Two patients come through the practice without the word dystonia ever being said. The first is referred for stubborn neck muscle spasm that has yielded to nothing for two years. The second is a musician or a writer whose fingers give way, who has been diagnosed with an overuse tendinopathy and prescribed rest, without result. In both cases the examination contains one sign that settles matters, and it is free. This article sets out that sign, the classification in force since its 2025 revision, and the real place of physiotherapy around botulinum toxin injections, with the level of evidence each modality deserves and not one notch more.

• Updated August 2026
• Level clinical summary
• References 55, PMID or DOI verified
• Area neurology

What is a dystonia, and what does the classification in force say?

Dystonia has a reputation for obscurity that owes most to its vocabulary. Yet the definition fits in one sentence, and the classification in two axes. Knowing them is already knowing what to look for on examination.

The definition, word for word

The international panel that carries authority defines dystonia as a movement disorder characterised by abnormal movements, abnormal postures, or both, sustained or intermittent. Three details of this definition are worth keeping, because each of them is an examination criterion2 :

• Dystonic movements are typically stereotyped and repetitive. The same muscle, the same direction, the same movement. An antalgic muscle spasm varies; a dystonia repeats itself.
• They may be tremulous or jerky. An associated head tremor does not rule out dystonia, it is part of it.
• They are often initiated or worsened by voluntary action, and frequently associated with overflow movements. This is the most useful point in practice: asking the patient to do something brings the sign out, whereas passive examination can be near normal.

Stiffness that gets worse when the patient acts, and not when you move them, is almost never mechanical.

The classification was revised in 2025

The founding consensus dates from 2013 and proposed two axes1. Twelve years later the same panel published a revision, drawing the lessons of how that framework was actually used2. What has to be known about this revision fits in three lines. The definition is kept, with minor clarifications of wording. The two-axis structure is kept too, but revised. And the stated aim of the revision is to make uniform implementation easier, in other words to make the framework usable by clinicians who are not movement disorder specialists.

The vocabulary of distribution, which is the vocabulary of this article's title

The terms met in clinic letters describe the extent of the involvement, and nothing else:

• Focal dystonia : a single body region. Cervical dystonia, blepharospasm, writer's cramp and musician's dystonia are focal dystonias.
• Segmental dystonia : two or more contiguous regions, for example cervical and scapular.
• Multifocal dystonia : two or more non-contiguous regions.
• Generalised dystonia : the trunk and at least two other regions.
• Hemidystonia : one side of the body. It must prompt a search for a contralateral basal ganglia lesion.

Cervical dystonia, the commonest focal form in adults, has moreover received a diagnostic consensus of its own. It describes three presentations (idiopathic focal or segmental, genetic, acquired), recognises two levels of certainty for the idiopathic form, definite and probable, explicitly includes non-motor manifestations, and addresses the question of the relationship between cervical dystonia and isolated head tremor. Its authors state that a probable diagnosis is enough for clinical practice, the higher level being reserved for research3.

This distinction is not academic: focal dystonias in adults are treated first line with botulinum toxin injections, whereas generalised dystonias, often of earlier onset and often genetic, call for general pharmacological treatment and then, if that fails, deep brain stimulation39.

Why is a cervical dystonia taken for a muscular torticollis?

This is not a beginner's mistake. It is a system-level mistake, documented by whole cohorts, and the physiotherapist holds a particular place in it: they are very often the first professional consulted.

What the patient has been told before arriving

A study of more than three million members of a Californian integrated health system compared the diagnoses recorded in patients in the years preceding their diagnosis of cervical dystonia with those of matched controls. The results map out exactly the path of a patient who passes through our practices5 :

Two readings of this graph. The first is neurological: the odds ratio of 68 for essential tremor mainly signals that dystonic head tremor is regularly labelled essential tremor, and that confusion has been the subject of a published exchange between teams5. The second concerns us directly: neck sprain or muscle spasm, odds ratio 2.77. That diagnosis is our most ordinary referral letter.

The physiotherapist is on the front line, statistically

An Australian cohort of 49 patients followed in a specialist centre reconstructed their pathway by questionnaire. The mean delay between first symptoms and diagnosis was 6.8 years, with a range of 0 to 53 years. Patients had seen three doctors on average, up to nine. Only 40 % had consulted within the first six months, and only 10 % had been given the right diagnosis straight away. The GP's first referral was to a specialist other than a neurologist in 31 % of cases4.

And the figure that should appear in every entry-level curriculum: more than half of patients had first sought physical therapies, even before consulting their general practitioner4. A substantial share of those 6.8 years is therefore played out in our practices. The authors go on to conclude that the delay owes partly to poor knowledge of the disease among health professionals, and that better diagnostic competence would have had a substantial effect on this population's access to treatment.

The patient referred for stubborn neck muscle spasm who has responded to nothing for eighteen months is not a poor responder: they may simply be mislabelled.

A more recent Turkish national survey, of 789 dystonia patients treated with toxin, finds a diagnostic delay of the order of a year and 15 % of initial diagnoses wrong17. The gap between one year and 6.8 years deserves to be stated plainly: these two figures do not measure the same thing, in different health systems and with different data collection methods. What is constant from one study to the next is the existence of the delay, not its length.

What is the sensory trick, and how do you look for it in a session?

This is the most useful sign in this whole article. It costs nothing, it needs no equipment, it takes ten seconds, and an expert recommendation goes as far as writing that its presence confirms the diagnosis.

What it is

The manoeuvre known in French as the geste antagoniste, and called a sensory trick or alleviating manoeuvre in the English-language literature, is a voluntary manoeuvre that transiently relieves the dystonic posture or movement. The best-known form in cervical dystonia is a light touch to the cheek, the chin or the occiput on the side towards which the head deviates. The patient places two fingers, without force, and the head recentres. This is not a correction by force: a patient who has to push with the palm to straighten their head is performing a different, so-called forcible trick, and the literature distinguishes it from the classic trick49.

