REVIEW

CLINICAL PRACTICE

Myoclonus: An Electrophysiological
Diagnosis
Shabbir Hussain I. Merchant, MD,1,* Felipe Vial-Undurraga, MD,2 Giorgio Leodori, MD,3 Jay A. van Gerpen, MD,4 and
Mark Hallett, MD, DM (Hon)5

Background Many different movement disorders have similar “jerk-like” phenomenology and can
ABSTRACT: Background:
be misconstrued as myoclonus. Different types of myoclonus also share similar phenomenological
characteristics that can be difficult to distinguish solely based on clinical exam. However, they have distinctive
physiologic characteristics that can help refine categorization of jerk-like movements.
Objectives In this review, we briefly summarize the clinical, physiologic, and pathophysiologic characteristics of
Objectives:
different types of myoclonus. The methodology and technical considerations for the electrophysiologic
assessment of jerk-like movements are reviewed. A simplistic pragmatic approach for the classification of
myoclonus and other jerk-like movements based on objective electrophysiologic characteristics is proposed.
Conclusions Clinical neurophysiology is an underutilized tool in the diagnosis and treatment of movement disorders.
Conclusions:
Various jerk-like movements have distinguishing physiologic characteristics, differentiated in the milliseconds range,
which is beyond human capacity. We argue that the categorization of movement disorders as myoclonus can be
refined based on objective physiology that can have important prognostic and therapeutic implications.

The clinical approach to the diagnosis of movement disorders begins characterized by generalized jerk-like movements is another
with the initial clinical classification of the movements observed important movement disorder that shares similarities in terms of
based on phenomenology into a discrete number of categories: clinical characteristics.6,7 Many peripheral movement disorders
tremor, myoclonus, chorea, ballism, athetosis, dystonia, spasms, tics, such as clonus and fasciculations have jerk-like phenomenology.
motor stereotypies, and functional. The correct categorization of the In addition, myoclonus may coexist with other movement disor-
movement disorder, or syndrome identification, is a critical first step ders and more than 1 type of jerk-like movement disorder can
in arriving at the correct etiological diagnosis. In this article, we coexist in an individual. So many different movement disorders
address the methodology for improving the clinical characterization have similar jerk-like phenomenology and can be misconstrued as
of myoclonus, including the pitfalls and challenges. Myoclonus is myoclonus based solely on clinical assessment.
defined as a syndrome clinically characterized by sudden, brief “jerk- Although many different types of jerks might look superficially
like” movements. It can be further classified either as positive, when the same, the physiologic characteristics of various types of jerk-like
associated with an active muscle contraction, or negative, when asso- movements noted previously are distinct. They allow for a clear,
ciated with a brief pause in ongoing muscle contraction.1–3 objective, and unequivocal diagnosis. We argue that categorization
Should any jerk-like movement be called “myoclonus”? Tic of a movement disorder as “myoclonus” can be further refined by
disorders, chorea, ballism, dystonia, and functional movements electrophysiological characterization.7,8 Using objective physiology
can all have some jerk-like phenomenology.3–5 Tremors, in par- allows for correct, reliable, and reproducible characterization of
ticular when the frequency is high, can sometimes look like repeti- jerk-like movements as defined by their physiologic signatures and
tive jerk-like movements and, oppositely, rhythmic myoclonic takes away the errors associated with misclassifying various jerk-like
jerks can be misconstrued as tremor.3 Startle syndrome commonly movement disorders solely based on clinical phenomenology.

1
Department of Neurology, Medical University of South Carolina, Charleston, South Carolina, USA; 2Facultad de Medicina, Clínica Alemana Universidad del
Desarrollo, Santiago, Chile; 3IRCCS Neuromed, Pozzilli, Isernia, Italy; 4Department of Neurology, University of Alabama, Huntsville, Alabama, USA; 5National
Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland, USA
*Correspondence to: Dr. Shabbir Hussain I. Merchant, Division of Movement Disorders, Medical University of South Carolina, 208 B Rutledge Ave-
nue, MSC 108, Charleston, SC 29425; E-mail: [email protected]
Keywords: myoclonus; startle, jerk-like movements, electrophysiology.
S.H.I. M. and F.V.-U. have contributed equally to this manuscript and are cofirst authors for this article.
Relevant disclosures and conflicts of interest are listed at the end of this article.
Received 10 January 2020; revised 3 April 2020; accepted 3 May 2020.
Published online 17 June 2020 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/mdc3.12986

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986

489
© 2020 International Parkinson and Movement Disorder Society
 REVIEW MYOCLONUS: AN ELECTROPHYSIOLOGICAL DIAGNOSIS

Objective physiology can further help localize the central or periph- recording of various jerk-like movements. Based on these salient
eral origin of a jerk-like movement within the nervous system that clinical and objective physiologic characteristics, we propose a prag-
can have important prognostic and therapeutic implications.8–11 matic diagnostic algorithm for their evaluation and classification.
Clinical neurophysiology is an underutilized tool in the diagnosis
and treatment of movement disorders.12 We are not arguing against
the importance of a good clinical exam; however, correct identifica-
tion and appropriate categorization of jerk-like movements rely on
spatial and temporal discrimination of its characteristics in the milli-
Results
second range, which is far beyond human capacity.2,5,13–19
Myoclonus Syndromes
Cortical Myoclonus

Methods Clinical Characteristics. The clinical characteristics include spon-

taneous jerks commonly involving the face and distal extremities
Based on a relevant review of the literature, we summarize the clini- that are usually increased with action. Can be triggered by tactile
cal characteristics, phenomenology, physiologic characteristics, and or other somatosensory stimuli to the distal extremities.1
pathophysiology of the common myoclonus and other jerk-like Physiologic Characteristics. The physiologic characteristics include
movement disorder syndromes (summarized in Table 1). We briefly very short duration bursts (usually <30 milliseconds) with simulta-
review the methodology and technical considerations for the neous agonist–antagonist activity. The conduction velocity of the

TABLE 1. Summary of comparative clinical and physiologic characteristics of common jerk-like movements
Category of ‘Jerk-
like’ Movement Cortical Myoclonus Reticular Myoclonus Propriospinal Myoclonus Startle Reflex

