Summary: A new theoretical framework in neuro‑philosophy and cognitive neuroscience challenges the longstanding consensus on how the brain produces choices and actions. This perspective disputes the linear “sandwich model” that treats decisions as a distinct, centralized cognitive stage sitting between sensory perception and motor output. Instead, it argues that what we call decisions emerge from distributed, simultaneous interactions among sensory, sensorimotor, and motor systems tightly coupled with the body and environment.
The framework shows there is no single localized neural process that corresponds to decision‑making. Rather, apparent choices are emergent patterns produced by circular, real‑time loops of perception and action. This view calls for new experimental approaches that study decision-related behavior in embodied, ecologically valid contexts.
Key Facts
- The Sandwich Model Fallacy: The common intuition and many traditional theories depict cognition as a linear sequence—sensory input, a central decision stage, then motor output. This simple model is increasingly at odds with empirical evidence about brain function.
- No Dedicated Decision Center: Neuroscience identifies specialized neural networks for sensing and for producing actions, but it does not find a discrete, centralized neural mechanism that literally “makes decisions.”
- Action Selection Rather Than Command: Instead of a top‑down controller, behavior is better described as action selection driven by the interplay of sensory, sensorimotor, and motor dynamics.
- Decisions as Abstract Descriptions: Under a physicalist view, decisions are abstract, nonphysical descriptions of behavior—useful for explanation but not causal agents that can push muscles or change neurons.
- The Cartesian Theater Problem: Postulating an inner central controller leads to the philosophical paradox of a “miniature homunculus” inside the brain, an infinite regress that explains nothing about how neural systems actually operate.
- Robot Demonstration: A simple robot built from basic sensory and motor modules can produce apparent goal‑directed wall‑following behavior without any internal decision module, demonstrating how intentional‑looking actions can emerge from simple loops.
- Embodied, Ecological Methods Needed: Tom James of Indiana University argues cognitive neuroscience should move away from strictly linear models and adopt experimental designs grounded in embodied cognition and ecological psychology to study continuous sensorimotor coupling.
Source: Indiana University
A fundamental mismatch between intuition and mechanism
Tom James, professor in the Department of Psychological and Brain Sciences, points out a persistent mismatch between how people describe making choices and what neurobiology reveals. Common sense and many lab models treat decisions as an intermediate stage between perception and action, with each stage mapped to a separate brain process. We speak naturally in terms of desires, beliefs, and intentions causing our actions—so a cognitive decision stage feels real.
But when we examine brain organization, the neat mapping breaks down. Sensory systems and motor systems have identifiable neural circuitry; the supposed central decision process does not. James argues that behavior arises from ongoing sensorimotor interactions rather than from a discrete decision engine. He prefers the term “action selection” to capture how available sensorimotor loops bias behavior toward particular outcomes.
That is not to deny the usefulness of decision language. As James explains, the vocabulary of decisions is indispensable for everyday communication and scientific description. However, the presence of decision‑related concepts does not imply the brain contains an explicit decision‑making mechanism; behavior that is well described as the result of a decision can be produced without any such internal controller.
James lays out these arguments in his article “Sensorimotor Mechanisms of Decisions and Actions,” published in the Journal of Cognitive Neuroscience.
How the argument is framed
James adopts a physicalist stance: only physical processes—sensory input, neural dynamics, muscle activations—can produce physical behavior. Nonphysical, abstract entities like “decisions” cannot themselves exert causal force. He uses analogies to clarify the point. Just as the center of mass is a useful mathematical property that does not itself push an object, so too a decision is an abstract descriptor of coordinated activity but not an independent causal agent.
A second analogy compares high‑level terms (for example, “the university”) to the distributed workings they summarize. Saying “the university acted” is shorthand that omits the concrete events, meetings, and people involved. Similarly, speaking of decisions compresses complex sensorimotor and environmental interactions into a single, abstract term that conceals the processes researchers must study to understand behavior.
A third, empirical illustration uses a simple wall‑following robot. Built only from sensors, motors, and minimal sensorimotor couplings, the robot produces behavior that looks strategic and purposeful. Yet it contains no decision module. This example shows how apparent intentionality can arise from straightforward physical loops.
Positing a mysterious central controller revives the Cartesian Theater paradox: describing the brain as housing a decision maker merely replaces one mystery with another—a tiny person inside the head who would herself require a smaller decision maker, ad infinitum.
An experimental path forward
James concludes by urging methodological change. To study decision‑like behavior accurately, experiments should recreate the embodied, ongoing interactions between organisms and their environments. He recommends ecological testing setups that give participants agency, allow continuous sensorimotor updating, and capture how action selection emerges from active sensing and real‑time loops.
A: Human language and thought are optimized to describe behavior at an abstract level. Saying we “decide” efficiently summarizes coordinated sensorimotor processes. The feeling of making a choice arises from those coordinated processes, not from a single localized control structure.
A: It is the flawed idea that a little homunculus inside the brain watches and decides for the person. This explanation generates an infinite regress—each inner observer would require its own observer—so it fails to explain how cognition emerges from neural systems.
A: The robot demonstrates that behavior that looks intentional or strategic can emerge from simple sensorimotor loops without any built‑in decision logic. This supports the view that human-looking choices could likewise be emergent rather than centrally commanded.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full.
- Additional contextual details were added by editorial staff.
About this neuroscience research news
Author: Elizabeth Rosdeitcher
Source: Indiana University
Contact: Elizabeth Rosdeitcher – Indiana University
Image: Image credit: Neuroscience News
Original Research: Closed access. “Sensorimotor Mechanisms of Decisions and Actions” by Thomas W. James. Journal of Cognitive Neuroscience. DOI: 10.1162/JOCN.a.2484
Abstract
Sensorimotor Mechanisms of Decisions and Actions
Decisions are frequently treated as an intermediary stage between perception and action, but the extent to which this assumption is embedded across cognitive neuroscience varies. This perspective argues that decisions and decision processes do not causally produce actions. Instead, actions arise from sensorimotor processes—continuous, embodied loops that couple perception and movement.
After surveying different interpretations of the causal link between decisions and actions, the article advances the claim that decision processes are not necessary causal agents. It recommends experimental approaches that emphasize ecological validity and active sensing, allowing participants to exercise agency and continuously update sensorimotor routines while researchers measure the dynamics that give rise to action selection.