Hypotheses

What is a knot?

Seven answers, each at its strongest, and each run as a model from its own physiology in the same patch of body: watch how their knots would form, feel, let go and move, then read what each says and what its model shows.

Watch them

Choose a scene. Each modelled theory grows its own knots in the same 4 cm of the upper back, where its own anatomy puts them, and meets the same stress, breath and hand. Nothing here is fitted to anyone's reports; it shows what each mechanism would do, and it measures nothing.

tap a theory to see it up close, below

結 Perforators

閂 Vascular latch

点 Trigger points

膠 Densification

神経 Nerves

握 Motor switch

覚 Perception

a knotletting goa hand, a rollerattentionwhere stress is heldthe breath's wave and the stress run along the timeline

up close

From above: 4 cm of the upper back, the neck toward the top, the muscle's fibres running across.

Beneath the spot

What instruments would record there, through the scene

The reports, and the models

What people who work with knots report (the Introduction's observations, in the site's own words), and in how many of the plausible settings each theory's model does it. The models were built without the reports; this is the comparison that comes after. A full circle: in every setting; an empty one: in none; a dash: the model has no route to it.

	結 Perforators	閂 Vascular latch	点 Trigger points	膠 Densification	神経 Nerves	握 Motor switch	覚 Perception
They let go on a slow out-breaththirty slow breaths: some held knot lets go
Deeper ones let go to focused attentionattention without touch frees the knot at the spot, or sooner
A patient hand releases them, one at a timea resting hand frees the knot at the spot, under it or as it lifts, where the same breaths alone would not, or sooner
A foam roller releases themrolled over: the knot at the spot lets go while rolled or within a minute
A hot shower softens many at oncefive minutes of heat frees a quarter or more of the patch's knots beyond what the same breaths do			—		—	—	—
They gather where stress is heldset by the shared scale in every theory: not a finding	scale	scale	scale	scale	scale	scale	scale
They limit movement: a stretch meets a dull blockheld half an hour: the knot at the spot blocks a stretch
A pop, or a tingle across a patch of skin, as one goesa spark or a twitch at a release
When one lets go, another moves in nearbyafter one lets go: a new knot holds within 10 mm
A big one frees small ones nearby, not the reversea parent's release frees its children; a child's does not free its parent		—	—	—	—	—	—
One comes back to roughly the same placeafter one lets go: the same knot holds again
Working one side eases the other	—	—	—	—	—	—	—
Drinking water eases release	—	—	—	—	—	—	—
A release brings a wave of well-being	—	—	—	—	—	—	—
Layers peel apart, and fill in again	—	—	—	—	—	—	—

No model yet reaches the last four: mirroring, water, the well-being after a release, and the layers that peel and fill. The reports are rough impressions, not measurements, and the perforator model is the most detailed of the seven (the site's own proposal): a model with more parts can reach more reports, and the others, built out, may reach further. Where a model and a report agree it is a hint, not a proof; the measurement that would decide is on the Research page.

01 Perforators

proposed here · 2026view in atlas →

A knot is a perforator held shut: the small artery that rises with its veins, nerve and lymphatic through a ring in the fascia, closed by the muscle of its own wall and kept closed until tone falls or movement frees it.

What it says

The superficial fascia is joined to the layer beneath in two ways: by fine fibrous strands, and at the places where something passes through — a small artery with its veins, a cutaneous nerve and a lymphatic, rising together through a fibrous ring. Surgeons call these perforators, and a flap of skin will not lift until the perforators tethering it are dealt with.

Suppose a knot is a perforator held shut. The artery is smooth muscle, and a small artery with tone in its wall is a switch: the pressure inside pushes it open, its muscle pulls it closed, and because muscle is strongest at one length, over a band of tone both open and shut are stable. A surge of stress carries it shut; it stays shut when the stress falls back, and opens only when tone falls well below the level that closed it. Held shut for an hour or two, its muscle adapts to the new length, as smooth muscle does, and the knot sets: it holds even at rest, and when at last it lets go the vessel opens wide. The nerve is the spark: a shut vessel starves its patch of skin, and when blood returns the nerve fires and the patch lights up — a foot waking, in miniature.

The ring in the fascia is where the bundle is pinned and where the knot is felt; it is not the hold. It may be why these are the knots that hands and rollers find: the layers slide past a vessel pinned there, so pressure and movement would concentrate on its wall, against a firm plane. The hyaluronan around the bundle is a liquid, and a liquid resists motion, not position: too weakly to hold a vessel shut (the check). Movement matters in another way. Smooth muscle stretched and released in rhythm loses force at once, and small arteries widen after being squeezed, more after several squeezes than after one long one. The breath moves the tissue, so what it does to a knot need not be the same everywhere.

The fascia is a second system: the superficial fascia and the gliding plane beneath it, from the galea and the SMAS of the face to the membranous layers of the trunk and limbs. It is not what a knot is; it is what the knots hold down. The reports about the fascia — layers peeling apart along lines, hollowing, and a filling-in felt as re-stitching — may belong to it.

Where: Where a vessel–nerve bundle pierces a fascia: the deep fascia for major perforators, the superficial fascia for smaller ones.
Holds: The vessel itself. A small artery with tone in its wall has two stable states over a band of tone, open and shut: once stress carries it shut, it stays shut until tone falls well below the level that closed it. Held an hour or two, its muscle adapts to being shut, and the knot can come to hold at rest.
Releases: Tone falling below the reopening threshold, with calm or with a breath that lowers drive more than it raises it; or movement reaching the wall, since rhythmic stretch cuts smooth muscle’s force and small arteries widen after being squeezed. Returning blood reaches the patch and wakes its nerve: the spark.
Test: Blinded palpation against a blinded Doppler map: do tender points sit on perforators? Then laser-speckle imaging at a release: blood flow returning over the spot within seconds, absent at a sham site, and sooner after pressure than after calm alone. In a myograph, a small artery shut for four hours should stay shut at resting tone, where one shut for five minutes reopens.

more of the account

Made of: A small artery shut by the smooth muscle in its own wall, and beside it a cutaneous nerve whose patch of skin has run short of blood.
Timescale: Seconds to shut or open · held while tone stays in the band · set within an hour or two, held at rest.
Breath: Two routes, and which one does the work is open. Through drive: a deep in-breath tightens the skin’s small arteries within seconds, and they take about half a minute to ease, so an even breath keeps knots held; a breath that lowers drive more than it raises it lets them go, everywhere at once, the easiest first. Through movement: the breath moves the tissue, and movement lowers the wall’s force where it reaches, more in one place than another.
Travel: Small arteries conduct: a dilation travels along the vessel wall toward its feeding vessel at millimetres to centimetres a second, and flow shifts across the tree as a branch opens — so knots seem to move, and even out.
Age: With age resting sympathetic tone rises and the skin’s small vessels respond less, all over the body; knots should rise with them in number and in size — first where stress is held — and children should be nearly free of them. A knot held for hours sets: old knots are the persistent ones, and flush when they go.
Count: Up to one per perforator: about 100,000 small, 3,600 medium and 374 major — tens of thousands of micro-knots by midlife, matching the dozens to the square inch that some report.
Evidence: The hold is physically possible, and demanding: with measured wall strength and thickness, blood pressure and reflex timings, a small artery can hold itself shut, but in healthy vessels only while its tone stays well above rest; thicker walls, as in hypertension, bring that nearer rest. With smooth muscle’s adaptation to a held length, as measured in arterioles, a knot held an hour or two would come to hold at rest. A hyaluronan collar cannot hold one. That a knot is a perforator, and that its state travels along the tree, is untested.