A landmark review proposed ordering this family of manoeuvres, because it is broader than one might think: sensory tricks in the strict sense, motor tricks, imaginary tricks (thinking about the trick is enough in some patients), forcible tricks, and reverse sensory tricks, which worsen instead of relieving. Its authors stress that sensory stimulation is not necessarily the critical element, and propose that these tricks act by reducing an abnormal facilitation in the dystonic brain8. Their practical conclusion deserves quoting as it stands: it seems useful for patients to look for a possible sensory trick.

How many patients have one, and how much it improves them

A systematic review published in 2026 gathered 53 studies on the subject, 31 of them in qualitative synthesis. Its results are the ones to have in mind, with their limitations6 :

• Reported prevalence ranges from 13 % to 90 % depending on the series and the dystonia subtypes. That spread is not noise: it reflects definitions and methods of looking for the sign that vary from one team to another.
• In the largest registry cohort (1,477 patients), 68.7 % had an effective trick. That is the most robust value we have.
• Acute motor improvement reaches 30 to 50 % of the head deviation while the trick is being performed, but the effect is transient.
• Responding to the trick is associated with a shorter disease duration and a better response to botulinum toxin. An effective trick is therefore also a prognostic pointer.
• No domain met the criteria for a formal meta-analysis. The authors are explicit: the magnitude, the durability and the predictors of the effect remain inconsistently measured. So no figure for diagnostic sensitivity or specificity can be given.

One team reports, moreover, in the introduction to a functional imaging study, that more than 80 % of patients with cervical dystonia have their symptoms improved by a sensory trick49. That figure is quoted in passing, not produced by an epidemiological study: the value to keep is the one from the systematic review.

Its diagnostic value, as written in the recommendations

An Italian expert panel has formalised diagnostic recommendations for cervical, oromandibular and limb dystonia, explicitly aimed at clinicians unfamiliar with dystonia. Their architecture is simple. The core of the diagnosis is the patterned and repetitive character of the movements or postures. And then the sentence that interests us: if a sensory trick is present, it confirms the diagnosis of dystonia. In patients who do not show one, the clinical features of the conditions that mimic dystonia must instead be actively excluded7.

The authors themselves state that the reliability, the sensitivity and the specificity of these recommendations have yet to be demonstrated. This is an expert recommendation, not a diagnostic accuracy study, and it must be cited as such. That does not lessen its practical value: in the hierarchy of signs that can be looked for free of charge in the practice, this one comes first.

How do you separate cervical dystonia, muscular torticollis and mechanical stiffness?

The table that follows is the practical heart of this article. It is built for the real situation: a patient sitting in front of you, referred for something cervical, with no decisive imaging and no neurological opinion.

Two mimics worth knowing, because they reverse the error

Neglected congenital muscular torticollis. It is rare in adults, but it exists, and it resembles cervical dystonia to the point of being diagnosed as such. A published case describes exactly that situation: an adult with a sternocleidomastoid fibromatosis missed in childhood, labelled cervical dystonia, in whom physiotherapy and chemodenervation were largely ineffective because of fibrous replacement of the muscle. Treatment was surgical, with partial relief35. The lesson is the mirror image of the one running through this article: the error goes both ways, and a dystonia that responds neither to physiotherapy nor to toxin must prompt the diagnosis to be reconsidered.

Functional dystonia. It belongs to the field of functional neurological disorders and is distinguished from idiopathic dystonia by a cluster of clinical pointers: often sudden onset, a posture fixed from the start rather than progressive, distractibility, variability, incongruence with the known dystonic patterns43. A recent case in a guitarist illustrates the difficulty of this differential diagnosis in the context, precisely, of a presumed task-specific dystonia44. This diagnosis is not made in the physiotherapy practice, but knowing it prevents stubbornness and prompts referral.

Why is a task-specific dystonia treated as a tendinopathy?

Because the patient is a manual worker of high precision, because they have repeated their movement a great deal, and because the first hypothesis in front of an upper limb that gives way in a musician is overload. This confusion is not recent: it is exactly one hundred and fifty years old, and it is at the origin of the literature on lateral epicondylalgia.

A confusion one hundred and fifty years old, and documented

In 1873, Dr Ferdinand Runge published a paper entitled, in translation, "On the etiology and treatment of writer's cramp". That paper is universally cited as the first description of tennis elbow. Three German-speaking authors retranslated it in full in 2022 to make it accessible, and their finding is clear: the main object of Runge's work was not the epicondyle, it was writer's cramp, illustrated by four clinical observations. His conclusion, in 1873, was that careful assessment of the activities that hampered writing before the cramp appeared is the key to management38.

The first paper on tennis elbow is a paper on writer's cramp. The confusion between task-specific dystonia and overuse tendinopathy is written into the very foundations of our literature.

What separates the two, clinically

Three features make all the difference, and they are accessible from the history alone, with no additional investigation.

First feature: task specificity. A task-specific dystonia appears only during the highly trained movement. The pianist whose ring finger curls on the keyboard has a normal hand for buttoning a shirt. The accountant whose index finger stiffens on the pen holds chopsticks perfectly well. A tendinopathy, by contrast, shows up on anything that loads the tendon, whatever the task. The largest phenotyping work published to date reviewed 173 patients with focal task-specific dystonia of the musician's arm, 50 writer's cramps and 16 other arm dystonias, examined by a single clinician over twenty-five years: it finds preferential involvement according to instrument class, the precision hand in pianists and guitarists on one side, the power hand on the other37. In other words, the disorder follows the technique, not the anatomy of the load.

Second feature: the absence of pain. Musician's focal dystonia is described in the specialist literature as a painless movement disorder, specific to the task, which impairs fine motor control during instrumental playing9. A musician who describes a loss of control without pain is not describing a tendinopathy. One caution, however: this absence of pain is the rule, not a law, and a dystonia can coexist with a genuine overuse condition in the same patient. So the question to ask is not "does it hurt?", but "does it hurt, or does it no longer obey you?".

