Strain- and age-dependent divergence in mouse appetitive spatial learning and decision strategies

Animals rely on associative spatial memory to navigate toward previously learned, reward-associated goals. This reward-guided navigation is supported by the hippocampus and its interactions with cortical and subcortical regions: processes which are vital for integrating sensory cues and forming experience-dependent associations. In adulthood, hippocampal-dependent information processing is shaped by aging, reflecting changes in synaptic plasticity and neuromodulatory support. In parallel, mice with distinct genetic backgrounds show systematic differences in behavior, sensory function, and hippocampal plasticity. How mouse strain and age interact to affect spatial appetitive memories has not been well defined. Here, we trained juvenile (2-3 months) and mature (7-8 months) adult CBA/CaOlaHsd and C57BL/6 mice in daily training events, to perform a T-maze task with rewards available at a fixed location and decreasing probability in one maze arm. The task consisted of an initial deterministic phase in which a correct response was always rewarded, followed by a probabilistic phase during which reward omissions became increasingly prevalent. We measured correct choices, as well as latencies across training blocks, and combined trial-by-trial metrics with reinforcement learning modeling to assess decision strategies. We observed that CBA/CaOlaHsd mice displayed lower choice latencies than C57BL/6 mice, reached high performance earlier, and maintained better performance when the reward probability decreased. Age was associated with higher latencies and modulated both performance and decision policies in a strain-dependent manner. Moreover, CBA/CaOlaHsd mice displayed higher learning rates from positive outcomes and adopted strategies consistent with more robust context exploitation under uncertainty. C57BL/6 mice, by contrast, exhibited stronger omission-driven strategy switching compared to CBA/CaOlaHsd mice. Together, these findings reveal robust strain- and age-dependent differences in trial-by-trial decision policies and spatial learning performance.