Muscle Weakness Is Almost Always Inhibition

Find the nociception, restore the strength — the physiology behind it, the studies that confirm it, and seven demonstrations from the clinic.

by Chris Knudsen, DPT - Muscle IQ Physical Therapy · Orem, Utah · MuscleIQ.com

For practicing physical therapists.

 

The claim

When a muscle tests weak, the weakness is almost always inhibition — the nervous system has turned the muscle's drive down — not the muscle mechanically failing. Something is sending nociceptive input, and that input inhibits the muscle. Find the nociception, reduce it, and the strength comes back. The MRI shows the surgeon where to put the scalpel; the weak muscle shows the physical therapist where to treat.

This is Inhibitory Driver Mapping: we are finding muscle inhibition, and finding what is driving the muscle inhibition.

The established literature demonstrates that nociceptive input can inhibit motor output. Based on two decades of clinical observations, we propose that this mechanism accounts for a much larger proportion of clinically observed weakness than is generally recognized.

A scope note before the argument. This paper is about the weak muscle a physical therapist finds in an ordinary caseload — a musculoskeletal patient with an injured or irritable tissue, tested on the table. It is not a claim about every possible cause of weakness. A muscle can also be weak from causes this model does not cover: a myopathy, a motor neuron disease, a denervated muscle — where nothing turns back on because there is no drive left to restore. "Almost always" is meant within the population this paper is about, not as a statement about all weakness everywhere.

A second limit. Nothing here tells you which treatment to use — manual therapy, therapeutic exercise, modalities are your call. What the paper offers is the weak muscle test, and the three things it tells you: that the weakness is inhibition rather than a failed muscle; that inhibition you can reverse is inhibition you can rehabilitate; and where to aim the treatment.

This paper makes its case in this order:

• How the nociceptive signal travels from tissue to muscle.

• A weak muscle test is inhibition, not a failed muscle.

• Where the weakness starts and how far it reaches.

• Why damage and inhibition are not opposites, and why a tear or other structural finding does not equal the weakness — the imaging and strength research show both.

• The physiology of how nociception weakens a muscle.

• How a second nociceptive input restores it.

• What can generate the nociceptive input, and why it needs a condition to act on.

• Seven demonstrations from the clinic.

• What Inhibitory Driver Mapping lets you see.

• What the research settles, what is our own clinical claim, and what future studies should try to resolve.

 

The machinery, from tissue to muscle

Figure 1

 

Three things arrive at the muscle's motoneurons, from three different sources.

• Voluntary command from the cortex, down the corticospinal tract. This is what you're asking for when you say "hold it there."

• Background drive from the reticulospinal system — running whether or not the patient is doing anything, and set by the state of the brainstem. This is the tone level.

• Segmental input at the cord — spindle afferents and the interneurons around them.

The path in summary. A nociceptor fires in irritated tissue — swollen, compressed, inflamed, or injured, whether or not the patient feels it as pain — and its signal enters the dorsal horn. From there it can reach the muscle by three routes of increasing length, and each one ends on the same cell — the alpha-motoneuron, the last neuron before the muscle.

• The short loop. Straight onto interneurons in the cord, which act on the motoneuron directly. No brain involved.

• The brainstem loop. Up to the reticular formation, which sends pathways back down that set the background tone of the muscles — working through those same spinal interneurons.

• The cortical loop. Up to the cortex, which sends the corticospinal tract back down to those same motoneurons. Nociceptive traffic arriving at the cortex can change how much of a command leaves it.

A fourth route acts on the signal rather than on the muscle: the brainstem sends fibers back down to the dorsal horn that damp the incoming nociceptive traffic itself.

The alpha-motoneuron adds up whatever reaches it. How many of the muscle's motoneurons fire, and how fast, is its strength.

A weak muscle test is inhibition, not a failed muscle

When experimental muscle pain was induced in healthy subjects, maximal voluntary knee-extension torque fell, while the muscle's own contractile properties — assessed by twitch interpolation — were not affected. (Twitch interpolation is two readings: an electrical pulse delivered to the relaxed muscle, which shows how the muscle itself responds, and the same pulse added during a maximal effort, which shows how much of that the person is driving.) What the patient lost was drive, not muscle, though that method cannot separate a spinal from a supraspinal source [2]. This was healthy subjects with experimentally induced pain — a mechanism demonstration, not a clinical population — but it isolates the point cleanly: the muscle can fire; it has lost permission, not capacity.

Two clinical studies make the same point by taking the nociception away. In fourteen patients with shoulder impingement — nine of them with full-thickness rotator cuff tears — a subacromial injection of local anesthetic was followed by a retest five minutes later: thirteen of the fourteen tested stronger in abduction on manual muscle testing, and isokinetic peak torque rose by a mean of 48%. The strength gains did not differ according to whether a tear was present — though with nine tears among fourteen patients, that is a small comparison and a null result rather than a demonstration of equivalence [30]. In sixteen patients with chronic arthritis and clinically active synovitis of the knee, draining the joint raised quadriceps peak torque immediately (p=0.004); by fifteen days, after a corticosteroid injection as well, torque was up an average of 25% — a change the authors note is too fast to be muscle growth [31]. Both studies are small and uncontrolled, and neither blinded the patient. What they show is a muscle producing more force within minutes of the joint's input being reduced.

Our clinic tends to see a high amount of chronic pain patients. In the large majority of our patients, a weak muscle reverses the moment a second nociceptive input is added. A muscle that turns back on in seconds did not lose its capacity — the weakness was inhibition all along. This is our own clinical observation, not a controlled trial.

Where the weakness starts, and how far it reaches

Arthrogenic muscle inhibition is the reflex inhibition of the muscles around a joint after injury, swelling, or surgery: voluntary activation fails, whatever state the muscle itself is in. A recent review notes that it can limit recovery after total knee replacement despite technically successful surgery [29]. The joint was rebuilt and the muscle still would not turn on. It is common: among 300 consecutive acute anterior cruciate ligament injuries, 56.7% showed clinically detectable inhibition before surgery [32]. A review looked at three common knee problems — anterior cruciate ligament injury, anterior knee pain, and osteoarthritis. It surveyed three levels: the motor cortex, the corticospinal tract carrying its output down, and the motoneuron pool that tract lands on. Adaptations turned up at all three. Every one of the conditions showed altered excitability, reduced voluntary activation, and reduced force [33].

What we find in the clinic has two stages. A standing nociceptive source inhibits in its own region — that is the baseline weak muscle found at rest. Asking the patient to use that inhibited muscle — a straight-leg raise, a lunge, a turn into the painful direction — brings on the global inhibition. Using that muscle adds nociceptive input, and that input can reach the reticular formation and lower the background drive going out to muscles far from the source — so a muscle with nothing wrong with it tests weak.

Does arthrogenic muscle inhibition stay in the muscles around the joint? The definition says so — but that is where the term begins, not something a study established. The question has rarely been asked. Where it has, the answer has not been uniform: in the experimental knee-pain study cited below, tibialis anterior — at a different joint — did not change [40], though the stimulation in that study was set for the quadriceps. Across 107 patients with knee osteoarthritis — a population the arthrogenic inhibition literature claims as its own — Simis and colleagues found that worse cartilage damage went with more motor inhibition, recorded at a muscle of the hand [28]. Reported pain ran the other way in the same analysis, so what the remote inhibition tracked was the joint's condition, not the patient's pain. The study is cross-sectional, and its authors read the increased inhibition as compensation rather than as damage.

