Why this page is an instrument and not a diagram
A diagram of pain would be a lie of confidence. The honest picture is a set of very different kinds of evidence: anatomy that is settled, acute responses that are measured but not pain-specific, cohort associations that hold for groups and not for individuals, circuits known mostly from animal work, and datasets so new their pipelines have not run. So this page renders evidence classes rather than conclusions, and hands you the certainty control. Set it to consensus and most of the stage goes dark. That is not the visualization failing; it is the visualization working.
What the classes mean
Measured is solid geometry generated by this repository from an identified dataset, like the corticospinal streamlines in the spinal-cord lens. Atlas is translucent reference anatomy, here the HCP-MMP1 parcels. Association is a dashed, undirected line: a statistical relationship some study reported in some cohort, never an arrow, because correlation carries no direction. Schematic is dotted and says so: the ascending pain pathway is dotted because the tract atlas this site ships contains no measured spinothalamic tract, and drawing a confident one would be a fabrication. Pending marks science that is underway, including this site's own unbuilt data layers. Nothing pending gets painted.
The measured layers, arrived and still coming
Three measured layers are now built from open data by deterministic
scripts in this repository. CoSpine (OpenNeuro
ds005883, CC0): the Perception lens paints the group
response to noxious heat on the right hand in 39 healthy volunteers,
computed as a voxelwise one-sample t over the dataset's own shipped
per-subject contrast maps, with the display threshold in your hands; the
cervical-cord map from the same dataset is still to come, in its own
space. DEIPP (OpenNeuro ds006815, CC0): the
one-person lens plays real continuous ratings of spontaneous pain,
23 sessions per person over more than half a year; the decoding results
stay quoted from the paper, since reproducing them needs the raw fMRI.
Skull marrow TSPO (OpenNeuro ds007907, CC0): the
persistence lens, at the frontier tier, shows descriptive group-mean
maps of the immune marker on a calvarial dome meshed from the data, with
the published covariate-adjusted analysis carrying the inference.
Reference guide: pain, from receptor to treatment
1. What pain is. Pain is an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage. That definition, from the International Association for the Study of Pain, does three jobs at once: it makes pain an experience rather than a signal, it makes the emotional component constitutive rather than a complication, and it separates pain from nociception, the neural detection of damage. Nociception can occur without pain, under anaesthesia or in the heat of an emergency, and pain can occur without nociception, as in phantom limb pain. Neither is the other's measurement. (Raja et al., PAIN 2020, doi:10.1097/j.pain.0000000000001939.)
2. Detection. Nociceptors are pseudounipolar sensory neurons with cell bodies in the dorsal root and trigeminal ganglia. One axon branch ends freely in skin, muscle, joint and viscera, carrying transducer channels tuned to damaging heat and cold (TRPV1, TRPM8), reactive chemistry (TRPA1), acid (ASICs) and mechanical force; the other branch enters the spinal cord. Thinly myelinated A-delta fibres carry the fast, sharp, well-localised first pain; unmyelinated C fibres carry the slow, burning, diffuse second pain. After injury, inflammatory mediators lower nociceptor thresholds, which is peripheral sensitization: protective in the short term, a problem when it outlives its purpose. People born with a nonfunctional NaV1.7 sodium channel have nociceptors that will not fire, and do not experience pain at all, the cleanest human evidence that these cells gate the whole cascade. (Dubin and Patapoutian, J Clin Invest 2010, doi:10.1172/JCI42843; McDermott et al., Neuron 2019, doi:10.1016/j.neuron.2019.01.047.)
3. The cord. Nociceptive afferents terminate in the dorsal horn, which is a processing station, not a relay: local circuits, descending inputs and non-nociceptive afferents converge there and decide what ascends, the insight behind gate control theory and behind TENS. Even the ganglion itself filters: the branch point of the sensory axon can stop spikes from passing toward the cord (Hao et al., PLoS Biol 2023, doi:10.1371/journal.pbio.3001958). Projection neurons cross the midline and ascend in the anterolateral system to the thalamus and brainstem.