Phenomenology Spontaneous jerks more Spontaneous jerks Arrhythmic, usually Bilaterally

commonly involving the involving the entire flexor, brief jerks synchronous flexor
face and distal upper body that can also involving the trunk, response to a
extremities be evoked by hips, knees in a startling stimulus
somatosensory stimuli fairly uniform
to distal extremities pattern

Reflex physiology Cortical onset short Craniocaudal EMG Craniocaudal EMG Early eyelid blink
lasting EMG bursts discharge beginning discharge usually followed most
(simultaneous at the lower medulla beginning at the consistently by SCM
agonist-antagonist level of the with subsequent
discharge) with a abdominal muscles; cranio-caudal
cortical EEG correlate restricted to spinal propagation
on back averaging cord propriospinal
pathways

Afferent Spontaneous; but can be Spontaneous; can be Spontaneous, but can be Auditory; can also be
stimulus triggered by tactile or induced by induced by tactile or elicited by visual,
other somatosensory somatosensory auditory stimuli somatic, and
stimuli more commonly stimuli, touch, or vestibular
to distal upper muscle stretch of stimulations
extremities distal extremities

Pattern of muscle Mainly involving distal Both flexors and Mainly flexors Mainly flexors (but
activation; upper extremities and extensors extensors also noted
flexors vs face (which have a to be activated)
extensors large cortical
representation)

Site of origin Cerebral cortex Medullary reticular Limited to spinal cord Caudal pontine
formation (likely (note that most cases reticular formation
nucleus reticularis are functional) (nucleus reticularis
gigantocellularis) pontis caudalis)

Velocity of ~100 m/sec > 50 m/sec 5–15 m/sec ~30 m/sec

conduction of
bulbo-spinal
efferent
volley

EMG burst <50 milliseconds <50 milliseconds ~150–450 milliseconds >70 milliseconds
duration (can be longer)
1
EEG, electroencephalogram; EMG, electromyogram; SCM, sternocleidomastoid muscle.

490 MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986
S.H.I. MERCHANT ET AL. REVIEW

FIG. 1. Cortical myoclonus physiology. (A) EMG traces showing short duration (<30 milliseconds) simultaneous agonist-antagonist bursts
with cortical EEG spike correlate. (B) Median nerve sensory-evoked potential showing M wave followed by C-reflex with corresponding
EEG trace with giant p25 and n33 sensory-evoked potential components. EEG, electroencephalogram; EMG, electromyogram.

efferent volley is ~100 m/sec with cranio-caudal propagation. Corti- suggestive of site of origin being the sensorimotor cortex with the
cal spikes can be detected on electroencephalogram (EEG) preceding velocity of conduction suggestive of transmission along the
the jerks noted on electromyogram (EMG). There is an exaggera- corticospinal tracts. Physiologic characteristics of enhanced SEP
tion of the P25/N33 components of the median nerve-evoked and exaggerated long loop reflexes suggest hyperexcitability of the
sensory-evoked potential (SEP).15 Physiologic characteristics of corti- sensorimotor cortex. The cerebellum is also implicated in the gen-
cal myoclonus are illustrated in Figure 1. eration and propagation of cortical myoclonus based on aber-
Pathophysiology. The pathophysiology commonly involves rancies noted in the cerebello-thalamo-cortical pathways and the
body regions with the largest cortical homuncular representation presence of cortical myoclonus in certain ataxia syndromes.9,15

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986 491
 REVIEW MYOCLONUS: AN ELECTROPHYSIOLOGICAL DIAGNOSIS

Brainstem Myoclonus axial predominant jerks or involves multiple contiguous body

segments and is referred to as propriospinal myoclonus. No cra-
Clinical Characteristics. The clinical characteristics include axial nial nerve innervated muscules are involved. Generalized jerks
predominant generalized jerk-like movements beginning with commonly originate in the abdominal musculature or other mus-
the ascending order of cranial nerve innervated muscle activation cles innervated at the level of the thoracic spinal cord.14,24,25
and the descending order of limb and axial muscle activation. Physiologic Characteristics. The physiologic characteristics
They can be spontaneous and can also be induced by somatosen- of segmental myoclonus include burst duration ranges
sory stimuli, especially the muscle stretch of distal upper between 50 and 500 milliseconds without caudo-rostral
extremities.5,18 propagation. 26,27
Physiologic Characteristics. The physiologic characteristics The physiologic characteristics of propriospinal myoclonus
include short duration bursts (usually <50 milliseconds) begin- include long burst durations typically ranging from ~150 to
ning in the lower cranial nerve innervated muscles with subse- 450 milliseconds (can be even longer). Conduction velocity of
quent cranio-caudal propagation. The conduction velocity of the the efferent volley ranges from 5 to 15 m/sec limited to the
efferent volley is ~50 to 70 m/sec. Stimulus sensitive myoclonus spinal cord without any cranial nerve innervated muscle activa-
may also be seen.18,20 tion. Generalized jerks usually originate at the level of the tho-
Pathophysiology. The order of muscle discharge is suggestive racic spinal cord with subsequent cranio-caudal propagation
of origin in the caudal medullary reticular formation (nucleus limited to the spinal cord. The pattern of muscle activation can
reticularis gigantocellularis). The conduction velocity of the be inconsistent. Stimulus-evoked jerks can be of variable
efferent volley is suggestive of the reticulospinal tract as its latency and at times longer than voluntary reaction times.
pathway. No cortical discharge is seen preceding the muscle Bereitschaftspotential or event-related desynchronization of the
jerks, although there may be an associated EEG spike. EEG beta band may be seen prior to EMG bursts on jerk-locked
spikes most commonly are noted after the EMG of cranial back averaging of the EEG in many cases, suggestive of a func-
nerve innervated muscles; however, no clear EMG correlation tional syndrome.14,24,25
noted with back averaging.18 Most commonly etiology for Pathophysiology. Spinal segmental myoclonus can be local-
reticular myoclonus is anoxic damage to the medullary reticular ized to the regions affecting a spinal segment. Most cases of
formation.18,20 propriospinal myoclonus are deemed to be of functional origin;
however, some cases of organic/symptomatic propriospinal
Myoclonus–Dystonia myoclonus have been described.14,16,28

