Wednesday, September 28, 2011

Jabberwocky of the Day : John Elder Robison on the Male Female Autism Imbalance

As was first pointed out on Autism's Gadfly, John Elder Robison is once again talking about things that he doesn't seem to have the faintest clue about.  Last time, Mr. Robison was making completely inaccurate statements about CDD.  This time, Mr. Robison is talking about the the fact that autism is far more common in males than females and coming up with some rather far fetched ideas to explain the discrepancy.

As most of you probably know, autism is about four times more common in boys than girls.  This gender skewing was first noted all the way back in the 1940s when Kanner "invented" autism and has remained a constant property of autism even as as the number of children with autism has exploded from 1 in 10,000 in the 1940s to 1 in 100 today.

What you may not know is that a similar gender skewing is commonly seen in other childhood disorders as well, such as ADHD and intellectual disability.  The exact ratio of boys to girls changes with the disorder but in general boys are more likely to have a developmental disorder than girls.  So, in reality, there doesn't seem to be anything that different about autism in this regard.

The question is then why are boys more likely than girls to have development disabilities.  In my opinion, the general answer to that question is built into the question itself.  Obviously, there is some different between males and females that either makes males more susceptible or females less susceptible to these conditions.  Either the reason is genetic, biological, social, or (most likely) some combination of the three.

The trick has always been to identify the causes of the conditions and then to work backwards to determine why males are more affected than females.  But since, when it comes to autism (and ADHD), we don't have any real clue as to what the the cause of the condition is, we don't have any real clue as to why the difference is there.

But that doesn't stop Mr. Robison from coming up with some rather, uhm, interesting ideas about why this difference exists.  He first seems to suggest that families with children on the spectrum are more prone to having boys than girls -
If we assemble a collection of families in which there is at least one autistic child, that distribution of sons and daughters is not 50/50.  It favors the males.
But I have never run across any study that has even suggested that families who have children on the spectrum are more likely to have boys than girls.  This claim would go against the generally held (and proven) idea that the chance of having a boy or a girl is roughly 50/50.

It would be a fairly simple matter to survey a large number of families who have with multiple children where at least one child has autism and see if this were true.  But I would like to think that, if this claim were true, someone would have noticed the discrepancy and published on the subject by now.

In fact, the largest autism sibling study to date that was published just this year and included data on 664 sibling showed (roughly) the expected 50/50 breakdown between the genders.

Who knows, maybe Mr. Robison's proposed relationship has been demonstrated in other studies and I just missed it.  But if so, I think he needs to show us the data and not just make unsupported statements.

Next, Mr. Robison goes on to show an astounding ignorance of basic chance when he suggests that -
One explanation is that some parents have a son with autism and stop having children.  So the girls that might even the male/female ratio are never born.  I think that explanation may be true today, but what about the ages before modern birth control?
Frankly, I don't even understand how this could work.  It seems like Mr. Robison is suggesting that first born children are more likely to be a boy than a girl but that idea is just plain nonsense.  The chance of having either a boy or a girl in any given pregnancy is almost 50/50.  That fact is as true for the first child as it is for the last child.

Maybe he is suggesting that a first born boy is more likely to have autism than a first born girl but then the chances even out after the first child?  But that would not make any sense either.  If boys and girls both had the same chance of having autism then you would have the exact same chance of having a first born boy with autism as you would a first born girl.

Then Mr. Robison makes some rather strange claims about the "historical record" (i.e. pre 1940) of people with autism -
Critics might say that we don’t know how autism was distributed among the sexes a hundred years ago, and that’s true.  The autism diagnosis has only existed for sixty-some years.  Yet we do have strong anecdotal evidence.  Using that, some modern day people have “diagnosed” historical figures with autism based on what we know of them and their lives.  How many of those individuals are female?  Almost none.
Those “post-mortem diagnoses” are certainly subject to challenge and I’m sure some are even wrong.  That said, they can’t all be wrong and the male-female ratio in the known historical record of autism remains strikingly tilted toward the male side.
There are so many things wrong with this argument that it is hard to know where to begin.