Taylor GI 1987 Saint-Cyr M 2009 Burton AC 1951 Bolton B 1936 Mayrovitz HN 2026 Ljung B 1975 Clifford PS 2006 Martinez-Lemus LA 2004 Syyong H 2008 Segal SS 1986 Heine H 1988 Yoshinaga T 2025

What its model shows

What holds it. Above the band an open vessel snaps shut; once shut it stays shut until its tone falls below the band.

◐Felt from inside. The patch a shut vessel starves signals through its own nerve: a held, achy place, felt without touching, and a tingle as it lets go. A held, tender place, in 12 of 19 settings.
●Found by a finger. Nothing firm for a finger to find, only tenderness where it presses; where its sleeve has jammed, the skin does not slide over it. Its shut artery, about 0.31 mm across and 10 mm down, is felt at about 2% of what a finger can find (at most 13%, where the fat is thinnest): nothing firm, in all 19 settings.
◐Stiffness. Each vessel passes the fascia in a sleeve of sliding tissue that its flow keeps wet; shut, the sleeve dries, its hyaluronan crowds (cowman2015), and it can jam, pinning the layers there. Held half an hour more, the knot at the spot blocks a stretch in 12 of 19 settings; where it did not at first, the block comes after about 12 min. When such a knot lets go (its own drive calmed for a minute), its vessel opens after about 9 s; its sleeve slides again within ten minutes in none of 9 settings, and otherwise stays stuck until the layer is sheared (a roller, a moving hand).
●Under a resting hand. Never while the hand still presses, in all 19 settings. It lets go within seconds of the hand lifting, in 16 of 19 settings. A vessel pressed shut cannot reopen until the pressure lifts.
◐To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in 9 of 19 settings. Drive is not local: attention changes only how much the breath moves the tissue at the spot. If attention could aim the breath's calming at one place (a hypothesis), it would in 10 of 19 settings.
●What an instrument would record. Laser speckle over the spot sees the skin's flow rise from 0.00 to 1.44 of resting flow as it lets go (the median of 16 releases), while the sham stays at 0.92 to 0.97. A press flushes every place it presses, so the control is a pressed place with no knot.

its full portrait: 22 lines, from 32 settings

Where they are

●Where. Where a small artery comes up through the deep fascia, under the skin and fat, in trees: a parent with its children around it. They gather where stress is held most, in all 19 settings.

How many

◐How many. Small arteries come up one every 4-5 mm: 5.0 to the square centimetre. After a stressful moment about 14 hold in this 4 cm patch (0.9 to the square centimetre); in 13 of 32 settings none can hold at all. A body has on the order of 100,000 such vessels (estimated from their spacing; 374 major ones, taylor1987). (depends most on how slowly a sleeve's water follows its vessel's flow: likelier the lower it is)

How and when they form

●How they form. A vessel shuts when its tone passes the fold of its switch; a parent that shuts takes its children with it. They appear within seconds of a stressful moment (the first after about 3 s), in all 19 settings.
●While stress lingers. While the stress lingers they stay as they are, in all 19 settings.
◐Pressed or rolled. Rolling a quiet place brings none out, in 28 of 32 settings: pressure outside a vessel keeps it shut only while it lasts. In the rest it brings one out, which stays after it stops in none of 4 settings. (depends most on pressure a hand or roller brings to a knot: likelier the lower it is)
◐With moods. Over an hour of moods they hold steady, in 18 of 32 settings; some come and go in the rest. (depends most on how slowly a sleeve's water follows its vessel's flow: likelier the higher it is)

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 16 of 19 settings; where they do, about 80% of them, most often late in the in-breath (28% of releases), the typical one at breath 2; one within the first two breaths in 11 of 16 settings. The out-breath eases tone and the in-breath raises it again, so a knot lets go when the easing outlasts the rise.
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 9 of 19 settings, after about 16 s; across the patch about 14% let go (with thirty slow breaths, 74%). Small breaths move the tissue little and change drive little: what is left is attention concentrating their movement at the spot. (depends most on shut radius: likelier the higher it is)
◐Heat. With five minutes of heat, as in a hot shower or a sauna, and the same slow breaths, the knot at the spot lets go where the breaths alone would not, or sooner, in 9 of 19 settings, after about 80 s; across the patch about 100% let go (with the breaths alone, 74%). Heat lowers the tone of the vessels' walls (minson2001), by the same number as it lowers the latch's clamp; what differs is the hold: once tone falls below its band, a shut vessel opens. (depends most on the share of a conducted response that crosses a branch point: likelier the higher it is)
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through the sympathetic drive to the knot's own small artery. To free the typical knot it must fall by about 71% of the holding stress for 16 s, or 54% for a minute; nothing shorter than 16 s frees it. A slow out-breath lowers it by about 17% for a few seconds, a subtle breath by 6%, minutes of slow breathing by about 26%, sustained. So one breath frees it in none of 19 settings. Tone eases slowly when drive falls (the gasp reflex's recovery, fitted to one recording), so a calming has to last.
◐Rolling the knot. Rolled over, it lets go while the roller works, in 18 of 19 settings, after about 5 s. Each pass presses the vessel shut and lifts: it can reopen only in the lifts, while the squeezing weakens its wall's muscle (ljung1975) and the roller shears its sleeve.
●How it lets go. Nothing firm to go; the tenderness fades over a minute or so. As it opens, blood floods the starved patch.
◐A spark. As blood returns the vessel's nerve bursts: a tingle over the patch it feeds. 338 of the 423 releases in the breathing and hand scenes came with one, in 17 of 19 settings.

How they move

●After one lets go. When the knot at the spot lets go, others go with it within ten seconds in all 16 settings (about 9); within ten minutes a new one holds nearby in 6 of 16 settings, and the same one holds again in none of 16 settings (the first return after about 34 s). In a tree a parent's knot shuts its children, and its release lets most of them go.
●Parents and children. Free one child under a held parent (its own drive down to rest for a minute): the child itself lets go in 5 of 18 settings, its parent in none of 18 settings, and a sibling in none of 18 settings. Free the parent instead: it lets go in 15 of 18 settings, and 2 of its 3 held children go with it (the median). Conducted dilation fades along the wall over about 2 mm (segal1989), and a child's knot sits some 4-5 mm from its parent's.

Over time

◐Held for hours. When its stress ends, a knot held three hours lets go with it, in all 32 settings. If the wall's muscle adapts to being shut, as smooth muscle held at a new length does, an old knot outlasts its stress in 4 of 32 settings, while one held half an hour lets go in 17 of 32 settings. (depends most on how much of its water a sleeve loses when its vessel's flow stops (and can gain in the flood after): likelier the higher it is)

● in every setting · ◐ in some · ○ in none · — the account says nothing

02 Vascular latch · vasocomputation

Johnson · 2023view in atlas →

Held predictions are held vascular tension: a contraction sustained long enough engages smooth muscle’s latch state and freezes a pattern in place.

