Why do we feel pain?

Why do we feel pain?

Start with a question that sounds like philosophy and turns out to be clinical. Where does experience begin? For most of the last century the answer was the cortex, the folded outer layer that handles vision, language and thought. Consciousness was assumed to be a high achievement of a high brain region.

The neurologist Antonio Damasio argued the opposite. Working with patients whose brain injuries had blunted their feelings, he found that they made worse decisions, not better ones, and he came to see feeling as the ground that thought stands on. In his account, the inaugural acts of consciousness are not thoughts but feelings about the body: hunger, thirst, itch, warmth, cold, and pain. These arise from a dialogue between the body and the upper brainstem, a region older and deeper than the cortex, and they exist for one reason: to keep the body within the narrow range of conditions it can survive. Pain, on this view, is not a sensory report that happens to be unpleasant. It is a feeling in the strict sense, a body-regulating signal with its own agenda.

Pain researchers reached the same conclusion from another direction. The neuroanatomist A. D. Craig traced the pathways that carry information about tissue state from the body through the brainstem to the insula and anterior cingulate cortex, and found that pain travels with temperature, itch, hunger and breathlessness rather than with touch. He proposed in 2003 that pain is best understood as a homeostatic emotion: a feeling about the body's condition, with a motivation attached, in the same family as thirst. The field's own definition caught up in 2020, when the International Association for the Study of Pain revised it to an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage. The word resembling is doing deliberate work: pain can be present when damage is not.

So the short answer to why we feel pain is that we are the kind of creature that has feelings about its body, and pain is the one that says protect this. The longer answer is in what a feeling is and what it is for, which is where a recent public conversation between the consciousness scientist Anil Seth and the writer Michael Pollan turns out to be a useful map.

Is pain a feeling or a sensation?

Both words get used, and the difference matters more than it looks. A sensation, in the way neuroscientists use the word, is a reading of something: light on the retina, pressure on the skin, a sound in the ear. It can be accurate or inaccurate, and in principle you could check it against the world. A feeling is a reading of how the body is doing relative to what it needs, and it comes with two things a sensation does not: a valence and a pull.

Pain sits firmly in the right-hand column. It is always unpleasant, it always carries an urge to do something, and it travels the interoceptive pathways described in our guide to interoception, not the touch pathways. Seth, whose research centers on the embodied side of consciousness, puts the emphasis in the same place: the most basic thing experience is made of is not vision or hearing but the shapeless feeling of being a living body, and emotion and mood are built on that.

This reclassification has a practical consequence. A sensation can be wrong about the world, but it is at least trying to track the world. A feeling is trying to track the body's needs, and the brain's estimate of those needs depends on more than the tissue: on memory, on context, on what happened last time, on how safe the moment seems. Our guide on whether the brain feels pain covers the anatomy. The point here is simpler. Asking whether a pain is real is a sensation question. Asking what the pain is for is a feeling question, and it is the one that leads somewhere.

What does sentience have to do with pain?

Pollan's book on consciousness opens in an unusual place: with plants. He does this, he explains in the conversation with Seth, because plants are the least plausible case and therefore the clearest. Botanists have shown that a sensitive plant whose pot is shaken repeatedly learns to stop folding its leaves, and remembers the lesson for weeks. Plants respond to the sound of a caterpillar chewing by moving defensive chemicals to their leaves. Several species are knocked out by the same anesthetics that put people under for surgery. Whether there is anything it is like to be a plant, in the philosopher Thomas Nagel's phrase, nobody knows, and Pollan and Seth both say so. Neither claims plants feel pain, and a botanist Pollan quotes points out that pain would be a poor adaptation for an organism that cannot run away.

What the plant case does is strip sentience down to its skeleton. Pollan's working definition has three parts:

1. Awareness of the environment, inside and out: some registration of what is happening to the organism.
2. Valence: a registration of whether a change is good or bad for the organism.
3. Approach and avoidance: the ability to move toward the good and away from the bad.

Even single-celled organisms do this. Consciousness, in Pollan's phrasing, is the way humans do sentience: the same three-part structure, made vastly more elaborate by a body, a memory and a nervous system like ours. The structure does not go away as it gets elaborate. It is still there, underneath, in every feeling.