4. The brain. There is no pain cortex. Acutely painful events evoke a distributed response: S1 and S2 for where and what, the posterior insula for the body's state, the anterior insula and midcingulate for salience, urgency and action, plus thalamus, prefrontal cortex and cerebellum. Much of that response is shared with any sudden, biologically important event, which is why region-level activity cannot be read as pain (Iannetti and Mouraux, Exp Brain Res 2010, doi:10.1007/s00221-010-2340-1). Multivariate patterns can track acute evoked pain at group level (Wager et al., NEJM 2013, doi:10.1056/NEJMoa1204471), but no imaging measure is validated to read an individual's pain: across 49 systematic reviews, no clinically validated biomarker of chronic pain exists (J Clin Med 2026, doi:10.3390/jcm15020550). Self-report remains the reference standard.
5. Mechanism types. Clinically, pain is sorted by mechanism because treatment follows mechanism. Nociceptive pain arises from tissue damage with an intact alarm system. Neuropathic pain arises from damage or disease of the somatosensory system itself: shingles neuralgia, diabetic neuropathy, radicular pain. Nociplastic pain arises from altered processing without clear tissue or nerve damage, with fibromyalgia the canonical example. Real presentations mix. ICD-11 separately recognises chronic primary pain, where pain is the disease, from chronic secondary pain, where it accompanies another condition.
6. Chronification. When pain persists, the systems carrying it change. Longitudinal imaging of subacute back pain found that connectivity between medial prefrontal cortex and nucleus accumbens predicted whose pain persisted (Baliki et al., Nat Neurosci 2012, doi:10.1038/nn.3153), that the brain activity accompanying ongoing pain shifted from sensory toward emotional circuitry as pain became chronic (Hashmi et al., Brain 2013, doi:10.1093/brain/awt211), and that white-matter structure and corticolimbic anatomy related to risk (Mansour 2013; Vachon-Presseau 2016). Pooled across hundreds of studies, the most consistent convergence is on the insula (Zeng et al., J Pain 2025, doi:10.1016/j.jpain.2024.104740), yet network meta-analysis finds no single chronic-pain connectome, only low-certainty local effects with high heterogeneity (Butry et al., Front Neurosci 2025, doi:10.3389/fnins.2025.1672542). Both facts matter: chronic pain is reproducibly brain-related, and it is not one brain state. Emerging work extends the story beyond neurons: an immune-cell marker in skull bone marrow is elevated in chronic back pain and knee osteoarthritis (ds007907; medRxiv 2025.07.19.25331817), frontier-tier evidence that the immune system participates.
7. Modulation. The brain regulates its own input. Circuits from prefrontal cortex, hypothalamus and amygdala through the periaqueductal grey and rostral ventromedial medulla reach back to the dorsal horn and can amplify or suppress what ascends (Ossipov et al. 2014, doi:10.1097/SPC.0000000000000055). Expectation, attention, learning and context work partly through this hardware, which is why placebo analgesia is genuine relief with measurable brain correlates (Tetreault et al., PLoS Biol 2016, doi:10.1371/journal.pbio.1002570), and why the same injury hurts differently on different days.
8. Individuality. Two people with chronic pain, scanned across more than half a year, each yielded a personal model that tracked their own spontaneous pain and failed on the other person (Lee et al., Nat Neurosci 2026, doi:10.1038/s41593-026-02221-3). Precision measurement of individuals works; a universal readout does not exist.
9. Treatment. Because pain lives at many levels, treatment acts at many levels, and the honest organising question is where. Guidelines put movement, education, CBT and multicomponent self-management first for chronic musculoskeletal and chronic primary pain (APA 2025; NICE NG193), mechanism-matched pharmacotherapy for neuropathic pain, procedures and neuromodulation where selection criteria support them (AAPM 2026 consensus, doi:10.1093/pm/pnag076), and pain reprocessing therapy showed large effects in one randomized chronic-back-pain trial (Ashar et al., JAMA Psychiatry 2022, doi:10.1001/jamapsychiatry.2021.2669). A treatment changing the brain is evidence it works, never evidence the pain was imaginary.
The interactive instrument above renders exactly these claims, each with
its evidence class and cohort; the full record set, with limitations, is
data/pain-literature.json, and the datasets behind the site
are on the citations and credits page.