Audio-visual interaction in a speed estimation task
Single Paper Presentation

Markus Hofbauer
Department of Medical Psychology, Ludwig-Maximilians-University Munich

Sophie Wuerger
Department of Psychology, University of Liverpool

Georg Meyer
Department of Psychology, University of Liverpool

Kerstin Schill
Department of Medical Psychology, Ludwig-Maximilians-University Munich

     Abstract ID Number: 105
     Full text: Not available
     Last modified: March 31, 2003

Abstract
We investigated how humans combine auditory and visual motion speed information. The observer was seated in front of a 180-deg arc containing 35 equidistant LEDs and loudspeakers. The LEDs and loudspeakers were switched on and off such that object movement from left to right or vice versa was generated. The observer’s task was to predict the arrival time of the moving target in the centre of the arc by pressing a button. We used two different experimental settings: In one experiment we added noise to the signal by switching on a certain percentage of LEDs and loudspeakers randomly. In a second experiment the object was hidden for a certain distance before and after the target position. In both experiments we assessed the accuracy of the estimated arrival time for the auditory, visual and bimodal condition at five different speeds.
We found that the accuracy of the estimated arrival time improved when speed signals were provided in both modalities. We compared the experimental data with a model that is optimal in a statistical sense (Maximum Likelihood Integrator) and show that humans combine auditory and visual speed signals in a slightly sub-optimal fashion.