Stanford Pain Atlas

Pain neuroanatomy

Named white-matter bundles from the HCP1065 atlas, drawn as teaching plates. Pulse timing uses atlas median lengths.

Fig. 1. Brain. Population-average tracts and Atlas_ROIs nuclei. Pick a system in the legend below.

Loading anatomy 0%

Systems

Pick a system to see its parts, one at a time in the order of the name cards, and the bundles that connect them. Pick one or more systems. Connecting tracts stay off until a system is chosen. Click a system card or button to add or remove that system.

Story sketches

Hemisphere

Right matches a sagittal figure: right cortex, right paired nuclei, right named tracts. Brainstem and PAG stay whole (they sit on the midline).

Fibres between two structures

and

Tracts

Point of view

Conduction velocity

Colour the cortex by

Deep nuclei

How solid the cortex is

Fig. 2. Cervical cord. AMU7T occupancy at C5, shown as a filled disk. Dorsal toward the top.

C5 filled occupancy disk. Dorsal toward the top. Not a GPU clip of a hollow C1–C7 surface. No pulse timing.

Pain sketch (cord)

Site

C5 is an AMU7T Atlas filled occupancy disk (vertebral C5, not a spinal segment). It is a 2D occupancy face extruded 2 mm, not a GPU clip of a hollow C1–C7 surface. Axial view: dorsal toward the top; subject's right is on the left (radiological). Thoracic, lumbar, and brainstem / Sp5c are Book hatched schematics: PAM50 NIfTI is not shipped, and Brainstem Navigator is not used. A site change also sets the dorsal-horn canvas below.

Hemisphere

Left/right follow the occupancy files (subject left is negative x). Ghost cord and gray matter stay. The white-matter envelope mesh is kept off so it does not hide named columns. Book T/L/Sp5c schematics are not sided occupancy, so this control does not hide them.

Point of view

Highlight

Columns

Default highlight is spinal lemniscus and Hausman dorsal-horn parcels. Click a column to inspect its Atlas caption. These are not HCP1065 streamlines.

Dorsal-horn laminae and a short local circuit are a Book schematic below. They are not AMU7T parcels. Occupancy columns have no conduction velocity control: there is no atlas median path length to divide.

Fig. 3. Dorsal horn. Book cervical circuit (Rexed/Todd) by default. Atlas occupancy is a selectable view.

Orange excitatory · blue inhibitory · purple unknown sign. Rexed 1952 names; Todd 2010 circuit (mainly rat I–III).

Dorsal horn

Default is the Book cervical circuit (Rexed/Todd). C5 anatomy is a selectable Atlas view: AMU7T occupancy of the cervical dorsal horn, with Lissauer as the licensed stand-in for fine DRG afferents after they enter and Book DRG/rootlets aimed at that occupancy (no licensed DRG volume; PAM50_rootlets is not used). Lumbar H&E is Zhang et al. 2024 GEO GSM7778679 Visium tissue image (CC BY 4.0), downsampled and rotated so dorsal is toward the top; fiducial dots are the Visium array, not cells. Lumbar nuclei is a 3D point cloud from that H&E: hematoxylin blobs (not reconstructed neurons). Points are coplanar (true section 10 µm). No parenchyma volume is drawn. Automatic GM/WM did not recover the butterfly and is not shown. Lumbar IF is Fig 2F. Neither histology view is Nissl, C5, Rexed, or registered to AMU7T. Lamina names from Rexed 1952 (cat). Circuit from Todd 2010 (mainly rat, laminae I–III). Circuit views: Not MRI, not AMU7T, not HCP1065 as a lamina map. Outer/inner lamina II is Todd usage, not a Rexed 1952 split. No human synaptic counts. No pulse timing.

Hemisphere

Applies to the Atlas anatomy view. Book circuit drawings are not sided occupancy.

Circuit highlight

Laminae

Cells

Default highlight is the nociceptive I–II circuit. Click a lamina or cell for its Book caption.

