The representational capacity and inherent function of any neuron, neuronal population or cortical area in the brain is dynamic and context-sensitive. Functional integration, or interactions among brain systems, that employ driving (bottom up) and backward (top-down) connections, mediate this adaptive and contextual specialisation. A critical consequence is that neuronal responses, in any given cortical area, can represent different things at different times. This can have fundamental implications for the design of brain imaging experiments and the interpretation of their results. Our arguments are developed under generative models of brain function, where higher-level systems provide a prediction of the inputs to lower-level regions. Conflict between the two is resolved by changes in the higher-level representations, which are driven by the ensuing error in lower regions, until the mismatch is "cancelled". From this perspective the specialisation of any region is determined both by bottom-up driving inputs and by top-down predictions. Specialisation is therefore not an intrinsic property of any region but depends on both forward and backward connections with other areas. Because the latter have access to the context in which the inputs are generated they are in a position to modulate the selectivity or specialisation of lower areas. The implications for classical models (e.g., classical receptive fields in electrophysiology, classical specialisation in neuroimaging and connectionism in cognitive models) are severe and suggest these models may provide incomplete accounts of real brain architectures. Here we focus on the implications for cognitive neuroscience in the context of neuroimaging.