Speculation is tasty and rampant, but not real knowledge

July 24, 2007

Several summers ago I spent six weeks in another country, doing research in collaboration with autism researchers there.  We traveled around by subway, train, and cab, visiting multiple schools with autistic children.  Some were regular schools that had an autism unit; others were boarding schools for special-needs children or for children with autism specifically.  We worked with the higher-functioning kids with autism (who had at least some language) and with younger mental-age-matched kids without autism (because mental retardation is very common in autistic kids – up to 74% although I read something non-peer-reviewed on the Internet that claims as low as 50% – to compare autistic and non-autistic kids on cognitive stuff you need to equate them on mental age, not chronological age).

We did a variety of tasks with the kids, involving stuff like sorting cards, or hearing stories and answering questions, or pushing buttons in response to pictures of faces on a computer screen.  One of the weirder tasks, which was something for our collaborators, not for me and my prof, involved measuring kids’ heads with a measuring tape.  I thought it seemed kind of an odd approach at the time – most of what I was involved in was heavily cognitive.  And so I’ve been surprised to see that line of research actually panning out.

Here’s a recent instance of it (citation below).  Researchers found that, very clearly, children with autism had larger head circumferences.  What I thought was interesting was that head circumference above the 75th percentile was  “associated with more impaired adaptive behaviors and with less impairment in IQ measures and motor and verbal language development.”

So something strongly associated with autism was also associated with less mental retardation (but mental retardation is strongly associated with autism).  There’s some interesting stuff in there…  I’m tempted to hypothesize beyond the data, but I’m not going to.

Larger head circumference was also associated with allergy/immune disorders in the kids and their first-degree relatives.  The researchers speculate that this relates to immune dysfunctions that themselves cause or are associated with increased cell cycle progression (cannot dredge what that means up out of my brain) and/or decreased apoptosis (deliberate, clean cell death).

At the time of my first couple years in grad school, it was being speculated that for the quarter or so of autistic kids who have a noticeable period of skill loss (coincidentally around the time of the MMR shot, which has led to a whole lot of bad crap) lose their skills around the time that the brain undergoes a major reorganization in which unnecessarily neuronal connections are “pruned” or cut back.

So: Maybe immune dysfunction leads to lack of cell death in pruning and maybe outside of it too, and thus autism and larger head circumference?

That and a dollar will get you one-quarter of a latte, or almost a small iced tea, or a tiny little bit of the funding necessary to gain actual knowledge, or a fair bit of annoyance from someone more educated who already knows why you’re wrong.   Speculation is tasty and rampant, but actual knowledge comes from eliminating the untrue theories with evidence, and hammering on the true ones until we decide there’s not much point in continuing to do so.

Sacco, R., et al. (2007) Clinical, Morphological, and Biochemical Correlates of Head Circumference in Autism. Biol Psychiatry.

Culture and autism rates: Race, class, and various weirdness

July 17, 2007

Autism Diva has extensive hypotheses about variation in autism rates within Southern California.  This is a good example of how you can start to explain cultural variation: you say could it be this, could it be this, could it be this, and then you move on to asking how to test each of those hypotheses.

new visualization of comorbidity shows bipolar-schizophrenia-autism link, and much more

July 11, 2007

Columbia University researchers have created a visualization of disease comorbidity using health records from 1.5 million people (article and full-text paper). There’s a lot they’ll be able to do with this – look for genetic links, look to see if some conditions protect from other conditions, look for potential environmental triggers like bacterial or viral infections.

This sounds awesome in general – people can process complex information about associations much faster when represented visually than when they see a bunch of numeric correlations. (Way to go, information usability!)
They find that bipolar, autism, and schizophrenia are associated (quotes from the paper):

We estimate that {approx}20–60% of autism-predisposing variations also predispose the bearer to bipolar disorder, and 20–75% of autism-predisposing variations also predispose the bearer to schizophrenia. It is therefore extremely likely that there is a three-way positive correlation among autism, bipolar disorder, and schizophrenia, a correlation that probably arises from a genetic variation that predisposes to all three disorders.