What it says

Michael Edward Johnson’s Principles of Vasocomputation (2023) makes three proposals about vascular smooth muscle: that vasomotion compresses fragile neural patterns into definite states; that a vascular contraction clamps the local circuit and so acts as a medium-term memory — a specific prediction, in the language of active inference; and that a contraction held long enough engages the latch-bridge mechanism, freezing the circuit as a durable hyperprior, cut off from awareness and from updating.

Craving and clinging — taṇhā — then become the reflexive compression of experience into stable, controllable states, and latched tension is their residue. The latch may be a system of its own, closer to the clinging of the mind than to the knots a roller finds; or the two may prove to be views of the same knots. Johnson places latches wherever smooth muscle wraps a tube — vessels and hollow organs — so the atlas draws them in all of those places without drawing the organs: the small arteries of the skin and muscle, the great arteries of the trunk and limbs, the heart’s own vessels, and the walls of the airways, gut and bladder. The brain, at the centre of his account, is beyond this atlas.

Where: Vascular smooth muscle wherever it wraps a vessel — beside neurons in the brain, in the small arteries of skin and muscle throughout the body — and the walls of hollow organs.
Holds: The latch itself, and a spiral: reduced flow leaves less energy to unlatch.
Releases: Attention to the held tissue, cycles of heat and cold, meditation — noticing creates prediction errors that tease latches open — and psychedelics.
Test: Image vascular tension directly as it forms and dissolves; the model predicts that vasomotion nearly stops during deep meditative cessations.

more of the account

Made of: Smooth muscle in the latch-bridge state: cross-bridges that stay attached, holding tension at a fraction of the usual energy.
Timescale: Latches form and dissolve over seconds to minutes, and can persist for hours or years.
Breath: Indirect: through autonomic state and attention rather than a specific reflex.
Travel: Tension patterns are distributed; release follows the loosening of the predictions they hold.
Age: Latches accumulate from early childhood as we settle on predictions; each makes the system simpler and less alive.
Count: No count is given. Any small artery could latch — the skin’s included, so micro-knots too — and any hollow organ: millions of sites in principle, placed anywhere along the vessels. The atlas draws a sample.
Evidence: An integrative theory. In pressurised skeletal-muscle arterioles the latch looks more like sustained activation than an economical hold; the claims about the brain are open. Arterioles held constricted for hours no longer relax fully: a memory any vascular view can draw on, from the muscle adapting rather than the latch.

Michael Edward Johnson 2023 Michael Edward Johnson 2024 Hai CM 1988

What its model shows

What holds it. Clamped, the region is cut off from awareness and its prediction cannot update; bring enough awareness to it and the prediction lets go, and the latch after it.

●Felt from inside. Clamped, the region is cut off from awareness: a dull blind spot, hardly felt until attention or a hand brings it back, and then tender. Hardly felt, unattended, in all 32 settings.
●Found by a finger. Nothing firm for a finger to find: what stiffens is no larger than a small artery. Its clamped artery, about 0.20 mm across and 12 mm down, is felt at about 0% of what a finger can find (at most 1%, where the fat is thinnest): nothing firm, in all 32 settings.
○Stiffness. Nothing at the knot resists a stretch: what is clamped is a small artery.
◐Under a resting hand. It lets go under the hand, in 18 of 32 settings, after about 8 s. A hand brings the region into awareness; the prediction updates, and the latch lets go over tens of seconds.
◐To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in 31 of 32 settings. Attention brings the region back into awareness, and the prediction updates.
●What an instrument would record. Blood flow at the spot rises slowly, from 0.39 to 0.44 of resting flow over the half minute around its release (the median of 12), as the latch lets go; nothing changes in the muscle's own activity. Before it, the region is under-perfused and quiet to the senses.

its full portrait: 22 lines, from 32 settings

Where they are

●Where. Wherever smooth muscle wraps a small artery, in the skin and in the muscle beneath: regions 5 mm apart. They gather where stress is held most, in all 32 settings.

How many

●How many. Small arteries in the skin and muscle, each able to latch (64 representative): 4.0 to the square centimetre. After a stressful moment about 10 hold in this 4 cm patch (0.6 to the square centimetre).

How and when they form

◐How they form. Stress tightens a region's vessels and writes a prediction that holds them; the clamp keeps it written. They build during a stressful moment, over a minute or two (the first after about 29 s), in 25 of 32 settings. (depends most on how much stress tightens a region's vessels directly: likelier the higher it is)
●While stress lingers. While the stress lingers they stay as they are, in all 32 settings.
●Pressed or rolled. Rolling a quiet place brings none out, in all 32 settings: a roller brings a region into awareness for a moment, which lets a held one update, not a new one form.
◐With moods. Over an hour of moods they hold steady, in 31 of 32 settings; some come and go in the rest.

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 2 of 32 settings; where they do, about 53% of them, most often late in the out-breath (43% of releases), the typical one at breath 14; one within the first two breaths in none of 2 settings. A breath lowers the stress that writes the prediction and tightens the vessels; where the clamp could not keep the prediction without it, the prediction lets go, and the latch after it.
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 31 of 32 settings, after about 32 s; across the patch about 40% let go (with thirty slow breaths, 0%). Attention does the work; the breath's size matters little.
◐Heat. With five minutes of heat, as in a hot shower or a sauna, and the same slow breaths, the knot at the spot lets go where the breaths alone would not, or sooner, in 10 of 32 settings, after about 3 min; across the patch about 0% let go (with the breaths alone, 0%). Heat lowers the clamp's command, by the same number as a perforator's tone; the held prediction then has to let go too, which heat reaches only through the clamp. The account lists heat among its releases. (depends most on how much heat lowers the tone of a small artery beneath the skin and fat, as a share of maximal: likelier the higher it is)
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through stress at the knot (a calming aimed there). No fall in it frees the typical knot, even to rest for a minute. A slow out-breath lowers it by about 17% for a few seconds, a subtle breath by 6%, minutes of slow breathing by about 28%, sustained. So one breath frees it in none of 32 settings. Calm reaches the prediction only through the stress that feeds it; awareness updates it directly.
◐Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in 20 of 32 settings; while it rolls in 12 of 32 settings. Each pass brings the region into awareness for a moment. (depends most on how much a hand's pressure brings a region into awareness: likelier the lower it is)
●How it lets go. Nothing firm to go; the tenderness fades over a minute or so. The latch relaxes slowly once its command falls.
●A spark. No spark: the latch lets go over tens of seconds, a slow warming as blood returns.

How they move

◐After one lets go. When the knot at the spot lets go, others go with it within ten seconds in 14 of 18 settings (about 2); within ten minutes a new one holds nearby in 12 of 18 settings, and the same one holds again in 10 of 18 settings (the first return after about 3 s). Each region is held by its own prediction: the ones that go with it are the ones the same hand reached. One freed under a hand that lifts too soon can come back: its vessels, still clamped, write the prediction again.
—Parents and children. No trees: each region is held by its own prediction.

Over time

◐Held for hours. When its stress ends, a knot stays held in 8 of 32 settings, and lets go with it in the rest, over about 3 min (from 31 s to 32 min); how long it was held makes no difference. What decides it is whether the prediction's own command keeps the clamp past its halfway point once the stress has gone. (depends most on how much a clamped region is cut off from awareness: likelier the lower it is)

● in every setting · ◐ in some · ○ in none · — the account says nothing

03 Integrated trigger point

Travell · Simons · 1983–2004view in atlas →

A knot is a contraction knot in a taut band of muscle, at an overactive motor endplate that starves itself into an energy crisis.

