Researchers propose a new approach to neuroimaging studies. They suggest, rather than beginning with predefined behavioral functions and trying to assign associated brain regions, new studies should focus on selecting brain areas first and screen for potential behavioral associations.