Pract Neurol 9:102-109 doi:10.1136/jnnp.2009.172213
  • Neurological Rarity

The paroxysmal dyskinesias

  1. A-F van Rootselaar,
  2. S Schade van Westrum,
  3. M A J Tijssen,
  4. Demetrios N Velis
  1. Department of Neurology and Clinical Neurophysiology, Academic Medical Centre, University of Amsterdam, The Netherlands
  2. Epilepsy Institutes of The Netherlands Foundation (SEIN), “Meer en Bosch” Campus, Heemstede, The Netherlands
  1. M A J Tijssen, Department of Neurology, Academic Medical Center, PO Box 22660, 1100 DD Amsterdam, The Netherlands; M.A.Tijssen{at}


    The paroxysmal dyskinesias are a challenging group of movement disorders characterised by painless dystonic and/or choreiform movements. Lack of familiarity with their features and a normal neurological examination between attacks frequently cause diagnostic delays, or even the diagnosis of a non-organic disorder. They are classified by their mode of triggering, and also by the duration and frequency of attacks, effectiveness of medication, and any associated syndromes. Four subtypes are recognised: paroxysmal kinesigenic dyskinesia induced by sudden movement; paroxysmal non-kinesigenic dyskinesia precipitated by for instance alcohol or caffeine; paroxysmal exercise-induced dyskinesia triggered by longer lasting activity; and paroxysmal hypnogenic dyskinesia occurring during sleep. Here we will summarise the characteristics of the subtypes, discuss the differential diagnosis, genetic aspects and pathophysiology, and give practical advice on the diagnostic work-up and treatment.


    The “paroxysmal dyskinesias” are a group of movement disorders where the diagnosis is largely based on the history—the patients are generally completely normal between attacks. The paroxysmal movements are mainly dystonic and choreiform. Dystonia is defined as sustained abnormal muscle contractions, causing twisting and repetitive movements or abnormal postures, and chorea as “dance-like”, restless movements jumping from one body part to another. Most patients have a family history, but sporadic as well as symptomatic cases, secondary to some identifiable pathology, have been described.13 Lack of familiarity with these disorders and the normal neurological examination between attacks frequently causes diagnostic delays, even up to 20 years has been reported, or the incorrect diagnosis of a non-organic disorder. Because most of these patients have a very good response to medical treatment it is important to be familiar with their clinical presentation and we will illustrate the syndrome with a typical case.

    A CASE

    A 35-year-old man was referred to our hospital with what he called left-sided “cramps”. His symptoms had started two years earlier. These abnormal muscle contractions occurred several times a day and lasted about 20 seconds. Each attack started with aching feelings in the left axilla subsequently spreading over the left side of his body, followed by left-sided painless muscle contractions accompanied by involuntary movements of his left leg. His work as a carpenter was almost impossible because the attacks were mainly induced by lifting heavy objects. However, his ability to abort an attack due to vague premonitory symptoms enabled him to avoid awkward situations and keep working. His behaviour during his efforts to avoid the attacks was considered “weird” by his colleagues. His medical history was unremarkable and the family history was negative for neurological disorders. Physical examination was normal.

    After extensively being questioned, his wife was able to describe a major nocturnal problem; short lasting attacks with involuntary kicking leg movements 10 times a night which left her bruised in the morning. Indeed, the attacks occurred more frequently during the night than during the day. Over the previous two years he had even been transferred three times to the emergency department of a hospital during the night because he had tonic-clonic movements and his wife was not able to wake him up, suggesting epilepsy. CT brain scanning and several electroencephalograms (EEGs) were normal. The attacks were considered as non-organic and the patient was referred to our clinic for a second opinion.

    He was seen by a movement disorders specialist who concluded that the daytime attacks were very suggestive of paroxysmal kinesigenic dyskinesia, albeit non-familial, while the night-time attacks were suggestive of nocturnal frontal lobe epilepsy. He had a 24-hour EEG with video recording. During the day, the attacks were mainly triggered by lifting and consisted of sudden vocalisation, rising of his trunk and dystonic movements of his left leg (see online video clip 1), some others occurred at rest (see online video clip 2). The attacks lasted for about 20 seconds. During sleep similar attacks were seen starting with abrupt arousal without any post-ictal signs such as confusion or drowsiness (see online video clip 3). The EEG was unremarkable except for a run of 7–8/s theta waves of low voltage over the vertex became visible just after the attack (during the attack the EEG was obscured by EMG-artefacts, fig); these were not typical epileptiform discharges and might have been post-ictal changes. Rhythmicity was considered to be consistent with a possible epileptic attack, especially when seen during the clinically suspected attack. Furthermore, these EEG abnormalities might have reflected activity generated deep in the frontal cortex away from the skull, as has been postulated in patients with frontal lobe epilepsy and no or minor EEG changes during or after attacks. An MR brain scan was normal.


