Thread by @AutSciPerson

So,

I am going to talk about the #Spectrum10K study that is an online pre-print (assuming will be published soon)

that literally looks at the genes of autistic people.

And it's just as bad as a lot of other autism research.

Let me explain -
1/21

First of all, it's important to explain that they are correlating genetic scores (polygenic scores, which I'll get to later)

with checklists and questionnaires.

Please read this thread I wrote to understand why this is bad:

https://twitter.com/AutSciPerson/status/1422253444522520577?s=20

2/21

Now you may be wondering,

What are the questionnaires they are correlating these polygenic scores (the genetic scores) to?

Here's a list I found from the caption of Figure 3 of the paper.

NVIQ = non-verbal IQ,
VIQ = Verbal IQ,

3/21

DCDQ = motor coordination assessed by the Developmental Coordination Disorders Checklist, VABS = adaptive behaviour assessed by the Vineland Adaptive Behaviour Scales,
F1 = Insistence on sameness,
F2 = Social interaction
F3 = Sensory-motor behaviour,

4/21

F4 = Self-injurious behaviour,
F5 = Idiosyncratic repetitive speech and behaviour,
F6 = Communication skills,
RBS = Repetitive Behaviour Scale,
SCQ = Social Communication Questionnaire.

5/21

From Figure 3:
Phenotypes are: ADOS Social affect (ADOS SA) and restricted and repetitive behaviour (ADOS RRB); ADI verbal communication (ADI VC), social interaction (ADI SOC), and restricted and repetitive behaviour (ADI RRB); Repetitive Behavior Scale-Revised (RBS)

6/21

Social Communication Questionnaire (SCQ); [lists the 6 factors again] Vinelands Adaptive Behaviour Scales (VABS); Development Coordination Disorders Questionnaire (DCDQ); full-Scale IQ (FSIQ); non-verbal IQ (NVIQ); and Verbal IQ (VIQ).

7/21

Aside from the ADOS, I believe these are the 2 main questionnaires they're using from all the databases -

Repetitive Behaviour Scale,
Social Communication Questionnaire.

These questionnaires are parent-report. They're correlating parent scores to autistic kid's genes.

8/21

So what are they actually looking at?

They're looking at "polygenic scores" (PGS) - basically, they took a bunch of genetic data from autistic people, and tried to see if they shared the same genes. Autistic people with the same genes had "high autism polygenic scores."
9/21

Results -

This is what they found for those people -
"higher PGS for autism was associated with reduced count of co-occurring developmental disabilities "

The more common genes you had with other people in the autistic group, the less co-occurring conditions you had.

10/21

They did some other weird stuff because they had parent-report data for it - they looked at age of walking independently and age of first words.

They also have a lot of correlations with IQ which is a whole other host of problems.

11/21

And then it's strange to me that they think they can link behaviors to genes?

Especially including self-harm, "Idiosyncratic repetitive speech and behaviour" and "communication skills."

They're making a very large assumption that self-harm is genetic & tied with autism?
12/21

And of course most of this data relies on parent report which we know does not line up 100% with self reports in autistic people. So who knows what kind of correlations they would really find with that. Plus the questionnaires aren't written for autistic people.
13/21

They're taking a giant amount of data, using all of the questionnaires/anything they have access to to fish for correlations, and then saying that's autism.

14/21

Imagine saying "There's a haystack I need to search to find something in. I don't know what it is I'm looking for, but I'm sure I'll know it when I find it! If I had a lot of money and a much larger haystack maybe I could find it!"

The best analogy I can think of atm.
15/21

Are we sure they know what autism actually is?

I didn't see any sensory processing related surveys in there (I'm sure there's a few statements but it's not a separate questionnaire).
I don't see anything about masking..
I don't see anything about environmental factors..
16/21

And then there's the systemic issues.

Here's a fun one:
“To ensure interpretability across analyses, we retained only individuals who passed the genotypic QC, which included only individuals of European ancestries.”

So uh, they're only looking at white people.
17/21

What did Geschwind say about making sure there's "diversity" and getting black and hispanic people's DNA in June 2021?

This pre-print was posted on August 5th, 2021.

