Previous work from the Tsai lab (Canter et al 2019) identified the mamillary body (MB) as one of the first sites of amyloid deposition in 5XFAD model mice, a region that also correlates with dementia severity in human patients. Single cell RNA sequencing of the mouse MB identified 2 distinct neuronal populations within the MB, with segregated distribution, target projection, and unique electrophysiology. Analysis of these populations in the 5XFAD mice found that one of these populations, those found in the lateral MB (LM), are uniquely susceptible to hyperactivity and neurodegeneration, while the second population (medial MB, MM) is largely unaffected. The activity of the LM population also directly contributes to mouse performance in memory tasks. Additionally, using iterative direct-expansion microscopy (idExM) from the Boyden lab, we have identified intriguing patterns of amyloid associated with specific projections in the fornix, the white matter tract from the subiculum with axonal inputs to the MB. This grant proposes to investigate the links between amyloid and excitability changes in the MB and fornix, including development of the tools necessary to achieve this goal. The hypothesis to be tested in this application is that amyloid preferentially associates with the subiculum-LM projection and that these axons exhibit hyperexcitability. Aim 1 will map the connections between this new population of LM neurons and its upstream inputs from the hippocampus, using a newly developed in situ sequencing techniques, as well as exploring pathology in the white matter projection regions of this circuit in 5XFAD mice and human brain samples, using recently developed expansion microscopy. Aim 2 will characterize the source and location of hyperactivity found in the LM neurons through advanced voltage imaging, as well as expand this work to other mouse models of AD. Aim 3 will use optogenetics and pharmacological approaches to determine if specific aberrant circuit activity drives the pathology and behavioral changes seen in the AD model mice.