couldn't change the information.

It was thought that all
those structural changes

that happen in the brain
to support the storage of memory

were finished with in about six hours,

and after that they were permanent.

This is known as the consolidation view.

However, more recent research suggests
that making a memory

is actually more like writing
in a word processor.

We initially make the memory
and then we save it or store it,

but under the right conditions
we can edit that memory.

This re-consolidation view suggests
that those structural changes

that happen in the brain to support memory

can be undone,

even for old memories.

Now, this editing process
isn't happening all the time.

It only happens under
very specific conditions

of memory retrieval.

So let's consider memory retrieval
as being recalling the memory

or, like, opening the file.

Quite often, we are simply
retrieving the memory.

We are opening the file as read-only.

But under the right conditions,

we can open that file in edit mode,

and then we can change the information.

In theory, we could delete
the content of that file,

and when we press save,

that is how the file,

the memory,

persists.

Not only does this re-consolidation view

allow us to account for some
of the quirks of memory,

like how we all sometimes
mis-remember the past,

it also gives us a way to destroy
those maladaptive fear memories

that underly PTSD.

All we would need would be two things:

a way of making the memory unstable,

opening that file in edit mode;

and a way to delete the information.

We've made the most progress

with working out
how to delete the information.

It was found fairly early on
that a drug widely prescribed
to control blood pressure in humans,

a beta blocker called Propranolol,

could be used to prevent
the re-consolidation

of fear memories in rats.

If Propranolol was given
while the memory was in edit mode,

rats behaved as if they were no longer
afraid of a frightening trigger cue.

It was as if they had never learned
to be afraid of that cue.

And this was with a drug
that was safe for use in humans.

Now, not longer after that,
it was shown that Propranolol

could destroy fear memories
in humans as well,

but critically it only works
if the memory is in edit mode.

Now, that study was with
healthy human volunteers,

but it's important because it shows
that the rat findings

can be extended to humans,
and ultimately to human patients.

And with humans,

you can test whether destroying
the non-declarative emotional memory

does anything to
the declarative event memory.

And this is really interesting.

Even though people
who were given Propranolol

while the memory was in edit mode

were no longer afraid
of that frightening trigger cue,

they could still describe the relationship

between the cue
and the frightening outcome.

It was as if they knew
they should be afraid,

and yet they weren't.

This suggests that Propranolol
can selectively target

the non-declarative emotional memory

but leave the declarative
event memory intact.

But critically, Propranolol can only have
any effect on the memory

if it's in edit mode.

So how do we make a memory unstable?

How do we get it into edit mode?

Well, my own lab has done
quite a lot of work on this.

We know that it depends on introducing
some but not too much new information

to be incorporated into the memory.

We know about the different chemicals
the brain uses to signal

that a memory should be updated

and the file edited.

Now our work is mostly in rats,

but other labs have found the same factors
allow memories to be edited in humans,

even maladaptive memories
like those underlying PTSD.

In fact, a number of labs
in several different countries

have begun small-scale clinical trials
of these memory-destroying treatments

for PTSD

and have found really promising results.

Now, these studies need replication
on a larger scale,

but they show the promise
of these memory-destroying treatments

for PTSD.

Maybe trauma memories do not need to be
the hell from which we cannot escape.

Now, although this memory-destroying
approach holds great promise,

that's not to say
that it's straightforward

or without controversy.

Is it ethical to destroy memories?

What about things
like eyewitness testimony?

What if you can't give someone Propranolol

because it would interfere
with other medicines that they're taking?

Well, with respect to ethics
and eyewitness testimony,

I would say the important
point to remember

is the finding from that human study.

Because Propranolol is only acting
on the non-declarative emotional memory,

it seems unlikely that it would affect
eyewitness testimony,

which is based on declarative memory.

Essentially, what these
memory-destroying treatments

are aiming to do

is to reduce the emotional memory,

not get rid of the trauma
memory altogether.

This should make the responses
of those with PTSD

more like those who have
been through trauma

and not developed PTSD

than people who have never
experienced trauma in the first place.

I think that most people would find that
more ethically acceptable

than a treatment that aimed
to create some sort of spotless mind.

What about Propranolol?

You can't give Propranolol to everyone,

and not everyone wants to take drugs
to treat mental health conditions.

Well here Tetris could be useful.

Yes, Tetris.

Working with clinical collaborators,

we've been looking at whether
behavior interventions

can also interfere with
the re-consolidation of memories.

Now, how would that work?

Well, we know that
it's basically impossible

to do two tasks at the same time

if they both depend on
the same brain region for processing.

Think trying to sing along to the radio

while you're trying to compose an email.

The processing for one
interferes with the other.

Well, it's the same when
you retrieve the memory,

especially in edit mode.

If we take a highly visual symptom
like flashbacks and PTSD

and get people to recall
the memory in edit mode,

and then get them to do
a highly engaging visual task

like playing Tetris,

it should be possible to introduce
so much interfering information

into that memory

that it essentially becomes meaningless.

That's the theory,

and it's supported by data
from healthy human volunteers.

Now, our volunteers watched
highly unpleasant films --

so think eye surgery,
road traffic safety adverts,

Scorsese's "The Big Shave."

These trauma films produce
something like flashbacks

in healthy volunteers
for about a week after viewing them.

We found that getting people
to recall those memories,

the worst moments
of those unpleasant films,

and playing Tetris at the same time,

massively reduced the frequency
of the flashbacks,

and again the memory had to be
in edit mode for that to work.

Now, my collaborators have since
taken this to clinical populations.

They've tested this in survivors
of road traffic accidents

and mothers who have had
emergency Caesarean sections,

both types of trauma
that frequently lead to PTSD,

and they found really promising
reductions in symptoms

in both of those clinical cases.

So although there is still much to learn
and procedures to optimize,

these memory-destroying treatments
hold great promise

for the treatment
of mental health disorders

like PTSD.

Maybe trauma memories do not need
to be a hell from which we cannot escape.

I believe that this approach

should allow those who want to

to turn the page
on chapters of their lives

that they would prefer
to never have experienced,

and so improve our mental health.

Thank you.

(Applause)