Third feature: rest does not fix it. This is the criterion that should trigger the rethink. An overuse tendinopathy improves with a reduction in load, even partially, even transiently. A task-specific dystonia does not budge. A musician who has stopped for three months and finds the symptom intact at the first session back has not mismanaged their return: they do not have a tendinopathy.

Who is affected

About 1 % of professional musicians have a focal dystonia, and the risk is higher for the instruments that demand the most fine motor control. When the workload differs between the two hands, the dystonia preferentially affects the more heavily used one. Psychological studies find more perfectionist tendencies in these musicians than in healthy musicians, which has led to a model in which a largely genetic predisposition meets intrinsic and extrinsic triggering factors10. A survey of fourteen practitioners who had seen, cumulatively, more than two thousand dystonic musicians draws a converging profile: an unfavourable social environment and poor-quality instrumental teaching, a perfectionist, anxious and compliant personality, and obsessive, inadequate practice9. The authors themselves state that these factors rest largely on anecdotal data and call for controlled work: this is a profile, not a quantified risk factor.

Beyond musicians, task-specific dystonia affects writers (writer's cramp), some precision trades, and athletes. A systematic review devoted to sport retained 31 studies out of 7,000 identified, and its verdict is harsh: no standardised outcome measures, a low level of evidence for every option, including botulinum toxin and pharmacological treatments45.

What is really impaired in writer's cramp

A French study published in 2026 brings a useful detail for building an exercise. Twenty-three patients with writer's cramp were compared with twenty matched controls on three dexterity tasks. The patients wrote more slowly (97 letters per minute against 171, p < 0.001) and less legibly. But of the three tasks, only the one measuring finger independence showed a difference : finger selectivity was lower in the patients (median 0.84 against 0.89, p = 0.01). Maximum tapping speed and visuomotor force tracking did not differ (p > 0.44 for both). In the patients, the legibility of automatic writing correlated with digital selectivity (r = 0.50, p = 0.02), and rehabilitation improved both dexterity and legibility46.

The practical consequence is direct: it is neither strength nor speed that has to be worked, it is the selective activation of one finger independently of its neighbours. A strengthening or endurance programme, which would be relevant in a tendinopathy, misses the measured deficit entirely.

What do we know about its frequency and its spontaneous course?

Two questions that patients ask systematically, and that the data answer better than one might think: is it going to get worse, and can it disappear.

Frequency, with its acknowledged uncertainties

The reference meta-analysis retained sixteen original studies, fifteen of which reported a prevalence. Across the twelve studies based on clinic caseloads, the overall prevalence of primary dystonia is estimated at 16.43 per 100,000 (95 % CI 12.09 to 22.32). The three general population studies find higher values. The incidence of cervical dystonia was reported by only one of the sixteen studies, with a corrected estimate of 1.07 per 100,000 person-years (95 % CI 0.86 to 1.32). The authors explicitly conclude that their own estimate probably underestimates the real prevalence18.

The Californian study already cited provides the most solid incidence estimate, from 200 cases identified among 15.4 million person-years: 1.18 per 100,000 person-years (95 % CI 0.35 to 2.0), with a clear imbalance between women (1.81) and men (0.52), and an incidence that rises with age5.

These figures have a practical consequence that must be stated bluntly. Dystonia is rare. A physiotherapist in private practice will see a handful of cases in a career, not a handful a year. That is not a reason to look for it everywhere: it is a reason to think of it precisely in the situation where it hides, that is, chronic neck pain stubborn to everything, or a painless loss of control specific to one task.

The spontaneous course: the answer is rather reassuring

A German team followed 100 patients with idiopathic cervical dystonia, recruited from its botulinum toxin clinics, over a mean of 17.5 years, with semi-structured interviews. Its results redraw the prognosis16 :

Three conclusions to pass on to the patient, attributing them honestly to this study and to its limitations:

• Idiopathic cervical dystonia is a non-progressive disorder. The authors write that this result contradicts the fear, widespread among patients, of a constant and continuous decline in their condition. That is therapeutic information in itself.
• A minority form, of rapid onset, remits massively. Nineteen per cent of patients developed their symptoms in under six months, and 92 % of them had a remission, against 5 % in the slow form.
• Major psychological stress precedes almost exclusively the rapid form (63 % against 1 %). The authors propose that it acts as an epigenetic trigger on predetermined ground. This is a hypothesis, framed as such by the authors, not a demonstrated mechanism.

How does physiotherapy fit together with botulinum toxin?

This is the practical question every physiotherapist who receives a dystonic patient asks: at what point in the injection cycle should the work be done, and towards what goal. The literature answers in part, and where it stops has to be said clearly.

First, the ranking of the treatments

Botulinum toxin is the first-line treatment of focal dystonias in adults. The practice guideline of the American Academy of Neurology classed, in 2016, abobotulinumtoxinA and rimabotulinumtoxinB as established as effective and to be offered (level A) in cervical dystonia, onabotulinumtoxinA and incobotulinumtoxinA as probably effective and to be considered (level B)19. One point of honesty is needed here: the record of this guideline carries the label "RETIRED", meaning that it has been withdrawn from the Academy's active corpus. A withdrawn guideline is not a refuted guideline: the trials that underpinned it still exist. But citing it as the guideline in force would be inaccurate.

More recent documents confirm the ranking of toxin without calling it into question. A systematic review driven by practice questions concludes that the beneficial effect of toxin on various aspects of cervical dystonia is well established, while noting that robust evidence is still lacking on several concrete points: dose equivalence between formulations, optimal treatment intervals, injection approaches, and the value of electromyographic or ultrasound guidance20. A joint document from three Italian learned societies also covers the pharmacological, surgical and rehabilitation options in adult dystonia21.

