15-minute brainscan to detect autism

Post Reply New Topic

This is the link:
http://www.mrc.ac.uk/Achievementsimpact/Profiles/ChristineEcker/index.htm

In 2010 a brainscan has been developed at the King's College in London by Dr. Christine Ecker and team to detect autism in 15 minutes with an accuracy of 90%.
I also watched an interview with Dr. Ecker in German as she is German explaining that it would be a good diagnostic tool to detect autism in adults and milder forms of it, as they have special software which can identify brainstructures typical for autistic people.

Does anyone knows more about this or how this reaserch has developped?

I read that when you are living in the UK you can get a referral from your GP for a testing.
I guess it would be much more accepted for getting support than a behaviour-based diagnosis.

Would you do this scan if it was available?

Oh, where did you watch that interview? Can you give a link?



I'm not sure I would. I'd like to for curiosity's sake but my previous experiences with MRIs proved that I cannot keep still long enough and end up causing heavy motion artefacts in the scans.

http://www.youtube.com/watch?v=zlaeNxn4sAE&feature=related

Go to 20.30min.
I would like to do a scan like this, because maybe it could give me more insight in which areas are the most "affected".
I wonder, if it could explain the "severity" in different parts.

I'd be very worried about something with a 90% accuracy rate, which seems pretty low to me. It would miss some autistic people, which is a worry, and give false positives in others, which is a big worry. I don't think you could use this kind of technology on its own without a much better accuracy rate; and if you could only use it in conjunction with conventional diagnostic assessments, it'd be a waste of time and money. You might as well stick to the conventional tests.

I'm not against technology at all (in fact the work is interesting, as it may well advance our knowledge about autism), but with a condition this complex it seems to me that the techniques are nowhere near advanced enough to make a reliable diagnosis – and it's possible they never will be. I don't think I'd bother having my brain scanned just yet, anyway.

I don't know about the percentage of accuracy of conventional diagnostic assessment, but using it as a conjunction for milder forms of autism or adults, which have "learned to adapt" throughout their lifetime or women, which more often get misdiagnosed than men it might be a good conjunction.

I have a thing for pictures of brrrainzzz, so sign me up, scan!

Does anyone know if this would work for those of us with other brain conditions?

No, as far as I understood it, they based it on the brainstructure of autistic people and developed the software on the "abnormalities" of autistic people's brainstructure compared to non-autistic people.
Then they tested it on austistic and non-autistic people.

What about claustrophobic people? I know some people couldn't stand having a brain scan because of all the confining machinery.... Im not sure, but I think it would probably push me towards panic attack, even if the long term benefits of the scan greatly outweighed the negatives of having it done.

In general though, it sounds like a great method of diagnosing people, if they can get the accuracy of results way up to something with very few errors (missing autistic people is going to confuse them when they think they've finally found the answer to their life....)

I'd do it in a heartbeat, if only to contribute data to something that valuable. I enjoy enclosed spaces anyway. As a kid, I would sit in my closet with the light on and read.

I would wait on it. I'm certainly not going to be a guinea pig for that. Let's refine our science a little, shall we, before we start using it for official diagnosis.

The fact that it is quick and easy is another scare for me. NOTHING good comes quick and easy.

Lastly, I have a retainer in my mouth and probably can't do an MRI of my head anyway. I like my teeth attached to my skull. Though, I do love brain pictures, and am trying to search my CAT scans myself to notice anything different.

Yes, yes, I'm weird.

How are we going to refine the science without stickin' a bunch of people in there and finding the problems?

Also, while it might not be good as a lone diagnostic tool yet, it could be a great diagnostic helper. It could turn getting a diagnosis for many into a quick screening or two and a brain scan rather than multiple sessions of intense screening and multiple layers of interviews.

A lot of people get claustrophobic like that. Depending on how bad it is, doctors often administer a sedative/anti-anxiety medication like Valium to counter it.

Re. a 90% accurate brain scan: Sounds wonderful, but I'd also like to know how many NTs are misidentified as autistics. The false-positive rate can be pretty problematic.

If by 90% accurate they mean that 90% of the time the test will correctly identify an NT as neurotypical... well, that could be a problem. Think of it this way: Test a thousand people. Ten are autistic. The test accurately picks out nine of the ten autistics... but it also picks out 99 of the neurotypicals as autistic, by mistake. So if you got an "autistic" result on the test, you actually only have a one-in-twelve chance of being autistic...

Now, I'm hoping that the false-positive rate is better than that; but you're not going to be able to use this one test as a way to detect autism. It could be used to increase the evidence in the favor of an autism diagnosis, but it's not going to be perfect. No test ever is.

Yes, the false-positive rate can be problematic.
But I also do wonder about conventional assessment, what the accuracy is. Do there any numbers exist?
I also do wonder that if a brainscan could increase the evidence in the favour for an autism diagnosis if it would benefit especially adults diagnosed later in life to easier get access to any kind of acknowledgement and support.

Here is another link, which goes more into detail and discussing the false-positive:

http://www.hospitalmanagement.net/features/feature112624/

To tell what the validity of a test is, you have to compare it to something else known to properly measure the thing you're trying to measure. So the 90% figure they got for this test was probably "pre-existing autism diagnosis as confirmed by X", with X being probably professional opinions and/or one of the major PDD rating scales.

You could also look at whether you can depend on the test to come out the same way every time. That means the test-retest and interrater reliability of autism diagnosis; that tends to be quite good for all ASDs as a group ("Are you on the spectrum?"), but unreliable for specific ASDs ("Asperger's? Autism? PDD-NOS?). So that means if you are diagnosed with, say, Asperger's, by Doctor 1 at Time 1, then you are extremely likely to be diagnosed with something on the spectrum either by Doctor 2 or at Time 2 later on. You're not guaranteed to get another Asperger's label, though. Telling apart specific ASDs, even when experts are involved, seems to be not too much more accurate than if they used coin-flips. Even the speech delay isn't a dead giveaway; there's still the PDD-NOS/Autism distinction and the fact that almost all speech delays resolve into partial or full access to verbal ability.

Some places have open mri machines. I've seen ads for them on tv.

I think having a mri would make me very claustrophobic.