It's depressing that there's so much research on autism, but so much of it is this "correlates with autism" junk. Autism is not a known single-cause thing. It is more a group of observable behaviors that can be measured dimensionally, analogous to intelligence or fever or personality. There's plenty of research pointing to "the autisms" (the idea that many different things may produce autism) and also to "broad autism phenotype" (the idea that the autism spectrum continues right into the population as a whole, and diagnosable autism is just the 99th percentile marker or so of autistic traits).
It would be far more informative if autism research took dimensional measures: IQ, some rating of social behavior, some rating of unusual sensory response, some rating of medical conditions (such as immune, gastrointestinal, seizures, etc.), parent education/socioeconomics, etc. Then correlate those measures rather than just correlating "autism or not."
Would we expect "low IQ" or "high IQ" or "fever" or "introverted" to correlate with a bunch of stuff? Yes. Would it tell us much? Not really. We have to dig deeper before we'll be able to see which people a given correlation relates to, and why. Autism is the same way.
If you're trying to apply autism research to any particular individual you simply can't. Some people with autism are just very analytic or "think in pictures" or otherwise have a mind that's tuned for things other than social interaction. Some seem to be overwhelmed by sensory input in ways that are extremely disabling. Some seem to have a collection of medical issues where autism is only one of the effects. Someone with autism may be entirely disabled requiring 24-hour care or may be a billionaire genius.
Because of all these "correlate with autism" studies, we know that some unknown large percentage of young children who have enough observable autistic traits to be diagnosed will have trouble when older, but nobody can tell which nonverbal child is Einstein or Temple Grandin, and which nonverbal child will need 24-hour care forever.
So it's just ridiculous and almost useless to do research that correlates this single yes-or-no "has autism" with anything. It's time to unpack autism into component parts.
btw autism research that does unpack it tells us a ton about non-autistic people as well, since autism is likely just the extreme end of a dimension (or dimensions) we all vary along.