An arrangement in the brainstem loop could carry a reach that wide. In a review whose authors set out their own stimulation mapping in the decerebrate cat, sites in the dorsomedial reticular formation suppressed muscle tone on both sides of the body while sites in the ventromedial reticular formation augmented it; in one animal, a dorsomedial site in the medulla abolished muscle tone while a neighboring ventromedial site increased it, and in both directions the effect outlasted the stimulus [45]. Two limits the same authors state: laboratories disagree on where the inhibitory region sits — others placed it ventromedially — so what is settled is the existence of opposing systems, not their exact address; and no nociceptive input was applied in that work, which maps the machinery and its signs, not what recruits them.

The two-stage sequence itself is our clinical observation: we know of no study that compares a standing nociceptive source against the same source actively loaded and measures a muscle away from it.

Damage and inhibition are not opposites — and why the tear does not equal the weakness

None of that means damaged tissue has little to do with weakness, as if the inhibition arrived out of nowhere. Often the inhibition is triggered by damaged tissue. The tissue generates the nociceptive signal, and that signal is what turns the muscle down. A rotator cuff tear, a torn meniscus, a herniated disc — any of these can be sitting right there on the image and be the driver of the weakness, not by subtracting contractile tissue but by what it is sending. The knee osteoarthritis patients above make the point: worse cartilage damage went with more motor inhibition, on two separate measures [28] — an association across patients, not a change measured within one. Damage and inhibition rose together rather than standing in for one another.

That changes the question worth asking. It is not whether the weakness belongs to the tear or to the inhibition; it is whether the muscle can be turned back on. If it can, the weakness was inhibition, whatever the tissue looks like.

The mechanics. A tear removes some contractile tissue, but as long as there is enough intact muscle and enough surrounding synergists to compensate, that loss is often not enough to move the strength test on its own — once the irritation is reduced. Much of the weakness you measure can be inhibition rather than missing tissue — which is what the anesthetic and drainage studies above put numbers on [30, 31].

The scan does not match the symptom. Structural findings are common in people with no pain at all. In 230 knees of pain-free adults, 97% had at least one structural abnormality on MRI, including meniscal tears in 30% and cartilage abnormalities in 62% — with 41% showing moderate-to-severe lesions and 31% severe ones [3]. Elite volleyball players with no shoulder pain showed abnormal shoulder MRIs across the board, including partial rotator cuff tears [4] (a small study). And baseline lumbar MRI did not predict who would develop low-back pain over the following seven years [5]. A finding on a scan is common, and it is not the same thing as a symptom.

Where strength itself was measured, the tear does not track it as cleanly as you would expect — pain does. In 891 shoulders of women aged 64–87, rotator cuff tears of all sizes had no isolated effect on strength in those under 70, while age, pain and testing the non-dominant arm each reduced it. In the over-70s, full-thickness tears did reduce strength — by roughly 30% for small and 40% for large tears [6]. Age matters here, and the paper should not be read as saying a tear never costs strength. Patients with tears severe enough to require surgery tested near-normal on manual muscle testing before repair — external rotation graded 4.4 out of 5 in traumatic tears and 4.5 in atraumatic ones [7]. Over 15 years, unrepaired tears grew from a mean of 16.2 mm to 31.6 mm in the anterior-posterior direction — close to double — while the strength difference in favor of repair was 1.8 kg and did not reach statistical significance [8]. And in isokinetic testing before and after repair, pain correlated with the supraspinatus's initial peak torque while fatty infiltration grade correlated with nothing — though tear size did correlate with preoperative initial peak torque, so size is not irrelevant either [9].

The same split shows up in surgical populations. In 61 patients scheduled for shoulder surgery, marked weakness in abduction and external rotation was recorded across every tear size, including in the eleven patients found at surgery to have no tear at all — though in that same study one deficit did track tear size: abduction strength at 10° of abduction, where 20 of 25 patients with large or massive tears lost more than half their strength against the other side, compared with 1 of 11 with no tear [34]. Every other strength and range-of-motion measure in it was unaffected by tear size. And a Cochrane review of physical examination for lumbar disc herniation concluded that most individual tests, muscle weakness among them, perform poorly when used on their own; note that nearly all of its studies came from surgical populations, where herniation is far more common than in an outpatient caseload [35].

Injured tissue does not equal muscle weakness. The scan is a data point, not a verdict.

Why nociception turns a muscle down

If the tissue is not what fails, the signal it sends is what does — and that physiology is well established. Nociception is not the same event as pain: a nociceptor fires when tissue is irritated, swollen, compressed, inflamed, or injured, and that firing does not always reach the threshold to feel pain. When it does not, the nervous system is still acting on the signal while the patient feels nothing.

This has been shown with pain removed from the circuit entirely. The cat work cited here used decapitated, high-spinal preparations — no brain in the loop at all, and therefore no possibility of pain — and nociceptive afferent activation still inhibited the monosynaptic input from muscle spindle afferents to alpha-motoneurons [11] — the one-synapse connection running from the muscle's own stretch receptors straight onto the cells that fire it. That is the short loop on its own, with nothing above the cord to contribute, and it settles the narrow point: the motor consequence of nociception does not require the signal to be felt.

The same dissociation has been recorded in people. After medial meniscectomy, fourteen men were randomized to one of three arms — 10 ml of bupivacaine into the knee, 15 ml, or no injection, which is roughly five men per arm — with quadriceps inhibition measured as the drop in the muscle's electrical activity during a maximal effort, against each man's own preoperative reading. The 10 ml arm had less pain than the no-injection controls but the same inhibition; only the 15 ml arm reduced both. At three to four days the drop in that electrical signal was still large — a median of 75% below each man's own preoperative reading — while pain was only mild or absent, and at ten to fifteen days it was still 35% with little or no pain. The authors concluded that the inhibition was not simply due to perceived pain, but due, at least in part, to stimuli coming from the knee [36] (read from the published abstract; no full text is available).

The same separation shows up over longer stretches. People with recurrent back pain, tested while symptom-free, still activate the short fibers of the deep multifidus late compared with controls; and on the previously painful side the normal ordering within that muscle — its short fibers firing before its long fibers — is absent [37]. After anterior cruciate ligament injury, motor unit firing rates remain altered up to a year after surgery even as maximal strength increases [38]. In the back-pain study the patient's report and the state of the muscle came apart; in the anterior cruciate ligament study it was the strength number and the state of the muscle.

One distinction has to be kept clear, because the two get confused. "Below the pain threshold" means the nociceptors are firing and the patient does not feel it. It is not the same as a stimulus too weak to activate nociceptors at all, which is a different case entirely. A non-painful laser produced no change in motor-cortex output while a painful one did [12] — though whether that non-painful stimulus activated nociceptors at all is not established, so the study marks the distinction rather than settling it.

There is also a causal demonstration that an afferent class other than pain can drive this kind of inhibition. When group III/IV muscle afferents — the fibers reporting a muscle's mechanical and metabolic state during fatigue — were blocked with intrathecal fentanyl in cycling subjects, the fall in motor-cortex excitability that normally accompanies fatiguing exercise did not occur; with those afferents intact, it did. The authors attribute the effect to these sensory neurons facilitating a class of inhibitory interneurons within the cortex that use GABA, the brain's main inhibitory neurotransmitter [39]. Nothing in that loop requires pain.

