Surgical treatment of patients with drug-resistant hypermotor seizures


Address correspondence to Li Yongjie, MD, PhD, Beijing Institute of Functional Neurosurgery, Xuanwu Hospital, Capital Medical University, Beijing, China. E-mail:


Purpose: To describe the clinical, electrophysiologic, neuroradiologic, and histologic findings in our patients with drug-resistant hypermotor seizures (HMSs) and to evaluate the outcome of their surgical treatment.

Methods: Twenty-three patients were identified by criteria for drug-resistant HMS. Surgical treatment and presurgical evaluation modalities including semiology, magnetic resonance imaging (MRI), interictal/ictal scalp video-EEG (electroencephalography), and intracranial recording were analyzed retrospectively.

Results: The common seizure frequency of 60–300 per month was observed among 15 patients. Sixteen patients (69.6%) experienced auras such as fear and palpitation. Marked agitation was observed in 12 patients and mild agitation in 11 patients. Groaning/shouting and asymmetric posturing were common accompanying symptoms. Asymmetric posturing was observed more often in patients with mild agitation than in those with marked agitation (p = 0.027). MRI detected focal abnormalities in six patients. Intracranial recording was conducted in 16 patients. The origins of seizures were localized in mesial frontal cortex in four patients, dorsolateral frontal cortex in four patients, and mesial temporal cortex in two patients. The epileptogenic zones (EZs) were resected from the frontal lobe in 21 patients and from the temporal lobe in 2 patients. The follow-up ranged from 12–60 months. Seventeen patients (73.9%) had been seizure-free, 11 of whom had presented with marked agitation (11 of 12) and 6 with mild agitation (6 of 11) (p = 0.069). Histologic examinations demonstrated focal cortical dysplasia (FCD) in 18 patients.

Discussion: The HMSs can originate from both the mesial and dorsolateral frontal cortex, and occasionally from the temporal lobe. Patients with drug-resistant HMSs should be recommended for resective surgical treatment.

Hypermotor seizures (HMSs) are seizures characterized by complex movements of the limbs and trunk, which appear “violent” when occurring at high speeds. The complex motor manifestations imitate normal movements; however, they are inappropriate for the situation and usually serve no purpose. Sometimes the movements are associated with affective symptoms such as fear, anger, concern, and verbal manifestations with violent yelling and screaming (Williamson et al., 1985; Lüders et al., 1998). HMS is generally considered a special manifestation of nocturnal frontal lobe epilepsy (NFLE) (Provini et al., 1999), which is often considered relatively benign because it is nocturnal dominant, genetically based, and drug relieving (Scheffer et al., 1995; Oldani et al., 1998). Although seizures occurring during sleep are usually thought to have a lower disabling impact on the quality of life, patients with HMS often complain about the violence with potential for hurting and disturbing. Furthermore, a significant proportion of HMS is drug-resistant (Hirsch et al., 1994; Provini et al., 1999). Data about surgical treatment in patients with drug-resistant HMS are relatively limited (Olivier, 1995; Jobst et al., 2000; Schramm et al., 2002; Nobili et al., 2003, 2007; Rheims et al., 2008). Herein we describe the clinical, electrophysiologic, neuroradiologic, and histologic findings and the surgical results of 23 patients who met our selection criteria for drug-resistant HMS.

Patients and Methods


We retrospectively reviewed the data of 620 patients who consecutively received surgical treatment at the Comprehensive Epilepsy Center of Beijing between April 2001 and April 2008. Twenty-three patients (3.7%) were identified by the criteria for drug-resistant HMSs. There were 17 men and 6 women, with a mean age of seizure onset of 5.8 ± 4.4 (range 5 months to 15 years) and mean age at surgery of 17.3 ± 8.0 (range 5–33 years). All patients had intractable epilepsies despite appropriate anticonvulsant medication. Surgical resections were performed in all 23 patients.

Presurgical evaluation

Magnetic resonance imaging

Standard magnetic resonance imaging (MRI) was performed on a 1.5T unit (Siemens Magneton Vision; Siemens, Munich/Erlangen, Germany) with conventional spin-echo T1-weighted axial, sagittal, coronal; and T2-weighted axial sequences in all patients (section thickness 5 mm, image gaps 1 mm). In addition, the fluid-attenuated inversion recovery (FLAIR) images with 5-mm–thick axial sections and coronal sections perpendicular to the long axis of hippocampus were obtained. [18]fluorodeoxyglucose positron emission tomography (FDG-PET) was not examined in these patients.

