Abstract
We have previously shown that the degree of integration of auditory and visual motion signals depends strongly on the precise stimulus configuration. When auditory motion signals are generated by cross-fading white noise between two loudspeakers and the visual motion signal is a random-dot-kinematogram, the increased sensitivity for the detection of bimodal motion can be accounted for by probability summation as opposed to low-level motion integration (Wuerger, Hofbauer & Meyer, 2003, P&P, 65(8), 1188–1196). However, when the individual auditory and visual motion signals are strictly localised and furthermore spatially and temporally co-incident, thresholds for motion detection are predicted by subthreshold summation (Wuerger et al., 2003, JoV, 3(9), 771a). In the present experiments, we further investigated the amount of spatial asynchrony tolerated by this low-level auditory-visual integration mechanism. We measured motion detection thresholds for auditory, visual and bi-modal motion stimuli, which were presented along a horizontal arc, containing 31 co-localised LEDs and loudspeakers, spaced 5 degrees apart. In each trial two intervals were presented, one interval contained noise only (LEDs and loudspeakers were switched on randomly in each frame), the other interval contained an additional motion signal. The observer indicated which interval contained the signal. The auditory and visual motion had different spatial offsets ranging from 0 to 75 deg, moving at a speed of 30 deg/sec. Results and Conclusion: The largest increase in sensitivity for the bimodal condition was found when the auditory and visual motion signals were strictly co-localised. There is a systematic decrease in sensitivity with an increasing spatial offset between the auditory and visual motion. This can be interpreted as an “auditory-visual receptive field” or tuning function with a bandwidth (width at half gain) of about 25 deg visual angle.