Diverting attention from ongoing temporal events shortens perceived duration. This claim was actually made more than a hundred years ago: "Tracts of time ... shorten in passing whenever we are so fully occupied with their content as not to note the actual time itself" (James, 1890, p. 626). This might be why "time flies when we are having fun"; i.e., if we are enjoying some experience, we might not be attending to the time dimension very much and relatively few clock "pulses" are accumulated. In contrast, when we are bored by a lecture or some other occurrence and time seems to drag on and on, we may be attending quite a bit to the time dimension of events, and hence seem to feel that the experience is lasting quite a long time. Impressive experimental support for these statements has been gathered over the last century (e.g., for a recent review, see Brown, 1997). Accordingly, most current models of interval timing consider some form of attentional control in prospective time estimation, when one knows in advance that a time interval must be estimated. One common assumption is that a period of time is estimated through an accumulation of temporal information from an internal source, a process that requires attention (e.g., Hicks et al., 1976; Lejeune, 1998; Meck, 1984; Thomas and Weaver, 1975; Zakay and Block, 1996).

In internal clock models, one source of attentional control is in a pacemaker-accumulator system, where some perceptual representation of duration is built. For example, in Gibbon and Church's (1984) information-processing model of interval timing, pulses continuously emitted from an internal source of temporal information, a pacemaker, are transmitted to an accumulator. Pulses are not continuously transferred; a switch, located between the pacemaker and the accumulator, is closed when an organism is estimating time, which allows pulse transfer during timing. In general, the switch provides a constant source of variability (as opposed to the Poisson variability of the pacemaker and the scalar variability of memory and response thresholds) by closing at the beginning of a duration to be estimated and opening when it ends, interrupting pulse accumulation (Gibbon et al., 1984). Activation of the switch would be under attentional control (Meck, 1984), so that attention is necessary for accumulation to proceed. If attention is diverted from timing, opening of the switch interrupts pulse accumulation. This mechanism results in temporal information still being emitted but not accumulated, which will not contribute to the estimated duration. The net effect of interrupting pulse accumulation during timing of an interval is a shortening of its perceived duration.

Interruption in the accumulation process was inferred from effects of nontemporal processing on concurrent time estimation. Using dual tasks, numerous studies showed that perceived duration shortens as the duration of nontemporal processing increases (e.g., Brown, 1985, 1997; Fortin and Couture, 2002; Fortin and Massé, 1999; Zakay et al., 1983). In time production tasks, shortening of perceived duration results in produced time intervals lengthening with increasing duration of nontemporal processing. In many of these studies, pulse accumulation was assumed to be interrupted because attention was diverted from timing to execute some concurrent task. Interruption has never been manipulated directly and independently from concurrent processing, however, so its effects could not be dissociated from effects of concurrent tasks.

On the other hand, interruption itself was systematically investigated in nonhuman animal research using the peak-interval procedure, with temporary breaks or gaps in stimuli that animals learn to time. In a recent study, we modified an experimental task where visual or memory processing was interpolated in human time production (e.g., Fortin et al., 1993) by replacing the nontemporal task with an empty break (Fortin and Massé, 2000). As in peak-interval experiments with breaks, participants had to interrupt timing during the break in order to correctly produce the target interval. This modification allowed the effects due to interruption to be independent from those due to executing some nontemporal task. The results showed effects related to break location, which seemed actually to be more specifically related to expectation of the break occurrence. The objectives of this chapter are (1) to review data on effects of break location, and their interpretation in terms of attentional time-sharing; (2) to present new data from two experiments where further predictions derived from this interpretation — predictions relating expectation, accumulation of temporal information, and uncertainty — are tested; and (3) to summarize data showing that the effect of break anticipation is general and independent of specific stimulus conditions or time estimation methods.

0 0

Post a comment