Abstract
The novel tank diving test (NTT) is a widely used behavioral assay for evaluating anxiety-like behaviors in zebrafish; however, results often exhibit considerable variability across different experimental settings. In this study, we systematically analyzed various methodological factors influencing the outcomes of NTT and introduced refinements to enhance its reliability and reproducibility. We optimized the detection parameters for region entry and freezing behavior using logistic regression analysis, significantly reducing false-positive classifications caused by tracking artifacts. The impact of pre-test stress conditions-restraint and darkness-was assessed, demonstrating that restraint effectively decreased the variability in behavioral parameters, such as latency to enter the top half (LTTH) of the tank and frequency of entries (FE). Conversely, combining darkness with restraint induced abnormal behaviors, limiting utility of the test. The effects of temperature were also rigorously evaluated, revealing that even subtle deviations within 3 °C of the standard temperature of 26.5 °C significantly affected behavioral variability, and 26.5 °C was optimal for reliable outcomes. Furthermore, we demonstrated that net-chasing during fish handling significantly increased the freezing time, suggesting the adoption of funnel-based transfers to reduce stress artifacts. Finally, behavioral patterns during stable test conditions followed a Poisson process, enabling the estimation of optimal test durations. Overall, our proposed refinements help establish a standardized, robust NTT protocol that minimizes variability and enhances the assay's sensitivity and reproducibility to investigate anxiety behavior in zebrafish.