Item

Abstract

Sensory modalities do not act in isolation; they can influence one another. One intriguing example
is temporal ventriloquism, where auditory and visual signals presented asynchronously
appear to be closer in time than they are generated. The characteristics of this perceptual phenomenon
are still not completely defined. In particular, it is not clear whether the order of presentation
has an influence on the effect, whether there are temporal grouping effects for other
modalities, and whether this effect only occurs for certain asynchronies of the signals. This
study aims to fill this lack of knowledge by investigating how crossmodal temporal grouping
affects perceived duration when onset and offset of a stimulus are defined by signals of different
modalities in all combinations of auditory, visual, and tactile stimuli. To test this, a two
interval forced choice (2IFC) paradigm was used, where participants had to judge which of two
intervals was shorter. One interval consisted of a lasting sound (filled interval), the other interval
was defined by two signals of different modalities at the onset and offset (empty interval).
These two signals could be auditory (beep), visual (flash of a LED) or tactile (vibration on the
participants’ left index finger). Different stimulus onset asynchronies (SOA) between 100 and
900ms of the crossmodal interval were tested. Moreover, the order of the signals forming the
empty interval was reversed. The duration of the filled interval was 30 to 170 of the duration
of the empty interval. For each SOA of the empty interval, the point of subjective equality
(PSE) for the duration was computed, which is the duration of the filled interval perceived to
be equal to the empty interval. Results confirmed that for audiovisual intervals grouping effects
were more salient with larger SOAs between the signals defining the empty interval. When the
light preceded the sound grouping effects were also found for 500 and 700ms. For audiotactile
intervals, instead, grouping effects occurred between 500 and 900ms but only when vibration
preceded sound. Lastly, for visuotactile intervals, grouping occurred at SOA of 500ms but only
when light preceded vibration. From these results, we can conclude that audiovisual temporal
ventriloquist is only an instance of a more general crossmodal grouping effect that occurs with
various modalities. Moreover, our data indicates that this effect acts differently on specific
modality combinations. Temporal grouping is affected by both the time between the signals
and their order of presentation.