Feedback loops and circular causality: a grammar to restore harmony to physiological systems

Causal inference and its graphical representation using directed acyclic graphs provide invaluable tools for calculating effects and handling confounding in medicine. However, since biological systems involve complex dynamics characterized by numerous feedback loops, practitioners need to abstract out the inner workings of cycles in order to reason with a causal framework. This text presents a step beyond linear causality, toward a language that can adequately capture the circular complexities underlying human health. The most fundamental homeostatic mechanisms are inherently cyclic, including sleep-wake states, hormonal regulation, hemodynamics, respiration, blood cell production, eating, and digesting. In fact, numerous medical interventions aim to restore harmonious activity within such systems, reversing dysfunctional states by appropriately modifying certain elements. These include sleep hygiene, synchronized cardioversion, mechanical ventilation, glycemic periodic control with insulin. This study describes medical contexts using cyclic graphs, covering interventions such as control of hormonal cycles in reproductive medicine. Health practitioners can benefit from this enriched perspective, which parsimoniously describes the fundamental behavior of disease processes.