Introduction

The study of timing and time perception bridges durations ranging from milliseconds to days. Very long durations, such as circadian rhythms, appear to be governed by a periodic oscillatory process, sensitive to an external zeitgeber for reset and entrain-ment, that has extremely low variability and is used to time each single 24-h duration (e.g., Aschoff, 1984). In contrast, interval timing in the seconds-to-minutes range shows much greater variability, but is also highly flexible in terms of the durations that can be accurately timed (e.g., Gibbon et al., 1997; Hinton and Meck, 1997).

Many of the conceptualizations of interval timing proposed during the past 40 years characterized the ability to perceive the passage of time as the function of a specialized chronometric mechanism (e.g., Creelman, 1962; Fraisse, 1963; Franken-hauser, 1960; Gibbon et al., 1984; Killeen, 1984; Michon, 1967; Treisman, 1963). Most models of this type suggest five basic components (time base, gate, counter, memory, and comparator), but the attributes of these components vary across timing models. For example, the specification of the time base ranges from the mean number of events perceived in a given time interval (Frankenhauser, 1960) to pulses emitted by an internal pacemaker (Gibbon and Church, 1984; Treisman, 1963). Moreover, in some pacemaker models, the pacemaker-accumulator system is continuous in that distinct times, limited only by the base rate of the pacemaker, are represented (Gibbon et al., 1984), whereas in other models, the timer is categorical (Kristoffer-son, 1984), meaning that all durations falling into a particular range are equivalent and therefore indistinguishable.

One feature the clock-counter models proposed to date share, however, is that they do not explicitly account for the effects of signal characteristics on duration perception. Whether nontemporal signal characteristics influence timing is an important question because it is intimately related to the cognitive and physiological nature of the timing process itself. For example, whether the internal clock is modality specific, i.e., whether there are different clocks for different modalities or whether the clock is amodal, tells us something very basic about the timing system and its instantiation in both cognitive and neurophysiological terms. In addition, whether nontemporal signal characteristic effects are consistent across ranges of stimulus durations may be suggestive of whether durations in the milliseconds range access the same timing system as durations in the seconds-to-minutes range. There are arguments in the literature that both milliseconds and seconds range timing have the same underlying neural substrate (Gibbon et al., 1997), as well as arguments that the neural substrates for these two duration ranges are different (Ivry, 1996). In a more general vein, the influence of signal characteristics on duration perception may have diagnostic value for discriminating among models of timing. For example, Grondin (2001) suggested that if one accepts the claim that independence of temporal judgments from sensory characteristics supports the idea of a central timing mechanism, then the evidence that nontemporal marker characteristics influence duration discrimination might lead one to question whether there is a central timing mechanism. Clearly then, understanding the influence of nontemporal stimulus characteristics is important for developing accurate models of timing behavior.

It should be noted, however, that there are models of time perception and production that inherently account for the influence of nontemporal stimulus characteristics. For example, Ornstein (1969) rejected the idea of a specialized chrono-metric mechanism and suggested instead that the appreciation of elapsed duration is dependent on both the amount of information available during the time interval and the perceptual and memory processes involved in processing that information. Within this framework, time is an emergent property of information processing. One advantage of such a purely cognitive approach to temporal processing is that the effect of nontemporal attributes of the signal on perceived duration is explicitly incorporated into the model. Time is merely the amount of information processed, and because nontemporal characteristics are also processed, they must influence perceived duration. Based on a meta-analytic review, however, Block and Zakay

