Transforming sensory experiences into adaptive long-term memories
Long-term memory allows humans to exploit information encountered in the past in order to behave more adaptively in the present. However, our memories are not perfect snapshots of our sensory experiences. My research explores the cognitive and neural pressures that shape what sensory information is stored in long-term memory and how this information is stored in a way that benefits behavior. In my talk, I will highlight fMRI and intracranial EEG experiments that seek to determine (i) how the content of our memories is actively molded by competing memories, task factors, and intrinsic neural architecture; and (ii) how these memory transformations promote adaptive behavior. First, I will present empirical evidence that competition between similar memories can lead to targeted differentiation of hippocampal traces. In line with predictions from computational models, this change to hippocampal representations minimizes future memory errors. Second, I will highlight work examining the relationship between cortical activity during perception and later recall and provide evidence for systematic differences between the two that are imposed by visual system architecture. Finally, I will bridge these two research areas and discuss two complementary experiments investigating how mnemonic information is transferred between the hippocampus and cortex in support of efficient memory-guided behavior. These experiments establish a framework for developing a computational and biological understanding of the factors that influence the content of human memory and its relationship to adaptive behavior.