The real rhythm of the injections, and what is not known about it

A scoping review conducted by three reference authors in the field sets the frame: treating cervical dystonia with botulinum toxin usually requires injections every 3 to 4 months, for as long as the symptoms persist, which may mean a lifetime. And above all, there is at present no consensus on how to assess the duration of the effect. The estimation methods vary from one publication to another, most rest on artificial constructs designed for clinical trials and are not transferable to practice. The authors call for flexible reinjection intervals, adjusted to individual needs50.

The question "when to schedule the sessions in relation to the injections" has no established answer, because the upstream question, "how long does the injection work for", has none either.

What physiotherapy adds, in figures

The most recent meta-analysis included fourteen articles in its review, only two of which were meta-analysable. Its main result is clear: physiotherapy has a significant effect on pain reduction when it is used as an add-on therapy to the toxin injection, with a mean difference of 5.00 points (95 % CI 3.74 to 6.26). The authors also report a possible positive effect on severity, disability and quality of life, and conclude that the variety of intervention types and durations does not allow any particular type of physiotherapy to be recommended11.

Two earlier reviews go the same way with the same caution. The original systematic review by the same team already concluded that a benefit was possible in 201423, and a review restricted to randomised trials alone, six in total, concludes that add-on physiotherapy and active home exercise programmes appear useful, toxin remaining the treatment of choice, and calls for research on the dose-response relationship22.

The trials, one by one, with what they actually showed

This table reads badly if all the trials are put on the same line, because they do not ask the same question. Two different comparisons hide in it, and confusing them produces exactly the misreading heard on courses: "the trials show that physiotherapy is of no use in dystonia". That is not what they show.

So, which schedule?

No trial has compared two schedules against each other. What can be said with the available data comes down to this: the two trials that found a benefit from physiotherapy added to toxin both placed their programme in the weeks that followed an injection, and the two trials that found no difference were comparing two forms of physiotherapy with each other, not physiotherapy with its absence. These two findings do not contradict each other, they answer different questions.

The reasonable course of action, which must be presented for what it is, namely a rationale of usage and not a validated recommendation:

• Schedule a block of sessions in the weeks that follow the injection, as in the two positive trials. With the overactive muscle partly denervated, the antagonist and postural control can be worked in a window where the dystonic agonist no longer overwhelms them.
• Switch to a self-exercise programme when the benefit of the injection wanes, rather than piling up passive sessions at the end of the cycle.
• Do not assess the patient just before the reinjection and draw a conclusion about physiotherapy from it : that is the point in the cycle where the effect of the medical treatment is weakest. Counsell et al. made that methodological choice deliberately, so as not to measure the effect of the toxin; in the clinic it mainly means that an assessment done at that moment measures a trough.
• Do not promise to reduce the doses or to space out the injections. The trial that measured mean toxin dose as a secondary outcome found no difference12.

Why it might work: a mechanistic pointer

The trial by Hu et al. includes a physiological strand worth flagging. The authors probed sensorimotor plasticity with a paired associative stimulation paradigm in transcranial magnetic stimulation. In the physiotherapy plus toxin arm, that plasticity, abnormally high in dystonia, fell towards the values of healthy controls, and its change correlated significantly with the clinical improvement14. This is an argument of plausibility, in sixteen patients: it makes the hypothesis credible, it does not demonstrate it.

A second pointer comes from the assessment. Cervical sensorimotor control is measurably impaired in cervical dystonia: on a blindfolded head repositioning task, 24 patients made an error 1.5 degrees greater than that of 70 controls (p < 0.006), and, an interesting detail, they overshot the neutral position systematically where the controls stayed short of it. The error correlated with no disease characteristic. The authors conclude that this control is trainable and could form a target for treatment adjunctive to the injections26.

Which modalities for which level of evidence?

The table that follows rates each modality against the GRADE domains. One clarification straight away, because it changes how to read it.

What is sensorimotor rehabilitation really worth in task-specific dystonias?

This is the only field where the physiotherapist can be the main therapist and not the adjunct. All the more reason to be exact about the level of evidence, which is low, and about what that licenses saying to the patient.

The idea, and where it comes from

The reasoning is pathophysiological. In a task-specific dystonia the cortical representations of the fingers are disorganised and overlap, and some patients have impaired spatial sensory discrimination. If the disorder rests on maladaptive plasticity, then training can, in principle, redirect it. Three families of programme follow from this.

Pure sensory training. Ten patients with focal hand dystonia learned to read braille, 30 to 60 minutes a day for eight weeks. Patients and controls both improved their spatial acuity; the patients showed a significant difference on the Fahn dystonia scale, 60 % of them reduced the time needed to write a standard paragraph, and the improvement in sensory perception correlated positively with the clinical improvement51. The one-year follow-up covered only three patients, who were continuing to improve52.

Splint-based retraining (sensory motor retuning). Eleven professional musicians wore splints immobilising one or more fingers other than the dystonic one, which performed repetitive exercises in coordination with the others, an hour and a half to two and a half hours a day for eight consecutive days under supervision, then one hour a day for a year. Every pianist and every guitarist improved markedly. The three wind instrument players, whom the authors themselves describe as accidental placebo controls, did not improve53. That detail is crucial: the founding method has no control group designed as such, it has a control group found after the fact.

Structured motor retraining. An anatomy-based programme, developed for musicians, offers a progression of fine motor control exercises performed away from the instrument, before being transferred to playing. The pilot covered four patients followed for about twelve months: all improved, two returned to their previous level of performance31.

What the best available data say, and what they do not

The most recent work is a seven-year follow-up of a feasibility study of sensorimotor rehabilitation. All twelve original participants, seven musician's dystonias and five writing dystonias, were traced. All showed improvement in the clinical outcomes at one year and at seven years, with effect sizes of 0.25 to 0.93, and the interviews indicated that they were still using the strategies learned seven years after the programme28.