That last finding was read at the cortex — the cortical loop — and that study could name the level because it also stimulated lower down, at the cervicomedullary junction, and compared the two. The cortical measures that follow are readings of that whole route — so the size of a motor-evoked potential, the muscle's response to a single magnetic pulse delivered over the motor cortex, is set by the whole of it: cortical excitability, corticospinal transmission, and the excitability of the spinal motoneurons. Under experimental pain especially, the motor cortex can show more inhibition — higher short-interval intracortical inhibition (SICI), another marker of cortical GABA activity, though a different receptor population from the one in the fatigue study above — and that evoked response can get smaller. (This is clearest for acute experimental pain; in chronic pain the SICI picture is mixed — the chronic-low-back study below [14] found no SICI difference at all.) A systematic review and meta-analysis of 49 studies found experimental pain reduces corticomotor excitability, with large effects when the pain lasted milliseconds to seconds and moderate effects when it lasted minutes to hours; for pain lasting days to weeks there was no group-level effect [10]. The cortical measures are not uniform, and the lower limb is where they diverge. During experimental knee pain, quadriceps motor-evoked potentials increased rather than decreased, with no change in intracortical inhibition [40]; the same group found the same increase after filling a healthy knee with fluid [41]. Both papers read the rise as increased drive arriving above a motoneuron pool that is being held down below it — other studies using the same experimental pain model report falls in quadriceps force, in voluntary activation, and in the H-reflex (a spinal reflex tested electrically) [40] — which those authors read as inhibition of the motoneurons themselves rather than gating of the sensory nerve feeding them, since nociceptive input from the knee reduces that gating rather than increasing it. Neither Rice study measured strength itself, and the authors of the pain study say they cannot confirm the muscle was inhibited during their recordings [40]. Down at the cord, the substrate is direct: nociceptive and non-nociceptive afferents converge on common interneurons in the reflex pathways to alpha-motoneurons — nociceptive input is part of the flexor reflex afferent system, the group of inputs that drives withdrawal from something harmful — and monosynaptic input to those motoneurons was inhibited during nociceptive conditioning [11] (in the cat).

And the inhibition is not always confined to the spot that hurts. A painful input to one hand can reduce motor-cortex output on both sides of the brain [12] — and that study can name the cortex, because it stimulated the same people two ways: the response to a magnetic pulse fell, while the response to direct electrical stimulation of the corticospinal fibers did not. When people imagined thumb movements, experimental pain prevented the rise in corticospinal excitability that normally accompanies motor imagery — and it did so whether the pain was at the finger or at the knee, regardless of location [13]. Chronic low-back-pain patients show reduced corticomotor excitability measured at a hand muscle, unrelated to the back: a higher resting motor threshold and lower intracortical facilitation — the paired-pulse measure of excitatory circuits inside the motor cortex, the counterpart to the inhibitory one above — with no difference in evoked-potential amplitude or short-interval intracortical inhibition [14]. Each of those studies applied an experimental pain under laboratory conditions. The knee osteoarthritis patients are the exception — there the nociceptive source was the patient's own diseased joint, not a stimulus applied in a lab [28]. Deep muscle pain in one hand reduced motor output to a neighboring muscle in the same limb but did not cross to the opposite hand, and pain in the skin over that same muscle did nothing at all [42]. How far it travels is not fixed, and the size of the effect also tracks the input: of two nociceptive intensities only the higher reduced corticospinal excitability, and a subset of subjects showed an increase rather than a decrease [43]. What spreads, and how far, appears to depend on the kind of input and its intensity rather than on a single rule. This is why a nociceptive source in one place can show up as a weak muscle somewhere else.

Why a second input turns it back on

The nervous system also has a mechanism for switching the inhibition off, and it is the fourth route named at the outset — the fibers that act on the incoming signal rather than on the muscle. A second nociceptive input, from anywhere on the body, engages the body's own "pain inhibits pain" system — called conditioned pain modulation (CPM) when it is measured in people, and diffuse noxious inhibitory control (DNIC) in the animal work where it was first described — run through the subnucleus reticularis dorsalis (SRD), a nucleus in the medullary reticular formation. A review of that literature reports that the SRD projects through the dorsolateral funiculus of the spinal cord onto the dorsal horn, where ascending nociceptive fibers hand off, releasing noradrenaline, serotonin and the body's own opioids at that junction [16]. In the original work, noxious stimuli applied to various parts of the body powerfully inhibited the spinal neurons that carry nociceptive signals — the dorsal-horn convergent neurons — depressing their responses by 60–100%; non-noxious stimuli did nothing [15] (in the rat; reviewed in [16]).

Rubbing a sore spot makes it hurt less, and there are two ways that could work. The touch signal could interfere with the nociceptive signal down at the cord, before either one reaches the brain. Or the touch could travel up to the brain and change what the brain does with the nociceptive signal once it arrives. One study was built to tell those two apart.

Eleven healthy volunteers received two inputs to the right forearm: an electrical stimulus delivered into the skin, which drives the thin nociceptive fibers, and a mechanical touch, which drives the low-threshold receptors that report ordinary contact. The two were paired at controlled intervals. Because the two signals travel at different speeds, the interval decides where the touch can still act on the pain signal. Delivered together, the touch runs ahead the whole way — reaching the cord first, and reaching the cortex about sixty milliseconds before the pain signal, too early to meet it there — so a change seen under that timing has to have happened below the cortex. Delivered sixty milliseconds apart, the two arrive at the cortex together. That sixty-millisecond figure was not arbitrary: it was settled by comparing forty, sixty and eighty milliseconds against the calculated difference in arrival time at the cortex. Recording the magnetic fields the brain gives off — magnetoencephalography, a scalp recording of cortical activity — placed the pain response in the posterior insula and the touch response in the parietal operculum, two separate areas.

Timed for the cortex, the touch significantly inhibited the pain-related activity in the posterior insula. Timed for the cord, the pattern was different: both areas were modulated, rather than the insula alone. Reported pain fell significantly under both timings. Both conditions were recorded at the cortex, so the spinal label is the authors' inference from conduction timing rather than a spinal recording [46].

So relief of this kind is not carried by the spinal gate alone. There is a route in which the touch signal reaches the brain and suppresses how the nociceptive signal is processed once it is there. The authors read the posterior insula as the place that happens — not a passive pain center, but a region that integrates sensory information about the state of the body, and where one arriving signal can alter how another is represented.

A second input turning muscle inhibition off is now proven in humans. Two experiments by the same group, using the same procedure — hypertonic saline infused into a small muscle of the hand to produce pain — differ in one respect and give opposite motor results. In the earlier one, electrical stimulation of the nerve at the wrist ran in that same session, at a rate and intensity the same group's later paper reports as five hundred stimuli per block, sufficient to make the painful muscle visibly twitch [48]; motor output to that muscle was not suppressed while the person was in pain, and dropped only after the pain had resolved (twelve subjects; read at abstract level) [47]. In the later one, the sensory and the motor measurements were moved to separate days, so the sustained stimulation was not running during the motor recording — only four pulses at each time point, delivered to check the muscle's own response, though each was strong enough to contract it; this time motor output was significantly suppressed during the pain itself, by the third block of recordings (ten subjects, full text read) [48]. The authors' own account of the difference is that the stimulation had prevented, reduced or masked the drop. That stimulation was called comfortable by the subjects, but it was repeatedly contracting a muscle with hypertonic saline in it; by the definition used throughout this paper — an input counts as nociceptive because of the tissue it loads, not because of what the person reports feeling — it was a nociceptive input. Three limits, all material. The comparison runs across two studies with different subject groups rather than as a manipulation within one, and the authors of the later paper call the observation incidental and ask for a study designed to test it. What was measured is the size of a muscle's response to a magnetic pulse over the motor cortex — the whole-route measure described above — not force on a dynamometer, not a manual muscle-test grade, and recorded in a resting muscle. And on the authors' own reading there are two routes to it: one where the stimulation cancels the pain-related fall in cortical facilitation, and one where it raises the excitability of the spinal cells driving the muscle directly, by way of the nerve volley traveling back toward the cord and the sensory feedback from each evoked twitch. The second route would lift the measurement with the nociceptive traffic unchanged, so this pair of studies constrains the mechanism less than it constrains the timing.