Now look at pain with that skeleton in mind. Pain is sentience with the volume turned up. It is a registration of the body's state, with the strongest negative valence the system can produce, and the most urgent avoidance signal: stop, withdraw, protect, do not do that again. That is exactly what you would build if you wanted to keep a mobile, breakable animal alive. The botanist's point about plants cuts the same way for people. Pain is for action. It exists to change what the body does next, and it only makes sense in a creature that can do something about it.

The sensitive plant that learns to ignore the shaking shows the other half of the design. A system that registers threat has to be able to stop registering it when the threat turns out to be harmless, or it would spend its life folding up. Habituation, the quieting of a response to a repeated stimulus that leads nowhere, is as old as sentience itself. Chronic pain, as the next section argues, is partly a failure of that ancient skill.

Why does pain keep going after an injury has healed?

If pain were a sensation, a reading of the tissue, this would be a paradox. If pain is a feeling, a protective signal generated by a system that estimates threat, it is not a paradox at all. It is a system that has not updated.

Consider what happens after an injury that frightens someone: a back that gave out while lifting, a neck after a collision. For weeks the protective feeling is appropriate. The tissue is healing, movement really could make things worse, and the pain does its job. But the system producing the feeling is learning the whole time. The spinal cord and brainstem circuits become more responsive, a process described on our page on central sensitization. The brain's model of the back, in the language of predictive coding, acquires a confident prediction: this region is fragile, bending is dangerous. The person, sensibly, avoids the movements that hurt.

Then the tissue heals and nothing tells the system. Avoidance means the prediction is never tested, so it never has the chance to be wrong. Guarded movement produces its own new signals, stiffness and effort and fatigue, which the system reads as further evidence of threat. The fear-avoidance research that Johan Vlaeyen and Steven Linton summarized in 2000 describes the behavioral loop: pain leads to fear, fear to avoidance, avoidance to deconditioning and a smaller life, and all of it back to more pain. In the terms of this page, the avoid half of sentience has become permanent, and the habituation that should have quieted the signal never ran, because the stimulus was never allowed to repeat harmlessly.

This is why so much of chronic pain treatment looks, from the outside, like practising the thing that hurts. Graded exposure, graded motor imagery, somatic tracking and VR embodiment all give the system repeated, safe encounters with the feared stimulus so that the ancient quieting mechanism can finally do its work. Our guide on fear of movement covers the behavioral side; the page on unlearning chronic pain covers the retraining.

What is the smudged windscreen, and why does it matter for pain?

Pollan describes the experience that sent him toward consciousness in the first place. Ordinarily, he says, we look straight through our experience at the world, the way a driver looks through a clean windscreen. Certain states, he mentions psychedelics and meditation, smudge the glass, and for the first time you notice there is a windscreen at all: something between you and reality, shaping what gets through, and you start to wonder why it is this way and not some other way.

Pain is the most transparent windscreen there is. When your knee hurts, you do not experience a feeling about your knee generated by your brain. You experience a bad knee. The pain presents itself as a property of the joint, as plainly as its color or its shape, and the idea that it is something you are producing seems almost offensive. This transparency is not a flaw. A protective signal that announced itself as a construction would be a weaker protective signal. But it has a cost when the signal is wrong, because a feeling you take to be a fact about your tissue cannot be questioned.

A good deal of modern chronic pain care amounts to smudging the windscreen on purpose. Pain neuroscience education does it with information: once a person understands that pain is a protective output rather than a damage readout, each painful moment means something different. The research summarized in our commentary on fifteen years of Explaining Pain suggests that this reframing helps on its own for some people, and is more powerful when paired with movement. Somatic tracking does it with attention: observing a sensation with curiosity, as an event in the body rather than a verdict on it, puts a visible layer between the person and the pain. Acceptance and commitment therapy does it with language, teaching people to notice a thought like my back is ruined as a thought rather than a fact. And experiments by Katja Wiech and others show that these cognitive shifts are not cosmetic: attention, expectation and reappraisal measurably change how a fixed stimulus is felt and which brain systems respond to it.

