Perioral Reflex Myoclonias: A Controlled Study in Patients with JME and Focal Epilepsies


Address correspondence and reprint requests to Dr. Thomas A. Mayer at Kleinwachau-Saxonian Epilepsy Centre, Radeberg Wachaverstr. 30, D-01454 Radeberg. E-mail:


Summary: Purpose: Perioral reflex myoclonias (PORM) are obvious, frequent, but often unobserved focal seizures in different epileptic syndromes and the leading seizure type in reading epilepsy. PORMs remain often undiagnosed because the patients are not aware that these are epileptic seizures and fail to report them. Their semiology is not fundamentally different in various epileptic syndromes.

Methods: We studied the frequency of PORM in patients with juvenile myoclonic epilepsy (JME) compared with patients with focal epilepsies. Twenty-five patients with JME were investigated with a standardized neuropsychological test program and compared with 25 matched patients with focal epilepsies. Statistical significance was calculated by using Fisher's exact test.

Results: We found significant differences between the groups regarding both frequency of PORM and activation of epileptic discharges. These observations seem to indicate that PORM, like praxis-induced seizures, are typical traits in JME.

Conclusions: PORM are more frequent in JME compared with focal epilepsies. The distinction between focal and generalized epileptic ictogenesis may be less clear than is traditionally believed.

Myoclonic seizures are defined as short, isolated, or repetitive arrhythmic jerks in a part of the body or in the whole body. For some epileptic syndromes, they are the name-giving features [e.g., juvenile myoclonic epilepsy (JME) or progressive myoclonus epilepsy (PME)]. In other syndromes like primary reading epilepsy (RE), the main seizure type are myoclonic jerks involving the perioral muscles, triggered by reading (aloud and silently), and in 25–30% of the patients also by talking. Reading-induced perioral myoclonias were considered to be a more or less exclusive seizure type in primary RE (1,2). However, because we could observe perioral myoclonias increasingly frequently in JME, in which they are triggered by talking and less frequently by reading (3,4), but do not occur spontaneously, we questioned the specificity of perioral reflex myoclonias (PORM) (5). PORM are short, sometimes repetitive, abrupt myoclonias around the mouth, which will clearly be noticed by the patients, sometimes with interruption of reading and speaking. Most of these jerks appear strongly localized and do not change the side in individual patients. In some cases, PORM are bilateral.

Bickford et al. (6) distinguished primary and secondary RE. The primary form, in which only reading-induced seizures occur, belongs to the idiopathic localization-related epilepsy syndromes. Secondary reading epilepsies, in which spontaneous and reading-induced seizures (e.g. psychomotor seizures) occur, belong to different epilepsy syndromes (e.g. symptomatic focal epilepsies). Another proposal by Bickford (7) was to distinguish a specific (“primary”) form from a nonspecific (“secondary”) form. Because data from several patients were published with the co-occurrence of RE and JME (3,8), a discussion involved whether RE should be considered as an idiopathic generalized epilepsy syndrome.


The clinical symptoms of PORMs bear the characteristics of cortical reflex myoclonias (CRMs). CRM is related to a small area of the sensorimotor cortex; it typically involves only a few adjacent muscles (9).

In RE the common precipitating mechanism of PORM, best investigated is the formal act of reading (i.e. the transformation of graphemes into phonemes) (1,2). In patients with RE, the sensorimotor cortex is not hyperexcitable per se, because spontaneous seizures never occur. A hypothesis exists that the hyperexcitable neuronal network that subserves speech may drive the relative motor cortex effectively through a direct transcortical pathway (10,11). This assumed bilaterally hyperexcitable speech network, under reading provocation, would give rise to bilateral myoclonic jerks, which need not be necessarily symmetrical (12).

In an analysis of EEG-backaveraging in a 53-year-old patient who had experienced isolated myoclonias of the right face for >10 years, induced by talking, reading, and writing, no EEG correlation was seen (13).


