Diagnosing autism with brain imaging

A couple of weeks ago, I expressed doubt about a commercial company alleging to diagnose autism with an MRI, based on activity in an area that lights up when one thinks about oneself.

Some actual researchers are looking at a different area of brain testing that I trust more just because of where it comes from. The Science Daily article says the brains of kids with autism spectrum disorders responded an average of 11 milliseconds (about 1/100 of a second) after hearing a sound, regardless whether the kid had a language impairment.

"This delayed response suggests that the auditory system may be slower to develop and mature in children with ASDs," said Roberts. An 11-millisecond delay is brief, but it means, for instance, that a child with ASD, on hearing the word 'elephant' is still processing the 'el' sound while other children have moved on. The delays may cascade as a conversation progresses, and the child may lag behind typically developing peers."

A 2009 study by Roberts and colleagues sheds light on how changes in brain anatomy may account for the delays in sound processing. The study team used MEG to analyze the development of white matter in the brains of 26 typically developing children and adolescents. Because white matter carries electrical signals in the brain, signaling speed improves when neurons are better protected with an insulating sheath of a membrane material called myelin.

"(This) may reflect delayed white matter development in these children."

This is really cool, if it works out, and we figured they should be able to do something like this someday fairly soon.

But it brings up the usual problem. Imagine life in a child care center when you can hook a newborn up to a computer, and put a drop of blood on a lab on a chip, and find out immediately where they fall on all sorts of continua.

Suppose you're running a CDD infant program, and somebody brings in a 6-week-old infant (or whatever your minimum age is) with autism, and you're going to have that kid in your center 250 days a year until he (usually a he) enters kindergarten. How would you treat that kid differently as an infant?

I don't know the answer to that, but I think the field needs to have a conversation about best practices when we can diagnose autism at birth and tell which ones are likely to become alcoholics, or violently aggressive, or depressed, or to have panic attacks. Once we can put a newborn's drop of blood on a chip and get back probabilities of suffering a whole list of mental issues, what do we do about it?

This sounds like something somebody should hold a conference around.