Recent neuroscience experiments have brought inconsistent findings to light about the influence of neural activity in the alpha-frequency band (at ≈10 Hz) on the temporal dynamics of visual perception. Whereas strong alpha effects were found when perception was more based on endogenous factors, there were null-effects for alpha when perception relied more on objective physical parameters. In this Perspective, I open up a new view on neural alpha activity that resolves some important aspects of this controversy by interpreting alpha not as temporal processing of sensory inputs per se but above all as the observer's internal processing dynamics, their so-called perception sets. Perception sets reflect internally stored knowledge for how to organize and build up perceptual processes. They result from previous sensory experiences, are under top–down control to support goal-directed behavior, and root in pre-established neural networks that communicate through alpha frequency channels. I present three example cases from the recent neuroscience literature that show an influence of alpha-driven perception sets on the observer's visual-temporal resolution, object processing, and the processing of behaviorally relevant image content. Because alpha-driven perception sets can structure perception from its high-level aspects, like categories, down to its basic building blocks, like objects and time samples, they may have a fundamental impact on our conscious experience of the sensory world, including our perception of time itself.

Ongoing fluctuations in neural excitability and connectivity influence whether or not a stimulus is seen. Do they also influence which stimulus is seen? We recorded magnetoencephalography data while 21 human participants viewed face or house stimuli, either one at a time or under bistable conditions induced through binocular rivalry. Multivariate pattern analysis revealed common neural substrates for rivalrous versus nonrivalrous stimuli with an additional delay of ∼36 msec for the bistable stimulus, and poststimulus signals were source-localized to the fusiform face area. Before stimulus onset followed by a face versus house report, fusiform face area showed stronger connectivity to primary visual cortex and to the rest of the cortex in the alpha frequency range (8–13 Hz), but there were no differences in local oscillatory alpha power. The prestimulus connectivity metrics predicted the accuracy of poststimulus decoding and the delay associated with rivalry disambiguation suggesting that perceptual content is shaped by ongoing neural network states.