(1997) concluded that memory-based models such as Ornstein's were better suited for explaining retrospective temporal judgments, whereas attentional models are needed to explain prospective temporal judgments. Given this conclusion, the focus here is on prospective temporal judgments, and a specialized system of interval timing is assumed. More specifically, the literature is approached from the perspective of the scalar timing theory. The scalar timing theory (also referred to as the scalar expectancy theory), originally developed by Gibbon (1977) and subsequently modified and extended by Gibbon and his colleagues, is perhaps the most highly cited clock-counter model in the timing literature (for a discussion of this model's impact, see Allan, 1998), although more recent work suggests that oscillatory or coincidence detection systems may be the appropriate biological implementation of timing (see Malapani and Rakitin, this volume; Matell and Meck, 2000; Matell et al., this volume; Miall, 1989). Included with a precise mathematical description of the psychophysical nature of the time sense is a companion information-processing model (see Figure 8.1) that specifies a distinct multistage timing mechanism consistent with the components of internal clock models outlined above (Gibbon, 1977, 1991; Gibbon and Church, 1984; Gibbon et al., 1984).

Clock

Memory

Decision

Response

FIGURE 8.1 Three-stage information-processing model of scalar timing theory. (Adapted from Meck, W.H., Cognit. Brain Res., 3, 227-242, 1996.)

Response

FIGURE 8.1 Three-stage information-processing model of scalar timing theory. (Adapted from Meck, W.H., Cognit. Brain Res., 3, 227-242, 1996.)

In its original form, scalar timing theory does not include an explicit account of the influence of signal characteristics on the perception of time. One reason for this omission may be that the literature is somewhat contradictory with regard to the effects on temporal experience of nontemporal stimulus characteristics such as whether the signal is filled or empty; its intensity, modality, and pitch; and the cognitive load experienced during the task. For example, increasing signal intensity has usually, but not always been found to increase the perceived duration of the signal (e.g., Berglund et al., 1969; Goldstone and Lhamon, 1974; Goldstone et al., 1978; Steiner, 1968; Zelkind, 1973). Cognitive variables have also been shown to influence judgment of duration, but not entirely systematically. The presentation time of nonsense words is judged longer than equivalent duration real words (Avant et al., 1975), but more familiar words are judged as longer than unfamiliar words (Devane, 1974; Warm et al., 1964; Warm and McCray, 1969). Allan (1979, 1992) provides an excellent summary of the effects of a number of nontemporal signal characteristics on perceived duration. In the present chapter, the focus is on the effects of stimulus modality — specifically the auditory and visual modalities — on temporal experience.

8.2 MODALITY EFFECTS ON TIME PERCEPTION 8.2.1 Filled vs. Empty Duration Discrimination

The durations used in studies of filled and empty interval discrimination are generally on the order of tens to hundreds of milliseconds, and the empty intervals are usually demarcated by very brief stimuli on the order of tens of milliseconds in duration. Although, Rammsayer and Lima (1991) reported superior discrimination for 50msec filled auditory intervals, compared to empty intervals, subsequent work by Grondin (1993) showed that whether filled or empty intervals support superior discrimination is modulated both by the modality of the signals used to demarcate the temporal interval and the length of the temporal interval (for a review, see Grondin, 2001). For example, although Grondin (1993) found superior performance for empty, as opposed to filled, intervals for both auditory and visual stimuli using the method of single stimuli and a test duration of 250 msec, with 50-msec target intervals, empty intervals demarcated by visual signals showed superior performance, and there was no difference for the auditory modality. Grondin (1993) proposed an internal marker hypothesis to account for these results. The basic idea is that it takes longer for the internal trace of a timing signal to disappear and timing to stop than it does to initiate timing when a signal is presented. For filled signals, timing is initiated with signal onset and terminated when the trace of the physical signal is eliminated. In contrast, for empty intervals timing might begin when the trace of the physical signal has disappeared and ends as soon as the second signal is detected. The end result is that the duration actually timed for an empty signal is smaller than that for a filled interval. Easier discrimination for brief empty intervals follows from this because, according to Weber's law, the size of the change necessary for a detectable difference to be perceived increases with the size of the stimulus, in this case, the amount of time perceived for a filled or empty interval. With long durations, the effect of the timing onset-offset difference washes out and the discrimination advantage for empty intervals disappears. Finally, this hypothesis can account for the result that the difference between empty and filled intervals is larger in the visual than in the auditory modality if one assumes visual signals persist longer than auditory signals.

