We focus on the connection between internal processes of regulation and their influence in the formation of behavioral patterns, and how they serve to provide a basis from where a biological organism derives a relation of value with its own activity. We touch on ongoing work exploring simulation modelling approaches to try to understand more about the issues required to overcome in order to capture the circularity involved in these processes, with particular attention to learning and sequential decision making formalisms of the perception-action loop, such as reinforcement learning and active inference. We describe an experimental setup where an agent learns to act in order to maintain the set of constraints that emerge and unfold as a consequence of the particular niche it inhabits.