Perhaps the worst part is the idea that a post-mortem diagnosis is evidence of anything besides someone having way too much time on their hands.  You simply cannot diagnosis someone based on little snippets of what is published about their life.

It is hard enough to diagnosis an adult with high-functioning autism who is sitting in front of you with their parents and has a complete record of their development.  When you add in a span of hundred years or so, several changes in culture, a selective report of only parts of their life, and the inability to talk directly to the person, then making an accurate diagnosis becomes next to impossible.

Then the idea that "they can't all be wrong" is just silly.  Yes, every post-mortem diagnosis can be wrong and, given the challenges in making an accurate historical diagnosis, I would be willing to bet that almost all (if not all) of them are in fact wrong.

And calling a post-mortem diagnosis "strong anecdotal evidence" does a disservice to the word "strong" as well as the word "evidence".  The idea that speculation based on extremely limited historical information is actually "anecdotal evidence" let alone "strong" is just absurd.

But, for the sake of argument, lets assume that these absurd statements are true and that the historical record does have more males than females with autism.  Does that tell us that there were more males than females in the past or does it tell us that, historically, men were much more likely to be written about than women?

Anyone with even a cursory knowledge of historical literature should know the answer to that question.

Perhaps the only intelligent thing that Mr. Robison has to say on this subject is that idea that female fetuses with autism are less likely to survive -
Geri Dawson suggested another possible explanation for the male-female imbalance.  What if girl embryos are actually more susceptible to some factor implicated in autism, but in a different way?  The factor that produces autistic baby boys might result in unsuccessful pregnancies when the fetus is female.  The result – fewer baby girls with autism are born
If this idea repeated by Mr. Robison is correct then you would expect to see a noticeably higher rate of autism in children with prenatal complications.  Having a history of miscarriages means that you have a greater chance of having complications in future pregnancies.  But the most recent study of the topic suggested that there is no such broad association.

Furthermore, in some of the known genetic causes of autism the exact opposite happens - the male child is either more effected or has a lesser chance of surviving.  A male child with Rett Syndrome is very unlikely to survive while a female can because of her double X chromosome.  And in Fragile X, a female is somewhat protected against having the symptoms of the disorder, again because of genetic differences.

So, if anything, I think this idea might be correct if it was flipped.  Rather than female fetuses not surviving, it seems more likely that being female offers some sort of protection against developing autism as well as other development disabilities.

But again, maybe there is data out there that shows a higher rate of miscarriages in mothers who children have autism.  But to me it sounds like grasping at straws.  Perhaps it would be better to have some actual concrete data to support a claim rather than just idle speculation.

I think the most disturbing part of all of this isn't the nonsense itself but rather Mr. Robison's position at Autism Speaks.  It doesn't really matter than he has some really strange ideas about autism or that he seems to misunderstand some basic science.  We all have areas where we are less than knowledgeable but still like to pretend we know what we are talking about.

No, I think the most disturbing part is that Mr. Robison serves on the scientific advisory board of Autism Speaks and helps select what science they will fund.  If a non-expert like myself can see the gaping holes in his ideas then what does that say about the projects he would suggest funding?


  1. But if so, I think he needs to show us the data and not just make unsupported statements.

    That would certainly be a first for him, of course there is a first time for everything.

    No, I think the most disturbing part is that Mr. Robison serves on the scientific advisory board of Autism Speaks and helps select what science they will fund.

    What is even more disturbing to me is that he serves on NIH advisory boards and gets to decide what research is funded at taxpayer expense. Also the fact that my sources tell me there is a good chance he will be appointed to the IACC after Obama signs the CARA into law. Of course, this is still at rumor and not definite at this point.

  2. Sorry to be not-very-intellectual, but Robison is out of his depth and talking crap.