8.2.2 Are Auditory Signals Subjectively Longer than Visual Signals?

The question of whether signal modality influences the subjective experience of time has a long history in psychology. For example, Gridley (1932) traced the suggestion to compare the senses on accuracy in judging the same time intervals back as far as Czermak in 1857. Most subsequent work on signal modality and the subjective experience of time has been on the visual and auditory modalities. Sometimes modality effects have been examined directly, whereas in other cases, the experiment allowed the question to be addressed, but it was not the primary focus of the research. Although, the literature includes research with both human (e.g., Behar and Bevan, 1961) and nonhuman animals (e.g., Meck, 1991; Meck and Church, 1982), the review that follows is primarily restricted to experiments with human participants.

A number of investigators have examined modality differences in perceived duration (e.g., Behar and Bevan, 1961; Bobko et al., 1977; Brown and Hitchcock, 1965; Goldstone and Goldfarb, 1964; Hawkes et al., 1961; Walker and Scott, 1981; Wearden et al., 1998). Some of this work, utilizing a number of different paradigms, has indicated that auditory signals are judged longer than equivalent duration visual signals (e.g., Behar and Bevan, 1961; Goldstone and Goldfarb, 1964; Walker and Scott, 1981; Wearden et al., 1998).

Goldstone and Goldfarb (1964) conducted one of the most extensive analyses of modality differences. In a series of experiments, participants classified auditory and visual signal durations relative to either a social standard or a subjective standard. For the former case, the standard was 1 clock second on a scale ranging from very much less than 1 sec to very much more than 1 sec, whereas in the latter case, the signal duration was judged on a scale ranging from very, very short to very, very long. The presented durations ranged from 0.15 to 1.95 sec. For both the social standard and the subjective standard conditions the visual signals were judged shorter than equivalent duration auditory signals. This effect was obtained both when signal modality was a between-subjects factor and when it was a within-subjects factor. Although the authors also found that low-intensity lights seemed shorter than equivalent duration high-intensity lights, they claimed that intensity differences could not explain the modality effect because a separate experiment had failed to reveal significant effects of pitch or intensity on the perceived duration of auditory signals.

Behar and Bevan (1961) also found differences in the perceived duration of auditory and visual intervals. Their participants classified a series of probe durations (1, 2, 3, 4, and 5 sec) using an 11-category scale ranging from very, very, very short to very, very, very long. In two experiments there was an additional probe duration that was an outlier relative to the other probe durations (9 or 0.2 sec). They found that auditory durations were judged longer than objectively equivalent visual durations, although it is possible this difference may have been due to the extreme outliers influencing visual signal classification more than auditory signal classification. A subsequent within-subjects experiment, however, using the same basic design, but without the extreme outlier manipulation, indicated that auditory signals were judged about 20% longer than equivalent duration visual signals.