Video-EEG monitoring and seizure semiology

Interictal/ictal scalp electroencephalography (EEG) was recorded using a video-EEG monitoring system (DaVinci; Embla, Broomfield, CO, U.S.A.), with electrodes placed according to the international 10–20 system in all patients. The duration of video-EEG monitoring duration ranged from 2–15 days, and at least three habitual seizures were recorded. The HMSs were identified if the recorded motor manifestations involved predominantly irregular sequential ballistic movements (pedaling, pelvic thrusting, thrashing, rocking movements), increase in rate of ongoing movements, or inappropriately rapid performance of a movement. If the seizure recorded was not stereotyped, we would record more seizures to identify the typical seizures that were HMSs. The main ictal manifestations were categorized according to the International League Against Epilepsy (ILAE) classification (Blume et al., 2001) and Rheims’ classification (Rheims et al., 2008). They were listed as: (1) groaning or shouting; (2) asymmetric tonic or dystonic posturing; (3) marked agitation including body rocking, kicking or boxing behavior associated with sitting up behaviors; (4) and mild agitation characterized by horizontal movements or rotation of the trunk and pelvis. Auras that occurred during presurgical video-EEG monitoring or from their medical histories were documented for statistical analysis.

Intracranial EEG monitoring

Epileptogenic regions were preliminarily identified based on the combined data from MRI, ictal semiology, and interictal/ictal scalp EEGs. Inconclusive or inconsistent results were obtained in 16 patients. We further performed intracranial EEG monitoring on those patients. The placement of the grid or strip electrodes was guided by noninvasive examinations and intraoperative electrocorticography (ECoG). In 14 patients, electrode placement on the mesial frontal regions was applied to identify the possible originating or involvement of gyrus cinguli/mesial superior frontal gyrus. The frontal dorsolateral cortex was also covered by grid electrodes in all patients to identify the spreading relation between mesial and dorsolateral frontal cortex. In three patients, additional electrodes were also used to cover temporal or other extrafrontal lobes. Three or more habitual seizures were recorded on each patient during the intracranial EEG monitoring. Preoperative and intraoperative functional mapping was performed in five and three patients, respectively.

Surgery and surgical outcome

Resective microsurgery was conducted according to the results of the anatomic–electroclinical investigations and was aimed at removal of the epileptogenic zones (EZs). ECoG before and after cortex resection was performed in all patients to further tailor the resection. Three patients were awakened from general anesthesia to accomplish intraoperative functional mapping. All surgical complications, major or minor, were reviewed. All patients were followed postoperatively for 12–60 months (mean 24 months) to identify an outcome of either seizure-free or unfavorable.

Pathologic diagnosis

All the tissue sections from cortical resection were routinely processed to obtain detailed pathologic malformation of cortical development (MCD) according to the grading system of Palmini et al. (2004).


The demographic features of the 23 patients are summarized in Table 1. Seizure frequency was calculated according to the attacks 1 year before the operations. High seizure frequency of 60–300 per month was observed in 15 patients. None of our patients had a family history of nocturnal seizure. Two patients had a positive history of perinatal accidents, whereas one patient had a drowning history when he was 3 years old. The neurologic physical examinations were normal in all patients except for lower IQ (≤70) in three patients. Twenty patients had seizure attacks dominantly at night or during sleep; two patients has seizures in the day time, and one patient demonstrated no time difference. After several failed monotherapy or combined therapeutic attempts, the patients were referred for surgical intervention.

Table 1.   Demographic features and clinical manifestations
Patient No.GenderAge (years)Seizure frequency/monthNocturnal seizuresMRI abnormalitySeizure semiologySurgical removalHistologyFollowup (months)Outcome
Seizure onsetSurgeryAuraManifestationsSiteMesial F
  1. M, male; F, female; L, left; R, right; F, frontal lobe; T, temporal lobe.

  2. Seizure frequency/month: the average seizure number in a month in the 1 year before surgery.

  3. Manifestations: The main ictal manifestations were categorized into four groups: 1, groaning or shouting; 2, asymmetric tonic or dystonic posturing; 3, marked agitation including body rocking, kicking, or boxing behavior associated with sitting up behaviors; 4, mild agitation characterized by horizontal movements or rotation of the trunk and pelvis.

