TED2014

Wendy Chung: Autism — what we know (and what we don’t know yet)

Filmed:

In this factual talk, geneticist Wendy Chung shares what we know about autism spectrum disorder — for example, that autism has multiple, perhaps interlocking, causes. Looking beyond the worry and concern that can surround a diagnosis, Chung and her team look at what we’ve learned through studies, treatments and careful listening.

- Geneticist
At the Simons Foundation, Wendy Chung is working to characterize behavior, brain structure and function in people with genetic variations that may relate to autism. Full bio

"Why?"
00:12
"Why?" is a question
00:14
that parents ask me all the time.
00:15
"Why did my child develop autism?"
00:18
As a pediatrician, as a geneticist, as a researcher,
00:20
we try and address that question.
00:25
But autism is not a single condition.
00:27
It's actually a spectrum of disorders,
00:29
a spectrum that ranges, for instance,
00:32
from Justin, a 13-year-old boy
00:35
who's not verbal, who can't speak,
00:37
who communicates by using an iPad
00:40
to touch pictures to communicate
00:42
his thoughts and his concerns,
00:44
a little boy who, when he gets upset,
00:46
will start rocking,
00:49
and eventually, when he's disturbed enough,
00:51
will bang his head to the point
00:52
that he can actually cut it open and require stitches.
00:54
That same diagnosis of autism, though,
00:58
also applies to Gabriel,
01:00
another 13-year-old boy
01:02
who has quite a different set of challenges.
01:04
He's actually quite remarkably gifted in mathematics.
01:07
He can multiple three numbers by three numbers
01:10
in his head with ease,
01:12
yet when it comes to trying to have a conversation,
01:14
he has great difficulty.
01:17
He doesn't make eye contact.
01:20
He has difficulty starting a conversation,
01:22
feels awkward,
01:24
and when he gets nervous,
01:25
he actually shuts down.
01:27
Yet both of these boys
01:29
have the same diagnosis of
autism spectrum disorder.
01:30
One of the things that concerns us
01:35
is whether or not there really is
01:37
an epidemic of autism.
01:39
These days, one in 88 children
01:41
will be diagnosed with autism,
01:43
and the question is,
01:45
why does this graph look this way?
01:47
Has that number been increasing
01:49
dramatically over time?
01:51
Or is it because we have now started labeling
01:53
individuals with autism,
01:57
simply giving them a diagnosis
01:59
when they were still present there before
02:01
yet simply didn't have that label?
02:03
And in fact, in the late 1980s, the early 1990s,
02:06
legislation was passed
02:10
that actually provided individuals with autism
02:11
with resources, with access to educational materials
02:14
that would help them.
02:18
With that increased awareness, more parents,
02:19
more pediatricians, more educators
02:23
learned to recognize the features of autism.
02:26
As a result of that, more individuals were diagnosed
02:29
and got access to the resources they needed.
02:33
In addition, we've changed our definition over time,
02:36
so in fact we've widened the definition of autism,
02:40
and that accounts for some of
02:43
the increased prevalence that we see.
02:44
The next question everyone wonders is,
02:47
what caused autism?
02:49
And a common misconception
02:51
is that vaccines cause autism.
02:54
But let me be very clear:
02:57
Vaccines do not cause autism.
03:00
(Applause)
03:04
In fact, the original research study
03:11
that suggested that was the case
03:13
was completely fraudulent.
03:15
It was actually retracted from the journal Lancet,
03:18
in which it was published,
03:21
and that author, a physician,
03:23
had his medical license taken away from him.
03:25
(Applause)
03:28
The Institute of Medicine,
03:32
The Centers for Disease Control,
03:34
have repeatedly investigated this
03:35
and there is no credible evidence
03:37
that vaccines cause autism.
03:41
Furthermore,
03:44
one of the ingredients in vaccines,
03:45
something called thimerosal,
03:48
was thought to be what the cause of autism was.
03:50
That was actually removed from vaccines
03:53
in the year 1992,
03:56
and you can see that it really did not have an effect
03:57
in what happened with the prevalence of autism.
04:00
So again, there is no evidence
04:03
that this is the answer.
04:05
So the question remains, what does cause autism?
04:07
In fact, there's probably not one single answer.
04:11
Just as autism is a spectrum,
04:14
