Training before the development of overt neuropathology is required for full benefit of longitudinal training. In the initial longitudinal study (seeFig. 1) (experienced mice, denoted as E), mice were trained and tested before the development of overt neuropathology. To assess the importance of training before the emergence of overt Aβ and tau neuropathology, a separate group of mice were trained and tested at 6, 9, and 12 months after they had developed early Aβ and tau deposits (delayed training mice, denoted as DT). In addition, a separate group of mice were made to swim the daily mean only at 2, 6, and 9 months but were trained and tested as all other groups at 12 months of age. This yoked (denoted as Y) group was included to control for exposure to the handling and the learning environment in the absence of a learning contingency. Although the 3xTg-AD DT mice learned similarly to experienced mice (B,C), homozygous 3xTg-AD mice trained after the emergence of overt neuropathology were significantly impaired during probe trials relative to experienced mice at 9 and 12 months of age (F). Yoked mice were substantially impaired during acquisition, across all genotypes (A–C), perhaps attributable to a learned helplessness in response to the water maze. However, the NonTg yoked mice eventually acquired the task and exhibited no impairment during probe trials (D). In contrast, although the yoked 3xTg-AD mice eventually reached MWM criterion during training (B,C), they were significantly impaired during probe trials (E, F). *p < 0.05 with respect to NonTg mice;^ϕp < 0.05 with respect to the corresponding experienced group.