  4. Outcome: F, seizure-free; U, unfavorable.

  5. Case 2 had two surgical procedures.

2 (1)M219>300MostR-FNo1+3R-FYesFCDIIA12U
3M92260–120MostNoUpset, fear1+3R-FNoFCDIIA12F
6M0.53360–120MostNoPalpitation, dyspnea1+2+4R-FYesMMCD48F
10M62410–60MostNoPalpitation, dyspnea1+2+4L-FYesFCDIIB24F
23M51710–20MostR-HippocampusEpigastric sensation1+2+3R-TNoHippocampal sclerosis12F

Auras and seizure semiology

Sixteen patients (69.6%) were experiencing or had previously experienced auras: fear in seven patients and palpitation/dyspnea in six patients. Marked agitation (3) was observed in 12 patients and mild agitation (4) in 11 patients. These two dominant manifestations were accompanied by groaning/shouting in 19 patients [(1) + (3)/ (4)], by asymmetric tonic/dystonic posturing in 15 patients [(2) + (3)/ (4)]. The asymmetric posturing was observed in five patients with marked agitation (5 of 12) and in 10 patients with mild agitation (10 of 11). The difference in the number of patients with asymmetric posturing was significant between these two groups (p = 0.027).


The MRI results of six patients detected focal abnormalities: four located in the frontal lobe and two in the temporal lobe. The latter included neoplasm and hippocampal sclerosis. The other 17 patients (73.9%) showed nonlesional MRI results.

Electrophysiologic investigations

Interictal and ictal EEG

The findings of interictal (during both wakefulness and sleep) and ictal scalp EEG are listed in detail. Interictal EEG abnormalities were lateralizing and/or localizing in 9 patients, bilateral in 13 patients, and absent in 1 patient. Interictal epileptic abnormalities were dominant in frontal lobe in 10 patients, frontotemporal lobe in 10 patients, and temporal lobe in 2 patients. The EZs were finally localized in the frontal lobe in the two patients with temporal dominant interictal EEG abnormalities. Ictal EEG abnormalities were lateralizing and/or localizing in 10 patients. The earliest detectable EEG changes were diffuse flattening of background activity in seven patients, a rhythmic delta or slow activity in two, spikes or spike-and-wave activity, in six and small amplitude fast activities in five patients. EEG recordings during the attacks failed to disclose ictal epileptic activity because of muscular artifacts in three patients. Overall, the available noninvasive anatomic–electroclinical findings were considered adequate to localize consistently the EZs and to define a surgical strategy in seven patients.

Intracranial recordings

In 16 cases, a variable degree of inconsistency among the anatomic–electroclinical data, as to the localization of the EZ, indicated the need for intracranial recordings. Frontal lobe sample was employed in all intracranial explorations, including double frontal lobe sample in 11 patients. In addition, strip electrodes were inserted into longitudinal fissure to sample the regions of mesial cortex or orbitopolar cortex in 14 patients, and extrafrontal grid or strip electrodes were arranged to sample temporal cortex in 4 patients.

The earliest detectable EEG modifications recorded by the intracranial electrodes included focal low fast activities in nine patients, focal rhythmic spikes or spike-and-wave activities in 6, and focal rhythmic spikes followed by diffuse flattening in one patient. The ictal EEG modifications were first observed in mesial frontal cortex or orbitopolar cortex in four patients, dorsolateral frontal cortex in four patients, both mesial and lateral cortex synchronously in four patients, and mesial temporal cortex in two patients (Figure S1). The initial EEG modifications were not clear in the remaining two patients. In cases 4 and 23, the video-EEG recording showed that the complex behaviors were motivated only when the focal activities spread to the frontal region from temporal lobe in a few seconds.

Surgery and surgical outcome

All 23 patients underwent surgical resections: 21 in frontal lobe (12 left and 9 right) and 2 in temporal lobe. The surgical resections involved both the mesial and dorsolateral frontal cortex in 15 patients, dorsolateral frontal cortex alone in 3 patients, orbitopolar region in 3 patients, and anterior temporal lobe in 2 patients (Fig. 1). EZ localized by focal low fast activities or rhythmic spikes on a few electrodes was considered focal or limited; therefore, a focal resection was performed. Otherwise, the EZ was considered larger, and wider or lobar resection was reasonable. In case 2, seizures reoccurred a month after the right frontal lobe resection and gradually worsened in the following months. A second operation was performed to remove completely the residual orbital cortex 1 year later and that patients was seizure-free during our 1 year follow-up.

Figure 1.

Schematic drawings of resections (black-shadowed areas) performed in the 23 patients. Patients with seizure-free outcome are highlighted by “*”. Gray shadowed area indicates first resection performed in case 2.

