Social affiliation motives modulate spontaneous learning in Williams syndrome but not in autism.

The study compared children with autism and Williams syndrome without a typically developing control group matched for mental age, so it cannot say how much the imitation patterns are specific to these diagnoses. The analysed samples varied by experiment: Experiment 1 excluded three Williams syndrome children for unusable video; Experiment 2 excluded five autism and two Williams syndrome children for calibration or equipment failure; Experiment 3 excluded two autism children for equipment failure; Experiment 4 excluded six Williams syndrome and four autism children because they did not engage with the toys. Three children with Williams syndrome were tested at home, which may have affected behaviour, although the authors report that removing these three participants did not change the results. The order of conditions was not fully randomised, and the authors say this may have affected performance. The tasks were kept short and simple to avoid fatigue, so they may not capture learning in everyday therapy. Visual-motor complexity was not manipulated, so motor or visual demands may have influenced results. The Williams syndrome group was slightly older and had lower scores on some measures, and the authors say this may not have influenced the findings. The autism group did not show the playful-live imitation boost, but did show more imitation after playful video, so the effect of social engagement depends on task format.

Do not conclude that a more animated live model will improve imitation in autism. In the live task, children with autism did not imitate more after the playful model than after the neutral model, and the study measured lab imitation tasks, not a tested therapy.