Two further findings connect that pain machinery to motor output. In the cat, the SRD and the medullary reticulospinal neurons that set motor tone are interconnected through collaterals of their descending axons, with reciprocal excitatory connections — and the coupling is tightest in the cells that respond to noxious input, over 81% of which send collaterals to the pain-damping nucleus [1]. Two human studies tie the two systems together in people, and they run in opposite directions, which is why what they support is a coupling rather than a direction. In 41 healthy adults, greater motor corticospinal excitability went with more efficient CPM — the same "pain inhibits pain" damping described above, measured in people [17]. And in 33 women with chronic myofascial pain, the second nociceptive input was the opposite hand held in near-freezing water for a minute: the women whose pain did not drop during it — whose descending inhibition failed to engage — showed more intracortical facilitation (a ratio of 1.43 against 1.11) and larger motor-evoked potentials (1.93 against 1.40 millivolts) than the women whose pain did drop [18]. No group size is given here on purpose: that paper's abstract and its own tables state the two group sizes in opposite order, so the figures are reportable and the split is not. Both studies are cross-sectional, and neither measured force. Note also which way that second study's finding runs: failed pain inhibition went with more motor excitability, not less, and its authors read that increase as compensatory. What turns the muscle back on is not this coupling reversing its own sign; it is the second input damping the nociceptive traffic that was driving the inhibition.

 

Figure 2

 

One thing in the clinical pattern is not explained by anything cited above, and it is the part the method turns on. A single input drops every muscle that was testing strong and, at the same moment, restores the one that was testing weak. Nothing in the descending damping system selects a target — it is described as diffuse — and the two human studies linking pain modulation to motor output run in opposite directions.

What follows is offered as a question rather than a finding. The input may be doing only one thing, everywhere at once.

A nociceptive input turns down the drive to muscles, and not only the ones beside it — that is the spread set out above [12, 13, 14], and, where the input is an inhibited muscle being used, the reticular route named at the outset. That is the half of the pattern where a strong muscle goes weak.

At the same time, a second nociceptive input engages the brainstem pain-damping system described above, which reduces the nociceptive traffic already running from an existing source. That is the half where a weak muscle comes back.

On a muscle carrying no source there is nothing for the damping to subtract from, so only the added input lands. On a muscle held down by a standing source, the damping has something to act on. The input is the same in both places, and the damping it engages is diffuse. What differs is the state of the muscle receiving it — and on that reading nothing has to select a target.

There may also be a second route to the same result, and it turns up in an ordinary touch. In twenty healthy adults with nothing wrong with their hands, brief puffs of air on the index finger cut the drive to that finger's own muscle to 85% of its unstimulated value; puffs on the ring finger left it unchanged; puffs on all five fingers together cut it to 70%. No pain was involved — the participants described the stimulus as taps. Measured with paired magnetic pulses, the same touch also changed two circuits within the motor cortex itself: it raised the excitatory one, intracortical facilitation, and it released the longer-latency of the two inhibitory ones — the response conditioned through that circuit nearly doubled, from 22% of the unconditioned response to 41%, in 14 of 16 people, meaning that circuit was holding the muscle down substantially less. The short-interval inhibition named earlier, the one that rises under experimental pain, did not change [53]. In healthy hands the net drive still came out down. In a muscle already being held down, a released inhibitory circuit is a second way it could come back up — and that one does not require the damping system at all.

What would have to be true is one comparison nobody has made: that at a muscle carrying a standing nociceptive source, the reduction in that source's traffic outweighs the traffic the new input adds. Both halves are published separately. The two set against each other, at one muscle, are not. None of that changes what the reversal tells you in the room.

So when a second, genuinely nociceptive input is added and the weak muscle strengthens immediately (temporarily), we learn three things at once:

• Diagnostic: the weakness was inhibition. A structural deficit cannot switch back on in seconds.

• Prognostic: reversible inhibition means recovery is more probable than where there is no drive left to restore. A muscle you can turn on is a muscle you can rehabilitate.

• Localizing: the input that turned it on marks a starting point for where to treat.

In the clinic, that turn-on is a probe, not the treatment — it tells you what you are dealing with and how likely it is to move, but it does not, by itself, fix anything.

 

What can generate the nociceptive input — and why it needs a condition to act on

The first input has to be genuinely nociceptive too, and what can supply one is wider than a palpable, tender lesion. Many ordinary things can generate it — pressure on a tender point, loading or using an injured joint (a lunge, a squat, a straight-leg raise), a held position such as a spine kept in flexion, a movement such as turning a painful neck, a sustained gaze to one side, even bright light. What they share is that none of them is nociceptive on its own. Each becomes a nociceptive input only when there is already a condition for it to act on: light only in a patient with photophobia, a lunge only on an injured or irritable knee, a held gaze only when the neck is involved, spinal flexion only with a low-back source. In a person without that underlying condition, the same movement, position, gaze, or light largely does not do it — and where it does, it is the exception rather than the pattern. For eye movement that pattern is now published: the same eye-movement tests that provoked symptoms in a third to two-fifths of concussed adolescent athletes provoked symptoms in at most 9 percent of healthy controls [49]. (The condition need not be felt as pain: a source can sit below the pain threshold and still drive inhibition — see the spinal finding above [11].)

Named diagnoses belong on the same list. A meniscus tear; a surgical incision with hardware in place; a lumbar disc herniation at L4–L5; cervical stenosis; a lateral epicondylitis — each of these can itself generate the signal that drives the inhibition. The driver does not have to be something painful. It is a condition in the body that the nerves communicate to the brain, and on those terms a structural or neurological diagnosis qualifies the same way a tender point does.

The healthy-subject evidence is the other half of this. When the visual system was activated in healthy people — who have no photophobia — it did not drive a nociceptive response; if anything it reduced one [19]. That is what the model predicts in advance — no condition, no nociception, no inhibition.

Seven demonstrations from the clinic

Each of the seven is a test–change–retest sequence recorded from our practice, on the standard 0–5 manual-muscle-test (MMT) scale. They are demonstrations of the phenomenon, not a treatment protocol. The mechanism under all of them is the one traced in "Why nociception turns a muscle down" and "Why a second input turns it back on": a nociceptive input reduces the drive reaching the muscle — above the muscle itself [2], at the cord [11], and along the corticospinal route [10] — and that reduction is not confined to the painful area [12, 13, 14]; a second nociceptive input damps the nociceptive traffic that was driving the inhibition [15, 16], and in people a second input running during the pain kept the expected motor suppression from appearing [47, 48] — a comparison across two studies rather than within one, read as an evoked response, not force. That is not repeated case by case. Where a case turns on evidence not already given, the studies follow the case.

