Repeated behavioral testing and the use of summary measures reveal trait anxiety in preclinical rodent models

The reliability and validity of preclinical anxiety testing is essential for translating animal research into clinical use. However, the commonly used anxiety tests lack inter-test correlations and face challenges with repeatability. While translational animal research should be able to capture stable individual anxiety traits - the core feature of anxiety disorders - the conventional approach employs a single type of test at a single time, which primarily reflects transient states of animals that are heavily influenced by experimental conditions. Here, we propose a validated, optimized test battery capable of reliably capturing trait anxiety in rats and mice of both sexes. Instead of developing novel tests, we combined widely used tests (elevated plus-maze, open field and light-dark test) to provide instantly applicable adjustments for better predictive validity. We repeated these tests three times to capture behavior across multiple challenges, which we combined to generate summary measures (SuMs). Our approach resolved inter-test correlation issues and provided better predictions for subsequent outcomes under more anxiogenic conditions or fear conditioning. SuMs were also shown to be more sensitive markers of stress-induced anxiety following social isolation. Finally, we tested our method's efficacy in discovering anxiety-related molecular pathways through RNA sequencing of the medial prefrontal cortex. SuMs revealed four-times more molecular correlates of trait anxiety than transient states, highlighting novel gene clusters. Furthermore, 16% of these correlates were also found in the amygdala. In summary, we provide a novel approach to capture trait anxiety in rodents, offering improved predictions for potential therapeutic targets for personalized medicine. We also provide recommendations to enhance feasibility without compromising validity or animal ethics, tailored to various scientific goals.