What it says

Janet Travell and David Simons mapped hundreds of myofascial trigger points and their referred-pain patterns (The Trigger Point Manual, 1983). The integrated hypothesis, developed by Gerwin, Dommerholt and Shah (2004), proposes that excess acetylcholine at a motor endplate sustains a local contracture that compresses the nearby vessels; the resulting ischaemia and acidity release bradykinin, CGRP and substance P, which sensitise nerve endings and keep the loop going.

Its strongest support: ultrasound and elastography show discrete, stiff, hypoechoic nodules with abnormal blood flow (Sikdar 2009); microdialysis finds an acidic, mediator-rich milieu at active points (Shah 2005); and needle recordings show spontaneous activity at the nidus that mental stress turns up (Hubbard & Berkoff 1993; McNulty 1994).

Where: Inside skeletal muscle, at the motor endplate zone of a taut band.
Holds: An energy crisis: sustained contraction squeezes the capillaries, and the ischaemia withholds the energy needed to relax; released substances sensitise nearby nerves.
Releases: Sustained compression, stretch, dry needling or injection — often with a local twitch as it goes.
Test: Image taut bands (MR and ultrasound elastography), record spontaneous electrical activity at the nidus, and sample its chemistry by microdialysis.

more of the account

Made of: A patch of contracted sarcomeres — the contraction knot — with excess acetylcholine release and a sensitising chemical milieu.
Timescale: Seconds to minutes to inactivate; weeks for the factors that sustain it.
Breath: Indirect: relaxation lowers motor drive.
Travel: Pain refers in stereotyped patterns, and key trigger points maintain satellites.
Age: Accumulates with overload, posture and injury.
Count: Hundreds: a few regions in each muscle, at mapped sites — far fewer than the micro-knots some report.
Evidence: Widely taught, and key findings replicate — stiff nodules on ultrasound, spontaneous electrical activity, an acidic milieu — but diagnosis by touch is unreliable and the construct remains contested.

Janet G. Travell 1983 Gerwin RD 2004 Sikdar S 2009 Shah JP 2005 Hubbard DR 1993 McNulty WH 1994

What its model shows

What holds it. Starved by its own contraction, a contracted band cannot relax until its drive falls below the band, or pressure or stretch loosens it.

◐Felt from inside. Its acid, sensitising milieu stirs the muscle's nerves: an ache felt from inside. A held, tender place, in 19 of 22 settings.
◐Found by a finger. A small firm nodule on a taut band, about the size measured by ultrasound (0.16 cm², sikdar2009). About 15 mm down, in the muscle. A finger can find it in 6 of 22 settings where knots hold; in the rest it lies below the edge of touch.
●Stiffness. A contracture resists a stretch: the taut band itself. Held half an hour more, the knot at the spot blocks a stretch in all 22 settings.
◐Under a resting hand. It lets go under the hand, in 14 of 22 settings, after about 18 s; as it lifts in 1 of 22 settings. Sustained pressure lengthens the contracture slowly, over the tens of seconds of pressure release.
◐To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in 12 of 22 settings. Attention concentrates the breath's stretch at the spot. If attention could aim the breath's calming at one place (a hypothesis), it would in 7 of 22 settings.
●What an instrument would record. Elastography through the pressing probe sees the nodule's contracture fall from 0.57 to 0.46 over the tens of seconds around its release (the median of 11), under the hand; a needle in it finds the endplate still active (3.0 times its typical level before, 3.1 after): the press lengthens the band, it does not quiet the endplate. Nothing changes in the skin.

its full portrait: 22 lines, from 32 settings

Where they are

●Where. In the muscle, along the zone where its nerve enters it: a few candidate places, not everywhere. They gather where stress is held most, in all 22 settings.

How many

◐How many. Candidate places along the innervation zone, a few in each muscle: 0.3 to the square centimetre. After a stressful moment about 4 hold in this 4 cm patch (0.2 to the square centimetre); in 10 of 32 settings none can hold at all. A body has a few hundred such places (300 taken as representative). (depends most on how steeply relaxation needs energy: likelier the higher it is)

How and when they form

●How they form. Stress raises endplate activity; the contraction squeezes its capillaries, and the starved sarcomeres cannot let go: an energy crisis. They build during a stressful moment, over a minute or two (the first after about 58 s), in all 22 settings.
●While stress lingers. While the stress lingers they stay as they are, in all 22 settings.
●Pressed or rolled. Rolling a quiet place brings none out, in all 32 settings: a brief press adds a little ischaemia, not enough to start a crisis where none is near.
●With moods. Over an hour of moods they hold steady, in all 32 settings.

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 12 of 22 settings; where they do, about 75% of them, most often at the bottom of the breath, as the in-breath begins (30% of releases), the typical one at breath 16; one within the first two breaths in 7 of 12 settings. Relaxation lowers endplate activity, and the breath's movement stretches the band. (depends most on pressure that stops flow in the muscle's capillaries: likelier the higher it is)
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 9 of 22 settings, after about 3 s; across the patch about 0% let go (with thirty slow breaths, 25%). Small breaths stretch the band little; the calm they bring lowers endplate activity a little. (depends most on how much a breath moves the tissue at a knot: likelier the higher it is)
—Heat. The model gives heat no route: the account's releases are pressure, stretch and needling.
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through endplate activity at the knot. No fall in it frees the typical knot, even to rest for a minute. A slow out-breath lowers it by about 15% for a few seconds, a subtle breath by 5%, minutes of slow breathing by about 28%, sustained. So one breath frees it in none of 22 settings. The energy crisis holds by its own squeezed capillaries: it needs pressure or stretch, not calm alone.
◐Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in 17 of 22 settings; while it rolls in 4 of 22 settings, after it stops in 1 of 22 settings. Each pass is a brief press, not the sustained pressure that lengthens the contracture. (depends most on how steeply relaxation needs energy: likelier the higher it is)
●How it lets go. The firmness goes over minutes; the tenderness fades over a minute or so. The contracture relaxes as energy returns.
◐A spark. A release fast enough twitches: the local twitch response. 254 of the 357 releases in the breathing and hand scenes came with one, in 8 of 22 settings. (depends most on how much a stretch loosens the contracture: likelier the higher it is)

How they move

◐After one lets go. When the knot at the spot lets go, others go with it within ten seconds in 8 of 17 settings (about 2); within ten minutes a new one holds nearby in 7 of 17 settings, and the same one holds again in 8 of 17 settings (the first return after about 7 s). Nothing here links one trigger point to another: the key point and its satellites are not in the patch yet.
—Parents and children. Nothing in this account links trigger points in trees: its key point and satellites are not modelled yet.

Over time

—Held for hours. Its slow sustaining factors, over weeks, are not modelled yet.

● in every setting · ◐ in some · ○ in none · — the account says nothing

04 Fascial densification

Stecco · 2011view in atlas →

Knots are densified loose connective tissue: hyaluronan between the fascial layers turned viscous, so the layers stop gliding.

What it says

The Steccos’ group in Padua described the hyaluronan-rich loose tissue between fascial layers and proposed its densification as a source of myofascial pain (2011). Its claim is about glide: a viscous layer resists the sliding of the layers over one another, and ultrasound finds the loose sublayers thicker in chronic neck pain, thinning again with treatment (Stecco 2014).

