At every moment of life, the body is filled with various sensory signals. Organisms can process this data because the brain—through a process called "categorization"—transforms the messy, noisy, and information-rich world into concepts that people can understand and apply experience to deal with.
People's categorization of the world is actually quite flexible. In traditional neuroscience views, categorization is thought to occur at the final stages of sensory processing: the brain receives information, decodes it, and matches it with memory. But this view cannot explain why the brain processes similar sounds differently in different contexts.
In Nature Reviews Neuroscience, two leading neuroscientists, Lisa Feldman Barrett and Earl Miller, present their latest understanding of brain function and structure, addressing this issue. They collaboratively propose a new perspective describing how the brain continuously reconfigures its categories based on sensory input and memory.
Figure 1: Traditional view of categorization.
The researchers argue that the brain projects these categories onto the world based on the body's survival needs. Before people are conscious of their sensory impressions, the brain is already preparing the body to act in ways that maintain the body's physiological energy resources. Thus, the brain's predictions—rather than the cumulative effects of sensory information—ultimately shape and constrain how people categorize objects and features in the world.
In this framework, categorization does not occur in a specific brain region but across the entire organ and beyond. To understand how organisms form categories, neuroscientists must examine the nervous system comprehensively from head to toe.
Heavyweight Collaboration
For years, these two influential neuroscientists have worked in the same field but never directly collaborated. Miller has measured high-level electrical activity patterns in the brain to better understand the mechanisms behind a neural computation model called predictive coding. Predictive coding posits that our senses are predictions the brain makes about the environment, rather than passively received sensory signals. Under normal sensory conditions with no new input, the brain generates predictive signals about the environment (known as feedback signals), which dominate incoming sensory information (or "feedforward" signals). But if something unexpected occurs—an event that deviates from the predictive model—sensory feedforward signals, such as those reaching the visual cortex, conflict with predictive feedback signals.
Figure 2: Predictive signal flow in category construction.
This discrepancy leads to prediction errors. When sensory signals violate the predictive model, they may enter conscious experience as feelings of "surprise." In her work, Barrett applies the concepts of prediction and forecasting to theoretical understanding of emotions. In her view, emotional categories—such as fear, joy, and anger—resemble predictive "action plans" generated by the nervous system to activate behaviors that have been beneficial in the past. Traditionally, emotions were thought to be hardwired into specific neural circuits. Barrett's research has helped show that emotional categories are constructed from signals from both inside the body and the external environment. In 2025, Barrett contacted Miller, and they agreed to jointly construct a new categorization framework—going beyond the traditional file-cabinet model, incorporating their theories of prediction and adaptation.
Adaptive Categorization
As animals, humans must adapt to survive in change—whether from changing environments or changing bodies. Due to constraints on energy, time, and other resources, humans must take shortcuts to sustain survival, and this is exactly the role of categories. Barrett says that categorization functions to use past experiences similar to new contexts to guide behaviors that maintain the body's systems. It also includes a set of behavioral rules for how to use experience to interact with new situations. But how are categories formed? According to traditional views, over time, experiences build a set of generalized attributes that form a category. However, these scientists argue that before the senses recognize traditional categories, the brain uses memories of similar past situations to predict a set of behaviors appropriate for the new context. Imagine a person walking down a road when something touches their leg. In a safe environment, this might be nothing. But if they are in a completely unfamiliar forest surrounded by tall grass, they feel their breathing and heartbeat become more rapid than usual. This is a feeling of tension, because the brain has already recognized a potential threat in the environment. As a result, a touch identical to one that would be negligible in ordinary circumstances is categorized completely differently.
The scientists' research builds on anatomical, electrophysiological, and brain imaging findings from their laboratories and other studies over the years. They propose that categorization is a continuous predictive tool through which the nervous system directs behavior to maintain the organism's survival.
Figure 3: Intrinsic brain networks regulating homeostasis.
Evidence comes from the neuronal level. They cite studies showing how brain connectivity is biased toward predictive signals over sensory signals. For instance, connections between neurons that promote feedback flow of internal signals far outnumber those processing sensory feedforward signals. Even within the visual cortex, 90% of synaptic connections facilitate feedback signal transmission. This suggests the brain is structured to predict categories, perceptions, and behaviors, rather than to reactively categorize stimuli. Related link: https://onlinelibrary.wiley.com/doi/10.1002/cne.23458
According to these scientists, the interactions between information waves in the brain may explain why the same set of external sensory signals is categorized differently in different contexts.
Ubiquitous
A core component of Barrett and Miller's framework is the origin of predictive signals in the brain. In most neuroscience literature on visual categorization, cortical integration areas (located on the brain's outer surface) are considered the source of predictive expectations. However, they take a different approach: they treat the limbic core as the source of these predictive signals.
Figure 4: Limbic-sensory cytoarchitectonic gradient of the cerebral cortex.
The limbic core includes deep brain structures that combine signals carrying information from the body, senses, memory, and higher cognitive functions. This integrative network is both close to and highly connected with the hypothalamus. Because the hypothalamus helps convey the body's energy state to the cerebral cortex—the brain's outermost folded layer responsible for higher cognition like executive control—the authors hypothesize that the limbic core contributes to the brain's ability to anticipate the body's energy needs. The limbic core is the junction where internal and external signals compress and converge. You can think of the nervous system as two funnels shaped like a bow tie, meeting at the narrowest point—the limbic core, which is the point of highest compression. Just as visual information is compressed as it travels through the visual cortex and deeper into the brain, physiological signals are compressed as they travel from the body via the vagus nerve into the brain. They then reach the limbic system core, where they integrate with compressed sensory information. Together, these signals generate an appropriate category and its associated behaviors. Subsequently, depending on the signal category, the interaction between predictive feedback signals and compressed sensory feedforward signals can occur in any neural region.
Barrett and Miller urge people to abandon the idea that categories are stored like files in the brain. They argue that from the way the brain compresses sensory information to how it predicts and prepares the body to respond, categorization is already "internalized." It is a core organizing principle of the nervous system that, through memory, perception, and behavior, helps people survive better in an information-rich world.
Original article: https://www.quantamagazine.org/a-new-framework-for-how-the-brain-compresses-our-noisy-world-20260824/