Reward Modality and Task Complexity jointly shape learning Dynamics in mice

Abstract

Learning efficacy relies on balancing reward-driven motivation with the difficulty of reaching a goal. Disentangling these factors is essential for understanding learning dynamics. Spatial navigation is a standard paradigm for probing these processes in rodents, but variation in mazes, cues, and protocols makes it difficult to determine how reward properties, rather than task design, shape learning and exploration,exploitation strategies. To date, no studies have directly compared reward type and learning difficulty within a single behavioral context. Here, we used the Starmaze to examine how reward modality and task complexity jointly influence learning and the transition from exploration to exploitation. We created three versions of the same sequential egocentric task in which a fixed goal was rewarded with either a hidden escape platform (aquatic version), food pellets, or medial forebrain bundle stimulation. Task difficulty varied by increasing the number of required decision points from zero to three. Our results reveal distinct behavioral dynamics across reward modalities and provide a unified framework for studying reward-dependent learning in sequential navigation.