Cross-modal confidence judgements
55.26, Tuesday, 20-May, 5:15 pm - 7:15 pm, Talk Room 2
Pascal Mamassian1, David Alais2; 1Laboratoire des Systèmes Perceptifs, CNRS & Ecole Normale Supérieure, Paris, France, 2School of Psychology, University of Sydney, Sydney, Australia
An event defined by two sensory modalities has better precision than either modality alone (Ernst & Banks, 2002, Nature; Alais & Burr, 2004, Curr Bio). Is the original precision of one modality still available once bimodal combination occurs? We used a perceptual confidence judgement to address this issue. Perceptual confidence (the subjective sense of the accuracy of a perceptual decision) depends primarily on the encoded stimulus precision (higher confidence for higher precision). We reasoned that if participants no longer had access to visual precision in an audio-visual stimulus their visual judgments would be over-confident when auditory precision was high and under-confident when auditory precision was low. Stimuli consisted of flickering grey bars that were initially slightly brighter on the left side and then slightly brighter on the right side. Participants indicated on which side the bright bars appeared for longer (duration bisection task). While performing this visual task, they wore headphones and heard a sound that started in the left ear and finished in the right. Even though instructed to ignore the sound, the transition duration of the sound, short or long, increased or decreased their sensitivity on the visual task. Participants then made a confidence choice (Barthelmé & Mamassian, 2010, PNAS) between two stimuli with identical visual properties but differing in auditory precision. We found that participants could reliably make confidence judgements, and thus exhibited meta-perception. However, participants also presented a bias to be more confident when auditory precision was high than when it was low. Our results show that visual and auditory signals were combined and improved the temporal judgement, and that participants were fooled by the combined bimodal percept when asked to perform a unimodal confidence judgement.