18/21

And in terms of #Spectrum10K and data sharing, well, just look at the databases used. It's international.
"The Simons Simplex Collection, the Autism Genetic Resource Exchange, the AIMS-2-TRIALS LEAP sample, and SPARK (N = 29,782 samples)."

Going to add 10k to that eh?

19/21

FUNDING:

Remember Autistica's statement about Spectrum 10K?

Look at this giant list of funding -

Wellcome Trust (SBC)
Autism Research Trust (SBC)
NIHR..in Cambridge (SBC)
NIMH
AIMS-2 (SBC)

& AIMS-2 receives funding from:
Autism Speaks
Autistica
SFARI

And more.

20/21

This research seems to have already been done. The paper is already written. It is posted online as a pre-print and presumably should either already be submitted to a journal, or will be submitted soon.

And this is BEFORE advertising the #Spectrum10K study.

Be wary.

21/21

Actual link to the study -
medrxiv.org/content/10.110…

Here are some examples I've found from the Repetitive Behavior Scale (from a different paper)

Almost all of them basically just mean "autistic child stims."

So what they may end up finding is genes correlated to masking/suppressing stims.

I also found slides that explain the domains of the RBS-R. I don't know how up to date it is. But this is what it says:
Stereotypy: repetitive movement, such as hand flapping, making sounds, head rolling or body rocking.

24/

Compulsive behavior: inntended and appears to follow rules, such as arranging objects in stacks or lines.

Sameness: resistance to change; for example, insisting that the furniture not be moved or refusing to be interrupted.

25/

Ritualistic behavior: involves an unvarying pattern of daily activities, such as an unchanging menu or a dressing ritual. This is closely associated with sameness and an independent validation has suggested combining the two factors.

26/

Restricted behavior: limited in focus, interest, or activity, such as preoccupation with a single television program, toy, or game.

Self-injury: includes movements that injure or can injure the person, such as eye poking, skin picking, hand biting, and head banging.

27/

Now I don't see anything in those domains in the RBS-R that addresses sensory sensitivities or differences. Not one of them do (only one that mentions stimuli at all is "stereotypy" i.e. stimming).

28/

It's honestly amazing that they can put these questionnaires behind paywalls.

I don't see how research is transparent when at least half of the questionnaires you literally cannot access the statements to.

29/

The Social Communication Questionnaire claims (on their website) that it "helps evaluate communication skills and social functioning in children."

I have no idea if that's the case, because I can't actually see the statements. And I couldn't find the RBS-R statements either

30/

I'm not sure if I found the right thing but if I did, it's weird.

For example, "Does she/he ever use your hand like a tool or as if it were part of his/her own body (e.g., pointing with your finger or putting your hand on a doorknob to get you to open the door)?"

31/

"Does her/his facial expressions usually seem appropriate to the particular situation, as far as you can tell?"
"Does she/he ever have any complicated movements of her/his whole body, such as spinning or repeatedly bouncing up and down?"

31/

"Does she/he ever seem to be more interested in parts of a toy or an object (e.g., spinning the wheels of a car), rather than in using the object as it was intended?"

Many of these questions seem quite outdated, and again really only focus on stimming.

31/

I found some of the general domains or items the SCQ is testing in this table -
semanticscholar.org/paper/The-Util…

Some of it includes -
Offers to share
Shows & directs attention
Seeks to share enjoyment (what??)
Social smiling (what??)
Appropriate eye gaze
Attention to voice
[cont]
32/

Range of facial expressions
Imitative social play
Pointing to express interest
Head shaking to mean "no"
Head nodding to mean "yes"
Social chat
Complex body mannerisms

So.. how is that autism again?
Or is this just a very specific profile of an unmasking autistic child?
33/

Correlations to any sort of biological or physiological measurement are only as good as the questionnaire you're using to define a "trait" or "features."

I don't see how either of these questionnaires, based on what I know, encapsulate being autistic.

It's stereotypes.
34/34

My apologies - they're not "only looking at white people" because they are people "of European ancestries" - could still be nonwhite.

But uh,
they're definitely not trying to look at nonwhite people/people of color in general, that is what I mean.

(And no that little caveat doesn't make it better)