The dynamic effect of context on interval timing in children and adults.

Human reproductions of time intervals are often biased toward previously perceived durations, resulting in a central tendency effect. The aim of the current study was to compare this effect of temporal context on time reproductions within children and adults. Children aged from 5 to 7 years, as well as adults, performed a ready-set-go reproduction task with a short and a long duration distribution. A central tendency effect was observed both in children and adults, with no age-difference in the effect of global context on temporal performance. However, the analysis of the effect of local context (trial-by-trial) indicated that younger children relied more on the duration (objective duration) presented in the most recent trial than adults. In addition, statistical analyses of the influence on temporal performance of recently reproduced durations by subjects (subjective duration) revealed that temporal reproductions in adults were influenced by performance drifts, i.e., their evaluation of their temporal error, while children simply relied on the value of reproduced durations on the recent trials. We argue that the central tendency effect was larger in young children due to their noisier internal representation of durations: A noisy system led participants to base their estimation on experienced duration rather than on the evaluation of their judgment.