How can subjective cognitive states be modeled in animal decision making without treating heuristic rules as actual neural machinery?

A shortcut is a description of behavior, not a piece of the brain

A heuristic is a decision shortcut: a quick, economical way of choosing that usually gets a good-enough answer. In human settings these shortcuts are often studied as patterns in what people decide. Judges, for example, show measurable tendencies toward availability, confirmation, representativeness, and anchoring, and their self-reported use of the anchoring shortcut correlates with a separate test of reflective thinking [1]. Executives can likewise lean on a "superstitious heuristic," a shortcut built on superstitious belief that can produce puzzling strategic choices, especially under high uncertainty [3].

The important point for animal modeling is that these are descriptions of a choice tendency, not an account of the machinery underneath. A superstitious heuristic is defined by the belief it rests on and the decisions it produces, and it is explicitly distinguished from other shortcut families such as heuristics-and-biases and fast-and-frugal rules [3]. Nothing in that definition says a particular neural circuit implements the rule. The same caution applies when heuristics are discussed alongside algorithmic reasoning in human decision-making: the shortcut is a level of description, and treating it as a physical mechanism is a separate, stronger claim [2].

Emotion can carry the subjective state that a rulebook cannot

If heuristics are descriptions, what supplies the subjective, felt part of an animal's decision? One proposal puts emotion at the center. The hypothesis is that vertebrate decision-making happens in two steps: emotions compete to set the organism's current priority, and then the animal chooses the behavior that maximizes its imagined near-future emotional wellbeing, drawing on memory of specific past episodes [4]. Emotion here is described as a cognitive mechanism that can temporarily become a bodily state controlling the body and driving behavior [4].

This is a hypothesis about how decisions are organized, not a demonstrated circuit. It is presented as a proposed modeling approach, and the authors frame it as a way to capture subjective experience in models that can apply to animals in natural and artificial settings [4]. The value for the reader is conceptual: emotion gives the model something internal and felt to work with, while the observable behavior remains the thing being predicted.

What this buys a modeler, and where the evidence stops

The practical gain is that you no longer need to claim the animal stores a list of rules. You can let a subjective state—the strongest current emotional need—focus attention and set priority, and then let imagined emotional outcomes guide the choice [4]. A heuristic pattern such as "avoid the risky option" or "repeat what worked last time" can then emerge as a summary of those choices rather than as an installed program. That keeps the description and the mechanism at separate levels, which is exactly the confusion the question is asking how to avoid.

The evidence for this specific framework is a proposed model, not a body of animal experiments confirming it [4]. The human work on heuristics is likewise about reported and measured choice tendencies, such as judges' self-reported shortcut use and its modest correlation with reflective-thinking scores [1], or about a conceptually defined superstitious shortcut in executives [3]. Those studies support the idea that shortcuts can be described without neural claims; they do not test the emotional-currency model in animals. The review linking generative AI to heuristic and algorithmic thinking is also about human judgment, not animal cognition [2].

Where subjective states fit, and what stays open

The framework's distinctive move is to treat awareness of one's strongest emotional need as a higher level of agency than the unconscious robustness mechanisms found in all life forms, and to treat imagination-based prediction as the tool for maximizing emotional wellbeing [4]. That is a claim about a possible architecture of vertebrate decision-making, offered as a hypothesis. It does not establish that any particular animal experiences emotion in the way the model assumes, and it does not specify the neural implementation.

So the honest answer to the modeling question is a division of labor. Use heuristics as compact descriptions of choice patterns, as the human studies do when they measure shortcut use or define a superstitious shortcut [1][3]. Use a subjective variable such as emotion as the internal state that organizes those choices, as the vertebrate framework proposes [4]. Keep the two levels explicitly separate, and treat the emotional-currency account as a promising hypothesis still awaiting direct tests in animals rather than as settled machinery.

Sources used in this answer

Related research questions