A Data-Driven Approach to Uncovering the Neural Dynamics of Anxiety

Abstract

Anxiety is a behavioral state induced by low-threat, uncertain situations in which perceived danger is diffuse. The anxiety state is then accompanied by increased vigilance and risk assessment to one’s surroundings. Recent studies have shown that the brain regions responsible for encoding anxiety are widely located in the frontal cortex and extended limbic system; however, the network architecture responsible for hypervigilance has yet to be elucidated. Here, I propose to employ a data-driven method of using in vivo recordings of electrical activity across multiple brain regions concurrently as mice freely explore classic ethological anxiety-related behavioral assays and are administered pharmacological agents that modulate the anxiety state. Using novel machine-learning techniques, I have generated neural models that reflect the network-level activity engaged during the performance of these tasks. I have then validated the structure of this anxiety network in its ability to generalize to other anxiety-related tasks and models of disease. I anticipate that this strategy will discover an independent network that is correlated with anxiety-related behaviors. Thus, successful completion of the proposed work will lead to a network-level understanding of anxiety. Furthermore, the framework discovered through this study has the potential to facilitate the development of new revolutionary approaches for anxiety disorders.