“JME appears around puberty and is characterized by seizures with bilateral, single or repetitive, arrhythmic, irregular myoclonic jerks, predominantly in the arms. No disturbance of consciousness is noticeable” (14). Seizure triggers are sudden awakening, sleep deprivation, stress, and alcohol. Valproic acid (VPA) is the antiepileptic drug (AED) of choice. The prognosis is good, but some authors believe that in 90% of all cases, life-long AED therapy is required (15). The age at onset is typically between 12 and 18 years. The typical EEG features are generalized spike–waves and polyspike waves, but focal epileptic discharges appear in ≤36% (16). At least one third of the patients are photosensitive (17).

Panzica et al. (18) investigated the cortical myoclonus in JME with jerk-related back-averaging on ictal epochs. They concluded that the ultimate mechanism responsible for ictal myoclonic jerks in JME is largely similar to that sustaining cortical myoclonus in more-severe pathologic conditions such as progressive myoclonus epilepsy.

The aim of our investigation was to look for the frequency of PORM in JME and in focal epilepsies with a special test battery used during video-EEG recording. The hypothesis was that PORM are more frequent in JME than in other epilepsy syndromes and that PORM like photosensitivity are a typical trait of JME. This trait would establish a relation between JME and RE.


In a pilot phase, we looked for the frequency of PORM in general. Systematic questioning of all new outpatients seen by the first author between 1994 and 1998 (n = 600) identified 17 patients with PORM. In the same period, another seven cases of PORM were detected among the inpatients of our hospital. Of these 24 patients, 10 had symptomatic focal epilepsies, two had idiopathic generalized epilepsy (IGE) with grand mal on awakening, and 12 had JME. Patients who were newly diagnosed with RE in the same period were not included in this study. Because JME appeared to be the syndrome in which PORM are most frequent, we began to investigate this co-occurrence systematically. We conducted a questionnaire survey for all our ambulatory patients with the diagnosis JME to search for PORM and other reflex epileptic traits. The questionnaire was sent to 86 patients with a well-established diagnosis of JME.

The diagnosis of JME was given by either of two of the authors (T.M., P.W.) based on a case history including a description, by the patients and their relatives, of typical myoclonic seizures and, possibly, generalized tonic–clonic seizures and absences, and the presence of generalized spike–waves (SWs) or poly-SW in the interictal EEG. All patients were investigated with magnetic resonance imaging (MRI), which was normal in all cases. Sixty-five (75%) of the patients responded to the questionnaire. Not considering nonspecific facilitating factors of seizures like lack of sleep, 33 patients answered “yes” to questions about specific precipitating stimuli. Twenty-five of these 33 patients were available for a prolonged polygraphic video-EEG recording and were compared with a group of patients with focal epilepsies.

Comparison group

The comparison group consisted of 25 matched consecutive patients of our epilepsy center with different types of focal epilepsies (i.e. epilepsies without known propensity for reflex epileptic mechanisms). Nineteen patients in this group had temporal lobe epilepsy, four frontal lobe epilepsy, and two parietooccipital epilepsy. They were collected during 1 year. Seizure frequency in this group ranged from three seizures per week to one seizure per month.

We performed a standardized interview with all patients and a video-EEG investigation lasting 3 h as a minimum, including 5-min hyperventilation, intermittent light stimulation, and complex neuropsychological tests. For the latter, we used the protocol of Matsuoka et al. (19) with some modifications, as follows.

  • • Reading silently (modification, 15 min, standardized unknown difficult text)
  • • Reading aloud (modification, 15 min, standardized unknown English text)
  • • Speaking (modification: speaking in a stressful manner to the camera about epilepsy and medical history)
  • • Mental calculations, 10 tasks (e.g. 11 × 11, 125/5)
  • • Written calculation: 10 tasks (e.g. 15 × 67 × 23 × 48)
  • • Writing of standardized texts (modification: duration minimum, 15 min)
  • • Spatial constructions, to build up a tower (the Jenga tower), drawing figures (people, dog, cow), block-design test.
  • • Modification: Rubik's cube: to arrange one side of the cube in one color.

Between two tests, a period of rest always was given. The total investigation lasted ≥3 h, including a total of ≥1 h of interspersed rest. Interictal spikes were rated as induced if an increase in number of >200% occurred compared with the baseline resting condition. The results were statistically investigated with Fisher's exact test (extended version).



