Over the past decade or so, it has become clear that our brains continuously predict what sensory signals are likely informative for goal-directed action based on past experience and statistical regularities in the environment, and hence, that learning is a much more pervasive feature of attention than is commonly assumed. In this talk, I will present findings from several behavioral and EEG studies that reveal how, at the neural level, such probabilistic learning rapidly structures attention: what we attend to and perceive in the environment and what we ignore. I will then discuss the idea that meditation may provide a method to reduce the influence of past experience on current experience by modulating attention and how this may alter conscious perception. I will present some findings in support of this notion and end by discussing important avenues for future research.