We tested the hypothesis that perception of an alternative image in ambiguo
us figures would he manifest as high-frequency (gamma) components that beco
me synchronized over multiple scalp sites as a "cognitive binding" process
occurs. Far 171 combinations of data from 19 electrodes, obtained from 17 s
ubjects and 10 replicate stimuli, we calculated the difference in correlati
on between the response to first seeing an ambiguous figure and when the al
ternative percept for that figure became consciously realized (cognitively
bound). Numerous statistically significant correlation differences occurred
in all frequency bands tested with ambiguous-figure stimulation, but not i
n two kinds of control data (a reaction-time test to sound stimuli and a no
-task, mind-wandering test). Statistically significant correlation changes
were widespread, involving frontal, parietal, central, and occipital region
s of both hemispheres. Correlation changes were evident at each of five fre
quency bands, ranging up to 62.5 Hz. Most of the statistically significant
correlation changes were not between adjacent sites but between sires relat
ively distant, both ipsilateral and contralateral. Typically, these correla
tion changes occurred in more than one frequency band. These results sugges
t that cognitive binding is a distinct mental state that is reliably induce
d by ambiguous-figure perception tasks. Coherent oscillations at multiple f
requencies may re fleet the mechanism by which such binding occurs. Moreove
r, different coherent frequencies may mediate different components of the t
otal cognitive-binding process. (C) 2000 Academic Press.