Directed Acyclic Graphs for Causal Modeling of Musical Interventions in High-Tropical Grazing Dairy Cattle: A Causal Framework Review

Abstract

Musical exposure in dairy cattle has been associated with potential effects on milk production, behavior, and physiological regulation; however, available evidence remains heterogeneous and potentially affected by biological and environmental confounding. In high-tropical grazing systems, these responses are simultaneously influenced by stress physiology, disease, nutrition, management, climatic variability, and individual adaptation. This review integrated causal inference and graphical epidemiology to develop biologically grounded directed acyclic graphs (DAGs) representing hypothesized causal pathways between musical exposure and milk production, behavior, and heart rate variability (HRV). An analytical narrative synthesis of physiological, ethological, autonomic, and epidemiological evidence was conducted, and three independent DAGs were constructed using DAGitty version 3.1. Variables were classified according to their epidemiological roles. The production DAG identified stress as a major integrative node linking disease, nutrition, reproductive physiology, management, and milk yield. The behavioral DAG highlighted the roles of ecological organization, resource access, social interaction, and environmental instability. The HRV DAG treated HRV as beat-to-beat variation in cardiac intervals that provides a non-invasive index of autonomic nervous system (ANS) modulation, including sympathetic and parasympathetic influences, and indicated that responses may occur through indirect modulation of stress physiology and vagal activity. We conclude that musical interventions should be interpreted as context-dependent sensory exposures rather than isolated causal determinants; the proposed DAGs provide an explicit basis for selecting covariates, avoiding inappropriate adjustment, and designing future studies of dairy cattle welfare and production.