Major depressive disorder (MDD), or clinical depression, is a severe mood disorder associated with persistent sadness, anhedonia and other symptoms, like fatigue. While different rodent models of MDD have been developed, traditional behavioral tests (e.g., the tail suspension, forced swim test and sucrose preference tests) show low construct validity, limiting the translational value of these models. A new Nature Communications study introduces CLOSER (Contrastive Learning-based Observer-free analysis of Spontaneous behavior for Ethogram Representation), a self-supervised machine-learning platform that analyzes spontaneous mouse behavior, recorded in a semi-natural environment, with high precision. The framework, which incorporates 3D pose skeleton coordinates and kinematic features, allowed the identification of distinct behavioral patterns in chronically stressed mice across both sexes and different disease stages. Using CLOSER, the researchers were also able to quantify the pharmacological effects of monoaminergic and non-monoaminergic antidepressants on the psychomotor symptoms of the stressed mice. Altogether, CLOSER offers a standardized approach to assessing depressive-like phenotypes in mice, improving interpretation of behavioral data for future psychiatric drug development.
Original reference: Oh, H. et al. Nat. Commun. https://doi.org/10.1038/s41467-025-67559-x (2025)
This is a preview of subscription content, access via your institution
