Proactive Preparation, Monitoring, and Time: Children Don't Prepare for the Future

Children can be notoriously constrained to the present, but a fascinating article in JEP:HPP by Vallesi & Shallice shows exactly how strong that constraint can be: in a study with 4-11 year-olds, they show that only children older than about 5 years will take advantage of additional time provided for them to prepare for a simple task.

Among adults, this finding is part of the literature on "foreperiod phenomena." Classically, this refers to the finding that reaction times are disproportionately faster with longer intervals between a warning stimulus and an "imperative" stimulus (e.g., the onset of the word "Go!" in a simple reaction time task), but also to sequential effects, in which longer intervals on preceding trials produce a general downward shift in reaction times on the subsequent trial.

Vallesi & Shallice review the traditional theoretical interpretations of these findings, which are that classical foreperiod effects reflect the use of interval probabilities by the cognitive system to increase its preparation for an imperative stimulus as that stimulus becomes increasingly likely. Sequential effects arise because the most recent trial has a priviledged position in the calculation of this probability, such that participants expect the next trial to contain a similar interval, thus reaching "peak preparedness" earlier following those trials which had short intervals; a cycle of "repreparation-maintenance" maintains this preparedness across time, even if the current trial has a relatively long interval.

Recent work by Vallesi showed that transcranial magnetic disruption of the right dorsolateral prefrontal cortex eliminates the foreperiod effect among adults; therefore Vallesi & Shallice sought to compare these effects with those among children, whose prefrontal cortices are naturally underdeveloped.

Using a 60 trials of a simple reaction time task and foreperiods of 1, 3 or 5s, the first study showed that only the 4-5 year-olds failed to show foreperiod effects: their reaction times showed no sensitivity to the amount of time provided for them to prepare on the current trial. In contrast, the duration of the previous trial did have an effect: like adults, they were faster when following a trial that had a shorter interval. (A second experiment using within-subject longitudinal comparison was necessary for the 3-way interaction to reach significance). A third experiment replicated these results with different intervals (1, 2 and 3s), more trials (120 vs. 60), and a focus on the 4-, 5- and 6-year-olds, showing that the foreperiod effect develops gradually across these age groups.

The authors suggest these results refute single-process models of the two components to the foreperiod effect, and argue instead for a model in which previous trial duration has a gross influence on motivation or arousal for the second trial amogn 4-year-olds, and that a more explicit repreparation-maintenance cycle speeds responding with increased preparation intervals only among older children. The authors suggest these effects may be sensitive to the precise distributions or lengths of foreperiod intervals used in a study, if the child's cognitive system for preparation is merely slower than others.

To my mind it's hard to understand how children could require more than 5 seconds to prepare merely to press a button, and indeed previous work shows RTs reach a minimum at around 4 seconds, relative to 1, 2, 8 and 16. The alternative story is more plausible in terms of foreperiod distributions; previous work showing a larger foreperiod effect among 5-7 year olds used an exponential distribution of intervals, such that a wider range of intervals might have allowed children to show more preparation as a function of interval probability, given their possibly coarser representations of time, and their possibly more impoverished ability to calculate trial type probabilities.

Alternatively, and compatible with both Vallesi's TMS work as well as Donald Stuss's work on right frontal patients, it's possible that right frontal regions implement "checking" or "monitoring" for the occurrence of the stimulus as a function of its conditional probability given the elapsed time. According to this view, 4-year-olds may show no preparatory foreperiod effect because this circuit is not functioning - or because it is insufficiently precise to monitor conditional probabilities (due to poor representations of time).