Walker and Scott (1981) asked participants to reproduce the duration of a presented stimuli by pressing and holding down a response button. They used intervals of 500, 1000, and 1500 msec demarcated by either a 600-Hz tone, illumination of a 15 x 20 mm light, or both. They found that 1000- and 1500-msec tones were perceived as longer than separately presented lights of equal duration, but that there was no difference for 500-msec stimuli. When both modalities were presented together, the duration reproduced was longer than that for light presented alone, but there was no difference between the combined stimulus and tone alone. In this experiment, the tone stimulus remained on during the light gap period and vice versa, so the conditions are not comparable to those described above for empty and filled duration comparisons. A second experiment used the same basic design as experiment 1, but with two exceptions. First, the time intervals were gaps in otherwise continuous tone and light presentations. Second, the intensity of the auditory signal was reduced relative to the first experiment. They found that a silent gap in an otherwise continuous tone was perceived as longer than an equivalent gap in an otherwise continuous light. If the light and tone turned off together, however, the gap duration was judged equivalent to that of the tone-alone gap duration. Apparently, when both signals were presented together, the auditory signal dominated because the combined signal was perceived like an auditory signal. Interestingly, a third experiment demonstrated that a 500-msec light was judged longer than a 500msec tone if the intensity of the tone was sufficiently reduced. However, they did not obtain such a difference for either the 1000- or the 1500-msec stimulus. They suggested that the difference in auditory and visual duration perception might have been due to differences in the perception of the onset and offset of visual and auditory signals. If the auditory signal had a faster perceptual onset and a slower or equal offset to the visual signal, then the auditory signal would seem longer. However, they rejected this possibility based on the finding that offset detection is faster for auditory than for visual signals (Goldstone, 1968). Unfortunately, they did not offer an alternative explanation for the auditory-visual difference, and they did not explain why they obtained different effects for the 1000- and 1500-msec stimuli than for the 500-msec stimulus.

Wearden et al. (1998) addressed the issue of modality effects in duration perception using both temporal generalization and verbal estimation methods in combination with within-subjects designs. For both paradigms, the auditory stimulus was a 500-Hz tone and the visual stimulus was a 4 x 4 cm blue square. In the temporal generalization experiment, the target duration selected on each test block came from a range of 400 to 600 msec, whereas in the verbal estimation experiment, duration length ranged from 77 to 1183 msec. They found that auditory stimuli were judged as longer than equivalent duration visual stimuli in both tasks, and that the auditory judgments were less variable than those for the visual stimuli. They interpreted the results as indicating that an internal pacemaker accumulating subjective time ran faster for auditory signals than for visual signals. The difference in variability they attributed to a variability difference in the operation of the mode "switch," which allows the internal pacemaker counts to accumulate in memory.

Other work, however, has failed to find evidence of modality differences (e.g., Bobko et al., 1977; Brown and Hitchcock, 1965; Hawkes et al., 1961; Kagerer et al., 2002; Szelag et al., 2002). Bobko et al. (1977) failed to find a statistically significant modality effect in four experiments, although they did report a clear trend for auditory signals to be judged longer than visual signals. Verbal estimation and magnitude estimation with and without a standard comparison stimulus were used, and the durations ranged from 0.25 to 5 sec in steps of 0.25 sec. The design was between subjects in that each participant experienced only a visual signal, the letter x presented tachistoscopically, or an auditory signal, a 1500-Hz tone.

Brown and Hitchcock (1965) used a reproduction method in which participants were initially trained on an interval of one signal modality, auditory (A) or visual (V), and were subsequently asked to reproduce the interval when it was demarcated by a signal of the same or a different modality. The four resulting conditions were AA, AV, VV, and VA. A group of ten participants was tested in each experimental condition, but all participants reproduced a complete set of intervals that ranged from 1 to 17 sec in increments of 2 sec. Comparisons of the AA and VV conditions failed to reveal any effects of signal modality on reproduced duration. In addition, the comparisons of the VA vs. VV, AA vs. AV, AA vs. VA, and AV vs. VV conditions at all test intervals revealed only two significant effects: those for AA vs. AV at 3 and 7 sec. Moreover, these effects were in opposite directions, with AV eliciting longer reproductions at 3 sec and AA eliciting longer reproductions at 7 sec.

Hawkes et al. (1961) used production, reproduction, and verbal estimation methods and durations of 0.5, 1, 2 and 4 sec. Each individual test session used a single method and modality, meaning that each participant served in nine test sessions total. They failed to obtain any significant differences in duration judgments between auditory and visual signals across the range of durations and methodologies examined.

