The design of multimodal interfaces rarely takes into consideration recent data suggesting the existence of considerable crossmodal spatial and temporal links in attention. This can be partly explained by the fact that crossmodal links have been studied almost exclusively in spartan laboratory settings with simple cues and tasks. As a result, it is not clear whether they scale to more complex settings. To examine this question, participants in this experiment drove a simulated military vehicle and were periodically presented with lateralized visual indications marking locations of roadside mines and safe areas of travel. Valid and invalid auditory and tactile cues preceded these indications at varying stimulus-onset asynchronies. The findings confirm that the location and timing of crossmodal cue combinations affect response time and accuracy in complex domains as well. In particular, presentation of crossmodal cues at SOAs below 500ms and tactile cuing resulted in lower accuracy and longer response times.

Tactons, or vibrotactile icons, have been proposed as a means to communicate complex concepts to users and to support multitasking in environments involving numerous visual and/or auditory tasks and stimuli. This study investigated the role of processing code in the interpretation of tactons while performing concurrent visual tasks in such environments. Participants decoded tactons composed of spatiotemporal patterns of vibrations – requiring spatial processing – and interpreted one of two types of visual task stimuli – requiring either spatial or categorical processing – in a driving simulation. Compared to single-task performance, there was a significantly larger dual-task performance decrement when the tacton task was paired with the visual task requiring spatial (as compared to categorical) processing. The findings are consistent with the assertion of Multiple Resource Theory that interference between concurrent tasks is greater when these tasks involve the same processing code. They illustrate how distributing task-related information across modalities is beneficial but not sufficient to avoid task interference. A direct implication of the findings is to avoid the use of spatiotemporal tactons in environments which rely heavily on spatial processing resources, such as car cockpits or flight decks.