Where: The loose layers between and within the fasciae.
Holds: Viscosity: the layers stop gliding, and the strain falls on nerve endings in the fascia.
Releases: Deep friction (fascial manipulation), heat and movement; the change is reversible.
Test: Ultrasound: the loose sublayers are thicker in chronic neck pain and thin with treatment (Stecco 2014).

more of the account

Made of: Hyaluronan whose chains clump as the tissue turns acidic and cool — a gel instead of a lubricant.
Timescale: Minutes, with heat and friction, to weeks.
Breath: Not central.
Travel: Along sequences of fascia — myofascial units and chains.
Age: Rises with immobility and age, as the loose layers thicken.
Count: Tens to hundreds of patches, each centimetres across — regions rather than points.
Evidence: Plausible biophysics with supportive imaging; small trials.

Stecco C 2011 Stecco A 2014 Langevin HM 2011

What its model shows

What holds it. Left still, the layer builds until it jams; movement breaks it down. Across the band both hold: a jammed patch stays jammed under movement that keeps a fluid one fluid, and gives way only past the band, slowly and then all at once.

◐Felt from inside. The layers stop gliding, so movement pulls on the fascia's nerve endings where they are stuck: an ache or a pull when moving, quiet when still. A held, tender place, in 28 of 32 settings.
●Found by a finger. Nothing firm for a finger to find: what changes is how the layers slide. A finger that slides the skin finds it held there, over a patch a centimetre or more across.
●Stiffness. The jammed layer resists sliding: a stretch pulls against it. Held half an hour more, the knot at the spot blocks a stretch in all 32 settings.
●Under a resting hand. Never while the hand still presses, in all 32 settings. It does not let go within the scene, in 31 of 32 settings. A resting hand warms the layer a little and does not slide it; warmth thins hyaluronan by about 2% a degree (cowman2015). If the hand worked the skin in small circles instead of resting (deep friction, the account's own way), it would let go under the hand in 13 of 32 settings, after about 21 s.
○To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in none of 32 settings. The account gives attention no route to the layer: attention alone changes nothing here.
●What an instrument would record. Ultrasound that tracks the layers sliding over each other (as measured in the back, langevin2011) sees the sliding at the spot rise from 0.02 to 0.72 of a free layer's over the half minute around its release (the median of 13); the skin's flow and the muscle's activity do not change.

its full portrait: 22 lines, from 32 settings

Where they are

◐Where. In the gliding layer between the fasciae, wherever movement is least: broad patches, not points. They gather where stress is held most, in 25 of 32 settings. (depends most on how much stress stills the layer's movement (guarding): likelier the higher it is)

How many

●How many. Patches of the gliding layer, 5 mm across (64 representative); a knot is a connected region of stuck ones. After a stressful moment about 26 of them are stuck in this 4 cm patch, in about 2 connected regions (the largest about 6 cm²). The account counts tens to hundreds of such patches in a body, each centimetres across.

How and when they form

◐How they form. They are there before any stressful moment: a layer left still (a night) jams, and a day's movement frees it only where it moves enough. Left still, the layer rebuilds its structure and thickens; movement breaks it down again. Where stress stills the layer, it jams. A stressful moment adds new ones in 3 of 32 settings, the first after about 88 s.
●While stress lingers. While the stress lingers they stay as they are, in all 32 settings.
◐Pressed or rolled. Rolling a quiet place brings none out, in 31 of 32 settings: a roller shears the layer, which breaks its structure down: it frees, it does not make. In one setting it brings one out, which goes when it stops.
◐With moods. Over an hour of moods they hold steady, in 27 of 32 settings; some come and go in the rest. (depends most on how fast the skin warms and its small vessels answer: likelier the lower it is)

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 12 of 32 settings; where they do, about 6% of them, most early in the out-breath (54% of releases), the typical one at breath 16; one within the first two breaths in 1 of 12 settings. The breath's movement shears the layer a little, and its calm eases the guarding, so the layer moves more.
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 1 of 32 settings, after about 290 s; across the patch about 0% let go (with thirty slow breaths, 0%). Nothing here answers attention, and small breaths hardly move the layer.
○Heat. With five minutes of heat, as in a hot shower or a sauna, and the same slow breaths, the knot at the spot lets go where the breaths alone would not, or sooner, in none of 32 settings; across the patch about 0% let go (with the breaths alone, 0%). Warmth thins the layer only about 2% a degree (cowman2015).
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through the stress that stills the layer (its guarding). No fall in it frees the typical knot, even to rest for a minute. A slow out-breath lowers it by about 17% for a few seconds, a subtle breath by 6%, minutes of slow breathing by about 29%, sustained. So one breath frees it in none of 32 settings. A jammed patch gives way only when movement shears it past its band, and then slowly at first: calm helps only by letting the layer move.
◐Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in 20 of 32 settings; while it rolls in 12 of 32 settings. The roller shears the layer beneath it: past its band, a jammed patch gives way. (depends most on how long a still patch of the gliding layer takes to rebuild its structure: likelier the higher it is)
●How it lets go. Nothing firm to go; the tenderness fades over a minute or so. A jammed patch gives way slowly and then all at once, like an avalanche (coussot2002): the layers slide again.
●A spark. No spark: nothing in a layer that starts to slide makes one; the pull on its nerve endings eases.

How they move

◐After one lets go. When the knot at the spot lets go, others go with it within ten seconds in 2 of 4 settings (about 2); within ten minutes a new one holds nearby in none of 4 settings, and the same one holds again in none of 4 settings. Patches jam and give way on their own; only a shared cause, a change in movement or stress, moves several at once.
—Parents and children. No trees: the layer's patches are not branches of one another. The fascial chains along which the account has densification travel are not modelled yet.

Over time

◐Held for hours. When its stress ends a knot lets go, however long it was held, in 21 of 32 settings. What decides it is whether ordinary movement, once the stress has gone, shears a jammed patch past its band. (depends most on how long a still patch of the gliding layer takes to rebuild its structure: likelier the lower it is)

● in every setting · ◐ in some · ○ in none · — the account says nothing

05 Peripheral nerve sensitivity

Quintner · Cohen · 1994view in atlas →

Tender points are sites of nerve sensitivity — the nerve and its sheath — rather than a lesion in muscle.

What it says

John Quintner and Milton Cohen (1994) argued that the idea of myofascial pain misattributes to muscle what arises in peripheral nerves, and with Geoffrey Bove (2015) they criticised the circularity of trigger-point theory. The perforator view agrees about the nerve, and adds the vessel that travels with it.

Where: Nerves of the skin and muscle, especially where they pierce the fascia.
Holds: Sensitisation of the nerve, peripheral and central.
Releases: Nerve gliding and desensitisation; freeing an entrapped nerve with fluid.
Test: Nerve-specific tests at tender points; the response to freeing the nerve with fluid (hydrodissection).

more of the account

Made of: Sensitised nerve trunks, through the small nerves of their own sheaths (nervi nervorum).
Timescale: Variable.
Breath: Indirect.
Travel: Along the nerve’s territory.
Age: Entrapment and sensitisation rise with age and injury.
Count: Thousands: wherever a nerve of the skin pierces the fascia — about as many as the medium and major perforators it travels with.
Evidence: Entrapments where nerves pierce the fascia are established (anterior cutaneous and cluneal nerve entrapment); extending this to all knots is contested.

