Classification images reveal auditory influences on visual detection of temporal luminance deviation

Katsumi Watanabe, Recearch Center for Advanced Science and Technology, the University of Tokyo

Abstract
Changes in visual appearance by auditory stimuli have been widely reported (e.g., Stein et al., 1996). However, mechanisms underlying auditory effects on vision are largely unknown. Here we use classification image technique (e.g., Ahumada, 1996) to show that a sound modifies the visual strategies used for detection of temporal luminance deviation. A pedestal luminance increment (signal) was embedded in a sequence of Gaussian contrast noise. The signal timing was fixed. In 50% of trials, no signal existed. A staircase procedure adjusted the signal level. A 20-ms beep was presented at a predetermined time, which varied among sessions, relative to the signal frame. Subjects reported whether the sequence contained the signal in 2AFC. Classification images were clearly peaked around the moment of the visual signal. More intriguingly, another conspicuous peak was present at about 20-ms before the sound, when the sound occurred within an about 400-ms window around the visual signal. The present study thus demonstrates that (1) classification images reveal the subject's strategies of audio-visual interactions, (2) the auditory enhancement of perceived visual intensity may operate at high temporal precision (~20 ms), and (3) the temporal window of this audio-visual interaction is relatively large (~400 ms).

Not available

Back to Abstract