Comparing autistic and normal children along the dimensions of reinforcement maximization, stimulus sampling, and responsiveness to extinction.

The purpose of this study was to assess autistic children's ability to maximize reinforcement, sample among available stimuli, and respond to extinction. Responding to five concurrent reinforcement schedules (CRF, FR 2, FR 4, FR 7, and FR 11) was monitored and compared to groups of mental age-matched and chronological age-matched normal children. When a given child consistently selected a given schedule of reinforcement, it was then put on extinction, and both perseveration and responding to the remaining (intact) schedules of reinforcement were monitored. The results showed, first, that while normal children selected the richest schedule of reinforcement, autistics did not maximize reinforcement, typically selecting one of the thinner schedules available. Second, the autistic children sampled less, and less efficiently, than the normal children. Third, the autistic children proved to be much less responsive to extinction, perseverating longer than the normal children before switching to a new reinforcement schedule. Fourth, whenever extinction proved to be a problem in the autistics, it was quickly achieved by a simple change in the stimulus conditions. These results are discussed as possible "keystone" deficits, which may be causally related to many of the behavioral deficits of autism and which might be important target behaviors for treatment.