How the Brain Reads the Body: New Theory Reshapes the Science of Interoception After Early Adversity

Every heartbeat, every flutter of the gut, every shift in breathing sends a continuous stream of signals upward into the nervous system. How the brain handles that stream — sensing it, predicting it, and deciding what it means — may be one of the most consequential processes in human health. In a new perspective article published in Translational Psychiatry, Paul W. Savoca, Jennifer A. Silvers, and Bridget L. Callaghan of the University of California, Los Angeles argue that the field of interoception research has been tangled in a conceptual knot, and that untangling it could reshape how scientists understand both mental and physical illness, particularly in people who experienced adversity early in life.

Interoception, the process by which the nervous system senses, perceives, and models the internal state of the body, has increasingly been proposed as a transdiagnostic mechanism — a single underlying process that cuts across diagnostic categories and contributes to a wide range of mental and physical health disorders. But the construct is inherently multidimensional, spanning everything from the raw detection of visceral signals to conscious awareness of them, and this complexity has created persistent conceptual and measurement challenges. The result, the authors contend, is a field that sometimes talks past itself.

The clearest illustration of that confusion is a recurrent debate over whether more interoception is inherently beneficial. Some studies suggest that people who are more attuned to their bodily signals enjoy better emotional and physical health, while others find that heightened bodily awareness is associated with more anxiety, more symptom reporting, and worse outcomes. For years, researchers have traded apparently contradictory findings without fully resolving why heightened sensitivity to the body could be a blessing in one study and a burden in another.

Savoca and colleagues argue that this tension arises from ambiguity between two distinct constructs that the literature has tended to collapse into a single notion of interoception. The first they label interoceptive modeling: the underlying process by which the nervous system predicts and tracks sensory information arriving from the periphery in order to generate a model of the current or future state of the body. This is, in their framing, a fundamentally predictive enterprise. The brain does not passively receive bodily signals; it actively anticipates them, compares incoming sensory evidence against those anticipations, and updates its internal model of the body’s state accordingly. Good modeling, in this view, means that the brain’s predictions about the body track reality closely — that anticipated heartbeats match actual heartbeats, that expected gastric signals correspond to what the gut is actually doing.

The second construct is interoceptive categorization: the categorization and interpretation of interoceptive information. Once the nervous system has a model of the body’s internal state, that state must be given meaning. A pounding heart can be categorized as excitement, as fear, as the aftereffect of climbing stairs, or as the onset of a medical crisis. The same modeled signal can therefore support very different experiences and behaviors depending on how it is interpreted. Categorization is where bodily sensation meets language, memory, context, and belief — and where individual histories exert their strongest influence.

Crucially, the authors emphasize that these are not independent processes. Modeling feeds categorization, and categorization can, in turn, shape how the system models future signals. But keeping the two conceptually distinct, they argue, dissolves the paradox that has haunted the field. Health risks, on their account, are more likely to stem from poor interoceptive modeling, from maladaptive interoceptive categorization, or from some combination of the two — rather than from an excess of interoception broadly. Better modeling, properly understood, should always be beneficial for mental and physical health, because a model that accurately tracks the body gives the organism reliable information to act on. What harms people is not accuracy itself but either a model that misrepresents the body’s state or an interpretive layer that assigns threatening or otherwise maladaptive meanings to accurately modeled signals.

This reframing has immediate consequences for measurement. Many widely used interoception tasks and questionnaires were designed under the assumption that interoception is a single capacity that can be scored as more or less. If modeling and categorization are distinct, then a heartbeat-counting task, a confidence rating, and a self-report questionnaire about bodily awareness may be tapping different constructs entirely — which would help explain why correlations among these measures are often disappointingly weak. The framework suggests that researchers should specify which construct a given instrument targets, and that studies reporting heightened interoception in clinical populations should ask whether heightened modeling, heightened categorization, or both are actually at work.

To demonstrate the framework’s explanatory power, the authors turn to early-life adversity. Children who grow up in environments marked by abuse, neglect, poverty, or chronic instability encounter bodies and worlds that are unpredictable and, in some cases, genuinely dangerous. The authors propose that features of such early environments can shape both modeling and categorization processes in ways that confer vulnerability to later mental and physical health challenges. On the modeling side, an environment of chronic threat may bias the predictive system toward anticipating danger from the body itself — amplifying the gain on certain visceral signals, or calibrating predictions around states of high arousal. On the categorization side, early experiences supply the interpretive templates through which later bodily states are understood. A child who learned that a racing heart preceded harm may, years later, categorize ordinary exertion-induced cardiac changes as signals of impending catastrophe, a pattern recognizable in panic and anxiety disorders.

The adversity framework also offers a way to think about physical health. Allostatic load — the cumulative wear on physiological systems produced by chronic stress — could interact with interoceptive processes in both directions. Poor modeling might delay the detection of genuine bodily problems, from hunger and satiety disruptions to emerging illness, while maladaptive categorization might amplify symptom perception and healthcare-seeking in unhelpful ways or, conversely, lead individuals to dismiss signals that warrant attention. Because interoception sits at the interface of brain and body, the authors suggest it is a plausible mechanistic bridge between the well-documented psychological consequences of early adversity and its equally well-documented physical health consequences, from cardiovascular disease to metabolic dysfunction.

The implications extend to intervention. If maladaptive categorization is the primary driver of distress in a given individual, then treatments that target interpretation — cognitive reappraisal, exposure-based therapies, or retraining the emotional meaning of bodily sensations — may be most effective. If poor modeling is the problem, then interventions that improve the fidelity of bodily predictions, such as certain forms of body-focused mindfulness practice or biofeedback, may be more appropriate. Distinguishing the two constructs in the clinic could therefore help match patients to the treatments most likely to help them, and could guide the design of prevention programs for children exposed to adversity before maladaptive patterns consolidate.

The authors are careful to present their framework as a conceptual advance rather than a settled empirical conclusion. It generates testable predictions — that modeling accuracy and categorization style should dissociate in specific clinical populations, that adversity should leave distinguishable fingerprints on each construct, and that interventions should show construct-specific effects — and it invites the field to develop measurement tools capable of testing them. But if the distinction holds up, it resolves one of interoception science’s most stubborn puzzles with a deceptively simple insight: the question was never whether more bodily awareness is good or bad. The questions are whether the brain’s model of the body is accurate, and whether the meanings assigned to that model serve the person or betray them. For the millions of people carrying the physiological legacy of early adversity, that distinction could determine which treatments work, which risk factors matter, and ultimately how the dialogue between brain and body goes wrong — and how it might be repaired.

Subject of Research: Interoceptive modeling and categorization as distinct mechanisms linking early-life adversity to health outcomes

Article Title: Understanding interoceptive modeling and categorization following early-life adversity

Article References: Savoca, P. W., Silvers, J. A., & Callaghan, B. L. (2026). Understanding interoceptive modeling and categorization following early-life adversity. Translational Psychiatry. https://doi.org/10.1038/s41398-026-04517-1

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04517-1

Keywords: interoception, early-life adversity, predictive processing, mental health, Translational Psychiatry, UCLA, body-brain communication, anxiety, measurement, transdiagnostic mechanisms, childhood stress, intervention