Abstract <p>Repeated exposure to stressful stimuli is a fundamental aspect of neuropsychiatric disorder models, particularly in research on fear conditioning and active avoidance. However, the complex behavioral strategies animals employ during these tasks are often reduced to single metrics, which obscure the underlying cognitive processes. This study introduces novel mathematical modeling approaches to deconstruct the behavioral strategies of rats during an active avoidance test under repeated stress conditions. We analyzed behavioral data obtained in a two-way active avoidance task (50 trials) in 54 WAG/Rij rats. Two complementary models were used: a dual-exponential decay model for reaction times and a sigmoidal approximation for learning curves. The dual-exponential model effectively distinguished two latent factors that impede learning: an early passivity factor associated with initial reaction speed and a late endurance factor related to behavioral fatigue. The sigmoidal model measured the rate of learning (adaptability) and identified the transition point to consistent avoidance (adaptation point). Our findings show that these models offer a more nuanced quantification of cognitive processes – such as executive function, decision-making, and problem-solving – as well as individual traits like motivation and perseverance, compared to traditional analyses. This approach provides a powerful tool for dissecting the heterogeneity of stress-coping strategies and linking them to underlying neurobiological substrates, such as the absence epilepsy phenotype observed in WAG/Rij rats.</p>