Published on 12, July, 2020
This is my first post in this discussion.
It might soon be possible to discover if someone is Autistic by a brain scan.
There is information about the new brain scan on the web site and the articles were written last year in 2010. Anyone can look it up for themselves on the web.
The computer can tell small differences in the brain so tell if the person is on the Autistic Spectrum.
It would be much quicker than interviewing the person.
I do not know if a person would count as Autistic if they behaved as though they were but the brain scan did not show Autism.
People in that situation might be denied help if they did not count as Autistic.
In the 1980s I was told that the brain scans did not detect anything wrong with my brians. People who knew me were not impressed.
I have since done brain scans for research the idea being to find out by scanning many Autistic people if their brains are different from Non Autistic people.
That research might have helped them develop the new brain scan.
Do you think the new brain scan will be a good thing when it is developed?
What would you feel if the brain scan found that your were not Autistic?
David
This story is a very good illustration of the perils of science reporting. Most of the media stories would have been based on the Kings College press release which opens like this.
Scientists from the Institute of Psychiatry (IoP) at King’s College London have developed a pioneering new method of diagnosing autism in adults. For the first time, a quick brain scan that takes just 15 minutes can identify adults with autism with over 90 per cent accuracy. The method could lead to the screening for autism spectrum disorders in children in the future.
That all seems straightforward until you look at the actual paper which is open access so you can download it for free. I wrote "look at" deliberately. It is far too technical for most of us to understand. But I did learn something.
All the test subjects were male, adult, right handed, of normal intelligence, with no history of any psychiatric disorder or illness that might affect their brains (eg psychosis or epilepsy). They all had blood rests to rule out other disorders like fragile X. But even if the test eventually proves viable for women, children, people of low IQ, etc. it faces an even bigger obstacle.
The test was 90% sensitive and 80% specific. This means that 9 times out of to it will correctly identify autistic people taking the test. But it will only correctly identify non-autistic people 8 times out of 10. In the real world screening tests would be used on 1000s of people at a time. If you test 1000 people for autism you would expect to find 10 with autism. 90% sensitivity means the test would find 9 of them. But the test would also incorrectly find autism in 20% of the rest - 198 false positives.
It may turn out to be a useful research tool that adds to our understanding of autism. But it is going to be a long, long time - if ever - before this test is any use as a screen for autism.