Quintner JL 1994 Quintner JL 2015 Yoshinaga T 2025

What its model shows

What holds it. No switch: a sensitised nerve is felt while its drive is above threshold, and not below it. The drive follows the stress at once through the muscle around it, and about fourteen seconds late through the sympathetic drive; the nerve's sensitivity lasts weeks.

●Felt from inside. A sensitised nerve firing on its own: an ache or a burn at a fixed point, sometimes spreading along its branches. A held, tender place, in all 32 settings.
●Found by a finger. Nothing firm for a finger to find: a nerve of the skin is a millimetre across. Pressed, the point is tender, and past a level it tingles along the nerve's branches.
○Stiffness. Nothing at the knot resists a stretch; stretching a sensitised nerve makes it fire (dilley2005), an ache, not a block.
◐Under a resting hand. It does not let go within the scene, in 25 of 32 settings; under the hand in 2 of 32 settings. Pressing a sensitised nerve excites it: it is felt more under the hand, and past a level it tingles along its branches. A hand's 20-80 mmHg is also more than slows the blood leaving a nerve (20-30 mmHg, rydevik1981).
○To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in none of 32 settings. The account gives attention no route to the nerve: attention alone changes nothing here.
●What an instrument would record. A recording from the nerve would see its firing fall from 1.05 to 1.01 of the level felt as a knot over the half minute around its release (the median of 13), following the sympathetic drive about fourteen seconds late; a local anaesthetic at the point would silence it (as at entrapped nerves of the abdominal wall, boelens2013). The skin's flow and the muscle's activity do not change.

its full portrait: 22 lines, from 32 settings

Where they are

●Where. Only where a nerve of the skin pierces the fascia, beside the perforator it travels with: a few fixed points, about 20 mm apart here. They do not gather where stress is held, in all 32 settings: they sit at the nerves' piercings, wherever those are.

How many

●How many. Nerves of the skin where they pierce the fascia (4 in the patch): 0.2 to the square centimetre. After a stressful moment about 3 hold in this 4 cm patch (0.2 to the square centimetre). The account counts thousands in a body, about as many as the medium and major perforators.

How and when they form

◐How they form. Stress excites a sensitised nerve through the sympathetic drive, about fourteen seconds late (devor1994), and through the load of the muscle around it. They appear within seconds of a stressful moment (the first after about 11 s), in 17 of 32 settings.
●While stress lingers. While the stress lingers they stay as they are, in all 32 settings.
●Pressed or rolled. Rolling a quiet place brings none out, in all 32 settings: a roller presses what it passes; where no nerve pierces there, it excites nothing.
◐With moods. Over an hour of moods, knots come and go (some form and let go at least twice), in 30 of 32 settings.

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 20 of 32 settings; where they do, about 67% of them, most late in the out-breath (56% of releases), the typical one at breath 20; one within the first two breaths in 1 of 20 settings. A slow breath eases the load around the nerve at once and the sympathetic drive about fourteen seconds later. (depends most on how far minutes of slow breathing lower stress: likelier the higher it is)
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 1 of 32 settings, after about 186 s; across the patch about 0% let go (with thirty slow breaths, 25%). Nothing here answers attention; small breaths ease the drive a little.
—Heat. The model gives heat no route to a sensitised nerve.
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through the stress at the nerve (the muscle's load around it, and the sympathetic drive). To free the typical knot it must fall by about 88% of the holding stress for 8 s, or 38% for a minute; nothing shorter than 8 s frees it. A slow out-breath lowers it by about 17% for a few seconds, a subtle breath by 6%, minutes of slow breathing by about 30%, sustained. So one breath frees it in none of 17 settings. The sympathetic route answers only after about fourteen seconds (devor1994), so a calming has to last; the site itself stays sensitised for weeks (dilley2008).
●Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in all 32 settings. Each pass presses the nerve, which tingles along its branches; nothing frees it.
●How it lets go. Nothing firm to go; the tenderness fades within seconds. What goes is its firing: the site stays sensitised, and the knot is felt again, in the same place, when its drive returns.
●A spark. No spark at a release. Pressed, a sensitised nerve tingles along its branches: a tingle of pressing, not of letting go.

How they move

◐After one lets go. When the knot at the spot lets go, others go with it within ten seconds in 6 of 17 settings (about 1); within ten minutes a new one holds nearby in none of 17 settings, and the same one holds again in 15 of 17 settings (the first return after about 12 s). Each site answers its own drive; a shared stress moves them together, and a knot comes back where it was, at the same nerve.
—Parents and children. No trees of knots: a nerve's branches spread from its piercing, and pressing the piercing is felt along them.

Over time

●Held for hours. When its stress ends a knot lets go, however long it was held, in all 32 settings. Its firing follows the stress within a minute; the site's sensitivity outlasts it by weeks (dilley2008), so the knot returns with the stress, in the same place.

● in every setting · ◐ in some · ○ in none · — the account says nothing

06 Motor switch

Johansson · Sojka (the loop); Heckman · Gorassini (the latch) · 1991–2024atlas view soon

A knot is a few motor units that have latched on: spinal motor neurons that keep firing after the input that recruited them, kept on by the metabolites of their own contraction.

What it says

Spinal motor neurons are not simple relays. A brief input can start self-sustained firing that outlasts it, measured in people by Gorassini and colleagues (1998): persistent inward currents amplify synaptic input up to fivefold or more, set by serotonin and noradrenaline, and switching them off usually needs inhibitory input (Heckman 2008). Blocking serotonin’s 5-HT2 receptors shortens the self-sustained firing of human motor units (Goodlich 2024).

Håkan Johansson and Pavel Sojka (1991) proposed a loop by which muscle tension keeps itself going: the metabolites of a static contraction drive the muscle’s own sensors, and through them the same motor neurons, raising stiffness and metabolites again and spreading to other muscles. Needle recordings find electrical activity at tender points that mental stress turns up (McNulty 1994), and in a postural muscle motor units take turns (Minerbi & Vulfsons 2018).

Put together, and tested here as a seventh view: a knot is a few motor units latched on under stress, a contraction beneath the skin that lets go the moment they fall silent, like an unclenching. It lives in muscle, so it predicts less over bone. It says nothing of a tingle at a release, and nothing in it remembers years.

Where: Skeletal muscle, switched from the spinal cord: the knot is in the muscle, its switch in the motor neurons that drive it.
Holds: Persistent inward currents: once recruited, a motor neuron keeps firing on a fraction of the input that recruited it, more so under serotonin and noradrenaline, the chemistry of stress. The contraction’s own metabolites drive the muscle’s sensors, and through them the same neurons: a loop.
Releases: Inhibition: switching the currents off usually needs inhibitory input. The unit falls silent and its patch unclenches at once — perhaps as the drive falls on an out-breath, under held pressure, or to attention, if these inhibit.
Test: Surface EMG that picks out single motor units, over a knot as it lets go: does one fall silent at the felt unclenching? With ultrasound to place the bump in muscle.

more of the account

Made of: A small patch of muscle fibres held contracted by a few motor units firing steadily; nothing in the tissue is damaged.
Timescale: On within seconds under stress; off within about a second once inhibited; kept for minutes to hours by the loop.
Breath: Slow breathing lowers the descending drive and the monoamines that deepen the latch, so an out-breath can drop a shallow unit below its hold.
Travel: In a postural muscle, motor units take turns; when one falls silent another can take up its load, and the loop spreads to neighbouring units and muscles.
Age: Not by itself: nothing in the switch remembers years. It would need the muscle, or its supply, to change beneath it.
Count: Not yet counted: as many as the small groups of motor units a body keeps working at a low level.
Evidence: Its parts are measured in people — self-sustained firing (Gorassini 1998), its amplification by serotonin and noradrenaline (Heckman 2008), shortened by a serotonin blocker (Goodlich 2024) — and the loop is a published hypothesis (Johansson & Sojka 1991). That a knot is a latched motor unit has not been tested.

