This project proposes an administrative supplement to our research project titled “Deep LOGISMOS” (R01-EB004640) for further development of our general-purpose segmentation of multiple interacting objects and surfaces to efficiently facilitate quantitative studies of targeted-exercise effect on human hippocampal morphology and radiomic features. This supplement extends the applicability of Deep LOGISMOS + JEI to Alzheimer's disease (AD) research. Our work will focus on the role of aerobic exercise on decreasing the pace of cognitive decline in AD or perhaps preventing such a decline altogether. It is well known that left/right hippocampus, one of the deep anatomical structures in the brain, and its subregions play a critical role in AD development. It is equally accepted that management or possibly treatment of AD progression must start prior to emergence of severe symptoms. Research study data available from the currently active project R01-AG055500 “Exercise to improve hippocampal connectivity and learning in older adults” will be used for method development as well as validity assessment of the overarching hypothesis that moderate to vigorous intensity exercise affects hippocampal morphology and/or hippocampal radiomic features vulnerable to aging and AD more than lower-intensity training in a pre-Alzheimer cohort. The following main aims will be accomplished within 12 months of this research project supplement: 1) Develop and validate Deep LOGISMOS + JEI method for longitudinal 3D+time segmentation of left/right hippocampus and its six subregions in baseline/follow-up MR image pairs. 2) Develop statistical models capturing left/right/joint/subregional hippocampus morphology/radiomics changes between baseline and follow-up for each of the two studied exercise training groups. 3) Determine whether and the extent to which higher-intensity exercise training is more strongly associated with changes of the left/right/joint/subregional hippocampus morphology and radiomics. Our results will be significant because early prevention has the biggest impact. Determining how exercise counteracts mechanisms of AD-related dementia and cognitive aging leads to understanding how such plasticity is possible and informs prevention strategies. The proposed work is innovative because the new Deep LOGISMOS + JEI segmentation will enable our team to test how exercise affects cognition by bringing together sensitive measures of training induced changes in hippocampal morphology, hippocampal-dependent learning and memory processes, and novel conceptualizations for how to capture the physiological changes induced by exercise regimens.