Why two "distractors" are better than one: modeling the effect of non-target auditory and tactile stimuli on visual saccadic reaction time

Saccadic reaction time (SRT) was measured in a focused attention task with a visual target stimulus (LED) and auditory (white noise burst) and tactile (vibration applied to palm) stimuli presented as non-targets at five different onset times (SOAs) with respect to the target. Mean SRT was reduced (i) when the number of non-targets was increased and (ii) when target and non-targets were all presented in the same hemifield; (iii) this facilitation first increases and then decreases as the time point of presenting the non-targets is shifted from early to late relative to the target presentation. These results are consistent with the time-window-of-integration (TWIN) model (Colonius and Diederich in J Cogn Neurosci 16:1000-1009, 2004) which distinguishes a peripheral stage of independent sensory channels racing against each other from a second stage of neural integration of the input and preparation of an oculomotor response. Cross-modal interaction manifests itself in an increase or decrease of second stage processing time. For the first time, without making specific distributional assumptions on the processing times, TWIN is shown to yield numerical estimates for the facilitative effects of the number of non-targets and of the spatial configuration of target and non-targets. More generally, the TWIN model framework suggests that multisensory integration is a function of unimodal stimulus properties, like intensity, in the first stage and of cross-modal stimulus properties, like spatial disparity, in the second stage.