Identifying Autism from Neural Representations of Social Interactions

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0113879

by WhitneyLand • 12 years ago

55 38 12 years ago

Is autism hard to diagnose? It seems that if anything the important thing about this study is the factors it discovered that allowed the populations to be separable.

mikecb• 12 years ago

The purpose of this study would seem to be more to validate the utility and/or robustness of fMRI than a novel diagnosis scheme for Autism.

Toenex• 12 years ago

I've not read the paper fully but it does look an interesting piece of work. I did notice that the controls were aged and IQ matched which doesn't suggest they we looking to differentially diagnose Autism from other disorders which is likely to be the more useful clinical technique. Also, the subjects were adult and I would expect most diagnosis to be performed in a much younger age group. This isn't the first attempt to apply fMRI techniques to the understanding of autism, plenty of prior work here (http://goo.gl/oW41Vm).

Earlier in my career I did some fMRI and, whilst an interesting technique, I've never been convinced that it quite delivers on the claims. Firstly, you aren't looking at an electrical signal but rather changes in blood flow which are (reasonably) assumed to be coupled with areas of increased neural activity. This felt to me a bit like trying to debug a computer program by pointing a heat sensitive camera at your motherboard.

Also all fMRI experiments are characterised by a block design - periods of activation and inactivation - and the response is expected to correlate with these in some way. However, the brain is pretty busy and thus how does one select a baseline; what about the parts of the brain that just stay on?

duaneb• 12 years ago

Autism actually can be hard to diagnose, as there are many different flavors of symptoms. This should be read as more of a machine learning exercise in attributes (so yes, you're correct), not as a medical breakthrough.

splinterofchaos• 12 years ago

There exist very accurate tests for autism, but /getting/ a diagnosis can be difficult. A teacher spotted autistic symptoms in me back when I was ten and recommended I get tested. It wasn't until three years later that a couple thousand dollars got taken out of my college fund and I received a day's trial of tests to determine that I have it. Insurance did not cover the procedure.

Now-a-days, people have more understanding and awareness (this was a little over ten years ago), but it is still an expensive procedure, involving many specialists of different areas, sometimes neurologists and brain scans.

hoopism4• 12 years ago

There's a saying among those who become familiar with Autism.

"If you've met one kid with Autism, you've met one kid with Autism"

There are certainly clear cases... but the trouble is identifying the non-traditional and less severe cases. These methods are great, but with early intervention being so critical it's doesn't help with diagnosing younger children (from what I read... seems patient would need more cognitive ability).

As a parent who has recently run the gauntlet of diagnosis the most frustrating part of the lack of awareness for younger children. Based on all my reading early intervention can have a tremendous impact of future outcome (before 3y/o is really the sweet spot).

There's so much stigma on classifying young children early that parents are often reluctant to seek out diagnosis and treatment. My daughter was diagnosed at 15 months and has since made amazing strides.

If there was one thing I wish I could pass on to every parent it's that there is absolutely no harm or shame in doing diligence if you feel something isn't right. Therapies are play based and are great for even nuerotypical kids. If you suspect a delay or risk... talk to a specialist ASAP.

aidos• 12 years ago

This is very good advice.

My wife is an Early Years Speech and Language therapist, so she is regularly exposed to children with autism.

Your daughter is lucky to have a proactive parent. Even when cases are identified within the system, by nursery staff for example, it seems to be more common than not that parents become very defensive about the idea of seeking further help. And actually, as a parent, I can understand it - but having seen first hand the difference professional help can make, especially in those early years, I would do everything in my power to get expert advice.

In the UK we have the NHS, and even though the waiting lists can be long (because these are the areas that tend to be the first to go during budget cuts), help is freely available. And it's not just autism we're talking about. There could be all manner of things that can contribute to a delay in those early years. There are people who can literally make the world of difference to the rest of child's future, and they really want to help.

hoopism4• 12 years ago

In Massachusetts, and many other state, there is free therapy available to children under 3. That applies to any developmental delay.

I encourage any parent to make full use of these resources. Having experienced them first hand it's been nothing but positive.

jerf• 12 years ago

"In the UK we have the NHS, and even though the waiting lists can be long (because these are the areas that tend to be the first to go during budget cuts), help is freely available."

Help is also generally, if not universally (I just can't guarantee it since this is often state-level), available in the US for all manner of developmental issues. And, uh, I didn't have any problems with waiting lists, either, with either of my children.

(The portrayal of the US as some sort of vast health-care wasteland often has more roots in politics than reality, which is not to say that there aren't problems, and even in some cases severe problems, but nevertheless, the problems are often greatly oversold because it benefits certain people very much to do so, a fact that should never be forgotten in these discussions.)