Clinical Characteristics. The myoclonus–dystonia syndrome has

been most systematically studied in patients with DYT11 muta-
Startle Syndromes
tions, where the myoclonus is observed in the majority of Startle syndromes, although originating in the brainstem, deserve
patients and is variably associated with dystonia. Patients with a separate categorization as they are a specific aberrancy of nor-
DYT11 show asymmetric rest and action myoclonus with pre- mal human reflexes.6
dominant proximal limb and axial distribution.21,22
Physiological Characteristics. EMG burst duration is in the
50 to 100 millisecond range with longer duration bursts likely
Startle and Exaggerated Startle Reflex
related to dystonia.22 Studies using trancranial magnetic stimula- (Hyperekplexia)
tion (TMS), back averaging, SEPs, and C-reflex failed to demon- Clinical Characteristics. The clinical characteristics include bilat-
strate hyperexcitability of the sensorimotor cortex in patients erally synchronous and predominant flexor jerk-like movements
with DYT11, thus suggesting a subcortical generator.21,22 induced by a sudden, unexpected stimulus. Fairly symmetrical
Pathophysiology. Abnormal activity in the cerebello-thalamic activation is noted beginning in the muscles originating in the
and the striato-pallido-thalamo-cortical circuits probably plays a lower brainstem with subsequent cranio-caudal propagation.
central role in DYT11 pathophysiology; however, whether these Exaggerated startle reflex is characterized by nonhabituating and
mechanisms underpins myoclonus rather than dystonia is excessive motor response secondary to auditory and at times
unclear.23 Finally, the presence in some patients of a startle somesthetic and visual stimuli. It is characterized by a blink; con-
response and the consistent presence of abnormal brainstem tortion of the face; flexion of the neck, trunk, and abduction;
excitability as tested by blink reflex, suggests that stimulus- and flexion of the arm. Exaggerated startling can result in falls
evoked myoclonus is possible in DYT11, and this particular and injuries.6,7,13
manifestation is suggestive of a likely brainstem generator.21–23 Physiologic Characteristics. The physiologic characteristics
include early blinking referred to as the stimulus induced blink
reflex followed most consistently by activation of the
Spinal Myoclonus
sternocleidomastoid muscle (SCM) with subsequent cranio-
Clinical Characteristics. The clinical characteristics can involve caudal propagation. EMG burst durations are ~70 milliseconds
one or a few contiguous spinal myotomes and is referred to as with the conduction velocity of the efferent volley being ~30
spinal segmental myoclonus. It can also present as generalized m/sec. Predominantly involves a flexor muscle response;

492 MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986
S.H.I. MERCHANT ET AL. REVIEW

FIG. 2. Startle reflex physiology. Surface poly-electromyogram single trace of 5 muscles showing acoustically induced blink reflex at a
latency of 35 milliseconds followed by an sternocleidomastoid muscle discharge with subsequent cranio-caudal propagation. Abd,
abductor; Orb, orbicularis.

however, extensor muscle involvement has been reported. Exag- Methodological Considerations
gerated startle reflex is characterized by excessive and more wide-
spread muscle activation and excessive EMG burst amplitude with for Physiologic Evaluation of
a lower threshold for response and impaired habituation.19,29–31 Myoclonus Syndromes
Physiologic characteristics of startle reflex physiology are illustrated
The following electrophysiologic techniques can help with the
in Figure 2.
characterization of myoclonic movements: poly EMG, SEP and
Pathophysiology. Startle is a normal reflex deemed to origi-
“C” (cortical) reflexes, EEG correlates of myoclonus, and startle
nate in the caudal pontine reticular formation (nucleus reticularis
study. The selection of the study depends on the movement
pontis caudalis). Exaggerated startle reflex can be secondary to
characteristics, but generally it is good to start with a poly EMG
hereditary hyperekplexia caused by a variety of genes affecting
recording.
the presynaptic or postsynaptic glycine receptors. These receptors
are involved in the modulation of ligand gated chloride channels
involved in brainstem and spinal cord reflex responses. Abnor-
malities in this receptor complex results in brainstem and spinal
Poly EMG
cord disinhibition in hereditary hyperekplexia. Cases resulting Electrical activity of the muscles involved in the jerk are
from sporadic mutations and idiopathic hyperekplexia have also recorded with surface EMG. The correct selection of the muscles
been reported.7,19,31,32 Differences in startle reflex abnormalities to record is critical and depends on the clinical pattern observed
in parkinsonian disorders also provide some useful pathophysio- during examination. The more muscles recorded, the better the
logic insights. Startle reflexes are characteristically enhanced in temporal-spatial resolution, but this will increase the study time
patients with multisystem atrophy, whereas they are reduced in duration and time required for analysis of the data, a critical fac-
patients with progressive supranuclear palsy. The reflex responses tor to be considered in a clinical setting.1,17 Recordings should
are intermediate and not significantly different compared with be obtained from a pair of agonist–antagonist muscles because
healthy controls in patients with Parkinson’s disease or dementia their simultaneous activation is a helpful indicator of a centrally
with Lewy bodies. Degeneration of the nucleus reticularis pontis driven activity vs a peripheral injury that may also produce short
caudalis in progressive supranuclear palsy and disinhibition of this lasting (in the myoclonic range) muscle bursts.1 In our experi-
nucleus in multisystem atrophy could explain the differences in ence, the general distribution of the movement is also helpful;
reflex response.33 for generalized jerks involving proximal and more symmetric