What follows is the culmination of observations over two decades in the clinic doing these tests. The patient presents to the clinic with a complaint. The first MMT that shows weakness is a manifestation of a "first" nociceptive input, causing a local weakness. All other tests were strong. (If everything is weak everywhere you are dealing with something else entirely.) Sometimes you find two or three completely distant weak muscles. Sometimes groups of muscles test weak. Taking a full history from the patient will usually give you a heads-up as to where you might find a weak muscle. Let's assume for this example that one muscle tested weak and all other muscles tested strong. Where is the origin for the nociception for this muscle weakness? You have to make the call. Find it. A touch over the suspect tissue will cause a second nociceptive input. A second added nociceptive input will reverse that original weakness to a strength. The same input will make all the strong tests weak. Said from the muscle's side: a global inhibition trigger weakens every muscle that was testing strong, while any muscle that was already weak from a standing trigger goes strong instead. Said from the input's side: a new input affects two muscles differently. For the muscle that already carried a standing nociceptive source, the new nociception is a second nociception. For the muscle that carried nothing, the same new nociception is a first. Actively using a weak muscle is itself a second added nociceptive input, causing global muscle inhibition to previously strong muscles. Touching a tender spot during an active use of a previously weak muscle blocks the global muscle inhibition effect. At rest, touching that same tender spot causes global muscle inhibition, weakening previously strong muscles.

Case 1 — Active use of a painful knee causes inhibition beyond the knee. A patient presents with right knee pain (baseline hip flexion and hip abduction both 4−/5 — already inhibited at rest, before anything is loaded). Standing, shoulder abduction at 90° tests 5/5. The patient lunges forward onto the painful right knee, holding the table for balance; the opposite shoulder now tests 4−/5. The patient stands back up, unloading the knee, and the shoulder is 5/5 again. Loading the nociceptive source dropped an uninvolved muscle; unloading it restored it, immediately and fully.

Case 2 — A second tender point restores what a first one suppressed. Supine, with a painful right knee. The patient performs a straight-leg raise on that side; the raised leg tests weak — the local weakness — and a shoulder tested during the raise also goes weak, more so under a larger load. Pressing a genuinely tender hip-flexor point then restores both the shoulder and the straight-leg raise. The control matters: a non-tender or distant contact that generates no nociception does nothing. Only a genuine nociceptive input reverses it.

Case 3 — One neck spot restores both shoulders. Right-sided neck pain, painful on rotation to the right. Baseline: right shoulder abduction 4−/5, left 5/5 — the weak side is the side of the source. Rotating the neck into the painful direction drops the previously strong left shoulder to 4−/5. A specific spot midway up the right side of the neck, when pressed during that rotation, returns the left shoulder to 5/5. With the head facing forward again, pressing that same spot returns the baseline-weak right shoulder to 5/5. Remove the contact and the weakness comes back. One driver, both shoulders.

Case 4 — Closing the eyes restores strength after a whiplash. One week after a slow-speed, bumper-to-bumper collision, with headache, neck pain, and light sensitivity. Seated, eyes open, shoulder abduction at 90°: right 4−/5, left 4/5 — both already down at baseline, the involved side lower. With the eyes simply closed: right 4+/5, left 5/5 — both sides improved. Nothing was touched. Here the eyes-open state — the held visual load in a light-sensitive patient — is acting as the nociceptive input that drives the inhibition; removing it, by closing the eyes, lifts it. The stimulus was a position, not a palpated point.

What the studies show. In a light-sensitive patient, light can act as a nociceptive input — and the studies that show this describe substantially the same brainstem zone from three angles. Bright light drives firing in trigeminal pain neurons in animals, graded to intensity and abolished by anesthetizing the trigeminal nerve [20]; in people with light sensitivity, light activates the brainstem trigeminal relay, and that activation falls when the eye is anesthetized [21]; and in whiplash, where light is rated as painful, the convergence sits in the trigemino-cervical complex [22] — where the lowest part of the spinal trigeminal nucleus runs on into the dorsal horn of the upper cervical cord, so nociception from the face and eye and nociception from the upper neck arrive on shared cells. Consistent with that overlap, anesthetizing neck trigger points — an average of 3.8 lidocaine injections into the upper trapezius — cleared light sensitivity in nine of eleven affected patients while raising pain thresholds at remote sites; in ten separate healthy controls the same injection into the thigh changed no pressure threshold, though light sensitivity was not measured in that group [23]. From there, nociception inhibits muscle as nociceptive input does [10]. The control is consistent with the direction — in people without photophobia the same light does nothing [19] — and the one link still unmeasured, that light changes a muscle test at a distance, is taken up under future studies.

Case 5 — Actively loading a painful low back causes inhibition beyond the back. Low back pain, patient pointing to the area. Standing, shoulder abduction is 5/5. The patient bends forward, as if doing dishes, loading the painful low back; shoulder abduction drops to 4−/5. Holding that same bent position, contacting the lumbar paraspinal tissue near the reported pain returns the shoulder to 5/5 — the position never changed, only the input was added. (No non-tender control contact was run in this case, and the contact point's tenderness was not separately confirmed. No muscle near the low back was tested at rest, so this case records only the second stage.)

Case 6 — A lateral gaze can drive inhibition beyond the neck. Neck pain and headaches. Head fixed facing forward throughout. Baseline: right shoulder 5/5, left 4−/5 — the standing weakness. Holding the eyes to the right drops the strong right shoulder to 4−/5 and, at the same time, raises the weak left shoulder to 5/5; with the eyes back to center, both return to baseline; holding the eyes to the left does the same thing — the right drops to 4−/5 and the left rises to 5/5. The neck never moved; only the eye position changed. A lateral gaze, to either side, is itself the nociceptive input: on the strong right shoulder it drives inhibition and the shoulder tests weak, while on the left — which already carried a standing pain — it adds a second nociceptive input that cancels that inhibition and restores strength.

What the studies show. Eye movement with the head still changes neck-muscle activity, a coupling distorted in whiplash — nine chronic whiplash patients against eleven unmatched controls, tracking a moving target rather than holding a position [24] — so in a neck-involved patient a held gaze can load the neck as a nociceptive input, which then lowers corticomotor drive beyond the neck [10, 12]. Eye and neck nociception meet on the shared cells introduced under Case 4 [22]. On the already-inhibited left shoulder, the gaze acts as the second input and cancels it — consistent with the DNIC and motor-tone circuitry [1, 15]. "Can," not "does": the gaze is a driver only in a patient whose neck makes it one.

That eye movement provokes symptoms in an involved patient and largely not in an uninvolved one is established clinically — the vestibular and ocular-motor screen cited earlier, in sixty-four concussed adolescent athletes, with smooth pursuit, saccades and convergence the items doing the provoking [49]. Two things that screen does not do: it pools headache, dizziness, nausea and fogginess into one symptom score without reporting which of them drove a given item, so it establishes symptom provocation rather than nociception; and its items are eye movements, where this case holds an eccentric position. Loading the neck under a still head is a related maneuver rather than the same one. Patients with ordinary neck pain show significantly greater disturbance of smooth-pursuit eye movement under neck torsion than controls (20 patients against 20 controls) [50], and with the eyes closed, whiplash patients sway more than controls after the same torsion — the authors put it as torsion that "may lead to" the deficit (20 against 20) [51]. Cervical findings of this kind appear in idiopathic neck pain, whiplash and concussion alike, all three differing from healthy controls; on the eye-movement and position-sense tests the three patient groups did not differ from one another, though their postural balance did (17 to 20 per group) [52]. Three of those four were read at abstract level. None of them measured muscle strength, and the direction differs from the one used here: that literature reads neck afferents disturbing eye control, while this case reads a held gaze loading the neck.

Case 7 — Even after a stroke, part of the weakness is still inhibition. A patient presents six months after a stroke, with left-sided hemiplegia. Manual muscle testing of hip flexion: right 5/5, left 3/5. Assessment found two spots on the left quadriceps that, when palpated, changed the left hip flexion test to 4/5. The weakness did not resolve — this is genuine neurological damage and it remains genuine. But a portion of what would ordinarily be recorded as stroke deficit moved under a thumb, which means that portion was inhibition sitting on top of the damage.