The questionnaire, which was sent to 86 outpatients with JME, provided the following data:

Seven of the 33 (21.2%) positive responses concerned seizure precipitation by intermittent light stimuli. Twenty-one responses referred to “praxis” and comprised stimuli such as writing (eight), decision making (three), computer tasks and video games (six), calculations (six), thinking (eight), and playing the piano (one). Seventeen patients reported about PORM precipitated by talking (delivering a speech in five), and nine of them in addition by reading. Eleven patients reported precipitation by both praxis and talking/reading.

Video-EEG investigation

Twenty-five patients with JME were compared with 25 matched (1:1, age and sex) patients with focal epilepsies. The mean age of the JME patients was 29.8 ± 7.2 years (range, 17–42 years); the mean age of the comparison group was 30.0 ± 10.4 years (range, 17–52 years). Sex ratio was identical: 13 males versus 12 females in both groups. No seizures, especially no perioral myoclonias, were seen at rest.

Interictal EEG

Normal interictal EEG recordings were seen in 10 patients with JME and in 13 patients with focal epilepsies (all patients were taking AEDs). We saw a statistically significant difference (Fisher's exact test, two-tailed) in two items of the EEG recording (Fig. 1): Generalized epileptic discharges were seen in the JME group in 15 patients (Fisher's exact test, two-tailed, p < 0.001) and in five patients in the focal group (not clearly focal, but bilaterally synchronous with frontal maximum), whereas non focal discharges were found in patients with JME, but in 11 patients in the comparison group (Fisher's exact test, two-tailed, p < 0.001).

Figure 1.

Interictal EEG recording. JME, juvenile myoclonic epilepsy; HV, hyperventilation. *Statistically significant difference, p < 0.05, Fisher's exact test, two-tailed.

Special attention was paid to possible local onset of epileptiform discharges in the polygraphic EEG recording.

EEG activation

Activation induced by hyperventilation (HV) showed a difference in both groups: HV provoked generalized spikes and polyspike waves in seven patients with JME, but also in two patients of the focal group with a frontal maximum (probably as an expression of rapid secondary synchrony) (Fig. 2). This difference was only marginally significant (two-tailed Fisher's exact test: p = 0.074). Photosensitivity was seen in five patients with JME but never in the focal group (two-tailed Fisher's exact test: p = 0.022). A significant difference also was noted between the JME patients and the comparison group (Fisher's exact test, two-tailed, p < 0.01) comparing the precipitation of epileptiform discharges with different triggers during the EEG recording (10 patients with JME, two of the focal patients). Patients of the JME group (20%) showed epileptic discharges during reading, and four during speaking. These discharges were spikes and rapid spike–waves, and they were generalized or centrally localized, as seen in Figs. 3 and 4. Only one of the comparison patients showed epileptiform discharges during reading with spikes related to the left parietocentral region (Fig. 5). This difference is statistically significant, whereas the subanalysis of reading- and speaking-induced epileptic discharges was not significant, probably because of small numbers of patients (Fisher's exact test, p = 0.189, two-tailed; p = 0.095, one-tailed). Four patients showed epileptiform discharges during speaking, but none of the comparison group (two-tailed Fisher's exact test, p = 0.055). Praxis-induced (manipulating with Rubik's cube) EEG discharges were seen in four patients with JME, but in none of the comparison group (two-tailed Fisher's exact test, p = 0.05). Calculating (n = 1) and writing (n = 2) were only a little provocative in the JME group, and none of the patients in the focal group showed epileptic EEG discharges with this kind of precipitation.

Figure 2.

Induced epileptic discharges. Induction of epileptic discharges by reading/speaking/praxis of >200% per time unit in comparison to baseline EEG recording at rest.*Significant difference, p < 0.05, Fisher's exact test, one-tailed. **Significant difference, p < 0.001. Fisher's exact test, one-tailed.

Figure 3.

EEG activation during silent reading in a patient with juvenile myoclonic epilepsy.

Figure 4.

Induced epileptic discharges during reading loudly in a patient with juvenile myoclonic epilepsy. Speed, 30 mm/s; sensitivity, 10 μV/mm.

Figure 5.

