Clinical depression is characterized by multiple concurrent symptoms, manifesting as a complex heterogeneous condition. Although some well-established classical behavioral assessments are widespread in rodent models, it remains uncertain whether rats also display stress-induced depression-related phenotypes in a multidimensional manner, i.e., simultaneous alterations in multiple behavioral tests. Here, we investigated multivariate patterns and profiles of depression-related behavioral traits in male Wistar rats subjected to inescapable footshocks (IS) or no-shocks (NS), followed by a comprehensive battery of behavioral tests and ethological characterization. We observed generalized stronger intra-test but weaker inter-test correlations. However, feature clustering of behavioral measures successfully delineated variables linked to resilience and susceptibility to stress. Accordingly, a noteworthy covariation pattern emerged, characterized by increased open field locomotion, reduced time in the elevated plus maze open arms, lower sucrose preference, and increased shuttle box escape failures that consistently differentiated IS from NS. Surprisingly there is little contribution from forced swim. In addition, individual clustering revealed a diversity of behavioral profiles, naturally separating NS and IS, including subpopulations entirely characterized by resilience or susceptibility. In conclusion, our study elucidates intricate relationships among classical depression-related behavioral measures, highlighting multidimensional individual variability. Our work emphasizes the importance of a multivariate framework for behavioral assessment in animal models to understand stress-related neuropsychiatric disorders.
Keywords: Animal models; Anxiety; Depression; Machine learning; Multivariate analysis; Stress coping.
© 2024. The Author(s).