Attention is understood as inferring the level of uncertainty or precision during hierarchical perception. This paper substantiates this claim using neuronal simulations of directed spatial attention and biased competition. These simulations assume that neuronal activity encodes a probabilistic representation of the world that optimizes free-energy in a Bayesian fashion. Free-energy bounds surprise or the (negative) log-evidence for internal models of the world, so this optimization can be regarded as evidence accumulation or (generalized) predictive coding. Crucially, both predictions about the state of the world generating sensory data and the precision of those data have to be optimized. The paper shows that if precision depends on the states, one can explain many aspects of attention. It illustrates this in the context of the Posner paradigm, using simulations to generate both psychophysical and electrophysiological responses. These simulated responses are consistent with attentional bias or gating, competition for attentional resources, attentional capture, and associated speed-accuracy trade-offs. Furthermore, if both attended and non-attended stimuli are presented simultaneously, biased competition for neuronal representation emerges as a principled and straightforward property of Bayes-optimal perception. Keywords: attention, biased competition, precision, free-energy, perception, generative models, predictive coding.Attention is understood as inferring the level of uncertainty or precision during hierarchical perception. This paper substantiates this claim using neuronal simulations of directed spatial attention and biased competition. These simulations assume that neuronal activity encodes a probabilistic representation of the world that optimizes free-energy in a Bayesian fashion. Free-energy bounds surprise or the (negative) log-evidence for internal models of the world, so this optimization can be regarded as evidence accumulation or (generalized) predictive coding. Crucially, both predictions about the state of the world generating sensory data and the precision of those data have to be optimized. The paper shows that if precision depends on the states, one can explain many aspects of attention. It illustrates this in the context of the Posner paradigm, using simulations to generate both psychophysical and electrophysiological responses. These simulated responses are consistent with attentional bias or gating, competition for attentional resources, attentional capture, and associated speed-accuracy trade-offs. Furthermore, if both attended and non-attended stimuli are presented simultaneously, biased competition for neuronal representation emerges as a principled and straightforward property of Bayes-optimal perception. Keywords: attention, biased competition, precision, free-energy, perception, generative models, predictive coding.