Explore-exploit balance as a cross-species index of cognitive flexibility

About This Grant

Project Summary This application responds to PAR-25-035 “Building in vivo Preclinical Assays of Circuit Engagement for Application in Therapeutic Development”. Altered explore-exploit balance is a transdiagnostic target in serious mental health challenges, including schizophrenia and mood disorders. Addressing the mechanisms that govern transitions between explore and exploit behaviors are therefore a high-priority target for preclinical models and for testing interventions to alleviate cognitive deficits that contribute to poor prognosis and quality of life. In three Aims, we will uncover the key task, pharmacological, and neural parameters leading to alterations in explore-exploit tradeoffs in mice. Across Aims we will use two well-established touchscreen cognitive tasks and computational modeling to measure explore versus exploit behavior in mice on a trial-by-trial basis. In Aim 1 we will test further modifications to each task with the goal of identifying parameters that further increase the similarity between mouse and human decision making performance. In Aim 2 we will examine how current approved pharmacological treatments for depression and schizophrenia impact decision making, testing the hypothesis that these treatments push explore-exploit balance in opposing directions. Functional neuroimaging in humans reveals that depression and schizophrenia are each linked with distinct disruptions in activity in defined cortical networks, and explore and exploit decisions are associated with opposing patterns of activity in these cortical networks. In Aim 3 we will use a powerful preclinical technique, mesoscale imaging, to measure calcium activity of dorsal cortical networks during touchscreen decision making. We will test the hypothesis that we will see evidence for distinct network activity in explore versus exploit decisions that are cognate to components of human cortical networks, and that pharmacological treatments for depression and schizophrenia must impact these cortical networks to alter behavior. Collectively, these experiments will establish computationally amenable cognitive measures of explore-exploit balance for rodents, and their link to cortical network states, as a powerful new preclinical tool.