If Autism is a Result of Neurological Differences, then...

Post Reply New Topic

Why can't brain scans be used to diagniose autism?

Just Curious...

I mean there has to technology out there, if they are reporting that there are neuorological differences.

We all have a little neuro difference and there are many variations not just autistic.

There really hasn't been any scanning technology that could trace the connections in the brain. An MRI or a CAT scan is great if you have a tumor, burst blood vessel, bullet, or other non-microscopic thing going on in your brain. For all the decades (centuries?) of talk about "brain wiring" we really have had no way to actually see that wiring.

But there's a new technique that's being worked on called "MRI diffusion tensor imaging," which can literally map the wiring (white-matter connections) in the brain. On "60 Minutes" last week they showed a short piece about a doctor who is working on MRI/DTI (it's still new and there are still bugs/problems). He showed a scan of a normal person's brain -- specifically, some isolated part that is involved with language. Then he showed a scan of the same spot that he took of Temple Grandin's brain. It was dramatically different. TG's brain had many more connections to many more areas.

I gather the people developing MRI/DTI are still trying to make sure that they data they get is actually true/correct, so it's still in development. And even after they perfect it it will probably take years for researchers to figure out the patterns and what they mean.

Selected tracts of a normal brain (I can't find any of an autistic brain):

There has been some interesting results using fMRI, but it's a relatively new technology and still being tested--just google "fMRI autism" to see reports of the studies.

What about mild cases? Will they be distinguishable? We know that NT - non-NT border is fluid. Temple is severe case who developed advanced coping techniques.

this youtube vid shows fMRI's of temple grandin's brain-
[youtube]http://www.youtube.com/watch?v=aT0zCzCp6yY[/youtube]

Autistic brains are different from typical in almost every way. The parts are different sizes and shapes (subtly, and on average; you have to measure a lot of different brains to see the effect). They're different microscopically; different sorts of connections, more local ones, less long-distance ones.

The reason we can't use this as a diagnostic tool is that there are so very many things that can change the way your brain looks both on the large and small scales. It could be autism; it could be variation within the realm of typical; it could be some other disorder. Autism is associated with a different brain--but autistic brains don't look enough like each other that you can tell an autistic brain when you scan it. You can say, "Well, this brain is quite different from the typical; so this person is more likely than typical to have autism," but maybe that's only a matter of a 5% chance of having autism instead of a 1% chance. For all you know, you're looking at ADHD, or fetal alcohol syndrome, or a typical kid with an unusually shaped brain. Autistic people have a lot of variation in their brains. For that matter, NTs have a lot of variation, too. With all the possible configurations for both NT brains and autistic brains and the brains of people who have other disorders, things get pretty uncertain.

Brain scans are expensive, and right now, they're unreliable for detecting autism. A behavioral inventory--surveying parents and teachers and the child (if old enough)--is just a much more reliable way to do a diagnosis.

You'd think that an objective test like a brain scan would be more trustworthy for detecting autism; but oddly enough, in this case it's not true. The best way to detect autism is for a trained observer with experience with autism in the appropriate age group to make a diagnosis using behavioral observations from various environments and from various points in the person's history.

Very interesting and helpful responses, so thank you for everyone who contributed to this thread!