Knots: seven mechanisms, what each would do, and a recording that would tell them apart

1. The problem

Massage therapists, physiotherapists and contemplative practitioners describe the same thing in different words: discrete, tender places in the soft tissue, most often over the neck, shoulders and back, that can persist for years and yet let go within seconds. Accounts from contemplative practice and bodywork add detail: knots seem to answer to the breath, letting go on a slow out-breath; deeper ones to focused attention; a patient hand or a foam roller releases them one place at a time; a hot shower softens many at once; they gather where stress is held and limit movement; a release can come with a small pop or a tingle across a patch of skin, and another knot may move in nearby. These are impressions, not measurements, but they are specific enough to test.

The speed is the first clue. Collagen, scar and adhesion change over days and weeks, so whatever lets go within a breath must be a state the body is holding, not a structure it has built. At least seven candidates have been proposed or can be assembled from established physiology, and they disagree about where a knot is, what holds it and what releases it. None has been tested against a recording of a release, because none has been made.

Rather than choose among them, we ask each to show itself: to grow its own knots, from its own physiology, in the same piece of body under the same scenes, and to say how they would feel, respond and let go, and what an instrument would record. Where their knots differ in something a hand or an instrument can see, a measurement can decide.

2. Models and methods

2.1 One patch, the same scenes

Every model lives in the same 4 cm square of the upper back over the trapezius (skin, fat, the superficial and deep fascia, muscle), with stress held more toward the neck and at one place beside a spot where a hand rests and attention goes. Each places its own units where its anatomy puts them. The scenes are the same for all: rest, three minutes of stress, then a held level; thirty slow breaths (4 s in, 6 s out); the same with attention on the spot; subtle breaths with attention; a hand resting on the spot for a minute; a roller over a quiet place, and over the knot; five minutes of heat, as in a hot shower or a sauna; an hour of stress that wanders as moods do; half an hour more at the held level. Stress is one shared unit (0 rest, 1 a surge) mapped onto each theory's own inputs in a runner whose mapping is written out.

2.2 The seven models

Perforators. A small artery with tone in its wall is a switch: the pressure inside pushes it open, its muscle pulls it shut, and because muscle is strongest at one length there is a band of tone over which both states are stable [1, 2]. Parameters come from measured wall strength, thickness and pressure; the breath acts through the skin's sympathetic reflex [3] and through movement, which cuts smooth muscle's force [4]. Vessels form trees of a parent and its children sharing pressure, with conducted responses along their walls [5]. Each vessel crosses the gliding plane in a sleeve of sliding tissue [6] that its flow keeps wet; shut, the sleeve dries, its hyaluronan crowds (viscosity rising as the third or fourth power of concentration [7]), and it can jam. A knot is a vessel held shut by its own wall; a shut vessel starves its patch, and its nerve bursts as blood returns (the spark).

The vascular latch. Johnson's account [8]: a held prediction commands a region's small arteries; the contraction clamps the region and cuts it off from awareness, so the prediction cannot update. Formalised as a loop between the prediction and the clamp, set from the account's two claims (durable while unattended; not set by ordinary stress alone); attention or touch lets the prediction update. The latch-bridge state itself economises and does not remember [9].

Trigger points. The energy crisis [10]: endplate activity contracts sarcomeres, the contraction squeezes their capillaries, and starved sarcomeres cannot relax; sustained pressure lengthens the contracture over the tens of seconds of pressure release; a fast release twitches.

Densification. The gliding layer as a thixotropic material with a critical stress [11]: it rebuilds a structure at rest and loses it under shear, so between two drives both a jammed and a fluid state hold; a jammed patch gives way slowly and then all at once. Hyaluronan alone hardly changes with pH in the body's range [12] and does not gel [13], so a structure that jams must be more than hyaluronan.

Nerves. A nerve of the skin, sensitised where it pierces the fascia beside a perforator [14, 15], fires when pressed or stretched after a neuritis [16, 17] and when the sympathetic drive rises, about fourteen seconds late [18]. It has no switch: a knot is felt while its drive lasts.

The motor switch. Spinal motor neurons, once recruited, keep firing on a fraction of the input that recruited them, through persistent inward currents set by serotonin and noradrenaline [19, 20]; the metabolites of the contraction keep the loop going [21]. A knot is a latched unit; it unclenches at once when inhibited.

Perception. Central gain, arousal and attention turn ordinary input into a place felt as a knot, with a little hysteresis [22]. Nothing in the tissue changes.

2.3 What is felt and recorded

The same rules for all: a finger feels stiffness through skin and fat, weaker as (a/√(a² + d²))³ for a region of size a at depth d, above a threshold of touch; tenderness from each theory's own sensitiser; and what an instrument would record at the spot and at a pressed sham: skin flow by laser speckle, single motor units by surface EMG, stiffness by elastography, the layers' sliding by ultrasound, a nerve's firing.

2.4 Settings, scale and portraits

Each theory runs in 32 settings drawn by Sobol sampling across the ranges of its uncertain numbers. Every parameter carries its source and the quoted line, checked word for word against PubMed records, or is marked as a guess with a range. One shared rule sets each theory's scale (a stressful moment forms knots at the typical place; the held stress keeps them; the held stress alone does not form them); it is calibration, not a finding. Each theory's portrait is generated from the runs: a line holds in every setting (structural for that theory), in some (with the sampled number that best predicts where), or in none. The code and every number are open [23].