This result has to be read properly. Twelve participants, no control group, a seven-year follow-up on a feasibility study, and conclusions that the authors themselves frame with restraint: the benefits may be maintained, and the underlying mechanisms need investigating. An effect size of 0.93 in twelve patients with no control is not the same thing as an effect size of 0.93 in a randomised trial.

For the whole rehabilitation of task-specific dystonias we have case series and feasibility studies. Not a single randomised trial of useful size. That does not mean that nothing should be done; it means that it has to be said to the patient.

Reviews of the field converge. A 2024 scoping review retained 17 articles out of 190, identifying ten different rehabilitation approaches in twenty years, and concludes that no definitive intervention protocol exists, a multimodal approach being commonly recommended29. An earlier systematic review had analysed 36 publications with a similar finding30. Ten approaches, none compared with the others: that is the definition of a field where we do not yet know.

A biological pointer that makes the avenue credible

A French team explored, with magnetoencephalography, the somatosensory evoked fields of fingers I, II, III and V in healthy controls, in untreated patients with writer's cramp, and in patients who had recovered legible handwriting after rehabilitation. In the cortex controlling the dystonic limb, the size of the hand representation in the trained patients was similar to that of healthy controls and significantly different from that of the untrained patients, with a reorganisation of the digit maps that the authors describe as supra-normal. In the cortex controlling the non-dystonic limb, trained and untrained resembled each other, and the representation remained enlarged and disorganised32.

This is an elegant result and a strong hypothesis: prolonged, appropriate rehabilitation could induce long-term plasticity phenomena, lateralised to the cortex of the dystonic hand. It is also a cross-sectional comparison between groups, whose direction of causality is not established: we do not know whether rehabilitation remodelled the cortex, or whether the patients whose cortex was remodelling are the ones who recovered.

What is actually worked on, and why

The deficit measured in writer's cramp is a failure of digital selectivity, with no impairment of tapping speed or of force control46. That gives a precise target, and rules out two false trails.

• What has to be worked: the activation of one finger independently of its neighbours, over small ranges and small forces, with no pursuit of speed; sensory discrimination of the affected segment; the movement away from the instrument or away from the triggering task, before any transfer; grip modification, which showed a lasting reduction in writing pressure and grip force in 26 patients33.
• What is beside the point: analytical strengthening of the flexors and extensors, chain stretching, endurance work, shockwaves, any modality designed for a tendon. None has a rationale here, and they take up session time.
• What takes time: the published protocols are counted in months, not in sessions. Eight weeks of daily sensory training, eight intensive days then a year of daily maintenance, twelve months of retraining. A ten-session programme does not test the hypothesis, it grazes it.

What changes when the dystonia is generalised?

Generalised dystonias are rarer, often of earlier onset, often genetic, and their treatment pathway has nothing in common with that of a cervical dystonia. The physiotherapist has a part in it, but a different one.

Why toxin is not enough

The logic is arithmetic: botulinum toxin acts on the muscles injected, and the total dose is capped by tolerance. In a dystonia involving the trunk and several regions, all the muscles concerned cannot be covered. Toxin keeps a place for targeting the most disabling sites, but the mainstay treatment is general pharmacological, then surgical.

What deep brain stimulation has demonstrated

Two landmark trials established the efficacy of internal pallidal stimulation in primary generalised or segmental dystonia.

The first, French and multicentre, covered patients with primary generalised dystonia40. The second, German, randomised 40 patients with primary segmental or generalised dystonia between active stimulation and sham stimulation for three months, with blinded video assessment. At three months the change in the motor score of the Burke-Fahn-Marsden scale was minus 15.8 points (SD 14.1) under stimulation against minus 1.4 points (SD 3.8) under sham stimulation (p < 0.001). The benefit was maintained in the open phase, and the patients initially on sham stimulation benefited in their turn. Twenty-two adverse events occurred in 19 patients, including four infections of the stimulator site and one lead displacement; the commonest adverse effect was dysarthria39.

A 2026 review takes stock of the twenty years that followed. Its message is that individual response varies a great deal, that this variability reflects the heterogeneity of clinical presentations, of genetic background and of the brain networks involved, and that dystonia should be understood as a network disorder affecting cortico-striato-cerebellar circuits. The effects of stimulation would therefore result from modulating distributed circuits rather than a single anatomical target41.

The physiotherapist's role, without overstating the evidence

It has to be said directly: there is no randomised trial of physiotherapy in generalised dystonia comparable to those in cervical dystonia. What can be done falls under clinical reasoning and the general principles of neurological rehabilitation, and must be presented as such.

• Preventing contractures and fixed deformities. A dystonic posture held for years creates structural limitations which will respond to no neurological treatment. This is the best justified objective.
• Preserving functional capacity and independence, with the assistive devices and adaptations required.
• Supporting the perioperative period of deep brain stimulation. The clinical effect of stimulation builds gradually over months, and the programming period is also a period of motor relearning.
• Watching for respiratory and nutritional complications in severe forms with axial involvement.

One more experimental avenue deserves mention for what it is: a French team combined transcranial direct current stimulation of the cerebellum with a goal-oriented motor training programme, in five patients with cervical dystonia poorly controlled by toxin. The maximum improvement in the total TWSTRS reached 37 % on average after stimulation alone, non-significant (p = 0.147), and 53 % after stimulation combined with motor training (p = 0.014), with a longer duration of improvement (3.4 months against 1.4; p = 0.011). The authors themselves conclude that these results remain to be confirmed by a randomised placebo-controlled trial in a larger sample47. Five patients, open-label: this is a hypothesis to test, not a modality to offer.

What do real clinical cases teach us?

Four published observations, with their identifiers. Each illustrates a point the statistics do not convey: the variety of the sensory trick, the reversibility of the diagnostic error, and what rehabilitation carried over months can do.

How do you apply this in the clinic?

