Advancing translational phenotyping of mouse models for psychiatric research by automated home-cage monitoring

The translational validity of animal models in psychiatric research remains a persistent challenge, partly due to reliance on short-duration, experimenter-driven behavioural assays that capture limited and often artificial snapshots of complex phenotypes. Home-cage monitoring (HCM) systems offer a paradigm shift by enabling continuous, non-invasive assessment of mice within their familiar social and environmental context. These approaches facilitate longitudinal quantification of behavioural domains highly relevant to psychiatric disorders, including circadian rhythm disruption, social interaction, anhedonia-related behaviours, and stress responsivity. Advances in machine learning, sensor integration, and computational analysis now allow scalable, high-resolution phenotyping across extended timeframes, frequently in group-housed conditions. Recent collaborative initiatives have catalysed efforts towards methodological standardization and interdisciplinary integration, highlighting the potential of HCM to enhance both construct and external validity in preclinical models of mental illness. Importantly, continuous monitoring captures behavioural dynamics and intra-individual variability that are largely inaccessible to conventional tests, offering novel endpoints with potential translational relevance to human symptom trajectories. Nevertheless, challenges remain, including harmonisation of analytical pipelines, interpretation of high-dimensional data, and alignment of behavioural readouts with clinically meaningful constructs. This review evaluates the current state of HCM technologies in the context of psychiatric modelling, discussing their potential to refine phenotyping strategies and enhance the reproducibility and translational impact of mouse models.