Activation of epileptic discharges during reading in a patient with focal epilepsy. Speed, 30 mm/s; sensitivity, 10 μV/mm.

Ictal EEG

The ictal characteristics were, in most cases, rapid spike–waves with frontal maximum (Fig. 4), whereas in one patient with focal epilepsy, left parietocentral spike–wave activity appeared during silent reading (Fig. 5). Sometimes myoclonic jerks made EEG analysis impossible (Fig. 6). No overlap of focal and generalized discharges was noted in individual patients. The EEG pattern was not different in PORMs and in praxis-induced seizures (Fig. 7).

Figure 6.

Induced epileptic seizure (PORM) in a patient with juvenile myoclonic epilepsy. Speed, 30 mm/s; sensitivity, 10 μV/mm.

Figure 7.

Induced epileptic seizures (myoclonic jerk of the writing arm) in a patient with juvenile myoclonic epilepsy. Speed, 30 mm/s; sensitivity, 10 μV/mm.

Precipitation of PORM

In nine (36%) of the patients with JME, we recorded PORM induced by reading, speaking, and other neuropsychological activation (Table 3). Only one (4%) comparison patient showed seizures induced by reading (Fig. 8). The difference was statistically significant (two-tailed Fisher's exact test, p = 0.037). This result can be explained by the selection of patients with the questionnaire. PORM were not found in all patients with JME who had reported them in the questionnaire.

Table 3. Seizure semiology of JME patients
  1. PORM, perioral reflex myoclonias.

2FemalePORM Myoclonia at rest and photoinduced
3FemalePORMPORM Myoclonia of the writing arm 
4Female PORM 
7Female AbsenceMyoclonia 
8Male Absence 
Figure 8.

Induced epileptic seizures. *Significant difference, p < 0.05, Fisher's exact test, one-tailed. Speed, 30 mm/s; sensitivity, 10 μV/mm.

Precipitation of praxis-induced seizures

Six of the JME patients had praxis-induced epileptiform EEG discharges (generalized spike–waves), and two of them had praxis-induced seizures. One patient from the comparison group with induced seizures showed atypical myoclonias around the mouth and a complex partial seizure with impairment of consciousness as the main symptom.

The group of JME patients with PORM and/or praxis induction was not different from the group of JME patients without these. In terms of sex, age, seizure types, age at onset, or specific details of the medical history, we found no differences (Table 2). In both groups, photosensitivity appeared frequently (e.g. three of six patients with praxis-induced epileptic discharges were photosensitive)

Table 2. Comparison between JME patients with and without positive precipitation effect
 Positive precipitation effect
Yes (n = 10)No (n = 15)
Mean age in years. (±SD)28.8 ± 8.430.4 ± 6.5
Age of onset in years (±SD)15.1 ± 2.913.2 ±–3.7
Sex ratio men/women6/47/8
Family history6 (60%)3 (20%) 
 Perioral myoclonia6 (60%)5 (33.3%)
 GTCS9 (90%)13 (86.6%) 
 Absence seizures5 (50%)6 (40%)  
 Discontinuation of medication4 (40%)8 (53.3%)
 Photosensitivity3 (30%)2 (13.3%)


All 25 patients with JME but two were treated with AEDs, 18 with VPA, nine in monotherapy, and nine in combination therapy. Two patients received lamotrigine (LTG); two topiramate (TPM); and one phenytoin (PHT) in monotherapy. Combination therapies involved LTG, primidone (PRM), carbamazepine (CBZ), levetiracetam (LEV), phenobarbitone (PB), and ethosuximide (ESM).

Three of the patients with JME and PORM were not seizure free with VPA; only one did not respond at all, and the other patients were seizure free with VPA. Polymedication reflects that these were patients of a tertiary referral center and include a selection of patients who are more difficult to treat.

Two patients in the comparison group did not take any medication, 12 were treated with VPA, nine with CBZ, and seven with LTG, all in combination with VPA. Other medications were oxcarbazepine (OXC; three), LEV (two), gabapentin (GBP; two), TPM, PHT, clonazepam (CZP), and clobazam (CLB; one each).