This chapter is the user manual. It contains nothing that is not sourced above, and it clearly flags what belongs to usage rather than to evidence.

Three minutes that change the patient's pathway

Faced with any stubborn chronic neck pain, especially after three months of treatment with no lasting progress:

1. Look at the head, patient seated, unsupported, at rest. Note the direction of the deviation. Photograph or film it with the patient's consent, at every assessment: it is stereotypy over time that makes the diagnosis, and it is only visible by comparison.
2. Ask for an action. Have the patient walk down the corridor, write three lines, carry a bag. A posture that worsens with action and not with passive movement is a strong argument.
3. Look for the sensory trick, with an open question rather than a closed instruction. Have several points tried, and the merely imagined trick.
4. Go back over the history. How many professionals seen, for how long, which treatments and with what result. Neck pain that has passed through three practitioners without ever letting go is no longer ordinary neck pain.

And in front of a professional of movement whose upper limb gives way: the three questions from the corresponding chapter, in that order. Does it hurt or does it no longer obey you. Does it happen only during that task. Has rest changed anything.

Writing the referral so that it gets somewhere

This is a neglected point and it weighs heavily: the GP's first referral was to a specialist other than a neurologist in 31 % of cases4. A precise letter shortens the circuit. Include in it:

• The constant direction of the abnormal posture, and how long it has been evolving.
• The result of the sensory trick test, described precisely (which point, which hand, what effect, for how long).
• The worsening with action, with the action in question.
• The record of the treatments already given and their lasting ineffectiveness.
• The explicit mention of the hypothesis: "a cervical dystonia seems to me to need ruling out". The word has to be written; it is what points to the neurologist rather than to the rheumatologist.

What is done, once the diagnosis is made

Three errors not to make

• Promising correction of the posture. Physiotherapy has not shown that it lastingly corrects the dystonic deviation. What it has shown is an effect on pain. Selling anything else risks a breakdown of trust by the third session.
• Looking for THE right technique. Two independent randomised trials totalling more than two hundred patients compared a specialised technique with ordinary physiotherapy and found no difference on their primary outcome1213. The time spent choosing the method would be better invested in adherence to the home programme.
• Treating a task-specific dystonia as a tendinopathy. This is the error that costs most in lost months, and it is one hundred and fifty years old38.

Frequently asked questions

In the same neurological area

Bibliography (55 verified references)

Each reference has been checked against two independent databases: the MEDLINE record served by Europe PMC for the PMID, the author list, the journal and the pagination, and CrossRef for the DOI. The lines below are produced automatically from those records, never typed by hand.

