Ten years ago, neuroscientists began to study cultural phenomena by using functional MRI. Since then the number of publications in this field, termed cultural neuroscience (CN), has tremendously increased. In these studies, particular concepts of culture are implied, but rarely explicitly discussed. We argue that it is necessary to make these concepts a topic of debate in order to unravel the foundations of CN. From 40 fMRI studies we extracted two strands of reasoning: models investigating universal mechanisms for the formation of cultural groups and habits and, models assessing differences in characteristics among cultural groups. Both strands simplify culture as an inflexible set of traits and specificities. We question this rigid understanding of culture and highlight its hidden evaluative nature.