Only transient postoperative complications were observed, including mild contralateral motor deficits for a few days in four patients and epileptic status on the first day for several hours in one patient. Seventeen patients (73.9%) had been seizure-free since surgery (Engel’s class I), including one patient with two surgical procedures (case 2), 11 of the patients with marked agitation were seizure-free (11 of 12) and 6 of the patients with mild agitation (6 of 11) were seizure-free (p = 0.069). The difference of outcome was not significant in patients with or without asymmetric posturing (p = 0.369). As of the last follow-up, anticonvulsive drugs had been withdrawn from 3 seizure-free patients and had been tapered in 4 and unchanged in 16 patients.


Histologic examination of resected specimens disclosed focal cortical dysplasia (FCD) in 18 patients. According to the grading system of Palmini et al., FCDs were further graded as: FCDIA in two cases, FCDIB in five cases, FCDIIA in five cases, and FCDIIB in six cases. Of the 17 patients with seizure-free surgical outcome, their histopathologic evaluations showed FCDII (A and B) in eight patients (47%). The hippocampal sclerosis shown on MRI was confirmed by pathologic evaluation, and a focal abnormality in temporal lobe was identified as gangliogliomas.


Data on patients with HMSs receiving surgical treatment are limited. The manifestations and localization of EZ of HMS were demonstrated only in a few previous reports. The clinical features of our population match only some traits of reported patients with NFLE (Oldani et al., 1996; Provini et al., 1999). An autosomal dominant inheritance was found in 8–43% of patients (Oldani et al., 1998; Combi et al., 2004) and a familial recurrence of the attacks was found in 25% of patients (Provini et al., 1999). None of our patients had a family history of nocturnal paroxysmal or seizures. The different selection criteria may be the reason; our patients were selected through a surgical program that included (1) drug-resistant epilepsies, (2) complaints sufficiently severe to consider surgical treatment, and (3) EZs that could be localized. Therefore, only a special subset of NFLE was selected, including some daytime dominant epilepsy and extratemporal epilepsy. Daytime seizures might lead to many more complaints than nocturnal ones, and those patients were more willing to consider surgical treatment.

In Rheims’s study, the dominant ictal manifestations were differentiated by marked or mild agitation (Rheims et al., 2008) to understand and study better the complex motor activity. In addition to the typical irregular ballistic movements, groaning/shouting and asymmetric tonic/dystonic posturing were two other common manifestations in our patients. The asymmetric posturing was observed more often in patients with mild agitation than those with marked agitation in our patients. The majority of the patients were experiencing or had previously experienced auras such as fear, palpitation, dizziness.

Presurgical evaluation

The process of presurgical diagnostic evaluation is a challenge in patients with HMSs. This is probably due to the large size of the frontal lobe, to the complexity of the functional network involved in the generation of frontal lobe seizures, and to the frequent absence of definite information provided by anatomic–electroclinical findings (Munari et al., 1995; Nobili et al., 2007). Furthermore, this and previous reports demonstrate that HMS may originate from the extrafrontal lobe, in particular, the insular or temporal lobe (Nobili et al., 2004; Ryvlin et al., 2006).

The typical manifestations of HMS strongly suggest frontal lobe origin or its involvement, but help little in lateralizing the EZ. We are not sure whether the high nonlesional MRI ratio may indicate relatively minor pathologic abnormalities. This was supported by the pathologic findings that demonstrated 18 FCDs. The similar condition was also described in previous reports: the incidence of uninformative findings of MRI was 47.6% in Nobili’s report and 86% in Provini’s series (Provini et al.,1999; Nobili et al., 2007). Another reason for the high nonlesional MRI ratio in our study might be the relatively rough image scan (5-mm section thickness). It is conceivable that other special MRI technologies, such as special lamellar scanning and magnetic resonance spectroscopic imaging (Guye et al., 2005) may disclose more subtle structural abnormalities in these demanding cases. The interictal epileptic abnormalities were found dominantly in frontal or frontotemporal lobe (87%), which was consistent with frontal localization of EZ. But the epileptic discharges were often bilateral (57%). In addition, the localizing role of ictal EEG was often limited by burst of diffuse flattening and violent muscular artifacts.

