Reward value of prosodic features of language for autistic, mentally retarded, and normal children.

Researchers played a story read with lively ups and downs of voice.

They let autistic, mentally retarded, and typical kids work for more of the story.

The team watched who kept pressing the button to hear the prosody again.

Typical kids worked hard to keep the lively voice playing.

Autistic and mentally retarded kids gave up quickly; the voice was weak reinforcement.

Only the mentally retarded group later copied the sing-song style in their own speech.

Lui et al. (2026) pooled 23 studies and also found autistic children notice prosody less, but the gap shrinks when tests give fewer answer choices.

Henson et al. (1979) showed autistic kids can imitate musical pitches as well as peers, so the ear works; the reward value, not the skill, is missing.

Goldberg et al. (2016) used a similar button-press test and saw high-functioning autistic children value social play just like typical kids—social sounds can work if they are strong enough.

Together the papers say: autistic learners hear prosody fine, yet they will not work for it—pick stronger reinforcers when teaching language.

When you run mand or echoic trials, skip the bouncy teacher voice as the main prize. Pair new words with edibles, toys, or music the child already seeks. Save prosody cues for prompting, not payoff.