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986 493
 REVIEW MYOCLONUS: AN ELECTROPHYSIOLOGICAL DIAGNOSIS

activation, it is important to record from different levels of spinal affecting hand muscles) support the cortical origin of the
cord and cranial nerves to ascertain the temporal and spatial pat- myoclonus.1,17
tern of spread of an efferent discharge volley. We suggest record- Back Averaging: The idea of back averaging is to segment
ing from the following muscles: (1) orbicularis oculi (VII cranial the EMG-EEG data in equally sized windows using markers
nerve)/mentalis (VII cranial nerve), (2) masseter (V cranial placed on the myoclonic bursts as fiducials. Then an average of
nerve), (3) SCM (XI cranial nerve), (4) a pair of upper limb ago- all the segments is done. The signal that is in phase will survive
nist/antagonist muscles (biceps/triceps or flexor carpi radialis/ the average, the activity that is out of phase (noise) will cancel
extensor carpi radialis), (5) abductor polices brevis (APB), (6) tho- out. This is a more sensible way of looking for a cortical corre-
racic paraspinal muscles, and (7) Lumbar paraspinal muscles. late of the myoclonus.1,17
To avoid movement artefacts associated with brisk move- Bereitschaftspotential Recording (BP): One of the poten-
ments, it is advisable to scratch off skin debris using skin prep tials that can be looked for with the back-averaging technique
before placing the surface electrodes. Recommended filter set- are the BPs. BP recordings are useful for the diagnosis of jerk-
tings for recording include a high pass of 20 Hz and low pass of like movements of functional etiology. The BP is a slow rising
250 to 300 Hz, which will capture most of the frequency con- negativity coming from the supplementary motor area and
tent of the EMG activity. The sampling rate has to be at least premotor cortex, noted about 2 seconds before voluntary or
2 times the highest frequency recorded (to avoid aliasing and functional (psychogenic) movements.5
make for a faithful recording)1. In general, for the data visualiza- Initial recordings are obtained at “rest” while the involuntary
tion and analysis, it may be helpful to rectify (make everything movements are occurring. In our experience, close to 50 move-
upward) the data to improve the signal to noise ratio.21,11,17 ments are needed, the reason being that in general at least 20 to
25 movements are needed to be able to see a BP on the EEG
after back averaging. Suggested EEG filtering parameters include
SEP and C Reflex a high-pass filter to be as low as possible (we recommend using
0.01 Hz or, even direct current) and the low pass at 50 Hz. BP
These studies will show an electrophysiologic correlate for corti-
being a very slow wave, gradually rising over 2 seconds or even
cal hyperexcitability and cortical reflex myoclonus.1 They can be
slower will be filtered out if the conventional EEG filter parame-
recorded together. Surface EMG electrodes are placed on the
ters are used.35,36 If enough good and consistent recordings are
thenar muscles of the stimulated hand muscle with simultaneous
obtained after back averaging the whole data, the data can be
EEG recordings obtained from CP3 or CP4 electrodes (contra-
spilt into 2 sets to look for consistency. It is also recommended
lateral to the stimulated hand referenced to the linked earlobe).
to record the subject doing voluntary movements and use that
Electrical stimulation is delivered to the median nerve at the
data as a positive control for the technique.14,28,36
wrist level; a square wave pulse of 0.2 to 0.5 milliseconds in
For back averaging, a marker is placed on the EMG at the
duration at a frequency of 1 to 2 Hz at an intensity equivalent to
beginning of each muscle burst, then the data are segmented in
10% to 15% above the motor threshold.3 It is important to mon-
windows around that marker and then averaged. For focal jerks
itor that the thenar muscles are at rest during the recording, as
the placement of EMG marker is fairly straight forward. For
“C” reflexes can be evoked in healthy subjects if the activated
multifocal and generalized jerks, the poly EMG recordings are
muscles are stimulated in a tonically contracted state.1,17 An SEP
evaluated to ascertain the most consistent and earliest muscle
will be considered giant if the averaged amplitude of P25/N33 is
EMG activity and back averaging performed by segmenting the
greater than 10 uV4 (normally around 1 uV). A “C” reflex
EEG/EMG window with a marker placed at the onset of the
observed in resting state at a latency of 45 to 50 milliseconds is
EMG burst.1,10 It can be technically challenging and difficult to
considered abnormal and is an electrophysiologic correlate of
obtain BP if there are repetitive jerks occurring at high frequency.
cortical reflex myoclonus, suggestive of disinhibited motor corti-
Using a methodology similar to jerk-locked back averaging,
cal activity.1,17 A normal H or F response can also be seen prior
event-related desynchronization and event-related synchroniza-
to the C-reflex with a latency of around 30 milliseconds.20,34
tion can also be evaluated. These are essentially changes in the
power of certain EEG band frequencies after performing a
time-frequency transformation of the epoch of the data around
EEG Correlates of Myoclonus the jerk-like movement and comparing these changes to those
Simultaneous Recording of EEG and EMG: EMG record- noted at baseline. Event-related desynchronization of the beta
ings are performed with a similar methodology as described for (13–30 Hz) and mu rhythms (8–12 Hz), characterized by a
the poly EMG recordings. For EEG recordings, it is preferable reduction in the power up to 20% to 30%, can be noted over
to record from multiple electrodes placed on the frontal, sup- the sensorimotor cortices (C3/C4) about 1.5 to 2 seconds pre-
plementary motor area, sensorimotor cortex, and midline (Cz, ceding the movement. Event-related synchronization of beta
C1, and C2 referenced to the linked earlobe). Conventional fil- (13–30 Hz) rhythms is also noted in the first second after a vol-
tering parameters can be used for the EEG. EEG potentials pre- untary movement.37
ceding the EMG myoclonic activity, with a delay compatible Cortico-Muscular Coherence: When the myoclonic activ-
with a corticospinal conduction (~20 milliseconds in myoclonus ity has a very high frequency as observed in some forms of