    Electroencephalogram during an attack while sleep. (A) Slow waves during sleep. There is no evidence of epileptiform activity. The sudden transition to high-amplitude, irregular fast activity is caused by muscle artifact on waking up (video clip 3, sleep, 23:55:55). (B) Post-ictal rhythmic theta-activity over vertex (electrode Cz-ggr) and mid to right parietal regions (electrode Pz-ggr and P4-ggr) for 10 s. Settings EEG channels: common average reference; high pass filter 0.1 Hz (A) or 1 Hz (B); low pass filter 70 Hz; notch 50 Hz; scale vertical bar  = 5.00 μV; horizontal bar  = 1 s.

    A diagnosis of kinesigenic dyskinesia and nocturnal frontal lobe epilepsy was made and he was started on carbamazepine, 100 mg twice daily. The attack-frequency dramatically declined. The carbamazepine was changed to oxcarbazepine because of adverse effects and the attacks completely disappeared when he reached 1200 mg daily.


    Currently, the paroxysmal dyskinesias are mainly classified by their triggers and on the basis of the duration and frequency of the attacks, effectiveness of medication, and associated syndromes; four subtypes are recognised (table). They are all rare. Paroxysmal kinesigenic dyskinesia is the most frequently encountered type, with an estimated prevalence of 1 in 150 000, followed by paroxysmal non-kinesigenic dyskinesia with an estimated prevalence of one in a million. Paroxysmal exercise-induced dyskinesia is even less often encountered. Paroxysmal hypnogenic dyskinesia is now considered a form of nocturnal frontal lobe epilepsy, as we will discuss below.

    Table The paroxysmal dyskinesias


    Paroxysmal kinesigenic dyskinesia is characterised by the occurrence of painless dystonia triggered by sudden action involving the whole body. Diagnostic criteria include:2

    • an identified kinesigenic trigger for the attacks

    • short duration attacks (<1 minute)

    • no loss of consciousness or pain during attacks (just sometimes the dystonia can be painful)

    • usually age of onset between 1 and 20 years if no family history

    • normal interictal neurological examination

    • exclusion of other organic diseases such as multiple sclerosis, vascular events, metabolic causes, or head trauma

    • control of attacks with carbamazepine or phenytoin.

    Less common triggers include a restricted movement such as stretching an arm, startle and the intention to move. The threshold for attacks is lowered by anxiety and stress. Most patients have premonitory sensations such as a not-easy-to-explain general feeling and/or focal numbness or tingling where the involuntary movement begins.2 These can help patients to prevent or abort attacks by avoiding or stopping the movement. Attacks are often unilateral, but can alternate or be bilateral. They occur frequently, not uncommonly more than 20 times per day. When the first attack happens after the age of 20, and there is no family history, further investigations to exclude an underlying disorder are recommended. Remission or marked spontaneous improvement occurs in most patients in their twenties or somewhat later.

    Most of the reported cases are familial. Linkage to chromosome 16q11.2-q12.1 has been established in several families, referred to as the episodic kinesigenic dyskinesia 1 locus or dystonia genetic locus 10 (DYT10, OMIM 128200). A second locus has been mapped to 16q13-q22.1 termed episodic kinesigenic dyskinesia 2 (OMIM 611031). Because the episodic kinesigenic dyskinesia 1 and 2 loci have been excluded in a paroxysmal kinesigenic dyskinesia pedigree, a third locus is suspected.2

    In a number of pedigrees, patients and/or family members report infantile convulsions. It appears that loci for paroxysmal kinesigenic diskinesia, for the “infantile convulsions and paroxysmal choreoathetosis syndrome” and for “benign familial infantile convulsions” overlap, suggesting these are allelic disorders.


