"Why?" is a question

that parents ask me all the time.

"Why did my child develop autism?"

As a pediatrician, as a geneticist, as a researcher,

we try and address that question.

But autism is not a single condition.

It's actually a spectrum of disorders,

a spectrum that ranges, for instance,

from Justin, a 13-year-old boy

who's not verbal, who can't speak,

who communicates by using an iPad

to touch pictures to communicate

his thoughts and his concerns,

a little boy who, when he gets upset,

will start rocking,

and eventually, when he's disturbed enough,

will bang his head to the point

that he can actually cut it open and require stitches.

That same diagnosis of autism, though,

also applies to Gabriel,

another 13-year-old boy

who has quite a different set of challenges.

He's actually quite remarkably gifted in mathematics.

He can multiple three numbers by three numbers

in his head with ease,

yet when it comes to trying to have a conversation,

he has great difficulty.

He doesn't make eye contact.

He has difficulty starting a conversation,

feels awkward,

and when he gets nervous,

he actually shuts down.

Yet both of these boys

have the same diagnosis of autism spectrum disorder.

One of the things that concerns us

is whether or not there really is

an epidemic of autism.

These days, one in 88 children

will be diagnosed with autism,

and the question is,

why does this graph look this way?

Has that number been increasing

dramatically over time?

Or is it because we have now started labeling

individuals with autism,

simply giving them a diagnosis

when they were still present there before

yet simply didn't have that label?

And in fact, in the late 1980s, the early 1990s,

legislation was passed

that actually provided individuals with autism

with resources, with access to educational materials

that would help them.

With that increased awareness, more parents,

more pediatricians, more educators

learned to recognize the features of autism.

As a result of that, more individuals were diagnosed

and got access to the resources they needed.

In addition, we've changed our definition over time,

so in fact we've widened the definition of autism,

and that accounts for some of

the increased prevalence that we see.

The next question everyone wonders is,

what caused autism?

And a common misconception

is that vaccines cause autism.

But let me be very clear:

Vaccines do not cause autism.

(Applause)

In fact, the original research study

that suggested that was the case

was completely fraudulent.

It was actually retracted from the journal Lancet,

in which it was published,

and that author, a physician,

had his medical license taken away from him.

(Applause)

The Institute of Medicine,

The Centers for Disease Control,

have repeatedly investigated this

and there is no credible evidence

that vaccines cause autism.

Furthermore,

one of the ingredients in vaccines,

something called thimerosal,

was thought to be what the cause of autism was.

That was actually removed from vaccines

in the year 1992,

and you can see that it really did not have an effect

in what happened with the prevalence of autism.

So again, there is no evidence

that this is the answer.

So the question remains, what does cause autism?

In fact, there's probably not one single answer.

Just as autism is a spectrum,

there's a spectrum of etiologies,

a spectrum of causes.

Based on epidemiological data,

we know that one of the causes,

or one of the associations, I should say,

is advanced paternal age,

that is, increasing age of the father

at the time of conception.

In addition, another vulnerable

and critical period in terms of development

is when the mother is pregnant.

During that period, while the fetal brain is developing,

we know that exposure to certain agents

can actually increase the risk of autism.

In particular, there's a medication, valproic acid,

which mothers with epilepsy sometimes take,

we know can increase that risk of autism.

In addition, there can be some infectious agents

that can also cause autism.

And one of the things I'm going to spend

a lot of time focusing on

are the genes that can cause autism.

I'm focusing on this not because genes

are the only cause of autism,

but it's a cause of autism

that we can readily define

and be able to better understand the biology

and understand better how the brain works

so that we can come up with strategies

to be able to intervene.

One of the genetic factors that we don't understand,

however, is the difference that we see

in terms of males and females.

Males are affected four to one compared to females

with autism,

and we really don't understand what that cause is.

One of the ways that we can understand

that genetics is a factor

is by looking at something called

the concordance rate.

In other words, if one sibling has autism,

what's the probability

that another sibling in that family will have autism?

And we can look in particular

at three types of siblings:

identical twins,

twins that actually share 100 percent

of their genetic information

and shared the same intrauterine environment,

versus fraternal twins,

twins that actually share 50 percent

of their genetic information,

versus regular siblings,

brother-sister, sister-sister,

also sharing 50 percent of their genetic information,

yet not sharing the same intrauterine environment.

And when you look at those concordance ratios,

one of the striking things that you will see

is that in identical twins,

that concordance rate is 77 percent.

Remarkably, though,

it's not 100 percent.

It is not that genes account for all of the risk for autism,

but yet they account for a lot of that risk,

because when you look at fraternal twins,

that concordance rate is only 31 percent.

On the other hand, there is a difference

between those fraternal twins and the siblings,

suggesting that there are common exposures

for those fraternal twins

that may not be shared as commonly

with siblings alone.

So this provides some of the data

that autism is genetic.

Well, how genetic is it?

When we compare it to other conditions

that we're familiar with,

things like cancer, heart disease, diabetes,

in fact, genetics plays a much larger role in autism

than it does in any of these other conditions.

But with this, that doesn't tell us what the genes are.

It doesn't even tell us in any one child,

is it one gene

or potentially a combination of genes?

And so in fact, in some individuals with autism,

it is genetic!

That is, that it is one single,

powerful, deterministic gene

that causes the autism.

However, in other individuals,

it's genetic, that is,

that it's actually a combination of genes

in part with the developmental process

that ultimately determines that risk for autism.

We don't know in any one person, necessarily,

which of those two answers it is

until we start digging deeper.

So the question becomes,

how can we start to identify

what exactly those genes are.

And let me pose something

that might not be intuitive.

In certain individuals,

they can have autism

for a reason that is genetic

but yet not because of autism running in the family.

And the reason is because in certain individuals,

they can actually have genetic changes or mutations

that are not passed down from the mother

or from the father,

but actually start brand new in them,

mutations that are present

in the egg or the sperm

at the time of conception

but have not been passed down

generation through generation within the family.

And we can actually use that strategy

to now understand and to identify

those genes causing autism in those individuals.

So in fact, at the Simons Foundation,

we took 2,600 individuals

that had no family history of autism,

and we took that child and their mother and father

and used them to try and understand

what were those genes

causing autism in those cases?

To do that, we actually had to comprehensively

be able to look at all that genetic information

and determine what those differences were

between the mother, the father and the child.

In doing so, I apologize,

I'm going to use an outdated analogy

of encyclopedias rather than Wikipedia,

but I'm going to do so to try and help make the point

that as we did this inventory,

we needed to be able to look at

massive amounts of information.

Our genetic information is organized

into a set of 46 volumes,

and when we did that, we had to be able to account

for each of those 46 volumes,

because in some cases with autism,

there's actually a single volume that's missing.

We had to get more granular than that, though,

and so we had to start opening those books,

and in some cases, the genetic change

was more subtle.

It might have been a single paragraph that was missing,

or yet, even more subtle than that,

a single letter,

one out of three billion letters

that was changed, that was altered,

yet had profound effects

in terms of how the brain functions

and affects behavior.

In doing this within these families,

we were able to account for approximately

25 percent of the individuals

and determine that there was a single

powerful genetic factor