Neuroscientists Propose Brain Categorizes World Based on Body's Energy Needs, Not Just Senses

Neuroscientists Lisa Feldman Barrett of Northeastern University and Earl Miller of MIT have published a new framework in Nature Reviews Neuroscience that redefines how the brain categorizes sensory information. The traditional model holds that the brain passively receives sensory data, extracts features, and matches them to stored templates, much like a filing cabinet. Barrett and Miller argue instead that categorization is an active, predictive process driven by the body's immediate physiological needs and energy constraints.

The researchers propose that the brain projects categories onto the world to guide behavior that maintains energetic resources, rather than deriving categories solely from external stimuli. Before conscious awareness of a sensation, the brain prepares the body for action based on predictions shaped by past experience and current internal states such as hunger, fatigue, or arousal. This means identical sensory inputs — like a scratch on the leg or a rhythmic sound — can be categorized as a threat or a neutral event depending on the body's predicted needs.

Their framework integrates Barrett's work on allostasis — the predictive regulation of energy use — and Miller's research on predictive coding, which views perception as the brain's predictions dominating sensory signals unless a mismatch creates a prediction error. They hypothesize that categorization does not reside in a single brain region but operates across the entire nervous system, from the cortex to the body's periphery. Anatomical evidence supports this: feedback connections carrying internal predictions vastly outnumber feedforward sensory connections, even in early sensory areas like the visual cortex.

Electrophysiological data from Miller's lab shows that traveling brain waves carrying information about goals, plans, and bodily state interact with waves carrying sensory data. This interaction allows the same sensory input to be categorized differently based on internal context. For example, an overripe apple may be categorized as food when the body signals energy deficit, but ignored when satiated. The limbic core, deeply connected to the hypothalamus, is identified as a key source of these predictive signals, linking bodily regulation directly to cortical categorization.

Luiz Pessoa of the University of Maryland called the framework a fresh perspective, highlighting the focus on energetic constraints as fundamental to category structure. Sandra Reinert of University College London noted that internal context remapping category rules is intuitive and supported by rodent studies. Timothy Buschman of Princeton University emphasized that categorization serves the current task and the meaning an object holds for the organism's goals.

The authors acknowledge that their framework synthesizes existing anatomical, electrophysiological, and imaging research rather than presenting new experimental data. They argue that understanding categorization requires studying the whole nervous system in the context of the body's survival needs, moving beyond models that treat the brain as a passive decoder of an objective world.