Performance by Children with ASD, Developmental Delay, and Typical Development On Delayed Non-Matched to Sample Task

Background: Previous literature suggests that children with ASD experience difficulty on the Delayed Non-Matched to Sample task (DNMS). In the DNMS task, children are initially presented with one stimulus covering a well that contains a reward. After a time delay, they are presented with the initial stimulus paired with a novel stimulus. During this second presentation, the reward is placed under the novel stimulus. The child must choose the novel stimulus in order to retrieve the reward. Dawson and colleagues (2001) asserted that understanding the relationship between stimuli and reward, not impairment in visual memory, compromises the performance of children with ASD on the DNMS task.

Objectives: (1) To compare performance of preschoolers with ASD, age-matched children with non-ASD developmental delays (DD), and mental-age matched typically developing (TD) children on the DNMS task. (2) To explore moderators of performance.

Methods: Participants with ASD and DD were 36 to 54 months of age; the two mental-age matched TD groups were 11 to 13 months and 26 to 28 months of age. Group membership was verified via the ADOS, ADI and Mullen Scales of Early Learning. The Communication and Social Behavior Scales Developmental Profile (CSBSDP) was used to screen for red flags for ASD in the 11- to 13-month-old group since the ADOS is not valid for children of this age. Each child received 5 pre-trial presentations of the DNMS tasks with no reward. Participants were then administered standard DNMS trials with rewards and increasing time delays based on performance. Success was defined along a two-tiered continuum: Criterion one: eight out of ten or 12 out of 15 correct trials at a 5-second delay. Criterion two: four out of five correct trials at a 30-second delay.

Results: Preliminary data suggest that the inclusion of pre-trials within the DNMS task is unifying performance on the task across all groups. Additionally, children with ASD who meet the qualifications for criterion one are more likely to also meet the qualifications for criterion two than TD controls and children with DD. Results of a chi-square analysis revealed that children with ASD were more successful at reaching criterion two, given that they had met at criterion one than were TD controls and children with DD (² (1, N = 55) = 11.67, p<0.001, ÿ=0.46).

Conclusions: Preliminary data support Dawson’s previous findings that children with ASD who meet criterion one quickly will also meet criterion two, thus confirming that visual memory does not pose a barrier in achieving success on the DNMS task. The preliminary data presented above appears to also corraborate with Dawson’s finding that pre-trials with no rewards increases the performance of children with ASD’s success, albeit marginally (from 50% to 60%).