1. Albanese A, Bhatia K, Bressman SB, et al.. Phenomenology and classification of dystonia: a consensus update. Mov Disord. 2013;28(7):863-873. PMID 23649720. doi 10.1002/mds.25475.
2. Albanese A, Bhatia KP, Fung VSC, et al.. Definition and Classification of Dystonia. Mov Disord. 2025;40(7):1248-1259. PMID 40326714. doi 10.1002/mds.30220.
3. Albanese A, Bhatia KP, Cardoso F, et al.. Isolated Cervical Dystonia: Diagnosis and Classification. Mov Disord. 2023;38(8):1367-1378. PMID 36989390. doi 10.1002/mds.29387.
4. Bertram KL, Williams DR. Delays to the diagnosis of cervical dystonia. J Clin Neurosci. 2016;25:62-64. PMID 26601813. doi 10.1016/j.jocn.2015.05.054.
5. LaHue SC, Albers K, Goldman S, et al.. Cervical dystonia incidence and diagnostic delay in a multiethnic population. Mov Disord. 2020;35(3):450-456. PMID 31774238. doi 10.1002/mds.27927.
6. Mehta A, Shivashanmugam T, Bruno MK, et al.. Sensory Tricks in Dystonia: A Systematic Review and Nested Quantitative Synthesis. Brain Behav. 2026;16(7):e71575. PMID 42446153. doi 10.1002/brb3.71575.
7. Defazio G, Albanese A, Pellicciari R, et al.. Expert recommendations for diagnosing cervical, oromandibular, and limb dystonia. Neurol Sci. 2019;40(1):89-95. PMID 30269178. doi 10.1007/s10072-018-3586-9.
8. Ramos VF, Karp BI, Hallett M. Tricks in dystonia: ordering the complexity. J Neurol Neurosurg Psychiatry. 2014;85(9):987-993. PMID 24487380. doi 10.1136/jnnp-2013-306971.
9. Détári A, Egermann H. Musicians Focal Dystonia: The Practitioner's Perspective on Psychological, Psychosocial, and Behavioural Risk Factors and Non-motor Symptoms. Med Probl Perform Art. 2022;37(3):200-206. PMID 36053497. doi 10.21091/mppa.2022.3023.
10. Altenmüller E, Jabusch HC. Focal dystonia in musicians: phenomenology, pathophysiology and triggering factors. Eur J Neurol. 2010;17 Suppl 1:31-36. PMID 20590806. doi 10.1111/j.1468-1331.2010.03048.x.
11. Kassaye SG, De Hertogh W, Crosiers D, et al.. The effectiveness of physiotherapy for patients with isolated cervical dystonia: an updated systematic review and meta-analysis. BMC Neurol. 2024;24(1):53. PMID 38302911. doi 10.1186/s12883-023-03473-3.
12. Counsell C, Sinclair H, Fowlie J, et al.. A randomized trial of specialized versus standard neck physiotherapy in cervical dystonia. Parkinsonism Relat Disord. 2016;23:72-79. PMID 26723272. doi 10.1016/j.parkreldis.2015.12.010.
13. van den Dool J, Visser B, Koelman JH, et al.. Long-Term Specialized Physical Therapy in Cervical Dystonia: Outcomes of a Randomized Controlled Trial. Arch Phys Med Rehabil. 2019;100(8):1417-1425. PMID 30796919. doi 10.1016/j.apmr.2019.01.013.
14. Hu W, Rundle-Gonzalez V, Kulkarni SJ, et al.. A randomized study of botulinum toxin versus botulinum toxin plus physical therapy for treatment of cervical dystonia. Parkinsonism Relat Disord. 2019;63:195-198. PMID 30837195. doi 10.1016/j.parkreldis.2019.02.035.
15. Verriello L, Bellizzi F, Lettieri C, et al.. Optimizing treatment for cervical dystonia: botulinum toxin alone or combined with physical therapy?. Neurol Sci. 2026;47(1):69. PMID 41483253. doi 10.1007/s10072-025-08636-3.
16. Dressler D, Kopp B, Pan L, et al.. The natural course of idiopathic cervical dystonia. J Neural Transm (Vienna). 2024;131(3):245-252. PMID 38244034. doi 10.1007/s00702-023-02736-0.
17. Yilmaz R, Öksüz N, Ceylan M, et al.. Patient Burden in Dystonia Diagnosis and Botulinum Toxin Treatment: A Nationwide Survey in Turkey. Brain Behav. 2025;15(3):e70325. PMID 40021839. doi 10.1002/brb3.70325.
18. Steeves TD, Day L, Dykeman J, et al.. The prevalence of primary dystonia: a systematic review and meta-analysis. Mov Disord. 2012;27(14):1789-1796. PMID 23114997. doi 10.1002/mds.25244.
19. Simpson DM, Hallett M, Ashman EJ, et al.. Practice guideline update summary: Botulinum neurotoxin for the treatment of blepharospasm, cervical dystonia, adult spasticity, and headache [RETIRED]: Report of the Guideline Development Subcommittee of the American Academy of Neurology. Neurology. 2016;86(19):1818-1826. PMID 27164716. doi 10.1212/wnl.0000000000002560.
20. Contarino MF, Van Den Dool J, Balash Y, et al.. Clinical Practice: Evidence-Based Recommendations for the Treatment of Cervical Dystonia with Botulinum Toxin. Front Neurol. 2017;8:35. PMID 28286494. doi 10.3389/fneur.2017.00035.
21. Romano M, Bagnato S, Altavista MC, et al.. Diagnostic and therapeutic recommendations in adult dystonia: a joint document by the Italian Society of Neurology, the Italian Academy for the Study of Parkinson's Disease and Movement Disorders, and the Italian Network on Botulinum Toxin. Neurol Sci. 2022;43(12):6929-6945. PMID 36190683. doi 10.1007/s10072-022-06424-x.
22. Loudovici-Krug D, Derlien S, Best N, et al.. Physiotherapy for Cervical Dystonia: A Systematic Review of Randomised Controlled Trials. Toxins (Basel). 2022;14(11):784. PMID 36422957. doi 10.3390/toxins14110784.
23. De Pauw J, Van der Velden K, Meirte J, et al.. The effectiveness of physiotherapy for cervical dystonia: a systematic literature review. J Neurol. 2014;261(10):1857-1865. PMID 24413637. doi 10.1007/s00415-013-7220-8.
24. Werner C, Loudovici-Krug D, Derlien S, et al.. Study protocol: multimodal physiotherapy as an add-on treatment to botulinum neurotoxin type A therapy for patients with cervical dystonia: DysPT-multi-a prospective, multicentre, single-blind, randomized, controlled study. Trials. 2021;22(1):740. PMID 34696821. doi 10.1186/s13063-021-05705-8.
25. Jacksch C, Loens S, Mueller J, et al.. Impact of Physiotherapy in the Treatment of Pain in Cervical Dystonia. Tremor Other Hyperkinet Mov (N Y). 2024;14:11. PMID 38464912. doi 10.5334/tohm.867.
26. De Pauw J, Mercelis R, Hallemans A, et al.. Cervical sensorimotor control in idiopathic cervical dystonia: A cross-sectional study. Brain Behav. 2017;7(9):e00735. PMID 28948067. doi 10.1002/brb3.735.
27. Kassaye SG, Crosiers D, Gudina EK, et al.. The immediate effect of peripheral stimulation on cervical dystonia disease outcomes: single-case experimental design. Arch Physiother. 2026;16:34-43. PMID 41783380. doi 10.33393/aop.2026.3773.