In the database Medline, we found 525 published articles between 1950 and 2005 dealing with JME. Only a few articles dealt with special investigations on reflex epileptic traits (19–21). We found 378 articles about the combination of reading and epilepsy; few of them dealt in detail with RE, and only one dealt with PORM in patients with JME (8).

In this article, 18 patients with RE are described, four of whom also had a diagnosis of JME. Three of them were female patients. The age at onset was between 14 and 17 years; one patient was 46 years old at the onset of RE. The seizure trigger in all patients was reading; in three patients, it also was calculations. Listening to conservation, playing chess, or speaking was a trigger in individual cases. The interictal and ictal EEG showed generalized epileptic discharges in all patients; none of the patients was photosensitive. Two of them responded to VPA; partial control was seen in two patients treated with PHT and PRM.

Matsuoka et al. (19) investigated 480 patients with different epileptic syndromes with a specially designed neuropsychological EEG activation program. This program consisted of different tasks including reading and speaking (both very brief), arithmetic, writing, drawing, and a block-design test. In 133 patients, an inhibitory effect of the activation was seen in the EEG. In 38 patients, activation induced epileptic discharges, most of them in response to mental or written calculation (n = 21), writing of special standardized sentences (n = 26), or spatial constructions (n = 24). Only in two of the 38 patients reading was a specific trigger; in none of them did speaking induce seizures.

Concerning the relation of these findings to epilepsy syndromes, Matsuoka et al. (19) found that all but two of the patients who responded to activation had idiopathic generalized syndromes. Most of them (n = 22) had JME. Two patients had temporal lobe seizures. Considering the semiology of precipitated seizures, they found in most cases (n = 32) myoclonic seizures in both arms; 23 patients had generalized tonic–clonic seizures (GTCS), 19 had absence seizures, two patients had secondarily GTCS, and two had partial seizures. PORM were not described in any of the patients, probably because the patients read and talked too briefly in this protocol. Another reason for missing PORM in this article could be that patients and investigators were not aware of the probability of brief PORM, which often have no EEG correlate but only an electromyogram (EMG) artifact.

Inoue et al. (21) investigated patients with JME with a special test battery to search for praxis-induced seizures (PIS). Such seizures are triggered by a complex interaction of thinking with a manual task. Several of their patients had seizures while performing calculations with the traditional Japanese abacus, the soroban. This activity is a perfect example of praxis in the sense of a differentiated manual task guided by cognition, and the usually rapid performance enhances the stimulus. More than one fifth of the investigated 213 patients were photosensitive (n = 47). In 20 patients, typing or writing was provocative. In 16 patients, playing cards or chess precipitated myoclonic jerks in the active arm. Mental or written calculation was provocative in 15 patients, and complicated finger manipulation, in 11 patients. Only a few patients responded to playing video games (n = 6) or playing musical instruments (n = 4). PORM were reported in none of the patients.

All these patients have in common that some local motor performance, with which the patients are active, provokes local myoclonic phenomena or, as Bickford put it in his first description of RE, that “proprioceptive bombardment” results in “reflex firing through the same motor segment” (6). It is thought provoking that these minimal focal motor seizures occur both in a condition that is considered a “generalized” epilepsy syndrome” (JME) and a in “localization-related epilepsy syndrome” (RE). Furthermore, the reflex-like phenomena we observed here look substantially different from the typical bilateral brachial jerks of JME. The reflex epileptic seizure symptoms remain restricted to rather circumscribed areas. It is an intriguing question how rather direct, local reflex-like sensorimotor interactions fit into a “generalized” syndrome.

In our study we investigated JME patients who, in response to a questionnaire concerning specific reflex epileptic traits, had described such traits in their history. We tried to document these traits with video-EEG. This was not possible in all patients, and this may be due to the situation of investigation in which some patients are not sufficiently relaxed but also to other reasons. Some patients may have been mistaken when they reported trigger mechanisms. Almost all were taking AEDs, which were not interrupted for the investigation and may have suppressed the triggered responses. In addition, it seems that in many patients, these reflex epileptic traits, just as photosensitivity, are present only temporarily in a certain period. Quite a few patients had warned us that they had known PORM or praxis induction earlier but did not believe they still had it, so we cannot in absolute terms answer the question, how frequent are these traits in JME.