there's a spectrum of etiologies,
04:15
a spectrum of causes.
04:18
Based on epidemiological data,
04:19
we know that one of the causes,
04:21
or one of the associations, I should say,
04:23
is advanced paternal age,
04:25
that is, increasing age of the father
04:28
at the time of conception.
04:29
In addition, another vulnerable
04:32
and critical period in terms of development
04:34
is when the mother is pregnant.
04:36
During that period, while
the fetal brain is developing,
04:38
we know that exposure to certain agents
04:41
can actually increase the risk of autism.
04:43
In particular, there's a medication, valproic acid,
04:46
which mothers with epilepsy sometimes take,
04:49
we know can increase that risk of autism.
04:52
In addition, there can be some infectious agents
04:56
that can also cause autism.
04:58
And one of the things I'm going to spend
05:01
a lot of time focusing on
05:02
are the genes that can cause autism.
05:04
I'm focusing on this not because genes
05:07
are the only cause of autism,
05:09
but it's a cause of autism
05:12
that we can readily define
05:13
and be able to better understand the biology
05:15
and understand better how the brain works
05:18
so that we can come up with strategies
05:20
to be able to intervene.
05:22
One of the genetic factors that we don't understand,
05:24
however, is the difference that we see
05:27
in terms of males and females.
05:30
Males are affected four to one compared to females
05:32
with autism,
05:35
and we really don't understand what that cause is.
05:36
One of the ways that we can understand
05:40
that genetics is a factor
05:42
is by looking at something called
05:44
the concordance rate.
05:45
In other words, if one sibling has autism,
05:47
what's the probability
05:50
that another sibling in that family will have autism?
05:52
And we can look in particular
05:55
at three types of siblings:
05:57
identical twins,
05:59
twins that actually share 100 percent
06:01
of their genetic information
06:03
and shared the same intrauterine environment,
06:04
versus fraternal twins,
06:08
twins that actually share 50 percent
06:10
of their genetic information,
06:12
versus regular siblings,
06:14
brother-sister, sister-sister,
06:15
also sharing 50 percent of their genetic information,
06:17
yet not sharing the same intrauterine environment.
06:20
And when you look at those concordance ratios,
06:23
one of the striking things that you will see
06:25
is that in identical twins,
06:27
that concordance rate is 77 percent.
06:29
Remarkably, though,
06:33
it's not 100 percent.
06:34
It is not that genes account
for all of the risk for autism,
06:36
but yet they account for a lot of that risk,
06:40
because when you look at fraternal twins,
06:42
that concordance rate is only 31 percent.
06:44
On the other hand, there is a difference
06:48
between those fraternal twins and the siblings,
06:49
suggesting that there are common exposures
06:52
for those fraternal twins
06:55
that may not be shared as commonly
06:56
with siblings alone.
06:59
So this provides some of the data
07:01
that autism is genetic.
07:02
Well, how genetic is it?
07:04
When we compare it to other conditions
07:06
that we're familiar with,
07:08
things like cancer, heart disease, diabetes,
07:10
in fact, genetics plays a much larger role in autism
07:13
than it does in any of these other conditions.
07:17
But with this, that doesn't
tell us what the genes are.
07:20
It doesn't even tell us in any one child,
07:23
is it one gene
07:26
or potentially a combination of genes?
07:27
And so in fact, in some individuals with autism,
07:30
it is genetic!
07:33
That is, that it is one single,
07:36
powerful, deterministic gene
07:38
that causes the autism.
07:40
However, in other individuals,
07:42
it's genetic, that is,
07:44
that it's actually a combination of genes
07:46
in part with the developmental process
07:49
that ultimately determines that risk for autism.
07:51
We don't know in any one person, necessarily,
07:55
which of those two answers it is
07:58
until we start digging deeper.
08:00
So the question becomes,
08:02
how can we start to identify
08:03
what exactly those genes are.
08:05
And let me pose something
08:07
that might not be intuitive.
08:08
In certain individuals,
08:11
they can have autism
08:13
for a reason that is genetic
08:15
but yet not because of autism running in the family.
08:17
And the reason is because in certain individuals,
08:21
they can actually have genetic changes or mutations
