Why studying intermodal duration discrimination matters

A critical issue in the field of time perception is whether or not explicit judgments about time are processed by some internal clock mechanism. A subsequent issue is whether or not this clock, if any, is central (i.e., is the same for a large range of durations, for whatever way of marking the intervals to be processed). There are several ways of marking time, including the use of signals delivered from different sensory modalities. In other words, do we have sensory specific representations of time, or is there an amodal—central—mechanism (Bueti, 2011)? This fundamental question is addressed here with an emphasis on the discrimination of brief empty time intervals. More specifically, intermodal intervals are of interest, an intermodal interval being marked by two brief and successive stimuli delivered from different sensory modalities. The interest for the effect of modalities on perceived duration and sensitivity to time has grown recently. Researchers have reported that intervals marked by auditory signals are perceived as longer than time intervals marked by visual signals (Walker and Scott, 1981; Wearden et al., 1998; Penney et al., 2000; see Grondin, 2003), but this issue received recent attention in a context where auditory and visual signals marking time could be presented simultaneously (Gamache and Grondin, 2010; Hartcher-O’Brien et al., 2014). The relative duration of intermodal intervals also received attention. For instance, intervals marked by an audio-visual sequence are perceived as longer than intervals marked by a visuo-auditory sequence (Grondin and Rousseau, 1991; Grondin et al., 1996). Moreover, some intermodal experiments emphasized the role of markers’ length on perceived duration (Grondin et al., 2005; Kuroda et al., 2014), with both the lengthening of the first and second marker resulting in longer perceived duration (Grondin et al., 1996). Recently, Mayer et al. (2014) conducted an investigation involving intermodal intervals lasting from 100 to 900 ms, with combinations of auditory, visual and tactile (A, V, T) stimuli. They observed that when a sound serves as the first marker, in either an AV or AT sequence, duration is perceived as longer than in conditions where a sound serves as the second marker, as in a VA or TA sequence; but reported no ordering effect when tactile and visual signals were used together (TV vs. VT). Mayer et al. interpreted their results in terms of sensory latency (see also Grondin, 1993; Grondin et al., 1996), arguing that the summative distortion pattern they observed (by opposition to a multiplicative effect) is consistent with the hypothesis that there exists a central timekeeping mechanism, common for the processing of any intervals, independently of their markers’ modality (see also Hartcher-O’Brien et al., 2014, for a similar conclusion). However, when intermodal and intramodal intervals are randomized from trial to trial, the overall interpretation in terms of sensory latency is more disputable (Grondin and Rousseau, 1991). For instance, for the discrimination of intervals circa 250 ms, Grondin and Rousseau (see their Table 6) reported a condition where the second marker of an interval was tactile, and the first was T, V, or A. They reported that an AT interval was perceived as much longer than TT and VT intervals. This could have been interpreted as if the A signal was detected more rapidly when serving as the first marker. However, when the second marker is always visual and the first one A, V, or T, it is not the AV intervals that are perceived as the longest, but the TV ones. In other words, an explanation based exclusively on latencies finds serious limitations when both intraand intermodal intervals are compared. Even more critical from a theoretical perspective is the question relative to the discrimination levels (sensitivity) reached with intermodal conditions. The recent data reported by Mayer et al. (2014) are also interesting as they describe the discrimination levels. In the VA and AV conditions, the Weber fractions are roughly the same, and vary from 30% at 0.1 s to slightly above 20% at 0.9 s. The results are essentially the same when auditory and tactile stimuli combinations are used, with the exception that performances are generally better when the auditory marker is presented first, especially at 0.1 s (above 40% in TA). With visual and tactile signals, the Weber fraction varies roughly between 25 and 31%, with the discrimination being usually better when the visual signal is presented first, especially at 0.1. For the discrimination of intervals lasting about 250 and 1000 ms, Rousseau et al. (1983) reported about the same performance levels in the VA and AV conditions. Also, for intervals lasting 1000 ms, Grondin (2003) reported about the same discrimination levels in AT and TA conditions, and in TV and VT conditions; at 250 ms, performance were slightly better in AT than in TA, and were slightly better in VT than in TV. The stability of the Weber fraction over time reported by Mayer et al. (2014) is a bit surprising considering the data reported by Grondin (1996) for