494 MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986
S.H.I. MERCHANT ET AL. REVIEW

progressive myoclonic epilepsies, looking for the EEG correla-

tion in the time domain can become difficult. One way of deal-
Pragmatic Physiologic
ing with those cases is to run a coherence analysis between the Approach for Diagnosis of
EEG and the EMG signal. This technique is described in research
background, but in the experience of the authors it is difficult to Jerk-Like Movements
apply this in clinic because there is always some normal cortico- Jerk-like movements can be broadly classified based on phenom-
muscular coherence because of muscle contraction. The cutoff enology into spontaneous, action induced, or stimulus induced.
for which level of coherence is abnormal is not well For spontaneous and action-induced jerks, the next step is to
stablished.38,39 record from the muscles identified based on the clinical exam.
For stimulus-induced jerks, the appropriate stimulus can be pro-
Startle Response Recording vided as noted in the methodology and the onset latency of the
jerk-like activity can be noted. Spontaneous jerks/induced jerks
Startle is a reflex response to a potentially threatening stimulus can be further subclassified as being focal/multifocal or general-
and originates from the caudal pontine reticular formation with ized involving the axial musculature. The EMG recording should
subsequent rostro-caudal propagation.29,30 For recording the star- be based on the pattern of jerks. For focal jerks, limited record-
tle reflex, surface EMG electrodes are placed on the muscles ing of the involved muscles can be performed with surface
innervated by different cranial nerves and also the distal muscles EMG, with recordings to be obtained from an agonist–antagonist
to visualize the spatial and temporal characteristics of the EMG pair. For generalized jerks that commonly involve the axial mus-
discharge.29 The recording parameters are the same as described culature, poly EMG recordings should be performed as suggested
for poly EMG. We recommend recording from the following in the methodology. Based solely on the surface EMG record-
muscles: (1) orbicularis oculi (VII cranial nerve), (2) mentalis (VII ings, the jerk-like movements can be classified into 3 broad cate-
cranial nerve), (3) masseter (V cranial nerve), (4) SCM gories based on the duration of the EMG discharge: (1) <50
(XI cranial nerve), (5) a pair of upper limb agonist/antagonist milliseconds, (2) 50 to 100 milliseconds, and (3) >100 millisec-
muscles (biceps/triceps or flexor carpi radialis/extensor carpi onds. A summary of the suggested pragmatic approach for the
radialis) based on the clinical characteristics of discharge, diagnosis of myoclonus/jerk-like movements is illustrated in
(6) APB, (7) cervical paraspinal muscles, and (8) thoracic/lumbar Figure 3.
paraspinal muscles.
The rationale for recording 2 cranial nerve VII innervated
muscles, namely orbicularis oculi and mentalis, being that in EMG Burst Duration <50
acoustic-evoked startle there is a first response of the orbicularis
oculi that is thought to be an auditory blink reflex that is inde-
Milliseconds
pendent from the startle response. The auditory blink reflex also Characteristic of cortical myoclonus or myoclonus of brainstem ori-
does not habituate as the startle response, normally. Recordings gin. The distribution, pattern of spread, and conduction velocity
obtained from the APB are to note a very late response in this can help with further differentiation.
muscle around 100 milliseconds, which is another characteristic Focal/multifocal jerks predominantly involving the face and
of the startle response. The prolonged latency suggestive of rela- arms with simultaneous agonist–antagonist bursts are most likely
tively slower conduction velocity of the startle volley, originating consistent with a cortical myoclonus. Giant SEPs and C
in the caudal pons and arguing against the startle volley being responses may be present, especially when there is a stimulus-
conducted via the fast conducting cortico-spinal tracts as noted sensitive myoclonus.
for cortical myoclonus where the APB discharge latency is ~20 For generalized jerks with truncal (flexors + extensors)
milliseconds. involvement, a further analysis of the spatial and temporal char-
For recording the acoustic startle reflex, a tone burst of acteristics of the discharge can be performed from the poly EMG
125 dB intensity, 1500 Hz frequency, and 50 milliseconds dura- recordings. Onset of the EMG discharge noted in the lower cra-
tion is provided through headphones. This parameters are within nial nerves innervated muscles, with subsequent cranio-caudal
the National Institute for Occupational Safety and Health propagation involving the axial musculature, is most likely sug-
(NIOSH) recommended exposure limits (https://www.cdc.gov/ gestive of a brainstem/reticular myoclonus. An analysis of the
niosh/topics/noise/reducenoiseexposure/regsguidance.html). velocity of conduction of the efferent volley can then be per-
Ten stimuli are randomly given every 1 to 2 minutes. It is formed to further affirm the diagnosis. A simplistic way to ascer-
important to ensure that the stimuli delivered are unpredictable. tain the velocity of conduction can be done utilizing the cranial
For the sensory-evoked startle, an electrical stimulus of 0.1 nerve and spinal cord innervated paraspinal muscles. The relative
millisecond duration can be given in the supraorbital notch2 with distances between the muscles can be measured and divided by
an intensity about 3 times the sensory threshold. The recording the time duration between the subjacent paraspinal muscles dis-
setup is the same as for acoustic startle. For tactile-induced jerks charges noted on the poly EMG to arrive at the velocity of con-
using a similar methodology of recording, a tactile stimulus can duction. If the conduction velocity is >50 m/sec (can be up to
be delivered by tapping over the surface EMG to mark the stim- 70 m/sec), it is suggestive of brainstem/reticular myoclonus. No
ulus on the poly EMG trace.29,30,32 giant SEPs are noted (although not suggested to be performed

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986 495
 REVIEW MYOCLONUS: AN ELECTROPHYSIOLOGICAL DIAGNOSIS

FIG. 3. Schematic illustration for physiologic diagnostic approach for “jerk-like” movements. Based on surface EMG recordings, jerk-like
movements can be categorized into 3 major categories based on burst duration: (1) <50 milliseconds, (2) 50 to 100 milliseconds, and
(3) >100 milliseconds. BP, Bereitschaftspotential recording; ECR, extensor carpi radialis; EMG, electromyogram; FCR, flexor carpi radialis;
LLR, long latency reflex; SCM, sternocleidomastoid muscle; SEP, sensory-evoked potential.

for most cases). Cortical myoclonus has a faster conduction commonly characterized by an initial short-latency, auditory-
velocity, usually >100 m/sec, and also the temporal onset is in induced eyelid blink followed most consistently by SCM muscle
the higher cranial nerves with caudal propagation.1,5,40 discharge followed by subsequent cranio-caudal propagation. The
Peripheral nerve injury–related fasciculation and myokymia conduction velocity based on poly EMG recording is ~30 m/sec,
may produce short duration muscle bursts with duration <50 notably slower than brainstem/reticular myoclonus and faster than
milliseconds. These can be differentiated from cortical myoclo- a spinal myoclonus ~5 to 15 m/sec. A diagnosis of exaggerated
nus by performing simultaneous recordings from agonist– startle/hyperekplexia can be made if there is no habituation to
antagonist pairs.40 Peripheral nerve injury–related discharges are repetitive stimulation as per the methodology noted
characterized by independent firing of the muscles. Cortical previously.5,13,30
myoclonus on the other hand has characteristic simultaneous
agonist–antagonist activation. Needle EMG can be used for the
evaluation of fasciculations and myokymia, if ambiguities remain. EMG Burst Duration >100
Milliseconds
EMG Burst Duration 50 to Characteristic of voluntary movements, tics, or functional move-
ment disorders. Could be seen in organic spinal/propriospinal
100 Milliseconds myoclonus.
Characteristic of a startle reflex. Duration range can be seen in the Focal/multifocal jerks with burst durations >100 milliseconds
myoclonus of brainstem origin and rarely spinal myoclonus. The can be back averaged to cortical EEG to look for the presence/
distribution, pattern of spread, and conduction velocity can help absence of a BP. The presence of BP is suggestive of a likely
with further differentiation. functional jerk disorder or voluntary movement. Event-related
Generalized axial and flexor predominant jerks induced most desynchronization of the beta and mu rhythms in relation to the
commonly by auditory stimulus are most consistent with a startle jerks could also further supplement and improve the diagnostic
syndrome. The spatial-temporal characteristics of the poly EMG gain compared with BP alone for the diagnosis of functional
trace can provide further reaffirmation. Startle reflex is most jerks.41 Focal or multifocal jerks preceded by a premonitory