The biggest problem with these resources is people knowing they exist at all, and also, as mentioned by others, denial on the part of the parents. (My wife and I have generally avoided denial but we can certainly see it in our extended families.) My wife and I know of at least two other children that we believe are probably autistic or something (not trying to diagnose, just observing an awful lot of similarities to the other autistic children we know... you get to know many of them when you go to the meetups), but the parents won't even hear the word "autism", let alone have it checked.

musername• 12 years ago

Lacking comprehension of the problem and the causes, it seems reasonably carefull not to blindly admit your child into care based solely on the authoritive argument. Possibly, doctors cannot explain it propperly either and I'm under the impression that the causes aren't really known. Now, if the parent's lack of proficiency in childcare was responsible, that would be embareassing indeed, and if the percentage of autism

spdustin• 12 years ago

I'd also pass along, for those in the US, that an "official diagnosis" qualifies your child as having a learning disability, which means that you have a unique opportunity to form an actual team of people at your child's school who will be better equipped to manage your child's particular idiosyncrasies, and who will all make whatever modifications and accommodations are necessary to ensure your child gets the best education they can.

Of course, individual schools may still have a poor array of teachers, but in my experience as a father of a nine year old boy with autism, being a proactive and considerate parent who wants to be part of the school support team goes a long way to making the most of your child's time in school.

My son had made a comment that, taken literally, seemed like a threat to another student. He was, in fact, verbalizing one of the many random thoughts we all have but most of us filter before it becomes spoken aloud. The other child's parent rallied to have my child removed from THE SCHOOL (my kid isn't violent in any way, nor does he show anger physically) but the presence of an IEP [1] protected my son, and the school district administration came to his aid. My role as a member of his IEP team (as opposed to just being an ornery parent) made sure that my son's needs were met, and his behaviors were evaluated in the context of his differently-wired brain.

Be an advocate for your child. Don't be defensive, and don't be a dick. You'll be surprised at how many people sincerely want your child to succeed.

[1] https://www.understood.org/en/school-learning/special-servic...

Edited to add IEP link

pygy_• 12 years ago

Have you heard of this study [0] on the effectiveness of Bumetanide in autistic children?

The thinking behind the trial is brillant, and the results are impressive in some cases [1].

Bumetanide treatment for autism is currently undergoing a phase III trial in France.

————

0. http://www.nature.com/tp/journal/v2/n12/full/tp2012124a.html

1. https://www.youtube.com/watch?v=7pChAJqEiyY

lucasgw• 12 years ago

100% agree. As a parent of a child with Autism, I can't stress enough the importance of strong, early intervention. We received our diagnosis at 2yrs,8mo. Our son was frankly autistic by any definition. We dove in hard, and for many years had close to 40hrs of various therapies per week. (OT, PT, Speech/Language, Behavior, Social, etc.) He had an IEP, and we advocated, because you have to.

He's 11 now, is in a neurotypical classroom with no formal support and thriving. He (and we) still work hard at his challenges... but we all have challenges. Neuroplasticity is a very real thing, and strong early intervention is sooooo important.

valdiorn• 12 years ago

Sensitivity and specificity! That title is a bit misleading :)

I too could write a test that identifies autism 97% of the time:

return Random.Get(0, 100) < 97;

nicklo• 12 years ago

Yeah I wonder what the false positive rate is. Didn't see any mention of it though.

duwease• 12 years ago

I think it's just the headline that's misleading, the abstract says quite clearly: "Moreover, machine learning algorithms classified individuals as autistic or control with 97% accuracy from their fMRI neurocognitive markers."

The "or control" doesn't tell you exactly what the false positive rate is, but at least it sets the upper bound at 3%.

arjie2• 12 years ago

What they did is take a control group that's the same size of the autism group. Then they did what's called leave-one-out cross-validation (which is exactly what it says, train on the entire set leaving one out, then use that one for validation and do so all ways). The measure they chose (Accuracy) is not without its flaws, but it would not be susceptible to the one you mention.

For instance, if you took 100 negative and 100 positive and then ran your test, you'd expect an accuracy of something a little more than 50%.

cbhl• 12 years ago

Under this methodology, how meaningful would having more data points be with respect to statistical meaningfulness? The study only appears to use 17 people in each group (for a total of 34 people).

forrestthewoods2• 12 years ago

As if autism is a "thing" that you "have". HIV is a thing that you can have. We check your blood and say "yup, you got HIV in there" or "nope, you're clean". Autism, like many terms, isn't a thing. It isn't even a state. It's a description of a state. But it's not a complete description of that state. It's really just a loose description of part of a state. Every case is different. Every autistic person is unique.

harperlee2• 12 years ago

If it is useful, I don't know a lot about autism, and for me your comment is super cryptic. What are you trying to say? Isn't autism sort of a dimension? I've heard about "the autist spectre", and the prevalence in males, and that's the idea that I got.

pavel_lishin• 12 years ago

"The Autist Spectre" sounds a little like "The Holy Spirit".

Did you mean "autism spectrum"?