If so that’s extremely interesting, although I wonder how much of a link between autism and schizophrenia is due to the two being mistaken for each other (or perhaps the overlap in predisposing genes is why the two are mistaken for each other – our diagnostic categories are attempting to clearly delineate fuzzy categories).

Also, everything under the sun is associated with autism, apparently (I wonder whether there’s a causal relationship, and if so which way, or whether it’s third variables, or all of the above):

[A]utism, which typically manifests before the affected child is 3 years old, has a strong positive correlation with a number of neurological disorders, some of which have a late-age onset…: attention deficit, epilepsy, cerebral palsy, depression, schizophrenia, bipolar disorder, neurofibromatosis, Parkinson’s disease, and migraine. Our estimated significant overlap between autism and tuberculosis may indicate that both diseases are associated with genetic changes weakening the immune system.

They also mention associations between allergies/autoimmune disorders and autism, schizophrenia, and bipolar disorder. And here’s something totally new: female breast cancer is negatively associated with schizophrenia and bipolar disorder. They proposed an explanation involving schizophrenia and bipolar being associated with increased probability of abnormal cell death in some tissues, and breast cancer being associated with an increased probability of abnormal cell proliferation. And they mention that tamoxifen (a breast cancer treatment drug) can help treat bipolar disorder – I’ll try to follow up on that in a future post.

And the credits go to: Andrew Rzhetsky, David Wajngurt, Naeun Park, and Tian Zheng of the University of Columbia. And any unnamed undergraduate or graduate assistants.

(For anyone who’s requested an entry on a specific topic: I haven’t forgotten you, I’ve just been too busy to do background research because of moving this weekend, and have been doing stuff I could sit down and type out instead. Entries on schizoaffective disorder, kindling, worries about personality change on meds, and lots of stuff on culture coming up, among other things.)

autism/vaccines: go read Spiked

July 5, 2007

The online magazine Spiked has an issue just out on MMR and autism, with a lot of articles which are both detailed and reasonable.

Social skills programs for children with autism largely ineffective (meta-analysis)

June 26, 2007

A meta-analysis of 55 published research studies found that social skills programs for autistic kids were poor overall, though better if the programs were held in the normal classroom setting rather than a pull-out one. They were more likely to use and to maintain the new skills, and more likely to use them in other settings. The researchers recommend greater amounts of programming, targeting programs towards kids’ skill deficits, and checking to see whether the program was actually carried out as designed.

No commentary on this one – just general surprise that they were found so solidly poor.

more on science and autism/vaccines

June 16, 2007

The LA Times has a good article on the history of autism/vaccine research.

They raise an interesting point at the end, which is that vaccines do not “cause autism” but we can’t prove that they never cause autism in anyone ever (and don’t currently have the knowledge of etiology to say how likely it is or isn’t that they never cause autism in anyone ever). So scientists continue to look to see whether there might be a tiny, highly susceptible population.

I’m sincerely hoping that what happens is that 1) they do find such a population and 2) it is identifiable, so that 1) anti-vaccine believers might focus on a tiny effect that really happens rather than a large effect that does not, and 2) vulnerable people will be identified and prohibited from taking the vaccine and non-vulnerable people will be identified and strongly encouraged to take the vaccine and ideally will. And then all autistic and non-autistic children are that much less likely to get deadly childhood diseases.

U.S. Court of Federal Claims is an autism expert

June 12, 2007

Okay, not really. But apparently it’s being asked to decide that thimerosol causes autism.

Who cares about science (and extensive epidemiological studies finding no link)? Let the courts decide…

On a (thematically, but not morally) related link, my dad came home from a chemistry conference with a paper someone presented on a possible biomarker for autism (summary: lower levels of stuff in pee):

Scroll down to “Stercobilin: A Possible Biomarker for Autism?”

Dad says that stercobilin has [mumbo jumbo isomerish chemical-esque] to do with mercury or thimerosol or something. It makes sense to believe the extensive epidemiological studies over a hypothesized connection without further evidence, though.