Two cautions. First, seeing the windscreen does not make the view behind it disappear. Pain does not stop because you understand it; the point of noticing the layer is to gain purchase on it, and most of the work is then done by the body in movement. Second, this is not the same as deciding pain is in your head. The windscreen is real glass. The feeling is produced by a real nervous system responding to real tissue, real memory and real context. What changes is that it is no longer mistaken for the road.

Is the brain a computer running a pain program?

The metaphor is everywhere, and a version of it has reached pain care: the pain is a software bug, a glitch in the brain's code, and the treatment reinstalls the program. It is a tempting picture, because it makes learned pain sound fixable. Seth and Pollan spend a good part of their conversation taking the metaphor apart, and the reasons they give are worth carrying over.

A computer works because hardware and software are cleanly separable: the same program runs on any machine, and you can understand it without knowing anything about the chips. The brain has no such separation. Every memory is a physical change in a network of neurons. Neurons do not simply fire or not fire; they are bathed in chemistry that changes how readily they fire, and they take part in rhythms that have no digital equivalent. Your brain is physically different from anyone else's because your life has been different. Pollan quotes the mathematician Norbert Wiener on metaphors in general, and the line fits this one:

For chronic pain the correction cuts two ways, and both are good news once you see them. First, learned pain is physically real. It is not a line of bad code sitting on top of a healthy body; it is a nervous system that has changed its structure and chemistry in response to experience, which is what the research on neuroplasticity and central sensitization describes. Nobody should feel that their pain is less legitimate because it was learned. Second, because the change was made by experience, it is experience that revises it. Not a reinstall, not a single insight, but repeated embodied evidence: movements that were predicted to hurt and did not, sensations observed without alarm, a body seen doing more than the model expected. Our guide to embodied cognition explains why first-person bodily evidence carries more weight than information, and why approaches like virtual embodiment therapy are built around it.

Seth adds a point that is easy to miss. Consciousness is strikingly easy to lose; a great many things have to be going right, in body and brain together, for experience to be present at all. The same is true of the feeling of being safe in a body. Sleep, mood, stress hormones and the state of the autonomic nervous system all feed into the estimate that pain is built from, which is why pain is worse on a bad night and why a treatment plan that ignores them tends to stall. Our page on nervous system dysregulation covers that side.

What does this mean for treating chronic pain?

If pain is a protective feeling rather than a damage gauge, then the target of treatment is the system that produces the feeling: its estimate of threat, its failure to habituate, and the self it has built around a body it considers unsafe. Tissue still matters and still needs assessing. But once the tissue has been assessed and the pain has outlived it, working on the tissue alone is working on the wrong column of the table.

The three framings on this page each point to a piece of what effective care does. Because pain is a feeling, treatment attends to the body's internal state, through interoceptive retraining, sleep, and the settling of a system stuck on alert. Because pain is valence plus avoidance, treatment is built around graded, repeated, safe return to the avoided movements, so the system can discover that the threat has passed. Because pain is a transparent windscreen, treatment includes education and attention practices that let a person see the feeling as a feeling, and so gain some say in it.

Karuna's program, Karuna Virtual Embodiment Training (KVET)™, is built on these principles. Over 12 weeks at home, on a provided headset, a person does short daily VR embodiment exercises designed to give the brain safe, first-person evidence that the body can move more than it predicts, alongside weekly one-on-one sessions with a pain coach who works on the fear, the avoidance and the story. A clinician reviews every case before the program begins and oversees it throughout. The aim is less pain and more function; many people improve substantially and some do not. The outcome data Karuna has collected, with its limits spelled out, is on the how it works page, and the whole framework is set out in our guide to therapeutic neuroscience.

For the person in pain, the shift is from asking what is wrong with my back to asking what my protective system has learned and how it can be shown something new. For clinicians, it is a reminder that the patient's pain is a feeling in Damasio's sense, as real and as embodied as hunger, and that validating it and retraining it are the same project rather than opposed ones. If your pain has persisted despite medical evaluation and you want to know whether a program built on these principles fits your situation, Karuna offers a free consultation. A clinician reviews every case, and the care team will say plainly if it is not the right fit.