496 MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986
S.H.I. MERCHANT ET AL. REVIEW

sensory urge, sometimes with a short-duration BP, can be seen in myoclonus occurs in the leg muscles of affected patients while
tic disorders.42 standing, this is only while they are bearing their entire body
Generalized jerks, which are usually noted to originate at the weight.46 For example, having them lean forward onto a chair,
level of the thoracic spinal cord with subsequent cranio-caudal the leg myoclonus utterly extinguishes but rarely, if ever,
propagation limited to the spinal cord, are suggestive of a prop- emerges in the upper limbs.46 The myoclonic bursts of OM are
riospinal myoclonus. The conduction velocity based on poly typically 50 milliseconds in duration, and ostensibly never lon-
EMG is slower, usually 5 to 15 m/sec. The muscle most consis- ger than 100 milliseconds.46 They are of very low amplitude
tently noted to be activated based on poly EMG can be back and resemble trembling, which is how patients usually describe
averaged to cortical EEG for the presence/absence of BP, which them, but are often not visible. The myoclonus is incessant,
can further help characterize the jerks as being functional or occurring irregularly at approximately 5 to 7 Hz, with inter-
organic propriospinal myoclonus, respectively. The consistency posed tonic muscle activity lasting hundreds of milliseconds in
of the pattern of discharge noted per poly EMG is also an impor- affected muscles, which are overwhelmingly the tibialis ante-
tant variable to note—more inconsistencies and variabilities in riors > gastrocnemii, occurring often synchronously in homolo-
the pattern and spread of discharge noted in functional gous muscles.46,48 These homologous, synchronous bursts may
disorders.5,14,16,28,35 alternate semirhythmically between the tibialis anteriors and
gastrocnemii for several hundred milliseconds.46 The quadriceps
are rarely involved and paraspinals appear to be unaffected.46 As
originally noted,47 cerebral disorders are invariably present,
Knowledge Gaps with microvascular leukoencephalopathy being most common,
Jerk-like movements are also associated with certain dystonic followed by Parkinson’s disease and atypical parkinsonism.
syndromes. Based solely on phenomenology, some terms used Although systematic EEG back averaging has not been per-
interchangeably to refer to these syndromes are “myoclonic dys- formed on OM patients, many have undergone cortical SEPs
tonia or dystonic myoclonus” and “myoclonus-dystonia syn- and assessment for long-latency reflexes at rest, which have
drome.” The myoclonus dystonia syndrome has been most been invariably unremarkable.48 In total, the aforementioned
systematically studied in patients with DYT11 mutations, where characteristics, coupled with the absence of involvement of
the myoclonus more commonly occurs independent of the dys- facial muscles, rare cases unilateral OM secondary to contralat-
tonia. The physiologic characteristics of myoclonus noted in eral large vessel frontal lobe strokes,48 and in particular the
these patients are most consistent with those noted in patients striking lower extremity distribution of the myoclonus, suggest
with brainstem myoclonus.21,22,43 Limited systematic studies are that the generator of OM may be within the vertex of the fron-
available to characterize the physiology of myoclonus noted in tal lobes.46–48
other forms of dystonia. Another entity that has been described The overlap noted in the physiologic characteristics of subcor-
as dystonic myoclonus where dystonia and myoclonus occur in tical myoclonus as currently described in the literature does not
combination resulting in longer duration of EMG bursts.10 merit any distinctive classification of this entity. It is more mean-
ingful to describe the myoclonus based on the physiologic char-
acteristics. The similarities noted with the more established forms
“Subcortical myoclonus” is another term that is used to
of myoclonic disorders can provide useful insights into their
describe myoclonus associated with certain disorders. It has
pathophysiology.
been used to describe myoclonus in corticobasal syndrome
(CBS), where the physiologic characteristics of myoclonus The diagnostic sensitivity and specificity of many of the electro-
are most consistent with those noted in cortical myoclonus physiologic measures described are not known. The absence of
without a clear and consistent cortical correlate (no giant certain neurophysiologic features in certain jerk-like movements
SEPs or EEG correlate).9,38,43 However, whether myoclo- do not necessarily rule out the diagnosis. As an example, the
nus in CBS is cortical or subcortical remains unresolved. absence of cortical spikes and giant SEP do not necessarily rule
Arguments favoring a cortical origin are based on out cortical myoclonus that could be secondary to technical limi-
magnetoencephalogram (MEG) back averaging, latency of tations and limitations of scalp EEG.49 On a similar vein, cortical
exaggerated long-latency cutaneous reflexes with synchro- myoclonus may not be stimulus sensitive at times, and as such the
nous distal involvement, and paired pulse TMS studies C-reflexes and long latency reflexes may also be normal. There-
suggestive of cortical disinhibition.9,44,45
fore, the results of the electrophysiologic assessment need to be
interpreted in their totality and should serve as complementary to
Myoclonus noted in orthostatic myoclonus is also deemed a carful history and physical examination.
to be of subcortical origin; however, the site of its origin
within the neuraxis largely remains unknown.46 Orthostatic
myoclonus (OM) was originally recognized in the bilateral leg
muscles of older patients with a variety of cerebral lesions while
standing.47 A subsequent, retrospective study of a large number
Conclusion
of OM patients confirmed some of the findings in the original Many different movement disorders with a ‘jerk-like’ phenome-
report but revealed that, although the lower extremity nology share similar clinical characteristics.4,13 Based solely on

