Susceptibility to Optical Illusions Varies as a Function of the Autism-Spectrum Quotient but not in Ways Predicted by Local-Global Biases.

van Timmeren et al. (2016) gave 13 optical illusions to college students who had never been diagnosed with autism. Each student filled out the Autism-Spectrum Quotient, a self-report that measures autistic-like traits in the general population.

The team asked a simple question: do people with higher autistic traits see fewer illusions? If they do, it would support the popular idea that autism always means “local-first” vision.

Only two illusions became harder to see as autistic traits rose. The other eleven looked the same no matter the score. In plain words, most illusions ignored the trait level.

The result breaks the tidy story that “more autistic traits equals less global vision.”

Guy et al. (2019) looked at children and teens who actually have an ASD diagnosis. They found strong local-to-global interference that stayed stable with age. That finding seems opposite to A et al.’s null result, but the difference is the group: diagnosed kids versus neurotypical adults. Traits in typical adults do not guarantee the same perceptual pattern seen in diagnosed autism.

Fahmie et al. (2013) used the same trait-split method with a visual search task. High-trait adults were faster at finding targets, a positive result. Taken together, the 2013 and 2016 papers show that trait-linked advantages appear in search yet barely show up in illusions. Task choice matters.

Huang et al. (2025) pooled 15 brain-imaging studies. Autistic participants relied on visual cortex for both local and global tasks, while controls used parietal areas. The imaging data suggest a brain-level difference that simple paper-and-pencil illusions may not tap.

Do not assume every client with ASD will “see the trees but miss the forest.” Check the task first. Visual search drills, hidden-picture games, or figure-ground worksheets may reveal different patterns than illusion cards. When you pick visual materials, test them with the individual instead of trusting a one-size-fits-all “local bias” rule.