For the above reasons, intracranial recordings are usually required. It is difficult to sample fully the bilateral frontal regions with limited electrodes. The arrangement of the grid or strip electrodes was determined according to the results of anatomic–electroclinical evaluation. The common early EEG modifications recorded by the intracranial electrodes were focal small amplitude fast activities. These fast activities were difficult to record with scalp electrodes. This may explain the low localizing ratio of ictal scalp EEG to some extent. The ictal EEG modifications recorded from the mesial frontal cortex may be earlier or later than those recorded from the dorsolateral frontal cortex. This means that the HMS may originate from either mesial frontal cortex or dorsolateral frontal cortex. In other patients, the ictal EEG abnormalities appeared simultaneously in the mesial and lateral cortex. The EZ may not be accurately covered with electrodes or it might hide somewhere under the surface of cortex. In the latter condition, stereo-EEG (SEEG) may show the privilege.

Origin of HMS

In previous reports, the frontal origin of HMSs was suggested by a few studies (Morris et al., 1988; Nobili et al., 2003; Kellinghaus & Lüders, 2004). In addition, orbitofrontal, mesial frontal, frontopolar, and anterior cingulate hyperperfusion was demonstrated with ictal SPECT (Harvey et al., 1993; Hayman et al., 1997; Vetrugno et al., 2005). A more detailed observation showed that he EZ of the patient with marked agitation was centered mainly on the ventromesial frontal cortex, whereas the EZ of the patient with mild agitation was localized primarily within the mesial premotor cortex (Rheims et al., 2008). In our study, postoperative data showed that the dorsolateral frontal resection alone could also yield seizure-free outcome in a few selected patients. This result indicated that HMS might also originate in lateral frontal cortex. The asymmetric posturing was observed more often in patients with mild agitation, and their surgical outcomes were not as good as those with marked agitation. We wonder if the EZs of patients with mild agitation are likely localized in dorsal/mesial premotor cortex or supplementary motor area (SMA). Inadequate removal of this area may be responsible for the poor outcome in four of the five patients with asymmetric posturing, but further study is necessary for confirmation.

The extrafrontal origin of HMSs, such as temporal lobe and insula, was also demonstrated in our study and previous reports. Functional neuroimaging abnormalities were also observed in insula of some patients with HMS (Harvey et al., 1993; Hayman et al., 1997; Schlaug et al., 1997). Although intracranial recordings cannot give sufficient information in these cases to explain how an epileptic discharge starting from the temporal or insular lobe can produce hypermotor behavior, some authors agree that hypermotor seizures with extrafrontal origin may be the result of activation or deactivation of frontal circuitry network (Bartolomei et al., 2002; Nobili et al., 2004; Dobesberger et al., 2008). Intracranial investigation has suggested that complex motor automatisms usually appear when the ictal discharge in the temporal lobe became more spread or involved extratemporal structures such as the cingulate and the frontal regions. Our clinical electrophysiologic data illustrated this process clearly in cases 4 and 23. The next question is whether the cingulate or mesial frontal cortex is essential to originate HMS. Although the propagation of ictal circuitry was not very clear because of the necessarily limited number of cortical and depth electrodes, it is concluded that HMS may originate from either mesial or dorsolateral frontal cortex. Mesial frontal cortex may be a symptomatic zone involved in the process of discharge propagation whether the seizures rise from mesial or lateral frontal, temporal, or insular lobe. Although we did not find an obvious clinical difference between dorsolateral and mesial/polar ictal onset, the consciousness recovery seemed slower in patients with temporal ictal onset than in patients with frontal onset. We were not sure whether it could be explained by the difference of conscious disturbance in patients with frontal and temporal ictal onset.

Surgical outcome and histology

Both our study and some previous studies suggest the value of appropriate surgical treatment in patients with HMSs. However, our seizure-free ratio was increased by some other factors. For example, both the two patients with temporal origin and the relatively short follow-up (1 year) increased the seizure-free ratio in our small series when compared with surgical outcome of frontal lobe epilepsy (Ferrier et al., 1999; Lee et al., 2008). In addition, the favorable surgical outcome might be explained or extrapolated by the high incidence of cases with a histologic diagnosis of FCDII, according to some authors’ reports (Schramm et al., 2002; Tassi et al., 2002; Nobili et al., 2007). In these reports, Taylor’s dysplasia was considered more focal in nature than architectural FCD, and the onset of seizures documented by intracranial recording was likely to represent a relatively focal abnormal cortex.

Transient contralateral motor deficits were the common surgical complications. This morbidity often occurred in patients with removal of posteromesial superior frontal gyrus. Transient functional impairment of the SMA may be responsible for this disability (Krainik et al., 2001). In addition, the relatively low incidence of postoperative hemiparesis in our series may be explained by pre- or intraoperative functional cortical mapping.


We would like to thank Liang Qiao and Rong Cai for their review of the grammar in this article.


We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

None of the authors has any conflict of interest to disclose.