An audio-visual flash-lag effect

Abstract

PURPOSE: The flash-lag effect (FLE) has recently generated much activity in vision research. Here we broaden the scope of enquiry by examining a cross-modal, audio-visual FLE. METHODS: Motion stimuli were either translating gaussian blobs (t=0.64s; s=0.45deg; speed=35deg/s) or low-pass (<2kHz) auditory noise sources which were moved smoothly over the same trajectory by varying the sign and magnitude of interaural time differences. Flashed (20ms) gaussian blobs or auditory noise bursts served as ‘flash’ stimuli, which occurred approximately in the middle of the motion trajectory (randomly jittered). Adaptive staircases varied flash location and homed in on subjective alignment of the ‘flash’ with the instantaneous motion position. In 4 conditions, the flash (f) or motion (m) could be visual (V) or auditory (A): Vf/Vm, Vf/Am, Af/Vm, Af/Am. RESULTS: In all conditions, the ‘flash’ was perceived behind the moving stimulus, requiring a large advance to be perceptually aligned (i.e., the FLE). Effect size was speed dependent. In the Af/Am version, misalignment over 4 speeds increased linearly with a time constant of ∼170ms. In the Vf/Vm version, the time constant was ∼20ms. The cross-modal conditions, (Vf/Am, Af/Vm) had intermediate values of ∼70ms and ∼120ms respectively. CONCLUSIONS: Whether motion was auditory or visual, differential delays were ∼100ms longer for auditory flashes. Whether the flashes were auditory or visual, differential delays were longer by ~50ms for auditory motion. These effects suggest general sensory processing principles underlie the FLE. The symmetry of effects suggests linear combination of delays within and between sensory modalities.