Brief report: a comparison of statistical learning in school-aged children with high functioning autism and typically developing peers.

Urgelles et al. (2012) watched high-functioning autistic kids and typical kids play a computer game.

The game hid patterns in sounds. Kids had to guess what came next.

No one told them the rules; the researchers wanted to see if the autistic kids could still pick up the hidden patterns.

Both groups learned the sound patterns equally well.

Scores on the game did not line up with the kids’ current language test scores.

This means the brain skill that spots patterns is not the same skill that builds spoken language.

Baixauli et al. (2016) looked at 24 studies and found that high-functioning autistic kids tell weaker stories. That seems opposite to Jessica’s null result, but the tasks differ: storytelling is open-ended and social; the game was silent and rule-based.

Sipes et al. (2014) followed similar kids and showed that language level, not diagnosis, predicts prosody gains. Both papers agree that language ability inside the autism group matters more than the autism label itself.

Lifshitz et al. (2014) found that autistic kids use less inner speech during tricky tasks. Together with Jessica’s finding, it hints that different learning routes—implicit versus self-talk—can each stay flat or strong in autism.

If a child’s language is delayed, do not assume the basic learning machine is broken. You can still use pattern-based teaching, like serial mand training or artificial grammar games, because the kids can track probabilities just fine. Save extra practice for real-language use—story retells, conversation—not for the hidden rules underneath.