Gorassini MA 1998 Heckman CJ 2008 Goodlich BI 2024 Johansson H 1991 McNulty WH 1994 Minerbi A 2018

What its model shows

What holds it. Recruited at its threshold, a unit keeps firing down to a small share of it (its persistent currents), while serotonin and noradrenaline last.

◐Felt from inside. Its contraction and its metabolites are sensed by the muscle's own sensors: a clench or an ache, felt from inside. A held, tender place, in 30 of 32 settings.
●Found by a finger. A firm band along the fibres, 5-10 mm wide and a few centimetres long: a motor unit's territory contracted. Overlapping units make it firmer. About 15 mm down, in the muscle. A finger can find it in all 32 settings where knots hold; in the rest it lies below the edge of touch.
●Stiffness. A latched unit's fibres are contracted: they resist a stretch until it falls silent. Held half an hour more, the knot at the spot blocks a stretch in all 32 settings.
◐Under a resting hand. It lets go under the hand, in 22 of 32 settings, after about 10 s. A held hand excites the units under it for a moment and then inhibits them (which a hand does is not known; the other variant follows). If a held hand only excited the units under it, it would let go under the hand in 2 of 32 settings.
◐To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in 16 of 32 settings. Attention at a place inhibits its units.
●What an instrument would record. Surface EMG that picks out single motor units sees the knot's unit fall silent at the moment it lets go (7 units firing under the electrode before, 0 after, the median of 16 releases), and elastography sees the firmness at the spot fall from 8.9 to 6.3 times the edge of touch. The skin's flow does not change.

its full portrait: 22 lines, from 32 settings

Where they are

●Where. In the muscle, wherever its motor units lie: territories 5 mm apart and overlapping. They gather where stress is held most, in all 32 settings.

How many

●How many. Motor units whose territories lie under the patch (64 representative): 4.0 to the square centimetre. After a stressful moment about 18 hold in this 4 cm patch (1.1 to the square centimetre).

How and when they form

◐How they form. Stress recruits motor units; some latch on and keep firing after the stress eases, kept on by the metabolites of their own contraction. They build during a stressful moment, over a minute or two (the first after about 61 s), in 15 of 32 settings.
●While stress lingers. While the stress lingers they stay as they are, in all 32 settings.
◐Pressed or rolled. Rolling a quiet place brings none out, in 28 of 32 settings: the roller excites the units it presses, but where the drive is far below their threshold none latches. In the rest it brings one out, which stays after it stops in all 4 settings. (depends most on how far a firing unit squeezes its own supply: likelier the lower it is)
●With moods. Over an hour of moods, knots come and go (some form and let go at least twice), in all 32 settings.

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 17 of 32 settings; where they do, about 14% of them, most often at the bottom of the breath, as the in-breath begins (49% of releases), the typical one at breath 8; one within the first two breaths in 16 of 17 settings. The breath lowers the descending drive and the monoamines that deepen the latch.
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 6 of 32 settings, after about 11 s; across the patch about 0% let go (with thirty slow breaths, 5%). Attention inhibits the units at the spot whatever the breath's size; the breath itself adds little. (depends most on the loop: how strongly a unit's own metabolites drive it: likelier the lower it is)
—Heat. The model gives heat no route to the motor neurons.
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through the unit's input (a fall in drive, or inhibition, such as attention's). To free the typical knot it must fall by about 93% of the holding stress for 8 s, or 88% for a minute; nothing shorter than 8 s frees it. A slow out-breath lowers it by about 15% for a few seconds, a subtle breath by 5%, minutes of slow breathing by about 29%, sustained. So one breath frees it in none of 32 settings. The latch holds a unit on at a small share of the input that recruited it: the input must fall far below that.
●Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in all 32 settings. Each pass excites the units under it.
●How it lets go. The firmness goes at once; the tenderness fades over minutes. The unit falls silent and its fibres relax: an unclenching.
●A spark. No tingle at a release: nothing in a motor unit falling silent makes one.

How they move

●After one lets go. When the knot at the spot lets go, others go with it within ten seconds in all 22 settings (about 6); within ten minutes a new one holds nearby in 21 of 22 settings, and the same one holds again in 9 of 22 settings (the first return after about 4 s). Neighbouring units share the metabolites' drive, so a latched unit can hand its load to one nearby.
—Parents and children. Motor units are not arranged in trees: neighbours share only their metabolites' drive.

Over time

●Held for hours. When its stress ends a knot lets go, however long it was held, in all 32 settings.

● in every setting · ◐ in some · ○ in none · — the account says nothing

07 Perception — sensitisation and attention

clinical medicineview in atlas →

No special structure in the tissue: the spinal cord and brain turn ordinary signals into tender spots, overbreathing makes the sparks, and attention makes the map.

What it says

Central sensitisation — a rise in the responsiveness of the pain pathways in the spinal cord and brain — is documented across chronic pain conditions (Woolf 1983; 2011). Ordinary pressure on ordinary tissue comes to hurt, and tender spots appear without a lesion beneath them.

The same view answers the stranger reports. Overbreathing produces exactly the tingling sparks people describe; attention amplifies whatever it rests on; a muscle inhibited by pain feels detached; sleep loss distorts the body map. None of this makes the sensations imaginary: they are real, made by ordinary physiology rather than by a hidden structure. It is the physician’s first answer, and the account every other view has to beat.

Where: The spinal cord and brain: a knot is felt in the body but made in the nervous system.
Holds: Central plasticity, kept up by threat, attention and poor sleep.
Releases: Safety, sleep, movement, a change of attention, ordinary breathing — global rather than local.
Test: Widespread lowered pressure-pain thresholds, including at places never worked; and at a felt release, blood-flow changes that are global, equal at sham sites, and follow carbon dioxide rather than the felt event.

more of the account

Made of: Raised gain in the pain pathways, and perception shaped by attention, expectation, sleep and the chemistry of the breath.
Timescale: Instantaneous for percepts; days to months for sensitisation.
Breath: Overbreathing lowers carbon dioxide and causes tingling and sparks anywhere; slow breathing calms arousal and threat.
Travel: A percept can move anywhere in an instant; sensitisation spreads by spinal segment.
Age: Tied less to age than to history: injury, stress and sleep.
Count: No fixed number: as many places as sensitisation and attention make, and they can move.
Evidence: Central sensitisation is well established in many chronic pain states, and overbreathing reliably causes tingling. Neither explains why the same discrete spots return to the same places.

Woolf CJ 2011 Lindahl JR 2017

What its model shows

What holds it. Felt as a knot once it is loud enough, and still felt until it falls a little lower: a narrow band, in the nervous system alone.