Cases 4 and 6 make the plainest version of the point in the whole set: with nothing palpated, a sustained eye position is enough to create the inhibition and, on a muscle that already carries a driver, enough to cancel it.

Case 7 makes a different point, at the boundary. Genuine neurological damage is the kind of exception this paper named at the outset — damage that leaves less drive to restore — and it stays an exception. What the case shows is that the exception need not be clean: even where real damage is present, some of the measured weakness can be inhibition layered on top of it — and that layer is worth testing for rather than assuming away. A muscle graded 3/5 after a stroke is not necessarily 3/5 of surviving neural pathway.

What Inhibitory Driver Mapping lets you see

Inhibitory Driver Mapping helps me identify where the patient is weak. It helps me test what movements are triggering. It helps me see the nociceptive effect of certain movements or external stressors on the patient's muscle system. Once you learn these rules for how nociception and muscle inhibition work, any therapist can tell if any action is a positive or negative effect on the patient.

What the literature settles

Five things in this paper are well established. Nociception reduces motor drive above the muscle, and it can do so beyond the painful area — the motor-cortex and pain-modulation literature carries that for the corticospinal route, with the caveat already stated above that the lower-limb cortical measures do not all run the same way. It also does not require the patient to feel pain: at the spinal level, nociceptive input inhibits that same one-synapse connection from the muscle's stretch receptors onto its motoneurons, in a preparation with no brain attached [11]. Injured tissue does not equal muscle weakness — the imaging and strength data carry that. The nervous system has a built-in mechanism for reducing nociceptive signaling when a second noxious input arrives. And a joint's condition can suppress the drive to an intact muscle, which is what arthrogenic muscle inhibition names and what the rehabilitation literature already works on [29]. Those are published.

What sits on top of them — that a weak muscle is almost always inhibition, that the weakness found at rest and the wider drop under active use are two stages of one process, and that you can find the driver and switch it off in the clinic — is our clinical claim. It is built on established parts, but it is ours, and it should be read as such.

What future studies should try to resolve

The load-bearing science is the five items just listed. What remains open is the clinical pattern itself — and each open piece points to a study that could test it.

• Does the reversal hold under controlled conditions? The graded, reversible strength change is our observation at case-series standard, not a controlled trial. A study could test it formally: a blinded examiner, a standardized load or position, instrumented strength instead of a hand (a hand-held or isokinetic dynamometer), and the reversal tested against a sham second input, with patients as their own controls.

• Do the two stages behave as described? A study could measure instrumented strength at a muscle away from the involved region, first with the nociceptive source simply present and then with that source actively loaded, in the same patients. Nothing published compares those two conditions.

• Does nociception below a person's own pain threshold inhibit muscle in a conscious human? The cat work establishes the principle at the spinal level with no brain in the circuit [11], and the post-meniscectomy patients show that electrical drop persisting in conscious people whose pain had largely gone [36]. What has not been done is the controlled version: a nociceptive input titrated below the subject's reported pain threshold, with instrumented strength as the outcome — and a check that the stimulus is genuinely nociceptive rather than merely unfelt.

• Does light change strength at a distance, and by a nociceptive route? No study has measured a muscle test under light. A study could measure instrumented strength in whiplash patients across light conditions, then test whether anesthetizing a neck nociceptive source removes any light-driven weakness — the design Freeman and colleagues used to clear light sensitivity and raise remote pain thresholds [23], with a strength measure added. Two honest limits sit under this one: light sensitivity is not attributable to trigeminal nociception alone [25], and the nearest supporting findings — nociception producing side-specific light sensitivity [26], and a visual input raising pain at a remote site [27] — come from other conditions, not whiplash.

• Does the second input subtract more than it adds? A study could measure instrumented strength at a muscle with a standing nociceptive source and at a muscle without one, under the same added input, in the same session. That single design would test the account given above and would say whether the two directions are one mechanism or two.

• Can the manual test be replaced by an instrument? The strength test here is a clinical probe, not a laboratory-validated measurement. Trained examiners do agree closely with one another on graded manual muscle testing, and where they disagree they err toward over-estimating strength [44]. But that agreement was measured on whole grades, in intensive-care survivors and simulated patients, not on the half-grade within-session change in an outpatient caseload that this paper describes. The gaze design in the working notes below swaps the examiner's hand for a machine-recorded response, and would confirm whether the inhibition is objective and examiner-independent.

When a muscle tests weak, find what is driving it before you conclude the muscle is the problem. Almost always, the weakness is inhibition — and inhibition is something you can turn back on.

References

1. Leiras, R., Martín-Cora, F., Velo, P., Liste, T., & Canedo, A. (2016). Cat's medullary reticulospinal and subnucleus reticularis dorsalis noxious neurons form a coupled neural circuit through collaterals of descending axons. Journal of Neurophysiology, 115(1), 324–344. https://doi.org/10.1152/jn.00603.2015

2. Graven-Nielsen, T., Lund, H., Arendt-Nielsen, L., Danneskiold-Samsøe, B., & Bliddal, H. (2002). Inhibition of maximal voluntary contraction force by experimental muscle pain: a centrally mediated mechanism. Muscle & Nerve, 26(5), 708–712. https://doi.org/10.1002/mus.10225

3. Horga, L. M., Hirschmann, A. C., Henckel, J., et al. (2020). Prevalence of abnormal findings in 230 knees of asymptomatic adults using 3.0 T MRI. Skeletal Radiology, 49(7), 1099–1107. https://doi.org/10.1007/s00256-020-03394-z

4. Lee, C. S., Goldhaber, N. H., Davis, S. M., Dilley, M., Brock, A., Wosmek, J., Lee, E. H., Lee, R. K., & Stetson, W. B. (2020). Shoulder MRI in asymptomatic elite volleyball athletes shows extensive pathology. Journal of ISAKOS, 5(1), 10–14. https://www.sciencedirect.com/science/article/pii/S2059775421000973

5. Borenstein, D. G., O'Mara, J. W., Boden, S. D., et al. (2001). The value of magnetic resonance imaging of the lumbar spine to predict low-back pain in asymptomatic subjects: a seven-year follow-up study. Journal of Bone and Joint Surgery (Am), 83(9), 1306–1311. https://doi.org/10.2106/00004623-200109000-00002

6. Hinsley, H., Ganderton, C., Arden, N. K., & Carr, A. J. (2023). Relationship between shoulder abduction strength and rotator cuff tear in elderly women: a general population study. BMJ Open, 13(7), e071908. https://doi.org/10.1136/bmjopen-2023-071908

7. Godshaw, B. M., Hughes, J. D., Boden, S. A., Lin, A., & Lesniak, B. P. (2022). Comparison of functional outcomes after arthroscopic rotator cuff repair between patients with traumatic and atraumatic tears. Orthopaedic Journal of Sports Medicine, 10(10), 23259671221126551. https://doi.org/10.1177/23259671221126551

8. Moosmayer, S., Lund, G., Seljom, U. S., Haldorsen, B., Svege, I. C., Hennig, T., Pripp, A. H., & Smith, H. J. (2024). Fifteen-year results of a comparative analysis of tendon repair versus physiotherapy for small-to-medium-sized rotator cuff tears. Journal of Bone and Joint Surgery (Am), 106(19), 1785–1796. https://doi.org/10.2106/JBJS.24.00065

9. Lee, J. H., Park, J. S., & Jeong, W.-K. (2022). Importance of initial peak torque of the supraspinatus muscle during shoulder flexion. Clinics in Orthopedic Surgery, 14(2), 272–280. https://doi.org/10.4055/cios21133

10. Chowdhury, N. S., Chang, W.-J., Millard, S. K., et al. (2022). The effect of acute and sustained pain on corticomotor excitability: A systematic review and meta-analysis. The Journal of Pain, 23(10), 1680–1696. https://doi.org/10.1016/j.jpain.2022.04.012

11. Steffens, H., & Schomburg, E. D. (1993). Convergence in segmental reflex pathways from nociceptive and non-nociceptive afferents to alpha-motoneurones in the cat. Journal of Physiology, 466, 191–211.