28. Butler K, Sadnicka A, Edwards MJ, et al.. Long-term (seven-year) follow-up of sensory-motor rehabilitation therapy for task-specific focal hand dystonia. J Hand Ther. 2025;38(4):799-809. PMID 40360314. doi 10.1016/j.jht.2025.02.008.
29. Catellani I, Arcuri P, Vita F, et al.. An overview of rehabilitation approaches for focal hand dystonia in musicians: A scoping review. Clin Rehabil. 2024;38(5):589-599. PMID 38238941. doi 10.1177/02692155231225705.
30. Chiaramonte R, Vecchio M. Rehabilitation of focal hand dystonia in musicians: a systematic review of the studies. Rev Neurol. 2021;72(8):269-282. PMID 33851716. doi 10.33588/rn.7208.2020421.
31. Ackermann B, Altenmüller E. The development and use of an anatomy-based retraining program (MusAARP) to assess and treat focal hand dystonia in musicians-A pilot study. J Hand Ther. 2021;34(2):309-314. PMID 34176657. doi 10.1016/j.jht.2021.05.007.
32. Bleton JP, Vidailhet M, Bourdain F, et al.. Somatosensory cortical remodelling after rehabilitation and clinical benefit of in writer's cramp. J Neurol Neurosurg Psychiatry. 2011;82(5):574-577. PMID 20562399. doi 10.1136/jnnp.2009.192476.
33. Baur B, Fürholzer W, Jasper I, et al.. Effects of modified pen grip and handwriting training on writer's cramp. Arch Phys Med Rehabil. 2009;90(5):867-875. PMID 19406309. doi 10.1016/j.apmr.2008.10.015.
34. Rosset-Llobet J, Fàbregas-Molas S. Long-term treatment effects of sensory motor retuning in a pianist with focal dystonia. Med Probl Perform Art. 2011;26(2):106-107. PMID 21695359. doi 10.21091/mppa.2011.2016.
35. Uluer MC, Bojovic B. A Rare Cervical Dystonia Mimic in Adults: Congenital Muscular Torticollis (Fibromatosis colli), a Follow-up. Front Neurol. 2016;7:7. PMID 26869987. doi 10.3389/fneur.2016.00007.
36. Frucht SJ, George MC, Pantelyat A, et al.. Incobotulinum Toxin-A in Professional Musicians with Focal Task-Specific Dystonia: A Double Blind, Placebo Controlled, Cross-Over Study. Tremor Other Hyperkinet Mov (N Y). 2024;14:32. PMID 38948014. doi 10.5334/tohm.903.
37. Frucht SJ. Deep Phenotyping of Musicians' Upper Limb Dystonia. Tremor Other Hyperkinet Mov (N Y). 2025;15:31. PMID 40688733. doi 10.5334/tohm.1044.
38. Stegink-Jansen CW, Jung B, Somerson JS. Translation of Runge's 1873 publication "On the etiology and treatment of writer's cramp": The first description of "tennis elbow". Clin Anat. 2022;35(3):316-322. PMID 34967051. doi 10.1002/ca.23830.
39. Kupsch A, Benecke R, Müller J, et al.. Pallidal deep-brain stimulation in primary generalized or segmental dystonia. N Engl J Med. 2006;355(19):1978-1990. PMID 17093249. doi 10.1056/nejmoa063618.
40. Vidailhet M, Vercueil L, Houeto JL, et al.. Bilateral deep-brain stimulation of the globus pallidus in primary generalized dystonia. N Engl J Med. 2005;352(5):459-467. PMID 15689584. doi 10.1056/nejmoa042187.
41. Latorre A, Peall KJ, Horn A, et al.. Deep brain stimulation for the treatment of patients with dystonia: advances towards network-based personalised care. Lancet Neurol. 2026;25(8):781-788. PMID 42456686. doi 10.1016/s1474-4422(26)00192-4.
42. Albanese A, Sorbo FD, Comella C, et al.. Dystonia rating scales: critique and recommendations. Mov Disord. 2013;28(7):874-883. PMID 23893443. doi 10.1002/mds.25579.
43. Demartini B, Goeta D, Gambini O. A case of functional dystonia: clinical cues to differentiate between organic and functional dystonia. Neurol Sci. 2019;40(5):1071-1072. PMID 30478718. doi 10.1007/s10072-018-3654-1.
44. Giorelli M. Functional dystonia in a guitarist: diagnostic challenges and clinical reflections. Neurol Sci. 2026;47(3):294. PMID 41746437. doi 10.1007/s10072-026-08902-y.
45. Nijenhuis B, van Wensen E, Smit M, et al.. Treatment of task-specific dystonia in sports: A systematic review. Clin Park Relat Disord. 2024;10:100245. PMID 38456155. doi 10.1016/j.prdoa.2024.100245.
46. Bleton JP, Takyi RB, Verneau M, et al.. Specific components of manual dexterity are affected in patients with writer's cramp: an observational comparative study and preliminary rehabilitation report. J Rehabil Med. 2026;58:jrm45215. PMID 41954216. doi 10.2340/jrm.v58.45215.
47. Bleton JP, Cossé C, Caloc'h T, et al.. Combination of anodal tDCS of the cerebellum with a goal-oriented motor training to treat cervical dystonia: a pilot case series. Front Neurol. 2024;15:1381390. PMID 38746658. doi 10.3389/fneur.2024.1381390.
48. Vu JP, Lee HY, Chen Q, et al.. Head tremor and pain in cervical dystonia. J Neurol. 2021;268(5):1945-1950. PMID 33417005. doi 10.1007/s00415-020-10378-5.
49. Huang X, Wang M, Wang D, et al.. Cervical Dystonia with Classic Sensory Tricks and Forcible Sensory Trick Showed Different Functional Connectivity Alterations: A Functional Near-Infrared Spectroscopy Study. J Clin Med. 2026;15(12):4735. PMID 42355903. doi 10.3390/jcm15124735.
50. Castagna A, Jinnah HA, Albanese A. Duration of botulinum toxin efficacy in cervical dystonia clinical trials: A scoping review. Parkinsonism Relat Disord. 2024;125:107011. PMID 38909588. doi 10.1016/j.parkreldis.2024.107011.
51. Zeuner KE, Bara-Jimenez W, Noguchi PS, et al.. Sensory training for patients with focal hand dystonia. Ann Neurol. 2002;51(5):593-598. PMID 12112105. doi 10.1002/ana.10174.
52. Zeuner KE, Hallett M. Sensory training as treatment for focal hand dystonia: a 1-year follow-up. Mov Disord. 2003;18(9):1044-1047. PMID 14502673. doi 10.1002/mds.10490.
53. Candia V, Schäfer T, Taub E, et al.. Sensory motor retuning: a behavioral treatment for focal hand dystonia of pianists and guitarists. Arch Phys Med Rehabil. 2002;83(10):1342-1348. PMID 12370865. doi 10.1053/apmr.2002.35094.
54. Boyd JT, Fries TJ, Nagle KJ, et al.. A novel presentation of an ocular geste antagoniste in cervical dystonia: a case report. Tremor Other Hyperkinet Mov (N Y). 2013;3:tre-03-199-4624-1. PMID 24386606. doi 10.7916/d8416vsj.
55. Nishida D, Mizuno K, Takahashi O, et al.. Electrically Induced Sensory Trick in a Patient with Musician's Dystonia: A Case Report. Brain Sci. 2023;13(2):223. PMID 36831766. doi 10.3390/brainsci13020223.