Congruent Auditory Input Can Bias Bistable Vision without Voluntary Attentional Control

Abstract

Prior research presents conflicting evidence as to whether the ability of congruent auditory information to resolve visual ambiguity is dependent on voluntary attentional selection of congruent stimuli (Van Ee et al., 2009; Alais & Parker, 2006). We used novel speech stimuli to test whether audiovisually congruent visual stimuli were dominant in binocular rivalry even when subjects were not aware which visual stimulus was associated with the auditory stimulus, effectively barring subjects from deliberately selecting the congruent stimulus with attention (verified post-experimentally). The stimuli were derived from prior research demonstrating that speech sounds distort the visual perception of ellipses towards associated mouth shapes: hearing a /wee/ sound causes increased perceived horizontal elongation of a circle, whereas hearing a /woo/ sound produces vertical elongation (Sweeney et al., 2012). In the current study, subjects were presented dichoptically with two animations of ellipses that each continuously elongated horizontally and vertically in alternation. The stimuli presented to each eye differed only in that they (a) were luminance-matched red or green, and (b) alternated counterphase such that, at any time, the horizontal extension of one ellipse was equal to the vertical extension of the other. Simultaneously, repeating voiced syllables /wee/ and /woo/ were played in synchrony with the visual stimuli so that, throughout each trial, there was a shape-sound correspondence with the ellipses presented to only one of the eyes. Subjects were instructed to continuously report which changing ellipse was perceptually dominant throughout the trial, ignoring the auditory stimuli as distracters. Each subject participated in sixteen 32-second trials, counterbalanced for color of congruent stimulus and the eye to which it was presented. Perceptual dominance durations for the audiovisually congruent stimuli were significantly longer than those for the incongruent stimuli. These results demonstrate that phonologically congruent auditory input can bias bistable vision in the absence of voluntary attentional control. Meeting abstract presented at VSS 2013