UC Riverside

Strategies for optimized performance in any task do not spontaneously occur. They develop over time. Across learning, our first strategy may be a successful strategy. On the other hand, we may require multiple strategies to succeed in a task. In this body of work, we examine spatial and temporal behavioral and neuronal strategies relevant to reward driven decision making using a whisker-based selective detection paradigm where mice learn to selectively respond to a preferred (target) stimulus and selectively ignore a nonpreferred (distractor) stimulus. Through widefield calcium imaging of task-relevant sensory and motor cortices, we first identified comparable target and distractor stimulus encoding in sensory regions, followed by attenuation of distractor and propagation of target encoding in motor regions. We interpret this localized attenuation filter as afunctional, potentially reactive, neural strategy for the selection process after stimulus presentation and before a planned motor response. Then we found that a global cortical state of low activity and low variability in a prestimulus epoch predicted response outcomes. We interpret the global prestimulus profile as a preemptive, potentially proactive, neural strategy before stimulus presentation. Finally, longitudinal investigation across learning behavior revealed that mice used both object-based and temporal strategies to transition from naïve to expert performance in the selective detection task. Furthermore, we found that the transition strategies differed between male and female mice such that male mice overlapped their response and temporal strategies before improving their object-based performance and that female mice improved their performance through sequential temporal and object-based intermediate strategies. We find evidence that supports development of multiple strategies across learning, transitioning mice from suboptimal to optimal performance in the selective detection task. In conjunction with the neural findings of expert behaving mice, we can further our understanding of how behavioral strategies form across learning to maximize successful outcomes.