The spatial resolution of crossmodal attention: Implications for the design of multimodal interfaces

Previous research on crossmodal attentional orienting has reported speeded reaction times (RT) when the stimuli from the different modalities are in the same spatial location and slowed RTs when the stimuli are presented in very different locations (e.g., opposite sides of the body). However, little is known about what occurs for spatial interactions between these two extremes. We systematically varied the separation between cues and targets to quantify the spatial distribution of crossmodal attention. The orthogonal cueing paradigm [Spence et al. 1998] was used. Visual targets presented above or below the forearm were preceded by either vibrotactile cues presented on the forearm, auditory cues presented below the forearm, or visual cues presented on the forearm. The presentation of both unimodal and crossmodal cues led to a roughly monotonic increase in RT as a function of the cue-target separation. Unimodal visual cueing resulted in an attentional focus that was significantly narrower than that produced by crossmodal cues: the distribution of visual attention for visual cues had roughly half of the lateral extent of that produced by tactile cueing and roughly one fourth of the lateral extent as that produced by auditory cueing. This occurred when both seven (Experiment 1) and three (Experiment 2) cue locations were used suggesting that the effects are not primarily due to differences in the ability to localize the cues. These findings suggest that the location of tactile and auditory warning signals does not have to be controlled as precisely as the location of visual warning signals to facilitate a response to the critical visual event.