Some moments in our lives provide information that is useful in itself, but other information is most meaningful when combined across many instances, as we come to understand the regularities in the world. How do we extract such structured knowledge from our environment? Answering this question requires an understanding of the initial acquisition of this information as well as its stabilization and integration with existing knowledge structures over time and with sleep. Our research combines neural network modeling and empirical methods (fMRI, EEG, patient studies, behavior) to uncover learning algorithms and principles of how memories of regularities in the environment come to be represented throughout the brain.

​