Using 20 years of functional magnetic resonance imaging (fMRI) data from tens of thousands of brain imaging experiments, computational neuroscientists have created a geometry-based method for massive data analysis to reach a new understanding of how thought arises from brain structure.