At every moment there is something a person/animal is trying to do (a goal) and a reason they are trying to do it (a context for that goal). In the Affect Management Framework (AMF; Haynes-LaMotte, 2025), contextualized goals are constantly shifting in the brain, informed by the senses of the world and the body (vision, hearing, touch, taste, smell, interoception, and proprioception) as well as the semantic factors of meaningfulness, certainty, and agency.

Because our affect is attached to our goals, what contextualized goals we take on and how and when we choose to pursue or relinquish across similar situations can be described as different affect management policies.

You can find an overview of the AMF here. In this post, I hope to introduce some research speaking to the neuropsychological basis for contextualized goals:

Our contextualized goals are always updating as part of ongoing experience, and can be thought of as a feature of the intrinsic brain activity that is always engaged, regardless of what someone is currently doing (Sadaghiani & Kleinschmidt, 2013). In fact, a considerable amount of the body’s energy (about 20% in an adult human) goes to meeting the metabolic costs of momentary brain activity, with the vast amount of this coming from intrinsic activity that is always present, even when unconscious (e.g., while asleep, under anesthesia; Raichle, 2015a). In terms of the hierarchical Bayesian estimation central to Predictive Processing and Active Inference, Sadaghiani and Kleinschmidt (2013) describe that “these intrinsic activity fluctuations reflect the dynamic nature of the underlying internal model. This model does not remain locked in a stationary mode but stays malleable by continuously exploring hypotheses regarding future experience and action” (p. 382).

Along these lines, it is worth considering what factors determine whether something constitutes prediction and or prediction error at a given moment within the Predictive Processing perspective. For example, pain is often considered interoceptive prediction error in the model (Kiverstein, Kirchhoff, & Thacker, 2022), but while this generally seems true, it does not account for the existence of self-harm and disordered eating behaviors that use pain and hunger signals as an interoceptive evaluation of the goals they support (Armey, Crowther, & Miller, 2011; Hooley & Franklin, 2018; Kay et al., 2020; Swerdlow, Pearlstein, Sandel, Mauss, & Johnson, 2020).

I would argue that it makes sense for the two consciously describable factors that determine the prediction/error designation to be one’s (1) goals and (2) expectations (in fact, expectations are most often the framework used for intuiting what would be considered prediction error). Goals represent the momentary outcome that the individual is seeking within the context of their environment, and expectations represent a sense of what the likely consequences of an........

© Psychology Today