Atypical categorical perception in autism: autonomy of discrimination?

Soulières et al. (2007) asked adults with high-functioning autism to sort sounds and then tell apart very similar sounds.

The task looked simple: decide if two beeps were the same or different. The trick was that some pairs crossed a learned category line and some did not.

Neurotypical adults usually get a free boost when sounds sit in different categories. The team checked whether the autistic group got the same boost.

The autistic adults sorted the sounds into categories just fine.

When they had to judge tiny differences, they never got the usual category boost. Their ears worked, but top-down category knowledge did not speed them up.

In short, they heard the world literally, without the helpful lens that categories provide.

Irwin et al. (2022) saw the same missing boost in kids with ASD, but in audiovisual speech. Children ignored the speaker’s mouth and relied only on raw sound. Together the two studies show the gap starts early and holds into adulthood.

Clayborne et al. (2024) looks like a contradiction at first: they taught preschoolers with ASD to form categories quickly using equivalence training. The difference is method. Isabelle tested natural perception; Clayborne taught an explicit skill. The studies sit side-by-side: perception stays flat unless we directly train the category.

Kargas et al. (2015) also found poorer auditory discrimination in autistic adults, linking the degree of deficit to IQ and repetitive behaviors. The pattern is consistent: basic discrimination is weaker and the weakness ripples into broader functioning.

Expect clients to miss the helpful shortcuts that categories give. When you say “dog” and “puppy” mean the same thing, they may not feel it.

Teach categories overtly with equivalence or other explicit drills; natural exposure is not enough.

Keep auditory and visual instructions clear and redundant so small perceptual slips do not snowball into errors.