12. Valeriani, M., Restuccia, D., Di Lazzaro, V., et al. (1999). Inhibition of the human primary motor area by painful heat stimulation of the skin. Clinical Neurophysiology, 110(8), 1475–1480. https://doi.org/10.1016/s1388-2457(99)00075-9

13. Neige, C., Lebon, F., Mercier, C., Gaveau, J., Papaxanthis, C., & Ruffino, C. (2022). Pain, no gain: Acute pain interrupts motor imagery processes and affects mental training-induced plasticity. Cerebral Cortex, 32(3), 640–651. https://doi.org/10.1093/cercor/bhab246

14. Corti, E. J., Marinovic, W., Nguyen, A. T., Gasson, N., & Loftus, A. M. (2022). Motor cortex excitability in chronic low back pain. Experimental Brain Research, 240(12), 3249–3257. https://doi.org/10.1007/s00221-022-06492-7

15. Le Bars, D., Dickenson, A. H., & Besson, J. M. (1979). Diffuse noxious inhibitory controls (DNIC). I. Effects on dorsal horn convergent neurones in the rat. Pain, 6(3), 283–304.

16. Zhang, Z.-Y., Mei, J.-L., Rao, Y.-Q., et al. (2025). The role of the subnucleus reticularis dorsalis (SRD) in pain modulation: A literature review. Current Medical Science, 45(4), 745–754. https://doi.org/10.1007/s11596-025-00082-8

17. Granovsky, Y., Sprecher, E., & Sinai, A. (2019). Motor corticospinal excitability: a novel facet of pain modulation? Pain Reports, 4(2), e725. https://doi.org/10.1097/PR9.0000000000000725

18. Botelho, L. M., Morales-Quezada, L., Rozisky, J. R., Brietzke, A. P., Torres, I. L. S., Deitos, A., Fregni, F., & Caumo, W. (2016). A framework for understanding the relationship between descending pain modulation, motor corticospinal, and neuroplasticity regulation systems in chronic myofascial pain. Frontiers in Human Neuroscience, 10, 308. https://doi.org/10.3389/fnhum.2016.00308

19. Sava, S. L., de Pasqua, V., Magis, D., & Schoenen, J. (2014). Effects of visual cortex activation on the nociceptive blink reflex in healthy subjects. PLoS One, 9(6), e100198. https://doi.org/10.1371/journal.pone.0100198

20. Okamoto, K., Tashiro, A., Chang, Z., & Bereiter, D. A. (2010). Bright light activates a trigeminal nociceptive pathway. Pain, 149(2), 235–242. https://doi.org/10.1016/j.pain.2010.02.004

21. Choudhury, A., Reyes, N., Galor, A., Mehra, D., Felix, E., & Moulton, E. A. (2022). Clinical neuroimaging of photophobia in individuals with chronic ocular surface pain. American Journal of Ophthalmology, 246, 20–30. https://doi.org/10.1016/j.ajo.2022.09.020

22. Watson, D. H., & Drummond, P. D. (2016). The role of the trigemino cervical complex in chronic whiplash associated headache: a cross sectional study. Headache, 56(6), 961–975. https://doi.org/10.1111/head.12805

23. Freeman, M. D., Nystrom, A., & Centeno, C. (2009). Chronic whiplash and central sensitization; an evaluation of the role of a myofascial trigger points in pain modulation. Journal of Brachial Plexus and Peripheral Nerve Injury, 4, 2. https://doi.org/10.1186/1749-7221-4-2

24. Bexander, C. S. M., & Hodges, P. W. (2019). Cervical rotator muscle activity with eye movement at different speeds is distorted in whiplash. PM&R, 11(9), 944–953. https://doi.org/10.1002/pmrj.12059

25. Denuelle, M., Boulloche, N., Payoux, P., Fabre, N., Trotter, Y., & Géraud, G. (2011). A PET study of photophobia during spontaneous migraine attacks. Neurology, 76(3), 213–218. https://doi.org/10.1212/WNL.0b013e3182074a57

26. Drummond, P. D., & Finch, P. M. (2021). Photophobia in complex regional pain syndrome: visual discomfort is greater on the affected than unaffected side. Pain, 162(4), 1233–1240. https://doi.org/10.1097/j.pain.0000000000002118

27. Drummond, P. D. (2002). Motion sickness and migraine: optokinetic stimulation increases scalp tenderness, pain sensitivity in the fingers and photophobia. Cephalalgia, 22(2), 117–124. https://doi.org/10.1046/j.1468-2982.2002.00332.x

28. Simis, M., Imamura, M., de Melo, P. S., Marduy, A., Pacheco-Barrios, K., Teixeira, P. E. P., Battistella, L., & Fregni, F. (2021). Increased motor cortex inhibition as a marker of compensation to chronic pain in knee osteoarthritis. Scientific Reports, 11, 24011. https://doi.org/10.1038/s41598-021-03281-0

29. Kakavas, G., Malliaropoulos, N., Pruna, R., Maffulli, N., & Bikos, V. (2026). Rehabilitation of arthrogenic muscle inhibition in patients with knee osteoarthritis and after knee arthroplasty. Current Reviews in Musculoskeletal Medicine, 19(1). https://doi.org/10.1007/s12178-026-10038-7

30. Ben-Yishay, A., Zuckerman, J. D., Gallagher, M., & Cuomo, F. (1994). Pain inhibition of shoulder strength in patients with impingement syndrome. Orthopedics, 17(8), 685–688. https://doi.org/10.3928/0147-7447-19940801-06

31. Rice, D. A., McNair, P. J., Lewis, G. N., & Dalbeth, N. (2015). The effects of joint aspiration and intra-articular corticosteroid injection on flexion reflex excitability, quadriceps strength and pain in individuals with knee synovitis: a prospective observational study. Arthritis Research & Therapy, 17(1), 191. https://doi.org/10.1186/s13075-015-0711-5

32. Sonnery-Cottet, B., Hopper, G. P., Gousopoulos, L., et al. (2024). Incidence of and risk factors for arthrogenic muscle inhibition in acute anterior cruciate ligament injuries. The American Journal of Sports Medicine, 52(1), 60–68. https://doi.org/10.1177/03635465231209987

33. Sherman, D. A., Rush, J., Glaviano, N. R., & Norte, G. E. (2024). Knee joint pathology and efferent pathway dysfunction: Mapping muscle inhibition from motor cortex to muscle force. Musculoskeletal Science & Practice, 74, 103204. https://doi.org/10.1016/j.msksp.2024.103204

34. McCabe, R. A., Nicholas, S. J., Montgomery, K. D., Finneran, J. J., & McHugh, M. P. (2005). The effect of rotator cuff tear size on shoulder strength and range of motion. Journal of Orthopaedic & Sports Physical Therapy, 35(3), 130–135. https://doi.org/10.2519/jospt.2005.35.3.130

