A secondary working memory challenge preserves primary place strategies despite overtraining

Abstract

Learning by repetition engages distinct cognitive strategies whose contributions are adjusted with experience. Early in learning, performance relies upon flexible, attentive strategies. With extended practice, inflexible, automatic strategies emerge. This transition is thought fundamental to habit formation and applies to human and animal cognition. In the context of spatial navigation, place strategies are flexible, typically employed early in training, and rely on the spatial arrangement of landmarks to locate a goal. Response strategies are inflexible, become dominant after overtraining, and utilize fixed motor sequences. Although these strategies can operate independently, they have also been shown to interact. However, since previous work has focused on single-choice learning, if and how these strategies interact across sequential choices remains unclear. To test strategy interactions across sequential choices, we utilized various two-choice spatial navigation tasks administered on the Opposing Ts maze, an apparatus for rodents that permits experimental control over strategy recruitment. We found that when a second choice required spatial working memory, the transition to response navigation on the first choice was blocked. Control experiments specified this effect to the cognitive aspects of the secondary task. In addition, response navigation, once established on a single choice, was not reversed by subsequent introduction of a secondary choice reliant on spatial working memory. These results demonstrate that performance strategies interact across choices, highlighting the sensitivity of strategy use to the cognitive demands of subsequent actions, an influence from which overtrained rigid actions may be protected.