VLM• 12 years ago

That would be "spectrum" as its quite smooth and analog. Unlike a medical condition like a broken bone where you either got two pieces or one and in between is kind of unlikely.

Historically no test existed other than behavior observations. Oh look a cluster of behaviors, we'll call it a disease.

That leads to some severe cultural problems like just medicating the hell out of people until their vegetate. Not a very nice thing to do.

The highest functioning people aka closest to normal (whatever normal is...) tend to not appreciate being drugged into submission, eliminating any chance of higher level human functioning. (edited to emphasis the side effects of most medications for autism symptoms tend to be intensely punishing both physically and mentally)

Its rather like obesity, there's quite a range and how much someone has and how much someone needs to be forcibly "fixed" is open to quite a bit of opinion.

OP is correct in that they selected a group of very high functioning people and controls, so what if anything that would have to do with low functioning people is mysterious and its quite possible they have little in common other than symptoms / behaviors. So saying "we can detect autism" is kind of like saying "we can fix broken computers". Well, define broken computers narrowly enough and ...

On the other hand OP may not have read the article because OP is claiming "we can't" and similar, when "we now can". At least in a highly restricted setting with specially selected test subjects.

It does not appear to be a magic bullet along the lines of a glucose tolerance test being pretty effective at diagnosing diabetes.

spain• 12 years ago

I've heard about "the autist spectre"

Now I want to see someone draw this.

baldfat2• 12 years ago

Autism is a catch all just like Cancer is a catch all. It works to get you into the ball park.

Yes you are 100% right BUT also 100% wrong (just semantics I don't know if we really disagree) there are subsets of Autism and Autism is a thing scientifically and medically. Just like depression is a thing that you can't get a specific test for doesn't mean it isn't a thing. Having adopted two boys with a very mild form of Autism called Aspergers, most doctors use to use the term High Functioning Autism. You can also have sever Autism, AKA low functioning Autism. When working with children or parenting I can specifically look at where they are. So while every person is unique you can be categorized as high functioning or low functioning autism and given many other sub-categories. The only reason why I say this is that many people dismiss early diagnosis and then their child doesn't get early intervention. The earlier the interventions start the better the outcomes.

Interesting note is Korea was saying they have almost no cases of Autism. Then the medical and scientific research conducted in Korea showed they had just as many children with Autism they just went undiagnosed medically.

drzaiusapelord• 12 years ago

Interesting note is Korea was saying they have almost no cases of Autism.

The same way Iran has no homosexuals. The stigma in those cultures are so strong, the condition is denied or re-routed into something more culturally pleasing. In Iran homosexuals are often given the choice of lengthy prison sentences or gender re-assignment surgery. In Korea, autism is swept under the rug, until fairly recently. In Japan, kids fall into "Hikikomori" or NEET categories and never seem to get the help they need.

forrestthewoods2• 12 years ago

I agree with all the things you've said.

Gonna throw a similar but slightly inverted anecdote out there as well. My sister is a PhD psychologist who works with children. She had to significantly reduce the number of children she forwards for ADHD testing because 100% of the time the result was "positive" and they were medicated. And she was sending in really borderline cases "just to be sure". Thinking she was being safe and responsible for the child's well being. Now she has to act as a blocker and only forward children who she is really, truly confident need it.

Asparagirl• 12 years ago

I would be careful reading too much into one study. If you read their Methods section, you can see that they only picked adults with "high-functioning autism" which is so loosely defined as to be meaningless. Is it supposed to represent Vineland scores, or ADOS scores, or IQ, or "doesn't live in Mom's basement", or what?

But if you see that they also cut anyone from the study that has a family history of autism (even though autism is highly heritable), cut anyone with developmental delays (and almost all people with autism have or had delays in some realm, from speech to fine motor), cut anyone with anxiety disorder (extremely common in autistic people), then really you're left with probably a bunch of almost-entirely-guys with what used to be called Asperger's and is today (well, as of DSM-5) lumped back in with autism the broader category, for better or for worse.

Point being, be careful not to think this is a representative sample of a big, big category.

All that being said, I would love to see how these results stack up against another recent project based in San Francisco at UCSF that is doing fMRI scans of kids (not adults) with sensory processing disorder (SPD) but who do NOT have autism.

The project's lead, Dr. Elsya Marco at UCSF, is studying something similar: biological evidence of SPD and how it can be distinguished on scans from autism. The two disorders often go together, but she's trying to tease apart why some people are hypersensitive to sensory stimuli and yet DON'T have autism.

Study info: http://www.ucsf.edu/news/2013/07/107316/breakthrough-study-r...

And they're crowdfunding right now to raise money to do a bigger sample size for the fMRI scans: https://crowdfund.ucsf.edu/project/53f230b014bdf718b0b2608d

(Each brain scan they do costs $550.)

Identifying Autism from Neural Representations of Social Interactions

Related Stories