Szelag et al. (2002) examined duration processing in children. The child's task was to reproduce the duration of a standard auditory or visual stimulus. Specifically, a standard duration ranging from 1 to 5.5 sec or 1 to 3 sec, depending on the condition, was presented, and following a brief delay, the stimulus reappeared and the participant had to respond on the keyboard when the correct duration had elapsed. They found that auditory standards were more accurately reproduced than visual standards for 1-, 1.5-, and 2-sec stimuli out of a stimulus range of 1 to 5.5 sec, but they did not obtain any evidence that auditory stimuli were experienced as perceptually longer than equivalent duration visual stimuli. In any case, the authors interpreted the accuracy difference as due to the nature of stimulus presentation rather than to a modality effect per se. The light stimulus was presented on a screen 1.7 m from the participant, whereas the tone was presented over headphones. They suggested that the auditory stimuli were better attended than the visual stimuli and therefore were reproduced more accurately. Interestingly, both modalities were presented in the same test session, but in a blocked fashion. Moreover, as noted above, on each trial the participant was presented with a target duration that then had to be reproduced. The significance of this aspect of the design is discussed below.

Kagerer et al. (2002) examined reproduction of durations in a range from 1000 to 5500 msec in a group of patients with unilateral focal brain injuries and a group of normal controls. The specific stimuli used in the experiment, rather than the comparison of brain-injured and control participant groups, are of interest here. They used four stimulus categories, and participants were tested on a different stimulus category on each of four test days. Two of the stimulus types were auditory (simple, 500-Hz tone; complex, sound of running water) and two were visual (simple, green oblong shape projected onto the wall of the test room; complex, 19th-century painting). They failed to find any differences across the different stimulus modalities and degrees of complexity.

At first glance, the conflicting results among studies that obtained and those that failed to obtain modality differences appear difficult to reconcile because of the variety of experimental procedures used. However, closer examination reveals some commonalities in the pattern of results across experiments. In general, the design of the studies that obtained an effect of modality allowed or encouraged the comparison of both the auditory and the visual probe durations with a common representation of the standard interval, be it a preexisting representation such as 1 clock second or a standard comparison interval presented in the experiment. This could occur through the use of a within-subjects design in which both signal modalities were presented in the same test session and participants were likely to use the same representation of the standard duration for both the auditory and the visual probe durations (e.g., Walker and Scott, 1981; Wearden et al., 1998). In addition, comparison to a prior representation, such as 1 clock second, or on a scale from very short to very long means, in effect, that the auditory and visual signals are being compared to the same internal standard, a situation that is similar to that of a within-subjects design (e.g., Behar and Bevan, 1961; Goldstone and Goldfarb, 1964). Of course, even if a participant experiences both modalities in the same test session, the design could allow or require the participant to compare an auditory probe with a representation of an auditory standard and a visual probe with a representation of a visual standard (e.g., Szelag et al., 2002). Furthermore, if a blocked design is used, then it is unlikely that the participant would use the same representation of the standard for both auditory and visual signals (e.g., Hawkes et al., 1961; Kagerer et al., 2002). Obviously, a between-subjects design, depending on its specific details, often absolutely prevents the use of a common standard representation for both modalities (e.g., Bobko et al., 1977). The possible exception to this may be when participants are asked to make comparisons to an explicit social standard like 1 clock second or on a subjective scale from short to long. Assuming the average objective duration of the social standard is approximately equivalent across groups of participants, then the groups are effectively using a common memory representation to which both auditory and visual signal durations are compared (Behar and Bevan, 1961; Goldstone and Goldfarb, 1964). As noted above, one possible explanation of the modality effect reasons that the internal clock runs at a faster rate for auditory than for visual signals. Therefore, the accumulated clock value for a given duration will be larger when the signal is auditory than when it is visual. As a consequence, if the auditory and visual accumulations are compared to a common memory representation, or to each other, then an auditory signal will seem longer than an equivalent duration visual signal.

