Behavioral regulation relies on interacting forces and predictive models

We discuss how our recent neural network model of personality and motivation can explain many aspects of the regulation of behavior.

Contrary to approaches that focus on a goal-corrected, set-point, and discrepancy-reducing mechanism, we argue that many aspects of regulation can be understood in terms of two other mechanisms. First, many aspects of the stability and coherence of personality, as well as the dynamics of personality, can be understood in terms of the interaction of forces within organized motivational systems, and their interaction with features of the environment and interoceptive states, that identify an individual's current needs. This has been described as a settling point or equilibrium of forces model, rather than a set-point architecture. Second, regulation has been shown to depend also on the use of predictive models of the world, either learned or innate. Such models can be thought of as feedforward models, in contrast to the feedback models characteristic of set-point, goal-corrected systems.

We describe a neural network model of these processes that simulates the behavior over time and situations of an individual and shows how important regulatory processes can operate through a process of interactive forces and predictive models of the world.