There is a disconnect between neuroscience research on concept learning and representation (focusing on categories of concrete objects), and concept learning challenges in science education (which concern systems of causal relationships among objects and events). Bridging this gap will both inform theories of science learning and expand our understanding about how the brain learns and represents knowledge. We examine both literatures, point to where they converge and diverge, and offer paths forward for collaboration between science educators and cognitive neuroscientists to the benefit of each. In cognitive neuroscience, we describe advances in neuroimaging analyses based on cognitive models that directly relate cognitive processes to neural activity. Adapting science education content to support these model-based analyses offers unique opportunities to answer open questions about the cognition behind successful science learning. Further, adapting educational materials for use in neuroimaging may create useful learning exercises that offer unique affordances complementing typical classroom activities.