Recently, Penney et al. (2000) addressed this possibility directly by comparing timing performance when the participants experienced both auditory and visual signals in the same test session with timing performance when the participants experienced only one or the other of the signal modalities in a test session. In the latter case, the experiment was a between-subjects design because each group of participants experienced a single signal modality. They used a duration bisection task in which participants were trained with examples of the short and long anchor durations prior to the test session. Three possible anchor duration ranges (2 to 8 sec, 3 to 4 sec, and 4 to 12 sec) were used across the series of experiments. An additional feature of some of the conditions, not addressed in detail here, was the simultaneous presentation of auditory and visual signals on the same trial, but with asynchronous onset and offset. In other conditions, single auditory and visual signals were presented sequentially in the same test session. The major finding was that when auditory and visual signals appeared in the same test session and had the same objective duration anchor values, auditory signals appeared subjectively longer than equivalent duration visual signals, as illustrated in Figure 8.2. This was true for both the simultaneous and the sequential presentation conditions. However, when auditory and visual signals appeared in separate test sessions, there were not any differences

Duration (s)

FIGURE 8.2 Group response functions averaged across participants for visual and auditory timing signals presented in the same test session. The anchor durations were 4 and 12 sec. p('long') = probability of a long response. (Data replotted from Penney, T.B., Gibbon, J., and Meck, W.H., J. Exp. Psychol. Hum. Percept. Perform., 26, 1770-1787, 2000.)

Duration (s)

FIGURE 8.2 Group response functions averaged across participants for visual and auditory timing signals presented in the same test session. The anchor durations were 4 and 12 sec. p('long') = probability of a long response. (Data replotted from Penney, T.B., Gibbon, J., and Meck, W.H., J. Exp. Psychol. Hum. Percept. Perform., 26, 1770-1787, 2000.)

Duration (s)

FIGURE 8.3 Group response functions averaged across participants for visual and auditory timing signals. The auditory single and visual single response functions are from single-modality test sessions. The auditory same and visual same response functions are from the same two-modality test session. The visual different response function is from a two-modality test session in which a different anchor duration pair was used for the auditory signals. The anchor durations were 4 and 12 sec. p('long') = probability of a long response. (Data replotted from Penney, T.B., Gibbon, J., and Meck, W.H., J. Exp. Psychol. Hum. Percept. Perform, 26, 1770-1787, 2000.)

Duration (s)

FIGURE 8.3 Group response functions averaged across participants for visual and auditory timing signals. The auditory single and visual single response functions are from single-modality test sessions. The auditory same and visual same response functions are from the same two-modality test session. The visual different response function is from a two-modality test session in which a different anchor duration pair was used for the auditory signals. The anchor durations were 4 and 12 sec. p('long') = probability of a long response. (Data replotted from Penney, T.B., Gibbon, J., and Meck, W.H., J. Exp. Psychol. Hum. Percept. Perform, 26, 1770-1787, 2000.)

in the classification of equivalent duration auditory and visual signals. Crucially, an additional experiment demonstrated that the appearance of auditory and visual signals in the same test session was not sufficient to elicit the modality effect. It was also necessary for the anchor durations of the signals to be objectively equivalent. For example, when a series of auditory signals with an anchor duration pair of 3 to 6 sec appeared in the same test session as visual signals with an anchor duration pair of 4 to 12 sec, then there was no modality effect. That is to say, the response function for visual signals 4 to 12 sec in this condition was indistinguishable from that generated by participants who experienced visual signals only in the test session, auditory signals only in the test session, or the auditory response function generated by participants who experienced both auditory and visual signals in the test session, as illustrated in Figure 8.3.