08:23
that are not passed down from the mother
08:26
or from the father,
08:29
but actually start brand new in them,
08:31
mutations that are present
08:33
in the egg or the sperm
08:35
at the time of conception
08:36
but have not been passed down
08:38
generation through generation within the family.
08:40
And we can actually use that strategy
08:43
to now understand and to identify
08:45
those genes causing autism in those individuals.
08:47
So in fact, at the Simons Foundation,
08:50
we took 2,600 individuals
08:52
that had no family history of autism,
08:55
and we took that child and their mother and father
08:58
and used them to try and understand
09:01
what were those genes
09:03
causing autism in those cases?
09:05
To do that, we actually had to comprehensively
09:07
be able to look at all that genetic information
09:10
and determine what those differences were
09:13
between the mother, the father and the child.
09:15
In doing so, I apologize,
09:19
I'm going to use an outdated analogy
09:21
of encyclopedias rather than Wikipedia,
09:23
but I'm going to do so to try and help make the point
09:26
that as we did this inventory,
09:29
we needed to be able to look at
09:31
massive amounts of information.
09:32
Our genetic information is organized
09:35
into a set of 46 volumes,
09:37
and when we did that, we had to be able to account
09:40
for each of those 46 volumes,
09:42
because in some cases with autism,
09:44
there's actually a single volume that's missing.
09:46
We had to get more granular than that, though,
09:48
and so we had to start opening those books,
09:50
and in some cases, the genetic change
09:53
was more subtle.
09:55
It might have been a single
paragraph that was missing,
09:56
or yet, even more subtle than that,
09:59
a single letter,
10:02
one out of three billion letters
10:04
that was changed, that was altered,
10:07
yet had profound effects
10:09
in terms of how the brain functions
10:11
and affects behavior.
10:13
In doing this within these families,
10:14
we were able to account for approximately
10:17
25 percent of the individuals
10:19
and determine that there was a single
10:21
powerful genetic factor
10:24
that caused autism within those families.
10:26
On the other hand, there's 75 percent
10:29
that we still haven't figured out.
10:32
As we did this, though,
10:34
it was really quite humbling,
10:36
because we realized that there was not simply
10:38
one gene for autism.
10:40
In fact, the current estimates are
10:42
that there are 200 to 400 different genes
10:44
that can cause autism.
10:46
And that explains, in part,
10:48
why we see such a broad spectrum
10:50
in terms of its effects.
10:52
Although there are that many genes,
10:54
there is some method to the madness.
10:56
It's not simply random
10:58
200, 400 different genes,
11:00
but in fact they fit together.
11:03
They fit together in a pathway.
11:04
They fit together in a network
11:06
that's starting to make sense now
11:08
in terms of how the brain functions.
11:10
We're starting to have a bottom-up approach
11:12
where we're identifying those genes,
11:14
those proteins, those molecules,
11:16
understanding how they interact together
11:19
to make that neuron work,
11:20
understanding how those neurons interact together
11:22
to make circuits work,
11:25
and understand how those circuits work
11:26
to now control behavior,
11:28
and understand that both in individuals with autism
11:30
as well as individuals who have normal cognition.
11:32
But early diagnosis is a key for us.
11:36
Being able to make that diagnosis
11:39
of someone who's susceptible
11:41
at a time in a window
11:42
where we have the ability to transform,
11:44
to be able to impact
11:47
that growing, developing brain is critical.
11:48
And so folks like Ami Klin have developed methods
11:51
to be able to take infants, small babies,
11:55
and be able to use biomarkers,
11:57
in this case eye contact and eye tracking,
12:00
to identify an infant at risk.
12:03
This particular infant, you can see,
12:05
making very good eye contact with this woman
12:08
as she's singing "Itsy, Bitsy Spider,"
12:10
in fact is not going to develop autism.
12:12
This baby we know is going to be in the clear.
12:15
On the other hand, this other baby
12:18
is going to go on to develop autism.
12:20
In this particular child, you can see,
12:23
it's not making good eye contact.
12:25
Instead of the eyes focusing in
12:27
and having that social connection,
