Closed-loop reinforcement of memory reactivation in the human brain

About This Grant

PROJECT SUMMARY/ABSTRACT New experiences are cemented as long-term memories through processes that unfold after initial encoding. Evidence from animal models and human neuroimaging suggests that a key process that may be involved in generating a stable and enduring memory is the spontaneous reactivation of the neural ensembles that encoded the original event. While there have been robust demonstrations of spontaneous neural reactivation of past experiences during sleep and periods of quiet rest, it is unclear whether these effects are epiphenomenal or if they are causally important for memory consolidation. Critically, technology for testing the causality of reactivation is limited, in part because reactivation events are comprised of spatially distributed, intermingled patterns of neural activity that are difficult to manipulate with precision. Developing tools that can causally test the role of reactivation in memory stabilization is crucial for understanding the foundational principles of long-term memory formation and retention. In this project, we will establish proof-of-principle of a novel real-time neuroimaging (fMRI) approach—closed-loop neural reinforcement—to test the impact of reactivation on memory retention. This approach is uniquely capable of directly modulating distributed neural activity patterns like reactivation events. In a unique, four-session experiment, we will further determine the reliability of broader neural correlates of reactivation in the hippocampus and default mode network. In Aim 1, we causally test whether cortical reactivation promotes long-term memory retention. We will use a closed-loop neuroimaging protocol in which participants up-regulate memory-specific reactivation after encoding new information. This protocol includes a category manipulation to serve as a within-participant control: participants will encode visual events comprising scenes and faces, but only a scene- or face-sensitive cortical region (i.e. paraphippocampal place area or fusiform face area) will be reinforced. We predict that participants who successfully induce reactivation will exhibit increased memory after a delay of 24 hours, but only for the category of memoranda related to the reinforced cortical region. In Aim 2, we determine the stability of interactions between reactivation on network-level consolidation processes, as cortical reactivation is thought to be one component of a broad collection of neural dynamics that unfold after learning in service of memory stabilization. We plan to leverage our protocol’s within-subjects design to characterize the reliability of relationships between cortical reactivation and other putative consolidation processes, including hippocampal reactivation, hippocampal-cortical dynamics, and default mode network activity. By experimentally manipulating cortical reactivation to determine causal involvement in memory performance, our project will both advance theory and offer clinically-relevant information about whether reactivation could serve as a promising new target for enhancing consolidation processes in rehabilitation of memory impairment.