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986 497
 REVIEW MYOCLONUS: AN ELECTROPHYSIOLOGICAL DIAGNOSIS

clinical characteristics they can be mislabeled and misclassified as Advisory Boards of CALA Health and Brainsway. He has
myoclonus.5,14 However, different types of myoclonus and reflex research grants from Allergan for studies of methods to inject
disorders have distinctive physiologic characteristics.10,15 Refin- botulinum toxins, Medtronic, Inc. for a study of DBS for dysto-
ing the characterization of these disorders using objective physi- nia, and CALA Health for studies of a device to suppress
ology makes this classification more meaningful. It also helps tremor.
ascertain the site of origin within the motor system. We propose
a simplistic diagnostic approach for the objective physiologic
diagnosis of myoclonic disorders, functional movement disorders
with a jerk-like phenomenology, and startle/startle-like reflexes.
As per the proposed algorithm, most jerk-like movement disor- References
ders can be recognized based on surface EMG recordings. Based 1. Shibasaki H. Electrophysiological studies of myoclonus. Muscle Nerve
2000;23:321–335.
solely on the duration of EMG discharge and pattern of spread
2. Caviness JN, Brown P. Myoclonus: current concepts and recent
to the different muscles, we can easily make the correct diagno- advances. Lancet Neurol 2004;3:598–607.
sis. These techniques can be easily incorporated into clinical 3. Toro C, Pascual-Leone A, Deuschl G, Tate E, Pranzatelli MR,
practice as a diagnostic test, and the methodology of recording Hallett M. Cortical tremor. A common manifestation of cortical myoclo-
nus. Neurology 1993;43:2346–2353.
can be adapted to suit most neurophysiology laboratories. Cor-
4. Caviness JN. Myoclonus. Parkinsonism Relat Disord 2007;13(suppl 3):
rect characterization of myoclonus can have important patho- S375–S384.
physiologic and therapeutic implications. One of the biggest 5. Merchant SH, Vial F, Leodori G, Fahn S, Pullman SL, Hallett M. A
challenges for the diagnosis and treatment of functional move- novel exaggerated "spino-bulbo-spinal like" reflex of lower brainstem
origin. Parkinsonism Relat Disord 2018;61:34–38.
ment disorders remains getting the patients to accept the diagno-
6. Bakker MJ, van Dijk JG, van den Maagdenberg AM, Tijssen MA. Startle
sis.24,28 Having an objective physiologic test for diagnosis can syndromes. Lancet Neurol 2006;5:513–524.
help facilitate the acceptance of the diagnosis, which can have 7. Dreissen YE, Bakker MJ, Koelman JH, Tijssen MA. Exaggerated startle
implications on prognosis and health care expenditures.24,50■ reactions. Clin Neurophysiol 2012;123:34–44.
8. Shibasaki H. Physiology of negative myoclonus. Adv Neurol 2002;89:
103–113.
9. Mima T, Nagamine T, Ikeda A, Yazawa S, Kimura J, Shibasaki H. Path-

Author Roles ogenesis of cortical myoclonus studied by magnetoencephalography. Ann

Neurol 1998;43:598–607.
10. Shibasaki H. Neurophysiological classification of myoclonus. Neurophysiol
(1) Research Project: A. Conception, B. Organization, Clin 2006;36:267–269.
C. Execution; (2) Manuscript Preparation: A. Writing of the First 11. Tassinari CA, Rubboli G, Shibasaki H. Neurophysiology of positive and
Draft, B. Review and Critique. negative myoclonus. Electroencephalogr Clin Neurophysiol 1998;107:181–195.
S.H.I.M.: 1A, 1B, 1C, 3A, 3B 12. Zutt R, Elting JW, van Zijl JC, et al. Electrophysiologic testing aids diag-
nosis and subtyping of myoclonus. Neurology 2018;90:e647–e657.
F.V.: 1A, 1B, 1C, 3A, 3B
13. Andermann F, Keene DL, Andermann E, Quesney LF. Startle disease or
G.L.: 1B, 3B hyperekplexia: further delineation of the syndrome. Brain 1980;103:985–997.
V.J.A.: 3B 14. Erro R, Bhatia KP, Edwards MJ, Farmer SF, Cordivari C. Clinical diag-
M.H.: 1B, 3B nosis of propriospinal myoclonus is unreliable: an electrophysiologic
study. Mov Disord 2013;28:1868–1873.
15. Avanzini G, Shibasaki H, Rubboli G, et al. Neurophysiology of myoclo-
nus and progressive myoclonus epilepsies. Epileptic Disord 2016;18:11–27.
16. Esposito M, Erro R, Edwards MJ, et al. The pathophysiology of symp-
Disclosures tomatic propriospinal myoclonus. Mov Disord 2014;29:1097–1099.
17. Shibasaki H, Hallett M. Electrophysiological studies of myoclonus. Muscle
Ethical Compliance Statement: The authors confirm that Nerve 2005;31:157–174.
approval by an institutional review board and patient consent 18. Hallett M, Chadwick D, Adam J, Marsden CD. Reticular reflex myoclo-
nus: a physiological type of human post-hypoxic myoclonus. J Neurol
was not required for this work. We confirm that we have read Neurosurg Psychiatry 1977;40:253-264.
the Journal’s position on issues involved in ethical publication 19. Brown P. Neurophysiology of the startle syndrome and hyperekplexia.
and affirm that this work is consistent with those guidelines. Adv Neurol 2002;89:153–159.
Funding Sources and Conflicts of Interest: The authors 20. Rektor I, Kadanka Z, Bednarik J. Reflex reticular myoclonus: relation-
ship to some brainstem pathophysiological mechanisms. Acta Neurol Scand
declare there are no conflicts of interest relevant to this work. 1991;83:221–225.
M.H. was supported by National Institute of Neurological Dis- 21. Popa T, Milani P, Richard A, et al. The neurophysiological features of
orders and Stroke (NINDS) intramural program. myoclonus-dystonia and differentiation from other dystonias. JAMA
Neurol 2014;71:612–619.
Financial Disclosures for the Previous 12 Months:
22. Roze E, Apartis E, Clot F, et al. Myoclonus-dystonia: clinical and elec-
S.H.I.M. serves as a contractor for New Touch Digital Inc. and trophysiologic pattern related to SGCE mutations. Neurology 2008;70:
has received fees for his time for consultation. M.H. holds pat- 1010–1016.
ents for an immunotoxin for the treatment of focal movement 23. Welter ML, Grabli D, Karachi C, et al. Pallidal activity in myoclonus
dystonia correlates with motor signs. Mov Disord 2015;30:992–996.
disorders and the H-coil for magnetic stimulation; in relation to
24. Erro R, Edwards MJ, Bhatia KP, Esposito M, Farmer SF, Cordivari C.
the latter, he has received license fee payments from the National Psychogenic axial myoclonus: clinical features and long-term outcome.
Institutes of Health (from Brainsway). He is on the Medical Parkinsonism Relat Disord 2014;20:596–599.