12:29
looking at the mouth, looking at the nose,
12:31
looking off in another direction,
12:34
but not again socially connecting,
12:36
and being able to do this on a very large scale,
12:39
screen infants, screen children for autism,
12:41
through something very robust, very reliable,
12:44
is going to be very helpful to us in terms of being
12:47
able to intervene at an early stage
12:50
when we can have the greatest impact.
12:52
How are we going to intervene?
12:56
It's probably going to be a combination of factors.
12:58
In part, in some individuals,
13:01
we're going to try and use medications.
13:03
And so in fact, identifying the genes for autism
13:05
is important for us
13:07
to identify drug targets,
13:09
to identify things that we might be able to impact
13:11
and can be certain that that's really
13:13
what we need to do in autism.
13:15
But that's not going to be the only answer.
13:17
Beyond just drugs, we're going
to use educational strategies.
13:19
Individuals with autism,
13:23
some of them are wired a little bit differently.
13:24
They learn in a different way.
13:27
They absorb their surroundings in a different way,
13:28
and we need to be able to educate them
13:31
in a way that serves them best.
13:33
Beyond that, there are a lot of individuals
13:36
in this room who have great ideas
13:38
in terms of new technologies we can use,
13:40
everything from devices we can use to train the brain
13:43
to be able to make it more efficient
13:46
and to compensate for areas in which
13:47
it has a little bit of trouble,
13:49
to even things like Google Glass.
13:51
You could imagine, for instance, Gabriel,
13:53
with his social awkwardness,
13:55
might be able to wear Google Glass
13:57
with an earpiece in his ear,
13:58
and have a coach be able to help him,
13:59
be able to help think about conversations,
14:01
conversation-starters,
14:04
being able to even perhaps one day
14:05
invite a girl out on a date.
14:08
All of these new technologies
14:10
just offer tremendous opportunities
14:12
for us to be able to impact
14:14
the individuals with autism,
14:16
but yet we have a long way to go.
14:19
As much as we know,
14:21
there is so much more that we don't know,
14:23
and so I invite all of you
14:25
to be able to help us think about
14:28
how to do this better,
14:30
to use as a community our collective wisdom
14:31
to be able to make a difference,
14:34
and in particular,
14:36
for the individuals in families with autism,
14:37
I invite you to join the interactive autism network,
14:40
to be part of the solution to this,
14:43
because it's going to take really a lot of us
14:45
to think about what's important,
14:48
what's going to be a meaningful difference.
14:51
As we think about something
14:52
that's potentially a solution,
14:54
how well does it work?
14:56
Is it something that's really
going to make a difference
14:57
in your lives, as an individual,
14:59
as a family with autism?
15:02
We're going to need individuals of all ages,
15:03
from the young to the old,
15:06
and with all different shapes and sizes
15:08
of the autism spectrum disorder
15:10
to make sure that we can have an impact.
15:12
So I invite all of you to join the mission
15:14
and to help to be able to make the lives
15:16
of individuals with autism
15:19
so much better and so much richer.
15:21
Thank you.
15:23
(Applause)
15:25

▲Back to top

About the Speaker:

Wendy Chung - Geneticist
At the Simons Foundation, Wendy Chung is working to characterize behavior, brain structure and function in people with genetic variations that may relate to autism.

Why you should listen

Wendy Chung is the director of clinical research at the Simons Foundation Autism Research Initiative, which does both basic and applied science to serve people affected by autism spectrum disorders. She's the principal investigator of the foundation's Simons Variation in Individuals Project, which characterizes behavior and brain structure and function in participants with genetic copy number variants such as those at 16p11.2, which are believed to play a role in spectrum disorders.
 
Chung also directs the clinical genetics program at Columbia University. In assessing and treating kids with autism spectrum disorders and intellectual disabilities, she uses advanced genomic diagnostics to explore the genetic basis of neurological conditions. She thinks deeply about the ethical and emotional questions around genetic medicine and genetic testing.