The Effect Of Sensory Modality In Children

There is ample evidence that auditory stimuli are judged longer than equivalent visual stimuli, and visual stimuli shorter than auditory ones. For example, by varying contextual factors (e.g., intensity, color of lights, level of practice) as well as the timing procedures (e.g., production, reproduction, verbal estimation, pair comparison), Goldstone and Lhamon (1972, 1974) observed that subjects systematically judged the auditory stimuli longer than the visual ones, when they shared a common duration. The robustness of this phenomenon led Goldstone and Lhamon (1972, p. 626) to conclude that the auditory-visual difference is a "fundamental property of human temporal processing" (see Penney, this volume).

Recently, in the context of scalar timing theory, Penney et al. (1998, 2000) have succeeded in explaining the effect of sensory modality in time judgments by using a variant of the temporal bisection procedure in which auditory and visual stimuli of the same physical durations were presented during the same sessions. The result was that psychophysical functions produced by human adults were shifted toward the left for auditory stimulus durations, compared to visual ones, showing that the former are judged longer than the latter. Furthermore, this leftward shift of the temporal bisection functions was proportional to the stimulus duration value to be estimated. That is to say, the auditory-visual difference in the bisection functions increased as the duration values increased. This proportional effect with the auditory-visual duration values supports the assumption that differences in clock speed were the main cause of the modality effect on temporal bisection judgments. Notably, the internal clock runs faster for auditory than for visual stimuli. Thus, for the same objective duration, more pulses are accumulated for auditory signals than for visual signals, and the subjective time seems longer. Using other interval timing tasks (e.g., temporal generalization, verbal time estimation), Wearden et al. (1998) obtained similar results, consistent with this clock speed interpretation.

Within a developmental perspective, we might assume that if the pacemaker-accumulator system is functional at an early age, the modality difference observed in human adults in temporal bisection would also be found in children, whatever the age groups tested. Therefore, the following experiment was designed to test the modality effect on temporal bisection performance in children aged 5 and 8 years, as well as in adults (Droit-Volet et al., submitted). The subjects were given two sessions of temporal bisection, one per day. Each session was composed of pretraining, training, and testing. In the first session, the subjects were presented the stimulus in the same sensory modality for both the pretraining, training, and testing phases, either visual or auditory (visu/visu and audi/audi). In the second session, they were presented stimuli in the same modality as in session 1 for the pretraining and training phases, but in another modality for the testing phase (visu/audi and audi/visu). Therefore, session 1 required a within-modal comparison and session 2 a cross-modal comparison. In session 2, the subjects were thus required to make temporal judgments relating an auditory comparison stimulus to a visual standard for one group (visu/visu-visu/audi group), and a visual comparison stimulus to an auditory standard for the other group (audi/audi-audi/visu group).

The temporal bisection functions obtained for the 5-year-olds, 8-year-olds, and adults for both session 1 (visu/visu or audi/audi) and session 2 (visu/audi or audi/visu) are shown in Figure 7.7. Consistent with our previous results, the psy-chophysical functions were orderly in all age groups, although flatter in the 5-year-olds than in the 8-year-olds and the adults. In addition, in each age group, the temporal bisection functions shifted to the left for the auditory stimuli, compared to the visual stimuli; hence, bisection points were lower. Thus, the subjects judged the auditory stimuli as being longer than the visual stimuli. This was observed for each transfer modality order, both from auditory toward visual (audi/audi-audi/visu) and from visual toward auditory (visu/visu-visu/audi), despite a less marked auditory-visual difference in the last condition for the two oldest age groups (for more details, see Droit-Volet et al., submitted). Both the statistical analyses and the modeling of our data attributed this modality effect to differences in the speed of the internal clock, with the clock running faster for the auditory than for the visual stimuli. The developmental version of the scalar timing model presented above, and adapted for this experiment, found similar differences in the clock rate for the adults and the children aged 8 years. In these two age groups, the visual clock rate was on

Visu/visu-visu/audi group

Audi/audi-audi/visu group

5-year-olds

Visu/visu-visu/audi group

200 300 400 500 600 700 800 Stimulus duration (ms)

200 300 400 500 600 700 800 Stimulus duration (ms)

6 0.8 2 0.7 s 0.6 J2 0.5 t3 0.4 ö 0.3 •2 0.2 Is 0.1 & 0

- Audi/Audi

- Audi/Visu

200 300 400 500 600 700 800 Stimulus duration (ms)

-year-olds 0.8

Adults

200 300 400 500 600 700 800 Stimulus duration (ms)

200 300 400 500 600 700 800 Stimulus duration (ms)

FIGURE 7.7 Psychophysical functions obtained from the 5- and 8-year-olds and the adults in a bisection task in the visu/visu-visu/audi group and the audi/audi-audi/visu group for the within-modal comparison sessions (visu/visu, audi/audi) and the cross-modal comparison sessions (visu/audi, audi/visu). (From Droit-Volet, S., Tourret, S., and Wearden, J., submitted.)