498 MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986
S.H.I. MERCHANT ET AL. REVIEW

25. Esposito M, Edwards MJ, Bhatia KP, Brown P, Cordivari C. Idiopathic 39. Caviness JN, Adler CH, Sabbagh MN, Connor DJ, Hernandez JL,
spinal myoclonus: a clinical and neurophysiological assessment of a Lagerlund TD. Abnormal corticomuscular coherence is associated with
movement disorder of uncertain origin. Mov Disord 2009;24:2344–2349. the small amplitude cortical myoclonus in Parkinson’s disease. Mov Disord
2003;18:1157–1162.
26. Ahn JE, Yoo D, Jung KY, Kim JM, Jeon B, Lee MC. Spinal myoclonus
responding to continuous intrathecal morphine pump. J Mov Disord 40. Tunc S, Bruggemann N, Baaske MK, et al. Facial twitches in
2017;10:158–160. ADCY5-associated disease—myokymia or myoclonus? An electromyog-
raphy study. Parkinsonism Relat Disord 2017;40:73–75.
27. Wong SSC, Qiu Q, Cheung CW. Segmental spinal myoclonus compli-
cating lumbar transforaminal epidural steroid injection. Reg Anesth Pain 41. Beudel M, Zutt R, Meppelink AM, et al. Improving neurophysiological
Med 2018;43:554–556. biomarkers for functional myoclonic movements. Parkinsonism Relat Dis-
ord 2018;51:3–8.
28. van der Salm SM, Erro R, Cordivari C, et al. Propriospinal myoclonus:
clinical reappraisal and review of literature. Neurology 2014;83:1862–1870. 42. Karp BI, Porter S, Toro C, Hallett M. Simple motor tics may be pre-
ceded by a premotor potential. J Neurol Neurosurg Psychiatry 1996;61:
29. Brown P. Physiology of startle phenomena. Adv Neurol 1995;67:273–287.
103–106.
30. Brown P, Rothwell JC, Thompson PD, Britton TC, Day BL,
43. Thompson PD, Day BL, Rothwell JC, Brown P, Britton TC,
Marsden CD. New observations on the normal auditory startle reflex in
Marsden CD. The myoclonus in corticobasal degeneration. Evidence for
man. Brain 1991;114 ( Pt 4):1891–1902.
two forms of cortical reflex myoclonus. Brain 1994;117 (Pt 5):
31. Matsumoto J, Fuhr P, Nigro M, Hallett M. Physiological abnormalities 1197–1207.
in hereditary hyperekplexia. Ann Neurol 1992;32:41–50.
44. Chen R, Ashby P, Lang AE. Stimulus-sensitive myoclonus in akinetic-
32. Brown P, Rothwell JC, Thompson PD, Britton TC, Day BL, rigid syndromes. Brain 1992;115 ( Pt 6):1875–1888.
Marsden CD. The hyperekplexias and their relationship to the normal
45. Pal PK, Gunraj CA, Li JY, Lang AE, Chen R. Reduced intracortical and
startle reflex. Brain 1991;114 ( Pt 4):1903–1928.
interhemispheric inhibitions in corticobasal syndrome. J Clin Neurophysiol
33. Kofler M, Muller J, Wenning GK, et al. The auditory startle reaction in 2008;25:304–312.
parkinsonian disorders. Mov Disord 2001;16:62–71.
46. van Gerpen JA. A retrospective study of the clinical and electrophysio-
34. Gilmore RL. Sensory evoked potentials in clinical practice. Semin Neurol logical characteristics of 32 patients with orthostatic myoclonus. Parkin-
1990;10:185–195. sonism Relat Disord 2014;20:889–893.
35. Meppelink AM, Little S, Oswal A, et al. Event related desynchronisation 47. Glass GA, Ahlskog JE, Matsumoto JY. Orthostatic myoclonus: a
predicts functional propriospinal myoclonus. Parkinsonism Relat Disord contributor to gait decline in selected elderly. Neurology 2007;68:
2016;31:116–118. 1826–1830.
36. Terada K, Ikeda A, Van Ness PC, et al. Presence of Bereitschaftspotential 48. Hassan A, van Gerpen JA. Orthostatic tremor and orthostatic myoclonus:
preceding psychogenic myoclonus: clinical application of jerk-locked weight-bearing hyperkinetic disorders: a systematic review, new insights,
back averaging. J Neurol Neurosurg Psychiatry 1995;58:745–747. and unresolved questions. Tremor Other Hyperkinet Mov (N Y) 2016;
6:417.
37. Pfurtscheller G, Lopes da Silva FH. Event-related EEG/MEG synchroni-
zation and desynchronization: basic principles. Clin Neurophysiol 1999; 49. Pillai J, Sperling MR. Interictal EEG and the diagnosis of epilepsy.
110:1842–1857. Epilepsia 2006;47(suppl 1):14–22.
38. Grosse P, Kuhn A, Cordivari C, Brown P. Coherence analysis in 50. Carson A, Hallett M, Stone J. Assessment of patients with functional
the myoclonus of corticobasal degeneration. Mov Disord 2003;18:1345–1350. neurologic disorders. Handb Clin Neurol 2016;139:169–188.

MOVEMENT DISORDERS CLINICAL PRACTICE 2020; 7(5): 489–499. doi: 10.1002/mdc3.12986 499