Brief abstract: Most neuroimaging studies average across heterogeneous individuals while they perform one or two tasks, often under the (often implicit) assumption that all participants perform the tasks using similar cognitive strategies and overlapping neural circuits. There is substantial evidence that these assumptions are often violated, and so group-averaged results might not only be imprecise and incomplete, but also fundamentally incorrect. Additionally, If the goal is to explain why people behave differently, then averaging a small number of neural measures across people is likely to miss aspects of neural architecture that differ across people and that might be relevant to explain individual differences in behavior. My talk will discuss an alternative approach aimed at addressing these issues, which I will refer to as dense, person-specific functional neuroimaging.