Sound-induced illusory flashes: issues for a psychophysiological investigation

The first clear case where auditory information radically affects unambiguous visual perception was documented some years ago (Shams, Kamitani and Shimojo, 2000). Varying the number of auditory beeps presented coincidently with visual flashes was found to vary the number of visual flashes perceived. The illusory effect is phenomenological rather than qualitative, is stable with respect to many experimental variables and as such is thought to reflect an extensive property of polysensory mechanisms in the brain. Classic time-locked averages (ERPs) and event-related spectral perturbation (ERSP) are compared for this data set. Although the behavioural data indicated that the illusion occurred on 53% of one-flash/two-beep trials, no differences were found in the ERPs between illusion and non-illusion trials. Examination of ERSP plots revealed that illusion trials showed a transient increase in gamma-band activity (40-45Hz) at around 170ms, while non-illusion trials showed a decrease in gamma-band power at the same time, and a transient increase in beta (20-30Hz) power at around 100 ms. Although only small ERP differences were found between the illusion and non-illusion trials, non time-locked dynamic changes in brain activity were found to exhibit different patterns between illusion and non-illusion trials