Linguistic and Non-Linguistic Semantic Processing in Individuals with Autism Spectrum Disorders: An ERP Study.

Bassett-Gunter et al. (2017) recorded brain waves while adults with and without autism read word pairs. Some pairs matched, like 'coffee-cup.' Others broke the rule, like 'table-cup.'

The team looked at two brain peaks: an early one that shows fast, automatic meaning links, and a later one that shows conscious checking. No one had to speak; the words just flashed on a screen.

Both groups ended up with the right answer, but their brains took different roads. Wave size was the same; only the timing and scalp map changed.

Adults with autism reached meaning through a slower, back-of-brain route. They still understood; they just needed a beat longer to slot the pieces together.

Qi et al. (2025) reviewed ten eye-tracking studies and saw the same lag in kids. Together, the papers form a line: prediction is weak in autism across ages.

Kamio et al. (2007) once reported missing automatic priming in high-functioning adults. Bassett-Gunter et al. (2017) did not see a vanishing wave, only a late shift. The gap is likely method: Yoko used spoken words and short timing; L used written pairs and gave more time.

Audras-Torrent et al. (2021) pooled brain scans and found weaker, scattered semantic networks in autism. The ERP delay seen here fits that picture—less efficient wiring forces a longer second look.

When you give verbal instructions, pause after key terms and repeat in new words. That extra second lets the late-integration system finish its work. Watch for understanding cues, not speed; the answer may come on their second beat, not yours.