My goal is to understand how humans produce purposeful, adaptive behavior. The main ingredient for adaptive behavior, in all animals, is memory: we understand the world around us by matching the flow of incoming sensory information to previous experience. Importantly, by retrieving past episodes that resemble our present situation, we can predict what is likely to happen next, thus anticipating forthcoming stimuli and advantageous responses learned from past outcomes. Hence, I am interested in how the brain generates predictions about the world. However, unlike many other animals, humans can also produce adaptive behavior that runs counter to our learning history. For instance, we are able to switch from life-long driving on the right side of the road to driving on the left side during a trip to the UK. This capacity to use contextual information (“I’m in London”) to override habitual responses in favor of temporarily more goal-conducive actions is referred to as “cognitive control”, and it greatly enhances the flexibility of human behavior. Cognitive control requires the formation of temporary memory ensembles that link responses to stimuli in novel ways; this is often referred to as a “working memory”, and conceptualized as strategically attending to a select set of currently task-relevant representations. However, the mechanisms that govern this interplay between attention and memory remain poorly understood; our research aims to improve this situation. In my lab, we address the above questions using behavioral, computational, neuroimaging (e.g., fMRI) and neuro-stimulation (TMS) techniques.