This is a population average over 1,065 young adults. Individual anatomy varies. Tractography can invent connections and miss real ones (Yeh 2022). These are not axons. A millimetre total under an action is the sum of each bundle's own median streamline path length, not one anatomical path. Path length is not the Euclidean span of the named structure (the anterior commissure is the extreme case).

What this atlas file does not contain. The ingested release hcp1065_avg_tracts_trk.zip (SHA-256 344aad43…779c67) has 87 named .trk.gz files. That HCP1065 zip has no spinothalamic tract, no forceps minor, no medial forebrain bundle, no trigeminothalamic tract, and no mammillothalamic tract. The corpus callosum is one file. Pain sketches therefore start at thalamus or trigeminal root. Mood sketches use uncinate, cingulum, anterior thalamic radiation, and fornix. Anterior thalamic radiation is not the slMFB (Bracht 2015; Denier 2020).
Sensory versus affective pain. Superior thalamic radiation is used here as a discriminative body-map stand-in. The HCP1065 file also carries motor thalamocortical fibres (VL/VA to M1), so the label is incomplete. Affective unpleasantness is sketched with anterior thalamic radiation, cingulum, and uncinate. Face pain is drawn crossed (left trigeminal file to right superior thalamic radiation) because second-order trigeminal nociception decussates; the crossing itself is not in the atlas.
Three MNI templates. Streamlines are Yeh 2022 ICBM 2009a. The glass cortex is HCP S1200 fs_LR, MNI152NLin6Asym. The cerebellum remainder is MNI152NLin2009cAsym. The TrackVis affine maps file millimetres to RAS millimetres. It does not convert one template into another. Laterality checks cannot detect a several-millimetre template offset.
Uncinate stems. HCP1065 ships a whole uncinate per hemisphere. Bhatia 2018 found FA reductions in subgenual and polar stems, not the lateral stem. This page cannot isolate those stems.
S1, S2, and limbic cortex. Those tints are Glasser HCP-MMP1 parcels already on the surface, not new anatomy. S1 is 3a, 3b, 1, and 2: a strip, not a homunculus. S2 is opercular OP1, OP2-3, and OP4. Limbic cortex is named cingulate, orbital, entorhinal, parahippocampal, and hippocampal surface parcels. It is not one organ. Amygdala and accumbens are not on this cortical map.
Pain-map cortex. The default cortical tint is Glasser MMP1 parcels grouped to match a teaching figure: S1, S2, M1, SMA, TPJ, ACC, PFC, and insula. Those are parcel stand-ins, not Brodmann drawings and not a homunculus. Parcel hue follows that figure (S2 is purple with PFC/ACC/insula), not the lamp colour of the system that lists the parcel. Isolation dims other tissue. It does not recolour a parcel to the lamp.
Deep nuclei. Thalamus, caudate, putamen, accumbens, amygdala, hippocampus, and brainstem are HCP Atlas_ROIs (Open Access Data Use Terms). PAG and locus coeruleus are Harvard AAN atlas v2.0 (CC0). Rostral ventral medulla is not in either atlas and is not drawn. Thalamus and accumbens atlas meshes stay unfilled. Hand-painted midline lids from /modeling are shown with those two structures.
Cervical cord columns. The second scene is AMU7T occupancy meshes (MIT) of Lévy/Hausman parcels in straightened PAM50-space millimetres, shown as a filled C5 cross-section (extruded 2 mm). That disk is not a GPU clip of a hollow C1–C7 surface. The spinal lemniscus parcel does not isolate the spinothalamic tract and is not 7 T tractography. It is not an HCP1065 file. HCP1065 still has no spinothalamic tract. Thoracic, lumbar, and brainstem / Sp5c are labelled Book hatched schematics on the same picker (PAM50 NIfTI is not shipped, not Navigator). The third canvas defaults to the Book cervical circuit (Rexed/Todd). C5 anatomy remains a selectable Atlas view: AMU7T occupancy of the cervical dorsal horn, with Lissauer as the licensed stand-in for fine DRG afferents after they enter and Book DRG/rootlets (no licensed DRG volume; PAM50_rootlets is not used). Cord sketches have no pulse: occupancy meshes have no atlas median path length.

Data were provided in part by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University.