    In patients with paroxysmal non-kinesigenic dyskinesia and a mutation in the myofibrillogenesis regulator-1 (MR-1) gene, the dyskinesia is precipitated by, for example, alcohol, caffeine or stress, and less often by exercise, fatigue or cold.3 The attacks can usually be prevented or aborted by sleep. Diagnostic criteria based on the characteristics of patients with a proven MR-1 gene mutation include:

    • hyperkinetic involuntary movement attacks with dystonia, chorea or combinations of these, typically lasting 10 minutes to 1 hour

    • normal neurological examination between attacks

    • onset of attacks in infancy or early childhood

    • precipitation of attacks by caffeine and alcohol consumption

    • usually family history of the movement disorder.

    The frequency of attacks is between one a week to several in a lifetime. Many patients describe premonitory sensations, and speech involvement has been reported often as well. In familial cases a mutation in the MR-1 gene at the 2q35 region has been identified (OMIM 18800, DYT8). A second locus linked to the 2q31 region has been reported in one family but with a different clinical presentation (OMIM 611147).3

    Bruno et al compared the clinical features of eight kindreds with MR-1 mutations to those of six kindreds with dyskinesia that was not clearly induced by movement or during sleep, but lacking MR-1 mutations.3 Patients without MR-1 mutations had a more variable age at onset. Alcohol was never reported as a precipitant and emotional stress less often so, whereas exercise and also fatigue were more often precipitants. Attacks could consist of ballism. Seizures occurred in the MR-1 negative group but not in the MR-1 mutation positive kindreds. Migraine occurred less often in the mutation negative families. Anti-epileptic drugs seemed to be more effective and benzodiazepines somewhat less effective compared to the MR-1 mutation positive group.


    This is a rarer and less well documented group. Attacks are triggered by longer lasting physical activity like walking or running for perhaps 10–15 minutes.13 They are usually dystonic and appear in the body part involved in the exercise and last for 10–15 minutes after stopping the exercise. An association with familial epilepsy has been reported, although rarely. Linkage-analysis in a family with “Rolandic epilepsy—paroxysmal exercise-induced dyskinesia—writer’s cramp syndrome” revealed a cluster of genes at the pericentromeric region of chromosome 16, near the paroxysmal kinesigenic dyskinesia locus. Recently, in three unrelated families with childhood onset exertion-induced dyskinisia and epilepsy, three different heterozygous mutations were identified in a gene encoding a major glucose transporter in the blood-brain barrier (gene map locus 1p35-p31.3, solute carrier family 2 member 1, SLC2A1; glucose transporter GLUT1; DYT17; OMIM 612126).


    The attacks occur mainly during the night and a typical one has been described as follows:1 the patient awakens with a cry and has involuntary dystonic and ballistic thrashing movements lasting up to 45 seconds not necessarily accompanied by EEG-abnormalities. Several attacks can occur each night. These attacks, also named paroxysmal nocturnal dyskinesia or nocturnal paroxysmal dystonia, are now recognised to be autosomal dominant nocturnal frontal lobe epilepsy; these patients can also have other types of attacks such as paroxysmal arousals and episodic nocturnal wanderings. Gene map loci are 20q13.2-q13.3 (ADNFLE type 1), 15q24 (type 2), 1q21 (type 3), and 8p21 (type 4), all encoding or close to a neuronal nicotinic cholinergic receptor subunit.


    • Hyperekplexia should be considered when a patient has a sudden cramping movement induced by startle. These patients have an excessive startle response followed by short-lasting generalised stiffening that causes them to fall “as stiff as a stick”, without loss of consciousness, from early childhood.

    • Startle epilepsy is another consideration; an asymmetrical tonic epileptic seizure is induced by startle and may last half a minute. The underlying cause is usually perinatal asphyxia and the patients also generally have a hemiparesis and learning disability.

    • The episodic ataxias (EA) can also resemble the paroxysmal dyskinesias. In EA1, the attacks are brief, triggered by startle and are associated with subtle myokimias in the face or hands. In EA2, the attacks last for hours to days, are induced by stress or exertion and usually downbeat nystagmus is present. Both EA1 and EA2 are ion channel disorders and respond to acetazolamide. Interestingly, patients with EA-1 can have paroxysmal kinesigenic dyskinesias as well.

    • Somatoform disorders can be difficult to differentiate but clues are inconsistencies in the symptoms, and late onset.

    • Juvenile myoclonic epilepsy, characterised by myoclonic jerks occuring mainly within the first couple of hours after awakening, more often involving the arms than the legs with typical EEG changes.


    • REM sleep behaviour disorder is a parasomnia characterised by the loss of normal skeletal muscle atonia during REM sleep with prominent motor activity accompanying dreaming. It can be idiopathic or associated with neurological diseases, particularly the synucleinopathies.