3. Results

3.1 What physics rules out

A liquid resists motion, not position: at its most viscous, hyaluronan around a vessel exerts a drag 12 times weaker than the pressure holding the vessel open [24, 25]. A patch of skin warms with a time constant of at least 44 s even at maximal skin flow, so warming cannot release a knot within a breath. Warmth thins hyaluronan by only about 2% a degree [7]. A vessel pressed shut cannot reopen until the pressure lifts. The sympathetic drive reaches a sensitised nerve only after about fourteen seconds [18]. And healthy small arteries hold themselves shut only over a band of tone (0.62–0.93 of maximal, rest 0.19); held shut at 4× resting tone, their muscle adapts to the new length and the knot sets, holding at rest, after about 1.7 h [26, 27].

3.2 The trace at release

The seven models part most sharply at the moment of release (Table 1). In every setting of each model the trace is its own; one recording at a release would separate them.

3.3 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. For each model we found the least calming of the knot's own drive, held for a given time, that frees the typical knot (Figure 1). In no model does a single breath free the typical knot through drive; minutes of slow breathing do in some. The breath's movement, and attention where a model gives it a route, do what a single calming cannot.

Figure 1. The least calming, as a share of the holding stress, that frees the typical knot (thick) and the easiest quarter (thin), against how long it lasts; grey bars: what a slow out-breath, a subtle breath and minutes of slow breathing do.

3.4 The perforator view in detail

Trees. 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 [5], and a child's knot sits some 4-5 mm from its parent's.

Stiffness. Each vessel passes the fascia in a sleeve of sliding tissue that its flow keeps wet; shut, the sleeve dries, its hyaluronan crowds [7], 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).

Rolling and warmth. 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 [4] and the roller shears its sleeve. 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 [28], 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.

Figure 2. Two of the switches. Left: a perforator's wall, open or shut over a band of tone. Right: the gliding layer, fluid or jammed over a band of movement: a jammed patch stays jammed under movement that keeps a fluid one fluid.

3.5 The reports and the models

Only after the models had been run did we set practitioners' reports against them (Table 2). The perforator model, the most detailed of the seven, reproduces the most; attention alone points to the latch and the motor switch; the layers' stiffness is shared by densification, trigger points, the motor switch and the perforators' sleeves; four reports reach no model yet.

4. Discussion

More than one kind of knot. No single model reproduces every report. A perforator held shut, with its sleeve, accounts for the spark, the release with a roller or a lifting hand, stiffness that outlasts the release until the layer is moved, and larger knots freeing smaller ones; heat, once the breaths' own share is taken out, frees about as many latch knots as perforator knots. It does not account for knots that let go to attention alone, which a latched clamp or a latched motor unit would. The simplest reading is that practitioners meet at least two kinds of knot, and a recording should classify each release rather than crown one theory.

Perforators and latches: one muscle, two holds. Both are vascular smooth muscle held contracted, and heat reaches both through the same number in these models. They differ in what holds them: a perforator by its wall's mechanics (a band of tone over which open and shut are both stable), the latch by information (a prediction that keeps its vessels clamped and the region out of awareness). They need not compete: a held prediction could be the command that keeps a perforator inside its band. Such a knot would let go to attention and to a roller alike, and spark as it went: the combination is the next model to build.

Peeling and filling. Some describe the fascia beneath the skin peeling apart along lines, leaving places that feel hollow, and then a filling-in felt as re-stitching. The perforator view suggests where to look: the perforators and their sleeves are the stitches that tether the superficial fascia to the deep; if many sleeves free at once and the returning flood fills the gliding plane, the layers could part along the lines between tethers, and close again as the fluid is taken up and the sleeves re-hold. This is not modelled; the fascia as a system of its own is the next model to build.

What would refute each. A structural line in a model, one that holds in every setting, is what a measurement can refute. A perforator knot that lets go while a hand still presses it; a latch knot that lets go with no rise in flow; a trigger point with no change in the muscle's stiffness; a densified patch whose layers slide as before; a nerve's knot a local anaesthetic cannot touch; a motor-switch knot with no unit falling silent; and, for perception, any change in the tissue that times itself to the felt release.

Limits. Most numbers specific to knots are guesses, marked and sampled; the models' traits say which of them decide each result. Three of the seven formalisations (the latch, densification, nerves) are ours, not yet checked by those who hold them. The shared scale gives every theory knots to study, by construction. The reports are rough impressions. A model shows what a mechanism would do with these numbers; it cannot show that knots are that mechanism.

Safety. Deep somatic and contemplative work can cause severe and lasting difficulties [29, 30]. Nothing here is a practice or a protocol for anyone to try; any recording belongs under ethics review, with ordinary breathing, comfortable pressure, and nothing on the front of the neck.

5. The measurement

A pilot of twelve participants with three to five knots each, about ninety minutes, gives forty to sixty release events, each marked by the participant with a button. At each: laser speckle over the knot and a pressed sham [31, 28]; high-density surface EMG over the upper trapezius, decomposed into single motor units [32, 33]; and ultrasound of the layers' sliding [34], before, at, and after the release and after the shoulder moves; with a blinded Doppler map of perforators under the knots. The predictions are fixed in advance by the models' runs. The protocol is on one page: the pilot.