Navigating in an uncertain environment requires attention to focus on the relevant locations and features. Despite being intuitively easy to grasp in terms of its effects, the concept of attention has been theoretically elusive. Recently, the Bayesian framework has been brought to bear on the concept of attention. In this framework, spatial attention is understood as inference about which spatial locations are likely to be relevant in the near future. Bayesian principles can be applied to estimate uncertainty of such beliefs. Theoretical computational modelling work proposes that uncertainty can be either expected or unexpected from the individual’s perspective. Expected uncertainty relates to unreliability of predictive relationships within a familiar context. Unexpected uncertainty, on the other hand, is caused by unexpected changes in the environment. Such theoretical conceptualisation of expected and unexpected uncertainty could clarify the elusive nature of spatial attention; yet, a core question is how the attention networks in the brain represent these two conceptually distinct forms of uncertainty. Strikingly, the experimental evidence investigating this issue is scarce. The current project aims to systematically investigate the neural dynamics of expected and unexpected uncertainty and how it drives human behaviour. Computational modelling will be combined with state-of-the-art neuroimaging methods to understand the role of uncertainty in attentional orienting.