200 300 400 500 600 700 800 Stimulus duration (ms)

Auditory

200 300 400 500 600 700 800 Stimulus duration (ms)

Auditory

200 300 400 500 600 700 800 Stimulus duration (ms)

Audi/Audi Audi/Visu

200 300 400 500 600 700 800 Stimulus duration (ms)

200 300 400 500 600 700 800 Stimulus duration (ms)

Audi/Audi Audi/Visu

200 300 400 500 600 700 800 Stimulus duration (ms)

FIGURE 7.7 Psychophysical functions obtained from the 5- and 8-year-olds and the adults in a bisection task in the visu/visu-visu/audi group and the audi/audi-audi/visu group for the within-modal comparison sessions (visu/visu, audi/audi) and the cross-modal comparison sessions (visu/audi, audi/visu). (From Droit-Volet, S., Tourret, S., and Wearden, J., submitted.)

average 10.5% slower than the auditory one. By contrast, the magnitude of this difference in clock speed was relatively larger in the 5-year-olds than in the 8-year-olds and the adults. It is therefore reasonable to suppose that clock speed runs even slower for the visual than for the auditory modality in the youngest children.

In the present study, other interesting developmental trends were found. The 5-year-olds judged the visual stimuli as shorter than the auditory stimuli, but their temporal judgments were also more variable for the visual stimuli, contrary to the 8-year-olds and the adults. Indeed, in the 5-year-olds, the psychophysical functions were flatter and the Weber ratio greater for the visual than for the auditory stimuli. According to scalar timing theory, clock speed could be a cause of this variability difference in modalities for the youngest children. Indeed, this theory suggests that the temporal judgments depend on a count of pulses generated by a Poisson source that emits pulses at random, but at some constant mean rate (Gibbon, 1977, p. 284). On the basis of the mathematics of a Poisson pacemaker, Gibbon (1977) explained how a slower pacemaker rate produces more variable temporal estimates. Consequently, whether the clock speed was slower for the visual stimuli than for the auditory stimuli in the 5-year-olds, their temporal discrimination was more variable and their bisection functions flatter, with a larger Weber ratio.

However, it has been argued that differences in pacemaker speed are not the main source of variability in a timing system (Gibbon et al., 1984). In agreement with this, Penney et al. (1998, 2000) assumed in their explanation of the modality effect a low level of variability in temporal integration. In fact, according to scalar timing theory, the main source of variance is in the memory-encoding process, which we are going to discuss below. In other interval timing procedures (e.g., temporal generalization and verbal estimation), Wearden et al. (1998) observed that adult's duration judgments, like those of the youngest children in temporal bisection, were also more variable with visual than with auditory stimuli. These researchers suggested that variation in clock speed is not the only difference affecting the processing of temporal information when subjects compare auditory and visual stimuli. Notably, the modality effect on the variability of duration judgments can be caused by greater variance of the switch latency for the visual than for the auditory stimuli. In order to dissociate these two effects (clock speed and switch latency) on the variability of duration judgments, Wearden et al. (1998) conducted an experiment in which they combined both the sensory modality of the presented stimulus durations (visual or auditory) and the presence or absence of a 5-sec train of clicks, presumed to increase the clock speed. In this experimental condition, they observed that the presence or absence of clicks changed the length of duration estimates, as did the modality of the stimulus, but only the modality effect produced significant differences in the coefficient of variation of duration judgments. Thus, Wearden et al. (1998) concluded that the greater variance observed in the latency to open and close the switch, used to start and stop the clock when timing signals, explained the increased variability in duration judgments for visual signals compared to auditory signals.

In our temporal bisection task, only the youngest children were more variable in their duration judgments for the visual than for the auditory signals. We can therefore assume that the processing of the duration of visual stimuli requires relatively more attentional effort from younger children. Indeed, the visual stimuli require a continuous focus of attention on the physical source of presentation of visual information (i.e., computer screen) in order to begin the processing of duration at the onset of the stimulus. In contrast, the processing of the duration of auditory stimuli requires less attentional effort, these stimuli being more directly perceived (e.g., Meck, 1984). As the younger children have limited attentional capacity and are easily distracted (for a review, see Dempster and Brainerd, 1995), the variability of the latency to close the switch would be greater for the visual than for the auditory stimuli. The modeling of our data provided support for this idea by showing that the memory representation of the standard durations in the youngest children was relatively less precise for the visual than for the auditory stimuli.

The greater sensitivity to duration for auditory than for visual stimuli in young children suggests a sort of primacy of audition over vision in the processing of temporal information. This is an old idea, already put forward by studies in infants on the perception of temporal characteristics of speech sound and rhythms (for a review, see Pouthas et al., 1993; Droit-Volet, 2000b). For example, Eimas et al. (1971) have shown that infants can distinguish elementary speech segments (i.e., Pa, Ta, Ba, Da) only on the basis of their temporal acoustic characteristics. Lewko-wicz (1989, 1992) has also shown that infants are able to detect changes in frequency of rhythm, but that this ability emerges in the auditory modality before it emerges in the visual modality. He stated that audition is specialized in the processing of temporal information, whereas vision is specialized to detect motion. Friedman (1990a, p. 87) added that the "early sensitivity to the temporal characteristic of sound may be a special biological adaptation that allows infants to process information about speech."

Although there is a form of primacy of audition over vision in the processing of temporal information, the present study, as the previous one, demonstrates that the clock mechanism underlying time perception is functional at an early age. Indeed, as in human adults, the internal clock of children appears to run faster for auditory than for visual signals. However, this does not preclude the possibility of developmental changes at this level, such as an increase in pacemaker speed. Nevertheless, the results of our studies using the temporal bisection and generalization tasks suggest that the main sources of developmental changes in timing behavior are probably elsewhere than in the basic functioning of the internal clock.

0 0

Post a comment