●Felt from inside. Felt from inside is all it is: a place the nervous system turns up. A held, tender place, in all 32 settings.
●Found by a finger. Nothing firm for a finger to find: nothing in the tissue changes. What a finger meets is an ordinary structure, felt more.
○Stiffness. Nothing at the knot resists a stretch: nothing in the tissue changes.
●Under a resting hand. Never while the hand still presses, in all 32 settings. It does not let go within the scene, in 25 of 32 settings; as it lifts in 7 of 32 settings. A hand adds input and draws attention: the place is felt more under it.
◐To attention, without touch. With attention resting on the spot and no touch, the knot there lets go where the same breaths alone would not, or sooner, in 15 of 32 settings. Attention first turns up what is felt there, then, with the breath, quiets it.
●What an instrument would record. Nothing in the tissue changes: skin flow, muscle activity and stiffness stay as they were. Only the report changes: what is felt at the spot falls from 2.43 to 0.75 of the level felt as a knot (the median of 7 releases).

its full portrait: 22 lines, from 32 settings

Where they are

◐Where. Wherever the body map tells places apart, about 10 mm on the back: broad and soft-edged. They do not gather where stress is held, in 29 of 32 settings: they sit where ordinary input happens to be loudest, and arousal raises every place at once.

How many

●How many. Places on the body map, about 10 mm apart: 1.0 to the square centimetre. After a stressful moment about 6 hold in this 4 cm patch (0.4 to the square centimetre).

How and when they form

●How they form. Arousal turns up what is felt, most where stress is held and the muscles guard. They appear within seconds of a stressful moment (the first after about 1 s), in all 32 settings. Some places are felt as knots even at rest, before any stress, in 11 of 32 settings.
◐While stress lingers. While the stress lingers they stay as they are, in 29 of 32 settings.
◐Pressed or rolled. Rolling a quiet place brings none out, in 30 of 32 settings: a press is felt while it lasts. In the rest it brings one out, which stays after it stops in 1 of 2 settings.
◐With moods. Over an hour of moods they hold steady, in 18 of 32 settings; some come and go in the rest.

How and when they let go

◐To slow breathing. Over thirty slow breaths some let go, in 14 of 32 settings; where they do, about 57% of them, most often early in the out-breath (40% of releases), the typical one at breath 12; one within the first two breaths in 4 of 14 settings. A slow breath eases arousal and is felt as safety; the in-breath is felt a little less.
◐To subtle breaths, attending. With subtle breaths, a third the size of a slow breath, and attention resting on the spot, the knot there lets go in 13 of 32 settings, after about 300 s; across the patch about 15% let go (with thirty slow breaths, 0%). Attention with safety quiets a place whatever the breath's size. (depends most on how fast attention with safety quiets a place: likelier the lower it is)
—Heat. The model gives heat no route; that heat calms arousal is not modelled.
●What a breath would have to do. Whatever its pattern, a breath reaches this knot through arousal, everywhere at once (perception has no drive of its own at a place). No fall in it frees the typical knot, even to rest for a minute; the easiest quarter need about 99% of the holding stress taken away for 16 s. A slow out-breath lowers it by about 17% for a few seconds, a subtle breath by 6%, minutes of slow breathing by about 28%, sustained. So one breath frees it in none of 32 settings. What is felt follows arousal and attention; sensitisation keeps a place loud after arousal falls.
◐Rolling the knot. Rolled over for three minutes, it does not let go within the scene, in 29 of 32 settings; while it rolls in 3 of 32 settings. Each pass is felt: the place is louder while it is rolled.
●How it lets go. Nothing firm to go; the tenderness fades at once. What goes is what is felt.
●A spark. No tingle at a release: its tingling comes only from overbreathing, in the hands, face and trunk at once.

How they move

◐After one lets go. When the knot at the spot lets go, others go with it within ten seconds in 7 of 8 settings (about 1); within ten minutes a new one holds nearby in none of 8 settings, and the same one holds again in 1 of 8 settings (the first return after about 64 s). Attention moves to the next loudest place, and the loudest knot quiets the others.
—Parents and children. No trees: places on the body map are not branches of one another.

Over time

◐Held for hours. When its stress ends a knot lets go, however long it was held, in 20 of 32 settings. (depends most on how close ordinary input is to being felt as a knot: likelier the lower it is)

● in every setting · ◐ in some · ○ in none · — the account says nothing

How they differ

The same scenes, the modelled answers side by side. Where two columns part on something a hand or an instrument can tell, that is a test (see the measurement).

	結 Perforators	閂 Vascular latch	点 Trigger points	膠 Densification	神経 Nerves	握 Motor switch	覚 Perception
Felt from inside	a held, achy place	a dull blind spot	an ache	a pull when moving	an ache at a fixed point	a clench, an ache	all it is
Found by a finger	nothing firm	nothing firm	a small nodule on a band, often too faint	nothing firm; the skin does not slide	nothing firm; a tender point	a firm band along the fibres	nothing firm
Stiffness	a block, once its sleeve jams	no block	a taut band	the layers do not slide	no block	a contracted band	no block
Where	at the fascia, on its vessels	anywhere there are small arteries	in muscle, where its nerve enters	in the gliding layer; broad	fixed points, at nerves	in muscle, anywhere	anywhere; broad
How they form	within seconds	over a minute or two	over a minute or two	over stillness; a stressful moment adds some	within seconds	over a minute or two	within seconds
Slow breathing	74% let go (typical)	0% let go (typical)	25% let go (typical)	0% let go (typical)	25% let go (typical)	5% let go (typical)	0% let go (typical)
Attention, no touch	sometimes	often	often	no	no	often	sometimes
A resting hand	as the hand lifts (never under it)	under the hand	under the hand	not within a minute (never under it)	not within a minute	under the hand	not within a minute (never under it)
Rolling the knot	while rolled	not within the scene	not within the scene	not within the scene	not within the scene	not within the scene	not within the scene
Heat	sometimes	sometimes	no route	no	no route	no route	no route
A spark	a tingle over its patch	none: a slow warming	a twitch	none	a tingle when pressed	none	none at a release
What a recording shows	skin flow floods in	flow rises slowly	stiffness falls in the muscle	the layers slide again	the nerve's firing falls	a motor unit falls silent	nothing in the tissue

What a breath would have to do

However complex its pattern, a breath reaches a knot only through a few routes: the drive to its vessels or motor units, the tissue's movement, attention. So the models need its effect, not its pattern. For each theory: the least calming of the knot's own drive, aimed at it and held for a time, that frees the typical knot (thick line) and the easiest quarter (thin), in 32 settings; the grey bars are what a breath does. Perception has no drive of its own at a place, so its calming is of arousal, everywhere.

How it is made

Each theory is a model in its strongest form, every number taken from a quoted line of a paper or marked as a guess with a range. The same patch, the same scenes, the same hand and breath for every theory; each places its own units where its anatomy puts them. What is felt comes from one rule for all: a finger feels stiffness through skin and fat, fainter the deeper it lies. The films show one typical setting; the lines beside each theory speak for all 32. In every theory a stressful moment forms knots and the held stress keeps them: that is how stress is scaled, not a finding.

A model shows what a mechanism would do; it measures nothing, and it cannot say which theory is true. The scenes are things that happen to a model, not instructions. Deep somatic and meditative work carries real risks: see the site's moderation guidance.

one perforator up close research the plan every line, as a note the engine every number

run 6f86748a92fc · commit 119f6bd · 32 settings per theory