    • Other parasomnias such as night terrors and confusional arousals.

    • Periodic limb movement syndrome is characterised by involuntary movements of the legs during sleep and preventing deep sleep, usually after the age of 40.


    In paroxysmal exercise-induced dyskinesia a gene encoding a major glucose transporter in the blood-brain barrier has been identified, and it has been proposed that the dyskinesias result from an exertion-induced energy deficit causing episodic dysfunction in the basal ganglia. In both the paroxysmal kinesigenic dyskinesias and the paroxysmal exercise-induced dyskinesias genetic loci are near ion channel genes. This and the paroxysmal nature of the disorders gives rise to the hypothesis that a channelopathy might play a role as it does in the episodic ataxias and autosomal dominant familial epilepsy syndromes. Surprisingly, the MR-1 gene in paroxysmal non-kinesigenic dyskinesia leads to a defect in the stress response pathway, not a channelopathy.

    The pathophysiology of the paroxysmal dyskinesias remains unknown. There are several hypotheses. The first is that they are forms of reflex epilepsy with a focus in the basal ganglia or thalamus. Supporting this theory are the stereotypic attacks and their response to antiepileptic drugs. Alternatively, a primary basal ganglia disorder has been suggested because there are seldom any scalp EEG abnormalities (but epileptiform activity in subcortical grey matter is nearly impossible to identify on the scalp EEG). Also, an ion channel disorder might lead to changes in both basal ganglia and cerebral cortex.


    In clinical practice, in patients with a family history and age of onset before 20 years, no investigations are required, except for genetic studies. In patients older than 20 years and without a family history, secondary causes need to be pursued. Neuroimaging and laboratory studies are necessary to rule out multiple sclerosis which can present with paroxysmal dystonias, head trauma, tumour or a vascular lesion such as a lacunar stroke causing hemiballismus. Metabolic derangements should also be considered, especially hypo- and hyperglycaemia, hyperglycinaemia and hypercalcaemia.

    Because paroxysmal dyskinesias are rare no randomised trials are available. The therapeutic strategy is based on the subclassification and anecdotal experience. Fortunately this is a condition where various approaches can be tried to “see what works” for a particular patient, although in some cases patients may prefer to live with their strange attacks—once they know what they are—without any treatment. Patients with paroxysmal kinesigenic dyskinesia and patients with nocturnal frontal lobe epilepsy respond particularly well to low dose carbamazepine (100 mg twice a day) or phenytoin. This approach is less successful in paroxysmal non-kinesigenic dyskinia but clonazepam and benzodiazepines have been reported to be effective, especially in the MR-1 positive cases. In paroxysmal exercise-induced dyskinesia acetazolamide, levodopa and trihexiphenidyl have been reported to be effective in small series or individual cases.


    This patient illustrates that lack of familiarity with the paroxysmal dyskinesias leads to difficulties making the correct diagnosis. His symptoms were interpreted as a non-organic disorder at first. Since physical examination was unhelpful, a carefully taken history, knowledge of the disorder, and video recording of an attack were required for correct diagnosis. This patient had characteristics both of paroxysmal kinesigenic dyskinesia and paroxysmal hypnogenic dyskinesia/nocturnal frontal lobe epilepsy, although not quite fitting the criteria of the former as his age of onset was over 20. However, cases have been described with, for instance, onset at 40 years.2 It seems that the different types of the paroxysmal dyskinesias and related disorders are on a spectrum rather than being different disorders.

    Practice points

    • “Strange” hyperkinetic attacks can be a paroxysmal dyskinesia.

    • Interictal neurological examination is usually normal.

    • Interictal EEG is usually normal, and the ictal EEG can be normal as well.

    • (Home) video recordings are essential for diagnosis.

    • Ask the partner about any abnormal night-time activities, like kicking, during sleep.

    • A family history should be sought for, not only similar attacks, but also childhood epilepsy running in the family can point to a paroxysmal dyskinesia.

    • Absence of a family history and age of onset above 20 require additional investigations to exclude an underlying cause (such as a tumour).

    • Low dose carbamazepine (100 mg twice daily) is usually effective.


    We would like to thank Thijs Boerée for his help with the figure. This article was reviewed by Richard Davenport, Edinburgh, UK.



    Free sample
    This issue is free to all users to allow everyone the opportunity to see the full scope and typical content of Practical Neurology.
    View free sample issue >>

    Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.