Mechanisms Underlying Modality Effects In Time Perception

The following section outlines some of the evidence suggesting that there are latency differences in signal detection or differences in clock speed between the visual and auditory modalities. As noted above, both timing onset-offset latency and clock speed have been proposed as putative mechanisms underlying modality differences.

Evidence that the response time (RT) for an auditory stimulus is shorter than that for a visual stimulus, by about 40 msec, comes from experiments showing that in order for an auditory and visual stimulus to be considered simultaneous in a temporal order task, the onset of the visual signal must precede the onset of the auditory signal by an amount slightly larger than this RT difference (Jaskowski et al., 1990). The evidence for this modality difference is somewhat ambiguous, however, because Rutschman and Link (1964) also found a 40-msec RT difference, but in their experiment the auditory signal had to precede the visual signal for judgments of simultaneity on the temporal order task. Interestingly, Hirsh and Sherrick (1961) found simultaneity with no onset difference. In both the peak-interval procedure and the temporal bisection task (Meck, 1984; Penney et al., 1996), presentation of a brief stimulus (e.g., 0.5-sec visual cue) indicating the modality of the upcoming timing signal (e.g., continuous visual signal) reduced the latency to begin timing in rats. Conversely, if the cue incorrectly predicted the modality of the timing signal (e.g., visual cue followed by an auditory signal), the latency to begin timing was increased. Moreover, Penney et al. (1996) found that clonidine, a noradrenergic antagonist, increased the latency to initiate timing in rats. This effect was localized to the switch process of the internal clock model.

A large body of work in rats, and other animals, much of it interpreted within the framework of scalar timing theory, has addressed both the neural and neurochemical substrates of the pacemaker-accumulator component of the internal clock (for a detailed review of the pharmacological basis of the internal clock, see Meck, 1996). For example, Maricq and Church (1983) reported that the dopaminergic agonist methamphetamine increased the rate of the internal clock, whereas the dopaminergic antagonist haloperidol decreased the rate of the internal clock when rats timed intervals in the seconds range. The importance of the dopaminergic system for timing has also been extended to human participants (Malapani et al., 1998; Malapani and Rakitin, this volume; Raamsayer, 1997a, 1997b). Although there is strong experimental evidence that both the rate of the internal clock and the latency with which timing processes are initiated are affected by pharmacological manipulations, the question of whether behavioral manipulations are also effective in modulating pacemaker speed is more important for the present discussion. Current evidence indicates that they are. For example, introducing rapid stimulus modulations, i.e., tone clicks, within a time interval demarcated by asterisks presented on a computer screen increases the apparent duration of the interval (Burle and Casini, 2001; Treisman et al., 1990). The verbal estimation task employed by Treisman et al. used stimulus durations in the hundreds of milliseconds range and a click train frequency that varied from 2.5 to 27.5 Hz across experiments. The primary purpose of the experiments was to obtain support for the existence of an internal temporal oscillator — an oscillator that could be perturbed by external stimuli via an influence on the arousal levels of the participants.

Perhaps most relevant here, however, is the work by Wearden and colleagues. Penton-Voak et al. (1996) followed up the findings of Treisman et al. (1990) by demonstrating that preceding a timing stimulus by an auditory click train influences the perceived duration of the timed stimulus. In four experiments using a variety of methods (temporal generalization, pair comparisons, verbal estimation, and short-duration production) and probe durations ranging from 200 to 1000 msec across experiments, they found that preceding the timing signal with a click train, which ranged from 1 to 5 sec in duration and had a frequency of 5 or 25 Hz, changed the expected target time by about 10% in human subjects in the temporal generalization, verbal estimation, and production paradigms. Droit-Volet and Wearden (2002; Droit-Volet, this volume) examined the influence of visual flicker stimuli prior to the training signals in a visual temporal bisection task with 3- to 8-year-old children and anchor durations of 400 to 800 msec and 400 to 1600 msec. Compared to the no-flicker condition, visual flicker caused a leftward shift in the response functions, and this shift was proportional to the duration being timed, a result that is consistent with a pacemaker speed effect. Earlier work by Wearden et al. (1999) also showed that auditory click trains relatively sped up or slowed down the internal clock in a temporal bisection task with short and long standards of 200 and 800 msec. In one condition, the intermediate probe durations were preceded by clicks, but the standards were not, whereas in another condition, the standards were preceded by clicks and the probe durations were not. Preceding the probes by clicks shifted the response function to the left, whereas preceding the standards by clicks shifted the response function to the right. The results from these click train studies were interpreted as indicating an effect on the speed of an internal pacemaker. Specifically, the click trains caused the pacemaker to emit pulses at a faster-than-normal rate, possibly because of an increase in arousal level or the amount of attentional resources dedicated to timing. Moreover, auditory click trains elicited effects for both auditory and visual timing signals. This outcome is significant because it suggests that there is a common amodal timing mechanism that is influenced by the click train.

