Visual perception is periodic and structured in recurrent temporal frames. However, it is unclear whether these cycles simply reflect fluctuations of visual excitability, and/or organize perception into discrete moments of information integration in which two events falling within one frame are more likely to be perceived as causally linked. In 1981, Varela et al. probed temporal framing of perception by presenting two sequential flashes at a constant delay and found these flashes to be perceived as simultaneous or sequential depending on the phase of ongoing alpha (7-13Hz) oscillations. These results are still awaiting replication (VanRullen et al, 2014). We here tested periodicity of visual information integration in two behavioural experiments probing (i) simultaneity and (ii) apparent motion perception in double flash stimuli. We flashed two visual events either sequentially or simultaneously. In experiment 1, two dots were displayed in the left and right visual fields, while participants reported whether the events were perceived as sequential or simultaneous. In experiment 2, the two events consisted of random dot patterns that could be perceived as coherent motion, incoherent motion or glass patterns. These stimulus pairs were presented time-locked to a preceding sound (0-300ms delays), shown to reset the phase of visual brain oscillations (Romei et al., 2012). Spectro-temporal dynamics of behavioural time courses (hits) revealed periodicity in perception. Dots perceived as sequential (Group 1, n=19) and apparent motion perception (Group 2, n=19) showed identical, cyclic patterns in the 12Hz (alpha) frequency band (with a significant phase consistency across participants). In contrast, a cyclic pattern in the 4Hz (theta) band was observed in the control conditions (simultaneously presented stimuli), suggesting that the alpha cycles in visual perception go beyond reflecting fluctuations in visual excitability. These data suggest that perceptual frames in the alpha-band play a key role in spatio-temporal visual information integration.
Read full abstract