The hippocampus is critical for capturing rich, multimodal representations of experience and facilitating the long-term storage and later recall of these experiences. During sleep and pauses in behavior, the hippocampus can ?replay? prior experience ? reactivating the neural ensemble corresponding to the original experience in a time-compressed manner. During sleep, such replay is thought to underlie memory consolidation, while during behavior, replay is thought to additionally serve a more prospective role: contributing to planning or deliberation by retrieving stored memories in order to inform upcoming decisions. However, the content of replay neither solely reflects recent experience nor reliably predicts future behavior, leaving it unclear how exactly the representations of experience that are replayed relate to upcoming choices. Understanding the relationship between replay and behavior is particularly critical because abnormalities in replay and sharp wave ripples (SWRs; the network activity signature of replay) have been observed concurrent with impaired memory-dependent behavior in aging and diseases of aging. Establishing how replay content changes with aging, and whether these changes cause deficits in memory-guided behavior, has the potential to generate new therapeutic strategies to prevent or reverse memory impairment. In order to define how replay contributes to memory-guided decision-making in normal cognition and in the context of age-related memory impairment, we have developed a neurofeedback-based operant conditioning paradigm that targets SWRs. This paradigm provides rapid feedback contingent upon real-time detection of SWRs at a specific point during each trial of a spatial memory task, and results in substantially increased occurrence of SWRs in a trial phase- specific manner. Consequently, subjects experience more replay at the required trial phase, which occurs immediately prior to the choice point of a memory-dependent task. In addition to demonstrating that replay can be enhanced by neurofeedback, this behavioral paradigm provides an increased opportunity to link the content of replay with subsequent behavior. This paradigm lays the foundation for the three aims of this proposal: to define the relationship between replay and memory-guided behavior, to assess how this relationship changes with age, and to adapt the operant conditioning strategy to directly counter age- related replay dysfunction. I will complete these aims with the guidance of an exceptional mentoring team led by Loren Frank and including Carol Barnes, Uri Eden, and Karunesh Ganguly. During the mentored phase of the award at UCSF, I will conduct the proposed real-time feedback studies, gain expertise in using state-space models to capture and quantify replay content, scale experiments to efficiently examine larger cohorts of young and aged animals, and focus on professional development in order to facilitate a successful transition into an independent faculty position at an academic institution.
During aging and diseases of aging, the ability to use memory to guide behavior becomes impaired, often causing severe reduction in independence and quality of life. Our understanding of the neural processes that allow memories to shape ongoing behavior is limited, and we know even less about how they change over aging. This project will use large-scale multi-electrode hippocampal recordings and real-time manipulation of memory-related neural processes in order to understand how memories are used to guide behavior in both young adult and aged subjects and to develop methods to directly counter age-related dysfunction of memory processes.
