Most goal-directed activities require temporal integration and monitoring of action sequences (Fuster, 1993, 2002; Luria, 1966; Norman & Shallice, 1986). In general terms, monitoring is the process by which agents assess their environments and involves activities such as assessing the progress of initiated plans, finding out what time it is, and anticipating obstacles. For example, remembering to complete an action in the future involves a monitoring phase during which the individual must monitor for the appropriate cue to execute an action. However, most everyday activities involve multiple goal- directed tasks, and e.cient monitoring requires a strategy, or a scheme, for scheduling actions (i.e., when and how to monitor). In most situations, this strategy must balance the cost of monitoring against the cost of having inaccurate information about the environment, and deciding between these costs can be a complicated optimization problem (Atkin & Cohen, 1996).

Although monitoring is a necessary task for all agents, including children, insects, and robots, few studies have investigated how these agents actually behave when the goal or deadline is approaching. Yet the monitoring concept has connections to several areas, including operant conditioning (Ferster & Skinner, 1957), process control (Moray, 1986; Senders, 1983), and some areas of ethology (for a review, see Pyke, 1984). However, research on time-based prospective memory suggests certain regularities of monitoring behavior in children and adults (Ceci & Bronfenbrenner, 1985; Harris & Wilkins, 1982; for an overview, see Ma¨ntyla¨ & Carelli, 2006).

However, mechanisms underlying time monitoring and their relation to response accuracy (i.e., time-    based prospective memory performance) are less well understood.

Consistent with earlier studies (Ceci & Bronfenbrenner, 1985; Kerns, 2000), our findings suggest that school-age children also monitor strategically by increasing the rate of clock checking when the deadline is approaching. In contrast to Ceci and Bronfenbrenner’s (1985) U-shaped monitoring functions, we observed low levels of early clock checking during the firrst 5-min interval. The overall shape of a monitoring function probably is mediated by a number of task-related factors (see also Ma¨ntyla¨ & Carelli, 2006), but in the current study even the youngest children showed increasing clock checking during the first 5-min interval.

Furthermore, children and adults not only used the same general strategy for monitoring deadlines (i.e., interval reduction) but also showed comparable levels of time-based prospective memory performance. Both the 8-year-old schoolchildren and the 20-year oldundergraduates provided more than 80% of the target responses within 10 s. However, to obtain an equally high level of time- based prospective memory performance, the children relied on external time keeping more frequently than did the adult participants.

Even if one assumes that a variety of cognitive processes can be recruited to support prospective memory retrieval (McDaniel & Einstein, 2000), it is reasonable to argue that time-based tasks rely more heavily on executive control processes than do event-based tasks (Craik, 1986; Einstein & McDaniel, 1996). In most time-based tasks of prospective memory, intentions are triggered by time-    related cues that can be mediated by external factors (e.g., noticing a clock on the wall) or internal factors (e.g., time-related associations, internal clock). Compared with event-based tasks, these cues are more implicit, and self-initiated thoughts and monitoring are critical to successful performance in most time- based tasks of prospective memory.

Compared with shifting tasks, most updating tasks require maintenance of dynamic event information, which provides a temporal coherence for the observed event. The functional role of subjective sense of time is to reduce monitoring costs by initiating monitoring actions when more specific temporal information is needed closer to the deadline. In that context, even category-level temporal information (e.g., ‘‘not yet—soon—now’’) might be su.cient to minimize early clock checking and to reduce monitoring costs. This type of cognitively constructed, but functional, time might be mediated by processes related to the maintenance and updating of working memory contents. Following the notion that updating and retaining dynamic event information in working memory contributes to a sense of temporal continuity, individuals with efficient updating and inhibition functions would be able to rely on this temporal information when monitoring deadlines. In contrast, an individual with difficulties in temporary maintenance and elaboration of working memory contents may experience discontinuities in sense of time, leading to an earlier and more frequent reliance on external time keeping.