35. van der Windt, D. A. W. M., Simons, E., Riphagen, I. I., et al. (2010). Physical examination for lumbar radiculopathy due to disc herniation in patients with low-back pain. Cochrane Database of Systematic Reviews, (2), CD007431. https://doi.org/10.1002/14651858.CD007431.pub2

36. Shakespeare, D. T., Stokes, M., Sherman, K. P., & Young, A. (1985). Reflex inhibition of the quadriceps after meniscectomy: lack of association with pain. Clinical Physiology, 5(2), 137–144. https://doi.org/10.1111/j.1475-097x.1985.tb00589.x

37. MacDonald, D., Moseley, G. L., & Hodges, P. W. (2009). Why do some patients keep hurting their back? Evidence of ongoing back muscle dysfunction during remission from recurrent back pain. Pain, 142(3), 183–188. https://doi.org/10.1016/j.pain.2008.12.002

38. Schilaty, N. D., McPherson, A. L., Nagai, T., & Bates, N. A. (2023). Arthrogenic muscle inhibition manifests in thigh musculature motor unit characteristics after anterior cruciate ligament injury. European Journal of Sport Science, 23(5), 840–850. https://doi.org/10.1080/17461391.2022.2056520

39. Sidhu, S. K., Weavil, J. C., Thurston, T. S., et al. (2018). Fatigue-related group III/IV muscle afferent feedback facilitates intracortical inhibition during locomotor exercise. The Journal of Physiology, 596(19), 4789–4801. https://doi.org/10.1113/JP276460

40. Rice, D. A., Graven-Nielsen, T., Lewis, G. N., McNair, P. J., & Dalbeth, N. (2015). The effects of experimental knee pain on lower limb corticospinal and motor cortex excitability. Arthritis Research & Therapy, 17(1), 204. https://doi.org/10.1186/s13075-015-0724-0

41. Rice, D. A., McNair, P. J., Lewis, G. N., & Dalbeth, N. (2014). Quadriceps arthrogenic muscle inhibition: the effects of experimental knee joint effusion on motor cortex excitability. Arthritis Research & Therapy, 16(6), 502. https://doi.org/10.1186/s13075-014-0502-4

42. Le Pera, D., Graven-Nielsen, T., Valeriani, M., et al. (2001). Inhibition of motor system excitability at cortical and spinal level by tonic muscle pain. Clinical Neurophysiology, 112(9), 1633–1641. https://doi.org/10.1016/s1388-2457(01)00631-9

43. Neige, C., Brun, C., Gagné, M., Bouyer, L. J., & Mercier, C. (2020). Do nociceptive stimulation intensity and temporal predictability influence pain-induced corticospinal excitability modulation? NeuroImage, 216, 116883. https://doi.org/10.1016/j.neuroimage.2020.116883

44. Fan, E., Ciesla, N. D., Truong, A. D., Bhoopathi, V., Zeger, S. L., & Needham, D. M. (2010). Inter-rater reliability of manual muscle strength testing in ICU survivors and simulated patients. Intensive Care Medicine, 36(6), 1038–1043. https://doi.org/10.1007/s00134-010-1796-6

45. Takakusaki, K., Chiba, R., Nozu, T., & Okumura, T. (2016). Brainstem control of locomotion and muscle tone with special reference to the role of the mesopontine tegmentum and medullary reticulospinal systems. Journal of Neural Transmission (Vienna), 123(7), 695–729. https://doi.org/10.1007/s00702-015-1475-4 (review, incorporating the authors' own decerebrate-cat experiments)

46. Hayamizu, M., Hagiwara, K., Hironaga, N., Ogata, K., Hoka, S., & Tobimatsu, S. (2016). A spatiotemporal signature of cortical pain relief by tactile stimulation: An MEG study. NeuroImage, 130, 175–183. https://doi.org/10.1016/j.neuroimage.2016.01.065 (read at abstract level; full text paywalled)

47. Schabrun, S. M., Jones, E., Kloster, J., & Hodges, P. W. (2013). Temporal association between changes in primary sensory cortex and corticomotor output during muscle pain. Neuroscience, 235, 159–164. https://doi.org/10.1016/j.neuroscience.2012.12.072 (read at abstract level; full text paywalled)

48. Schabrun, S. M., Burns, E., & Hodges, P. W. (2015). New insight into the time-course of motor and sensory system changes in pain. PLOS ONE, 10(11), e0142857. https://doi.org/10.1371/journal.pone.0142857

49. Mucha, A., Collins, M. W., Elbin, R. J., Furman, J. M., Troutman-Enseki, C., DeWolf, R. M., Marchetti, G., & Kontos, A. P. (2014). A brief vestibular/ocular motor screening (VOMS) assessment to evaluate concussions: preliminary findings. American Journal of Sports Medicine, 42(10), 2479–2486. https://doi.org/10.1177/0363546514543775

50. Daly, L., Giffard, P., Thomas, L., & Treleaven, J. (2017). Validity of clinical measures of smooth pursuit eye movement control in patients with idiopathic neck pain. Musculoskeletal Science & Practice, 33, 18–23. https://doi.org/10.1016/j.msksp.2017.10.007 (read at abstract level)

51. Yu, L. J., Stokell, R., & Treleaven, J. (2011). The effect of neck torsion on postural stability in subjects with persistent whiplash. Manual Therapy, 16(4), 339–343. https://doi.org/10.1016/j.math.2010.12.006 (read at abstract level)

52. Majcen Rosker, Z., Kristjansson, E., & Vodicar, M. (2023). How well can we detect cervical driven sensorimotor dysfunction in concussion patients? An observational study comparing patients with idiopathic neck pain, whiplash associated disorders and concussion. Gait & Posture, 101, 21–27. https://doi.org/10.1016/j.gaitpost.2023.01.011 (read at abstract level)

53. Bao, S., Wang, Y., Escalante, Y. R., Li, Y., & Lei, Y. (2024). Modulation of motor cortical inhibition and facilitation by touch sensation from the glabrous skin of the human hand. eNeuro, 11(3), ENEURO.0410-23.2024. https://doi.org/10.1523/ENEURO.0410-23.2024

 

Draft Ideas — Proposed Studies

Working notes, not part of the paper's argument. References here are given by author and year rather than reference number.

Objectively confirming gaze-induced muscle inhibition. This is the instrument-based design referred to in the last bullet of "What future studies should try to resolve" — can the manual test be replaced by an instrument? — set out in full. With the arm held in 90° of shoulder abduction — the same position used in the clinic — deliver a single transcranial magnetic stimulation pulse to the motor cortex and record the shoulder muscle's evoked response while the patient cycles gaze: eyes straight → eyes right → eyes straight → eyes left. The pulse can be delivered during the active hold, so the muscle is tested while it is working, not at rest. Each gaze position is compared against that same person's straight-ahead reading — the patient is their own control — and the effort of the contraction is held steady so any change is attributable to the gaze and not to reduced effort.

A drop in the recorded response under right or left gaze would be an objective, instrument-measured confirmation of the inhibition seen on manual testing — independent of the examiner, the patient's effort, or expectation. This tests only whether the inhibition is real and measurable; it does not attempt to identify where in the nervous system the change occurs.

The measurement format is adapted from published work showing that eye position and eye movements change corticospinal drive to resting limb muscles (de Wit et al., 2020; Falciati & Maioli, 2017; Falciati et al., 2013); the proposed change is to measure the actively contracting shoulder muscle, in the clinical test position, as gaze is shifted.