Interestingly, reflex epileptic traits were occasionally seen also in the comparison group, so that their specifity seems to be not very high. The control group were matched in age and sex, and all of them had focal epilepsies. Frontal lobe epilepsies especially have similiarities in semiology to JME, which may have something to do with the positive results in the control group. An optimal design for this study would have been to investigate consecutively both untreated JME patients and matched controls, but such patients are seen too rarely to make such a study feasible.

VPA is the treatment of choice in JME (and also for PORM), but not all investigated JME patients responded to VPA. Three of these nonresponders had the combination of PORM and JME, which could support the results of the investigation of Genton et al. (22), who found that significantly more patients with the combination of different seizure types in JME (absences, myoclonic jerks, GTCSs) were therapy resistant. It is possible that in some patients, inappropriate treatment could have had a facilitating effect for the occurrence of PORM. Two patients were without any treatment, three had CBZ, and three, OXC in combination; both AEDs are ineffective in JME and can increase the frequency of myoclonic jerks.

An ongoing discussion (23) considers the relation of supposed “generalized” widespread bilateral discharges to bilateral or unilateral regional expression of clinical symptoms. This discussion perhaps should be pursued. One of the open questions in this respect is, in photosensitive JME patients, by what intermediaries the occipital cortical input of intermittent lights produces myoclonic jerks of the extremities. Still, these are bilateral and roughly symmetrical. New patients, especially with the association of different epilepsy syndromes, will give more information in the pathophysiology of epileptic seizures with bilateral symmetric expression (24).

These observations seem to indicate that the distinction between focal and generalized epileptic phenomena may to some extent be artificial. The more closely we look at the prototype of an idiopathic generalized epilepsy syndrome, JME, the more we must realize that it is indeed a syndrome that challenges our syndromic concepts.

For better understanding of the pathophysiology of RE, Koepp et al. (25) investigated by using EEG, MRI, and positron emission tomography (PET) in a patient with RE. MRI was normal, whereas (11C)-diprenorphine PET revealed periictal opioid binding decreased in both temporal lobes and the left frontal lobe, maybe due to an abnormal activity in the network subserving reading.

Archer et al. (26) investigated two patients with reading epilepsy by using functional MRI (fMRI). Spike activity overlapped with reading activity in the left middle frontal gyrus, a structure recruited during working memory cognitive tasks. The authors postulated that, because of a local structural anomaly, the spikes of reading epilepsy spread from working memory areas into adjacent motor cortex, activating a cortical subcortical circuit. It would be necessary for understanding the pathophysiology of PORM to investigate all patients with PORM during reading provocation to look for the brain activity in comparison to normal controls (27). The next step to show differences between both groups was investigated in a recent study and failed, probably because all seven investigated patients had an effective treatment before fMRI was done (28). None of our patients showed a structural abnormality as Archer (26) had shown.

In contrast to patients with progressive myoclonus epilepsies (PME), the perioral reflex myoclonias in JME patients are phenomenologically indistinguishable from those in primary reading epilepsy. Like these, they are much rarer, more uniform, and more difficult to provoke. They are epileptic myoclonias and often accompanied by a spike or spike–wave on the surface EEG. The action myoclonias seen in many PMEs are much more easily evoked, ubiquitous, and variable, and without EEG spikes.


PORM may represent a frequent but often unobserved focal seizure type in different epileptic syndromes, not only in RE. They are more frequent in JME compared with focal epilepsy syndromes. The three reflex epileptic traits (i.e., photosensitivity, praxis, and reading) in JME may be found alone or coexist in one patient. They thus seem to be genetically independent. PORM remain often undiagnosed because the patients are not aware that they are a symptom and fail to report them. Their semiology is not fundamentally different in various epileptic syndromes. PORM are symptoms in different epileptic syndromes, where they indicate restricted local epileptic activity. These observations seem to indicate that the distinction between focal and generalized epileptic ictogenesis may be less clear than is traditionally believed. VPA seems to be the AED of choice in most cases of both primary RE and of PORM in JME. The response of PORM in other epilepsy syndromes to AED remains to be investigated.