In accounting for their results, Penney et al. (2000) claimed that the auditory and visual signals drove an internal clock at different rates. They located this rate difference at the level of the mode switch within the scalar timing model (see Figure 8.1), rather than suggesting that the pacemaker "pulsed" at a faster rate for auditory than for visual signals. In their view, the pacemaker rate was the same for both auditory and visual signals, but the probability that the mode switch would be in the closed state was different for the auditory and visual signals. They suggested that for visual signals the mode switch could oscillate or "flicker" between an open and a closed state with the result that some number of pacemaker pulses or counts would be lost during the presentation of the visual signal. The degree of count loss would be proportional to the duration of the timing signal with the end result that the size of the modality effect would be proportional to the signal duration. In principle, the state of the mode switch would be determined by the ability of the signal to capture attention because the mode switch requires attention to be allocated in order to enter the closed state (see Droit-Volet, this volume; Fortin, this volume). Assuming auditory signals more easily capture attention than visual signals, and there is some evidence that auditory signals are processed automatically, whereas visual signals are subject to controlled processing (Meck, 1984; Posner, 1978), then they are better able to maintain the switch in the closed state. Therefore, for visual signals, many clock ticks would fail to accumulate, and consequently, the visual duration would seem subjectively shorter than an equivalent auditory duration if both signals were compared to a common memory representation. In truth, the experimental evidence available from this series of experiments does not allow for a differentiation between a change in pacemaker rate and a change in the efficacy of the mode switch, but there is evidence in the literature, discussed in detail below, that auditory and visual signals share a common pacemaker but have separate switch-accumulator modules (Rousseau and Rousseau, 1996). However, no matter what the underlying cause of the different temporal accumulations for auditory and visual signals, it is clear that the modality effect of auditory signals being experienced as subjectively longer than equivalent duration visual signals requires more than just differential clock rates for both the visual and auditory signal modalities. The modality effect also requires that the auditory and visual signals be compared to a common memory representation in order for a clock rate difference to manifest as a difference in perceived duration. As noted above, the common memory representation could be due to a common memory pool to which both auditory and visual training signals contribute to varying degrees (cf. Penney et al., 2000) or due to the use of a common social standard. In the conditions that revealed a modality effect in the bisection experiments described above, the subjects failed to maintain separate memory representations for the visual and auditory anchor durations. In other tasks, subjects may compare the current passage of time to some preexisting standard, such as 1 clock second. Clearly in this case also, a fast auditory accumulation and a relatively slower visual accumulation would seem different relative to the common comparison duration; therefore, auditory signals would seem longer than visual signals. Even in cases where the participants are not told explicitly what the comparison should be (e.g., when they are told to use a scale from very, very short to very, very long), it is probable that they would make the comparison relative to some preexisting mental representation even as they built up a sense of the range of durations presented in the specific experimental context (see Lustig, this volume).

The present model, however, does not easily account for the failure, described earlier, of Brown and Hitchcock (1965) to find consistent modality effects when comparing participants asked to reproduce signal durations using a different modality from that of the sample duration. For example, based on the preceding, one would expect the group of participants presented with an auditory sample and a visual test to reproduce longer durations than the group of participants presented with a visual sample and an auditory test; yet this did not happen. Interestingly, Droit-Volet et al. (submitted; Droit-Volet, this volume) recently found modality effects in a version of the temporal bisection task that shares critical features with the Brown and Hitchcock (1965) design. Specifically, participants were trained and tested with one signal modality (e.g., visual) and, on a subsequent test day, experienced the same signal (e.g., visual) in training, but a different modality signal (e.g., auditory) in the test. The results indicated that participants experienced auditory signals as longer than equivalent duration visual signals. As participants would have compared an auditory signal with a visual memory representation of the anchors or a visual signal with an auditory memory representation of the anchors, these results are consistent with the model outlined above.

Note, however, that a common memory representation combined with clock rate differences is not a requirement for differences in discriminability between auditory and visual signals, because clock rate differences alone will allow one signal to be more easily discriminated than another based upon the total number of pacemaker pulses accumulated within a given duration (Gibbon, 1977). Rammsayer and Lima (1991) raised this possibility as an account of differences in discriminability between empty and filled durations. In brief, the higher the pulse rate, the finer the discrimination between intervals that can be made because the delay between pulses is shorter.

0 0

Post a comment