Rousseau and Rousseau (1996) obtained strong evidence that a common amodal central pacemaker underlies timing of visual and auditory signals by using a task in which participants were not explicitly asked to time. In the most basic version of their stop reaction time task, participants experienced a unimodal sequence of brief auditory or visual stimuli that had a constant stimulus onset asynchrony (SOA). Participants were instructed to respond as quickly as possible when they thought the sequence had ended, but because the sequence length varied, counting stimuli did not allow participants to complete the task accurately. Therefore, although participants were not explicitly instructed to time, and the task was very different from the explicit timing tasks mentioned in this chapter, the task did require them to build up a representation of the SOA and compare the interval following each stimulus with this representation to determine whether to respond that the sequence had ended.

With simple unimodal sequences, Rousseau and Rousseau (1996) found that stop reaction times for visual sequences were 45 msec slower than those for auditory sequences across all SOA durations tested (250 to 1000 msec). This result is consistent with a modality-dependent switch-latency effect rather than a clock speed difference between auditory and visual signals, and suggests that both auditory and visual timing access a common pacemaker.

Some of their most interesting stop reaction time results, however, were obtained from a series of experiments that used bimodal single SOA sequences, bimodal multiple SOA (polyrhythmic) sequences, and bimodal polyrhythmic sequences. In the bimodal experiments, each sequence consisted of alternating auditory and visual stimuli. The results from the bimodal sequences indicated that participants processed these bimodal sequences as two concurrent unimodal subsequences; i.e., they timed the unimodal subsequences in parallel. In the unimodal polyrhythmic sequences, there were two SOAs, but all signals were identical (i.e., the same duration, frequency, and amplitude for all the auditory signals in the sequence, and the same central source and luminance for all the visual signals in the sequence). The unimodal polyrhythmic sequence experiments demonstrated that participants were able to time the auditory subsequences in parallel, but were unable to time the visual subsequences in parallel.

Rousseau and Rousseau (1996) interpreted their results within a multiple switch-accumulator framework developed by Church and Meck (Church, 1984; Meck and Church, 1984) to account for the ability of rats to simultaneously time

Pacemaker Switch Accumulator System


FIGURE 8.4 A modified scalar timing theory information-processing model that incorporates the modality-specific switch-accumulator modules proposed by Rousseau and Rousseau (1994, 1996). In this model a common amodal pacemaker is assumed, and inputs from modality-specific switch-accumulator modules feed into common memory and decision mechanisms.


FIGURE 8.4 A modified scalar timing theory information-processing model that incorporates the modality-specific switch-accumulator modules proposed by Rousseau and Rousseau (1994, 1996). In this model a common amodal pacemaker is assumed, and inputs from modality-specific switch-accumulator modules feed into common memory and decision mechanisms.

more than one signal. According to Rousseau and Rousseau (1996), there is a central pacemaker that subserves multiple switch-accumulator timing modules. The parallel timing ideas developed by Rousseau and Rousseau (1994, 1996) can be incorporated into a modified scalar timing information-processing (IP) model, as illustrated in Figure 8.4. In this modified IP model a common pacemaker is assumed so that modality effects may be manifested through either differences in switch closure latency or differences in the ease of maintaining the switch in the closed state only, and not via differences in pacemaker speed. Inputs from both modality-specific switch-accumulator modules are shown feeding into common memory and decision mechanisms. This is a requirement of the model given the claim made earlier that some modality effects are a consequence of comparing both visual and auditory temporal accumulations to a common comparison interval.

0 0

Post a comment