Multi-dimensional metrics of autism?

Post Reply New Topic

I'm curious if anyone has encountered any good, accessible multi-dimensional metrics for autism, in contrast to scalar metrics like AQ and EIQ and so forth. ... I guess I'm looking for something that sets aside the "how autistic are you" question in favor of "how are you autistic".

Ideally I'd like to find something like the Political Compass, but probably with a lot more than two axes. So for example, the axes might be things like fascination with patterns, difficulty reading emotional cues, sensitivity to bright lights and loud noises, and so forth...

Some reasons why I'm interested:

- It might allow us to get a handle on the ways that some of us cluster together ... from my readings so far, it doesn't seem like even experts in the field have much of a handle on trait distribution beyond "some autistic people exhibit some characteristics more strongly than others" -- but the distribution of traits must be at least somewhat clumpy, and it seems likely that this clumping may contain important clues about the specific strategies that can help us in life, as well as the underlying similarities and differences in our brain configurations. (And on that last note, here's a study that plots autistic neurotypes in 5-space: http://www.jneurosci.org/content/30/32/10612.full).

- Taking the angle between two such vectors would allow a (rough) scalar measure of how similar two people's personal experiences of autism are, which is something that it's not always easy to get a handle on intuitively.

- I just really like n-dimensional vectors. Can't get enough of 'em.

Anyway, I haven't found anything with a casual search, but I may be using the wrong searchterms... Would be interested to know if anybody has run across anything like this.

Well, a lot of autism scales have subscales. For example, the AQ has subscales of social skill, attention switching, attention to detail, communication and imagination, according to this article.

Thanks! That's actually kind of perfect for what I'm looking for. The subsets of course haven't been validated in the same way as the instrument as a whole -- you wouldn't want to use them for any kind of formal research purpose -- but it seems like this should do just fine for a rough heuristic. Now I just need to set up some little form or app or whatever to generate vectors from the AQ results...